- Email: [email protected]
Comparative rates of harms in randomized trials from more developed versus less developed countries may be different
Accepted Manuscript Comparative Rates of Harms in Randomized Trials from More-Developed Versus Less- Developed Countries May be Different Despina Contopoulos-Ioannidis, clinical associate professor, Xanthippi Tseretopoulou, registar in paediatrics, Megan Ancker, field physician, Juan N. Walterspiel, pediatric infectious diseases consultant, Orestis A. Panagiotou, postdoctoral fellow, Yvonne Maldonado, professor, PA John Ioannidis, professor PII:
S0895-4356(16)30028-2
DOI:
10.1016/j.jclinepi.2016.02.032
Reference:
JCE 9143
To appear in:
Journal of Clinical Epidemiology
Received Date: 23 April 2015 Revised Date:
4 January 2016
Accepted Date: 4 February 2016
Please cite this article as: Contopoulos-Ioannidis D, Tseretopoulou X, Ancker M, Walterspiel JN, Panagiotou OA, Maldonado Y, Ioannidis J, Comparative Rates of Harms in Randomized Trials from More-Developed Versus Less- Developed Countries May be Different, Journal of Clinical Epidemiology (2016), doi: 10.1016/j.jclinepi.2016.02.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript (with PAGE numbers)
ACCEPTED MANUSCRIPT
Comparative Rates of Harms in Randomized Trials from More-Developed Versus LessDeveloped Countries May be Different
1 2 3
, Xanthippi Tseretopoulou,
RI PT
Despina Contopoulos-Ioannidis, clinical associate professor
registar in paediatrics 4, Megan Ancker, field physician 5, Juan N. Walterspiel, pediatric infectious diseases consultant
6
, Orestis A Panagiotou, post-doctoral fellow
Department of Pediatrics, Division of Infectious Diseases, Stanford University School of
Medicine, Stanford, CA, USA
M AN U
1
2
Palo Alto Medical Foundation Research Institute, Palo Alto, CA, USA
3
Meta-Research Innovation Center at Stanford, Stanford, CA, USA
6
7
TE D
5
Leeds Teaching Hospital, NHS Trust, Leeds, UK Medecins San Frontieres, Paris, France
Mendocino Coast District Hospital, Fort Bragg CA, USA Division of Cancer Epidemiology & Genetics, National Cancer Institute, National Institute of
EP
4
, Yvonne
SC
Maldonado, professor 1 8 9, John PA Ioannidis, professor 3 8 10 11
7
Health, Bethesda, MD, USA
Department of Health Research and Policy, Stanford University School of Medicine, Stanford,
AC C
8
CA, USA 9
Senior Associate Dean for Faculty Development and Diversity, Stanford University School of
Medicine, Stanford, CA, USA 10
Stanford Prevention Research Center, Department of Medicine, Stanford University School of
Medicine, Stanford, CA, USA
1
ACCEPTED MANUSCRIPT
11
Department of Statistics, Stanford University School of Humanities and Sciences, Stanford,
CA, USA Address correspondence to: Despina Contopoulos-Ioannidis, MD; Department of Pediatrics,
RI PT
Division of Infectious Diseases, Stanford University School of Medicine, 300 Pasteur Drive,
AC C
EP
TE D
M AN U
SC
Room G312, Stanford CA, 94305, USA; e-mail: [email protected]
2
ACCEPTED MANUSCRIPT
Funding: There was no funding for this study. Conflicts of interest: There are no conflicts of interest to declare for DCI, XT, MA, OAP, YM and JPAI. JNW worked as a consultant for US Bayer A.G. in 2002-3. “Whistleblower" on Bayer
that are paid for directly or indirectly by Bayer A.G.
RI PT
A.G. Currently does not hold any Bayer A.G or subsidiarie’s stock, nor does he give any lectures
The views expressed in this paper are solely those of the authors and do not necessarily represent
SC
the official views of the National Institutes of Health or the US Department of Health and
AC C
EP
TE D
M AN U
Human Services.
3
ACCEPTED MANUSCRIPT
Abstract Objectives: We set up to evaluate the relative risk of harms in trials performed in less-developed versus more-developed countries.
RI PT
Study Design: Meta-epidemiologic evaluation using the Cochrane Database of Systematic Reviews. We considered meta-analyses with at least one RCT in a less-developed country and one RCT in a more-developed country. We targeted severe adverse events (AEs),
SC
discontinuations due to AEs, any AE, organ system-specific AEs, individual AEs and all discontinuations due to any reason. We estimated the relative odds ratio (ROR) of harms
under each category of harms.
M AN U
between more and less-developed countries for each topic and the summary sROR across topics
Results: We identified 42 systematic-reviews (128 meta-analyses, 521 independent RCTs). Summary sRORs did not differ significantly from 1.00 for any harm category. Nominally
TE D
significant RORs were found in only 6/128 meta-analyses. However, in 27% (35/128) of metaanalyses the ROR point-estimates indicated relative differences between country-settings >2fold. Considering also ROR 95% confidence-intervals, in 92% (118/128) of meta-analyses one
EP
could not exclude a 2-fold difference in both directions. Conclusions: We identified limited comparative evidence on harms in trials from these two
AC C
country-settings. Substantial differences in the risk point-estimates were common; the potential for modest differences could rarely be excluded with confidence.
Key words: Comparative Safety, Comparative Harms, More developed countries, Less developed countries, Randomized trials, Meta-analyses.
4
ACCEPTED MANUSCRIPT
Running title: Comparative Harms in RCTs from More vs Less Developed Countries What is new? •
Randomized trials on important clinical questions are frequently performed in less-developed
RI PT
countries, without long-standing clinical research tradition. We have previously shown that trials in such country-settings tend to report more favorable mortality outcomes. It remains unknown whether similar differences exist in the reporting of harms. We studied the
SC
comparative safety of medical interventions in randomized trials from more- versus less developed countries.
The identified comparative evidence on harms in trials from these two country-settings was
M AN U
•
limited. •
Not many nominally-significant discrepancies in relative-risks of harms between country settings were detected. However, substantial differences in the point-estimates of the risks of
TE D
harms were common and the potential for modest differences between the two country settings could rarely be excluded with confidence. What this adds to what was known?
Due to the limited comparative evidence, it remains unclear whether safety data from trials in
EP
•
less developed countries could be reliably used to guide policy development and clinical
AC C
decision making in more developed countries, but also less developed countries. What is the implication, what should change now? •
Evaluation of the comparative safety of medical interventions in trials from these two country settings should be further pursued and additional comparative evidence needs to be accumulated.
5
ACCEPTED MANUSCRIPT
Alternative methodological approaches, i.e. through analysis of comparative safety data reported within international multisite trials, between trial sites in more developed countries
EP
TE D
M AN U
SC
RI PT
and trial sites in less developed countries should be also considered.
AC C
•
6
ACCEPTED MANUSCRIPT
Introduction As the budget for and participation rates in randomized clinical trials (RCTs) in more developed countries are limited, it is becoming increasingly more common that clinical
RI PT
guidelines and clinical decision-making about important questions of health interventions and health care will depend on evidence from trials performed in less developed countries.[1] With the globalization of clinical research, emerging countries are increasingly more actively involved
SC
in clinical trials.[2-4] Asian and Latin America regions have recently shown the largest annual increase in the number of registered clinical trials [5]. Trials done exclusively in less developed
M AN U
countries often have low methodological quality. [6] Only 56% of 670 surveyed researchers from developing countries reported that their research had been reviewed by a local institutional review board [7] and only 11% of published clinical trials conducted in China in 2004 report that their study protocol was reviewed by an ethical review committee.[8] Some reports are raising concerns for underreporting of adverse events from studies
TE D
performed in developing countries.[9] For example, large differences in the reported rates of ciprofloxacin associated arthropathy were seen in children from North America compared to
EP
children from Latin America.[9] However, the frequency of this phenomenon has not been systematically studied. If results from trials performed in countries without a longstanding
AC C
tradition in clinical research will be used to guide clinical decision-making, an empirical largescale evaluation of these trends is needed. While there are potential benefits from the globalization of clinical research, it is important to evaluate whether results are similar and possible to extrapolate across different settings. In a previous evaluation [10] we assessed differences in mortality and primary efficacy outcomes in RCTs performed on the same topic in more versus less developed countries. We found that on average, trials in less developed countries tended to report more favorable results for the 7
ACCEPTED MANUSCRIPT
experimental intervention [10]. Sometimes genuine differences between country settings could explain differences in results; however, selective outcome reporting, publication, language and other biases [11-13] in the literature coming from less developed countries was considered more It would be important to assess whether major harms
RI PT
likely to explain these discrepancies.
outcomes are also similar or different in RCTs from more versus less developed countries.
Very often the number of patients studied in pre-licensure trials is small to allow the
SC
robust evaluation of both safety and efficacy outcomes. [14] I Individual studies and even individual meta-analyses are on average underpowered to detect differences in reported rates of
M AN U
clinically important adverse events from different country settings.[15] Therefore, we performed a large-scale meta-epidemiological evaluation of safety outcomes in trials from more vs. less developed countries. Methods
TE D
Definitions of countries
The categorization of countries into more developed and less developed countries was done as previously described in our earlier paper on comparative results for mortality and
EP
primary efficacy outcomes.[10] In brief we considered more developed countries to be those with both longstanding established marker economies and longstanding tradition in clinical
AC C
research as previously suggested.[16] Such countries included the United States, Canada, Australia, New Zealand, Israel, Japan and Western European countries. All other countries except for those in Eastern Europe were considered as less developed. We excluded RCTs from Eastern European countries as these represent another unique type of countries in transition. [17] Harm-related endpoints
8
ACCEPTED MANUSCRIPT
We targeted 6 main categories of harm-related endpoints [18], three of which were considered as primary endpoints because they combine adverse events of all types and they only include harms. We did not focus on mortality, as this was the focus in our previous paper [10].
RI PT
The primary endpoint categories were: severe adverse events; discontinuations due to adverse events; and any adverse event. The secondary endpoint categories were: organ system-specific adverse events; individual adverse events; and all discontinuations-due to any reason (in some
SC
studies, discontinuations due to harm might not have been separately reported, but they could have been included under such a broader study endpoint; this endpoint would then be relevant
M AN U
for harms, even though not fully specific). Under the categories of organ system-specific and individual adverse events we considered several subcategories (e.g. gastrointestinal adverse events, hematologic adverse events etc.; and headache, neutropenia etc., respectively). These endpoints were considered secondary because they either do not include all adverse events or
TE D
they count also some events that are not due to adverse events.
We also considered a composite primary-harms endpoint (combined primary-harms), where all three primary harm endpoints (severe adverse events, discontinuations due to adverse
EP
events and any adverse event) were considered together. Eligible meta-analyses
AC C
We included meta-analyses that quantitatively synthesized evidence on harm-related
endpoints and included at least one RCT from a less developed country and at least one RCT from a more developed country for the same compared interventions and the same type of harmrelated endpoint. We focused on RCTs performed exclusively in more developed countries or exclusively in less developed countries; international multicenter trials were excluded unless all participating sites were from the same country-setting.
9
ACCEPTED MANUSCRIPT
Search We screened a list of 131 systematic reviews from the Cochrane Database of Systematic reviews (last update August 27, 2012) that we had previously identified [10] as having RCTs
RI PT
from more developed and less developed countries for mortality. The reasons for this choice were that this selected group of systematic reviews would have been more likely to have included also evidence on harms from both country settings; and the questions addressed were
SC
likely to pertain to serious clinical conditions and diseases given that death was also an outcome of interest. Meta-analyses within the same systematic review for different compared
M AN U
interventions or for different types of harm-related-endpoints were considered as separate metaanalyses. Thus, each systematic review could have contributed data to more than one category of harm-related endpoints. Data extraction
TE D
From each eligible meta-analysis we extracted information on the compared interventions, the harm-related endpoint, the study (author, year), the number of patients with harm-related endpoints and the total number of patients in the experimental group and the control
EP
group respectively per study. No continues harms were encountered. Statistical synthesis of data
AC C
We synthesized the data in the following sequential steps as previously described [10 18
19]: First, for each RCT we calculated the odds ratio (OR) of harm-related endpoints for the experimental vs. the control intervention. When needed we coined estimates so that an OR>1 always corresponds to an increased risk of the harm-related outcome. If in the 2×2 table for each study (events/total number of subjects in the experimental arm and the control arm respectively) there were one or more cells with zero events, then a standard correction of 0.5 was added to all
10
ACCEPTED MANUSCRIPT
cells to allow for a meaningful calculation of the OR, as previously described.[20] Second, for each individual meta-analysis, under each category of harm-related endpoint, we calculated the summary OR (sOR) from all RCTs from the more developed countries and the sOR from all
RI PT
RCTs from the less developed countries using random-effect model for the meta-analysis; we also performed sensitivity analyses using fixed-effects model. [21 22] Random effects model assumes that the combined studies aim at identifying an average treatment effect, while fixed
SC
effects model assumes a common effect across the combined studies. Third, for each individual meta-analysis, and for each category of harm-related endpoint, we calculated the relative OR
M AN U
(ROR) between more developed and less developed countries by dividing the sOR from trials in more developed countries by the sOR from trials in less developed countries. Fourth, for each category of harm-related endpoint, we generated a summary ROR (sROR) between more and less developed countries by synthesizing the RORs from all the individual meta-analyses, under
TE D
each category of harm-related endpoint using random-effects models. A ROR>1 (or sROR>1) means that less developed countries reported less unfavorable results for the experimental intervention relative to more developed countries.
EP
For the combined primary-harms endpoint, we synthesized available RORs from all 3 primary endpoints; whenever RORs existed for more than one primary endpoint for the same
AC C
meta-analysis, we used the ROR that had the smaller variance. For the meta-analyses for organ system-specific adverse events and individual adverse events, we combined RORs only when they were targeting exactly the same types of organ system-specific or individual adverse events. We calculated the between meta-analyses heterogeneity under each category of harm-
related endpoint by using the I2 metric with its corresponding 95% CIs.[23 24]
11
ACCEPTED MANUSCRIPT
We calculated the number of meta-analyses where there was a statistically significant difference in the sOR between the two country settings (95% CIs of ROR excluding 1.00) and
the two country settings (ROR≥ 2.0 or ROR ≤0.50).
RI PT
the number of meta-analyses where the ROR suggested difference exceeding 2.0-fold between
In topics with statistically significant differences in harm-related endpoints between the two country settings, we also explored whether there were statistically significant differences in
SC
the baseline risks of these endpoints in the control group between meta-analyses from more developed and less developed countries. We synthesized the baseline risks per country setting by
M AN U
random effects using the Freeman-Tukey transformation method [25] for the meta-analysis of proportions and compared the results in the two country settings by chi-square test. We also examined whether there was evidence for small study effects using the Harbord’s test (considered to be significant for P<0.10). [26]
TE D
In sensitivity analyses we compared meta-analyses of U.S. trials only vs. trials from less developed countries. We also performed sensitivity analyses for the primary endpoints where we excluded trials from four Asian “tigers” (Hong Kong, Taiwan, Singapore, and South Korea) that
EP
evolved from less developed countries into advanced economies according to the International Monetary Fund in 1997, although their tradition of clinical research is still not as long-standing
AC C
as in the main more developed countries. All analyses were done in Stata SE12 (College Station, TX: StataCorp LP). We followed
the PRISMA [27] (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines for reporting. Results Eligible topics 12
ACCEPTED MANUSCRIPT
Of the previously identified 131 systematic reviews [10], 42 were considered eligible (Supplementary file 1) with a total of 128 separate meta-analyses: 12 meta-analyses for severe adverse events, 12 for discontinuations due to adverse events, 9 for any adverse event, 5 for
RI PT
organ system-specific adverse events, 89 for individual adverse events and 1 for all discontinuations due to any reason (Figure 1). From a total of 521 independent RCTs -after excluding 38 from mixed country settings, 5 with no reported country setting and 4 from Eastern
SC
European countries- we eventually analyzed 474 RCTs; 337 from more developed countries and 137 from less developed countries (Supplementary file 2). The topics of the 42 systematic
M AN U
reviews were related to infectious diseases (n=10), neonatal diseases (n=6), maternal-fetal conditions (n=6), cardiovascular diseases (n=5), neurologic diseases (n=5), gastrointestinal diseases (n=3), oncologic conditions (n=3), transplantations (n=2), rheumatologic diseases (n=1) and pulmonary diseases (n=1). The individual topics appear in Supplementary file 3.
TE D
The median sample size per RCT from more developed and less developed countries for the three primary harm endpoints was 175 and 278 respectively for severe adverse events; 151 and 377 respectively for discontinuations due to adverse events and 82 and 143 respectively for
EP
any adverse event (Supplementary file 4). The median adverse event rate per RCT in more developed and less developed countries was 3.5% and 3.4% respectively for severe adverse
AC C
events; 4.8% and 1.0% for discontinuations due to adverse events and 10.1% and 10.1% respectively for any adverse event (Supplementary file 4). Relative differences in individual meta-analyses (more-developed vs. less-developed) The clinical topics and the ROR estimates between the two country settings for the three primary endpoints and the combined primary-harms endpoint are shown in Table 1 and Supplementary file 3; the respective figures for the summary OR per meta-analysis according to 13
ACCEPTED MANUSCRIPT
country-setting and the ROR per meta-analysis, for each category of these harms are shown in Supplementary files 5-8. Respectively, the summary ORs per country-setting per meta-analysis and the ROR per meta-analysis for each of the secondary endpoint categories are shown in
RI PT
Supplementary files 9-11. The summary estimates (sRORs) for the different primary and secondary endpoints and the combined primary-harms endpoint are shown in Supplementary file
SC
12.
The summary sRORs did not differ significantly from 1.00 for any category of harm.
M AN U
When all harm-related endpoints were considered, nominally statistically significant relative differences between more developed and less developed countries were identified in 5% (6/128) of all meta-analyses (ROR>1 in two and ROR<1 in four) (Supplementary file 3). However, in 27% (35/128) of meta-analyses the relative differences in the summary ORs between more and less developed countries, based on the ROR point estimates, were larger than 2-fold. More
TE D
specifically, RORs≥ 2.0 were seen on 20 meta-analyses while ROR ≤ 0.50 in 15 meta-analyses (p=0.36) (Supplementary files 3 and 11).
EP
There was also large uncertainty left around the estimated RORs and in 92% (118/128) of meta-analyses, when the ROR 95% confidence intervals were also considered, one could not
AC C
exclude a 2-fold difference in harms in both directions (lower ROR CI≥0.5 and upper ROR CI≤2.0) (Supplementary file 11). Moreover, in 67% (86/128) of the meta-analyses, there were more than 10-fold differences between the upper and lower CIs of the RORs. More specifically this occurred in 67% of meta-analyses for severe adverse events; 75% of meta-analyses for discontinuations due to adverse events; 44% of meta-analyses for any adverse events; 40% of meta-analyses for organ system specific adverse events; and 69% of meta-analyses for individual adverse events. 14
ACCEPTED MANUSCRIPT
Results were similar when fixed effects were used to synthesize the sORs for the primary harm-related endpoints per country-setting within each meta-analysis. (Supplementary files 1315). In sensitivity analyses comparing only USA trials against trials from less developed
RI PT
countries 4/24 (17%) of meta-analyses for the primary endpoints showed nominally differences beyond chance for the comparison by the country setting (ROR>1 in one and ROR<1 in three) (Supplementary files 3, 16-18). Further sensitivity analyses excluding trials from the four Asian
SC
“tigers” also gave similar results for the primary endpoints and the combined primary-harms
M AN U
endpoint. RORs
Table 2 shows in more detail the 6 topics for which a nominally statistically significant difference was found between studies from less developed and more developed countries. For 2
TE D
topics, less developed countries reported less unfavorable results (RORs>1); while the opposite occurred in 4 topics (RORs<1). Studies from less developed countries had reported relatively less neonatal hematologic toxicities with zidovudine (vs. placebo) for prevention of HIV mother
EP
to child transmission; and relatively less diarrhea with third generation cephalosporins (vs. conventional antibiotics) for acute bacterial meningitis. While studies from less developed
AC C
countries had reported relatively more severe adverse events with the use of phosphodiesterase inhibitors (vs. placebo) for chronic obstructive pulmonary disease; relatively more adverse events of any type with stem cell therapy for acute myocardial infarction; relatively more gastrointestinal complications with preoperative chemotherapy (vs. surgery alone) for resected thoracic and esophageal cancer; and relatively more vomiting after the second dose of rotavirus vaccines (vs placebo).
15
ACCEPTED MANUSCRIPT
For the above 6 topics there were no statistically significant difference in the baseline risks of harms in the control groups (control rates) between trials from less and more developed
RI PT
countries; nor was there evidence of small study effects (Supplementary file 19). Discussion
Our study provided a bird’s eye view of the pattern of reporting of harms and harm-
SC
related endpoints in trials from the two country settings. We explored the relative differences in rates of harms across different country-settings using diverse approaches in order to obtain a
M AN U
more comprehensive appraisal. Our inferences need to be cautious given the large uncertainty in the relative incidence of harms for many of the examined topics. Due to the dearth of harm data in individual randomized trials but also in individual meta-analyses, it would not have been possible to identify significant differences between country-settings and only a bird’s eye view
TE D
and large scale evaluation across diverse topics was able to probe such country-setting specific harm differences. Overall, across all analyzed meta-analyses for all types of harm-related endpoints, we did not detect nominally significant discrepancies in the reported rates of harm-
EP
related endpoints between trials from less developed and more developed countries when results were summarized for each harm category. However, when the point estimates of the relative
AC C
differences of harms between more and less developed countries were considered, in 27% of the meta-analyses the relative differences were larger than 2-fold. Many meta-analyses had limited evidence and substantial uncertainty and when the uncertainty around the relative differences were also considered, differences in the relative odds of harms exceeding 2-fold in both directions could not be excluded in 92% of meta-analyses.
16
ACCEPTED MANUSCRIPT
We should acknowledge some study limitations. First, we did not screen the whole Cochrane Database for Systematic Reviews with harm-related endpoints. However, the 131 systematic reviews we screened was a group of meta-analyses that was highly likely to have
RI PT
included harm-related endpoints both from trials in more developed and less developed countries as they had already included trials from both country settings for mortality. By focusing also on mortality, our pool of these 131 screened systematic reviews was about clinically important
SC
medical conditions and interventions. Trials from more developed countries were the preponderance, as we had also observed in our assessments of mortality previously. [10]
M AN U
However, trials from less developed countries are expected to become more common in the future. Second, in the 128 individual meta-analyses, included in the 42 eligible systematic reviews, with comparative evidence on harm-related endpoints from the two country settings only a tenth had data on the primary harm endpoints; the majority had reported individual
TE D
specific adverse events. The clinical significance of these individual harms, without any additional information on their severity grading, is unclear.[28] This probably reflects poor reporting for harms and poor compliance with CONSORT standards of reporting, which is a
37].
EP
universal problem in randomized trials independently of the trial’s country setting [29-31] [32Third, we focused on trials performed exclusively in more-developed or less-developed
AC C
countries and did not analyze international multisite trials. It remains unknown whether the reported rates of harms in trials from other countries in economic and social transition and from study-sites from different country settings included in the same international multisite trials are systematically different. However, the vast majority of published reports of international studies included in the CDSR do not report results separately per country setting. In our database, less than 10% of the RCTs were international studies from mixed country settings. In Bayer A.G.’s
17
ACCEPTED MANUSCRIPT
ciprofloxacin patent extension studies the FDA noted an arthropathy rate of 21 % reported by investigators in the United States compared to none in a less developed country [9]. This Bayer/Quintiles database was further analyzed and divided into countries [38], and ultimately
conclusions are still pending.
RI PT
handed over to the US Department of Justice for a false claims investigation [39]; final Fraud must be ruled out whenever significant discrepancies
findings, rather than following a patient for adverse events.
SC
between investigator groups are found, as it is easy for an untrained investigator to report no
In summary, although we did not identify consistent over- or under-estimation of risks of
M AN U
harms between trials done in more versus less developed countries, substantial differences in the estimates of risks were relatively common and the potential for modestly large differences often could not be excluded with confidence. Collection and reporting of harms should be emphasized and standardized in all randomized trials, both from more developed and less developed
TE D
countries. More emphasis should be given to the standardized reporting of clinically important harms, such as severe adverse events or discontinuations due to adverse events. It is very important to have available the information required for the prompt identification of systematic
EP
differences in harms across different country settings. Differences may reflect genuine variability in the incidence of harms or differences in the reporting of harms. Reported rates of harms
AC C
may be influenced by the genetic profiles of populations, but may also depend upon ethical practices and monitoring systems in place in different countries. Disentangling these reasons would require that harms/harm-related endpoints are collected and reported transparently and in a harmonized way across randomized trials [28]. Country-setting discrepancies in the reported rates of harms should be routinely incorporated and explored in meta-analyses and systematic reviews. Moreover, regulatory agencies, such as the FDA or the EMEA, should enforce that such
18
ACCEPTED MANUSCRIPT
transparent harm-related information is provided per country and site for all randomized trials, including also multisite international trials. This may allow understanding how often and why some countries and sites have different results. Until such transparent comparative evidence on
RI PT
harms is available from trials in different country-settings, guideline developers and policy makers should cautiously extrapolate harm data from trials from one country setting to guide
AC C
EP
TE D
M AN U
SC
clinical decision making in very different country settings.
19
ACCEPTED MANUSCRIPT
Contributorship: DCI had the original idea and developed the study analysis plan; JPAI, XT, MA, OAP, JNW and YM critically commented on the study analysis plan; DCI, XT, MA performed data extraction; DCI and OAP performed the statistical analyses; DCI drafted the first
RI PT
manuscript draft and all authors critically reviewed it and approved the final version; DCI is the guarantor for this paper. Ethical approval: Not applicable.
SC
Transparency declaration: Dr. Contopoulos-Ioannidis (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that
as planned have been explained.
M AN U
no important aspects of the study have been omitted; and that any discrepancies from the study
Copyright/license for publication: The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, “a worldwide license to the Publishers and
TE D
its licensees in perpetuity, in all forms, formats and media (whether known now or created in the future), to: i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include within collections and
EP
create summaries, extracts and/or, abstracts of the Contribution, iii) create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the
AC C
inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) license any third party to do any or all of the above.”
20
ACCEPTED MANUSCRIPT
References 1. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials: race-, sex-, and age-based disparities. JAMA 2004;291(22):2720-6
RI PT
2. George M, Selvarajan S, S SK, et al. Globalization of clinical trials - where are we heading? Current Clinical Pharmacology 2013;8(2):115-23
3. Duke T, Fuller D. Randomised controlled trials in child health in developing countries: trends
SC
and lessons over 11 years. Archives of Disease in Childhood 2014;99(7):615-20.
4. Dandona L, Katoch VM, Dandona R. Research to achieve health care for all in India. Lancet
M AN U
2011;377(9771):1055-7
5. Drain PK, Robine M, Holmes KK, et al. Trail watch: global migration of clinical trials. Nature reviews. Drug Discovery 2014;13(3):166-7
6. Zhang D, Freemantle N, Cheng KK. Are randomized trials conducted in China or India
2011;64(1):90-5
TE D
biased? A comparative empirical analysis. Journal of Clinical Epidemiology
7. Hyder AA, Wali SA, Khan AN, et al. Ethical review of health research: a perspective from
EP
developing country researchers. Journal of Medical Ethics 2004;30(1):68-72 8. Zhang D, Yin P, Freemantle N, et al. An assessment of the quality of randomised controlled
AC C
trials conducted in China. Trials 2008;9:22
9. Division of special pathogen and immunologic drug products. Summary of clinical review of studies
submitted
in
response
to
a
pediatric
written
request.
http://www.fda.gov/ohrms/dockets/ac/05/briefing/20054152b1_03_03_Cipro%20medical.pdf Accessed 1/21/2015
21
ACCEPTED MANUSCRIPT
10. Panagiotou OA, Contopoulos-Ioannidis DG, Ioannidis JP. Comparative effect sizes in randomised trials from less developed and more developed countries: metaepidemiological assessment. BMJ 2013;346:f707
RI PT
11. Lorenzo C, Garrafa V, Solbakk JH, et al. Hidden risks associated with clinical trials in developing countries. Journal of Medical Ethics 2010;36(2):111-5
2002;21(19):2869-77
SC
12. Smith R. Publishing research from developing countries. Statistics in Medicine
13. Juni P, Holenstein F, Sterne J, et al. Direction and impact of language bias in meta-analyses trials:
empirical
study.
International
M AN U
of controlled
2002;31(1):115-23
Journal
of Epidemiology
14. Duijnhoven RG, Straus SM, Raine JM, et al. Number of patients studied prior to approval of new medicines: a database analysis. PLoS Medicine 2013;10(3):e1001407
TE D
15. Tsang R, Colley L, Lynd LD. Inadequate statistical power to detect clinically significant differences in adverse event rates in randomized controlled trials. Journal of Clinical Epidemiology 2009;62(6):609-16
gs-r,
EP
16. United Nations Statistics Division. Composition of macro geographical (continental) regions and
selected
economic
and
other
groupings.
AC C
2013.http://unstats.un.org/unsd/methods/m49/m49regin.htm.
17. Platonov P. Clinical trials in Russia and Eastern Europe: recruitment and quality. International Journal of Clinical Pharmacology and Therapeutics 2003;41(7):277-80
18. Lathyris D, Panagiotou OA, Baltogianni M, et al. Safety of medical interventions in children versus adults. Pediatrics 2014;133(3):e666-73
22
ACCEPTED MANUSCRIPT
19. Contopoulos-Ioannidis DG, Baltogianni MS, Ioannidis JP. Comparative effectiveness of medical
interventions
in
adults
versus
children.
The Journal
of Pediatrics
20. http://handbook.cochrane.org/ ChfsroiTCCVuM.
RI PT
2010;157(2):322-30
21. Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews. Annals of Internal Medicine 1997;127(9):820-6
SC
22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials 1986;7(3):177-88
M AN U
23. Ioannidis JP, Patsopoulos NA, Evangelou E. Uncertainty in heterogeneity estimates in metaanalyses. BMJ 2007;335(7626):914-6
24. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine 2002;21(11):1539-58
1950;21:607-11
TE D
25. Freeman MF TJ. Transformations related to the angular and the square root. Ann Math Statist
26. Harbord RM EM, Sterne JA. . A modified test for small-study effects in meta-analyses of
EP
controlled trials with binary endpoints. Statistics in Medicine 2006;25:3443-57. 27. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic
AC C
reviews and meta-analyses of studies that evaluate health care interventions: explanation
and elaboration. Annals of Internal Medicine 2009;151(4):W65-94
28. Ioannidis JP, Evans SJ, Gotzsche PC, et al. Better reporting of harms in randomized trials: an extension of the CONSORT statement. Annals of Internal Medicine 2004;141(10):781-8
23
ACCEPTED MANUSCRIPT
29. Papanikolaou PN, Ioannidis JP. Availability of large-scale evidence on specific harms from systematic reviews of randomized trials. The American Journal of Medicine 2004;117(8):582-9
medical areas. JAMA 2001;285(4):437-43
RI PT
30. Ioannidis JP, Lau J. Completeness of safety reporting in randomized trials: an evaluation of 7
31. Ioannidis JP. Adverse events in randomized trials: neglected, restricted, distorted, and
SC
silenced. Archives of Internal Medicine 2009;169(19):1737-9
32. Mukherjee SD, Coombes ME, Levine M, et al. A qualitative study evaluating causality
M AN U
attribution for serious adverse events during early phase oncology clinical trials. Investigational New Drugs 2011;29(5):1013-20 .
33. Wahab IA, Pratt NL, Kalisch LM, et al. The detection of adverse events in randomized clinical trials: can we really say new medicines are safe? Current Drug Safety
TE D
2013;8(2):104-13
34. Schroll JB, Maund E, Gotzsche PC. Challenges in coding adverse events in clinical trials: a systematic review. PloS One 2012;7(7):e41174.
EP
35. Huang HY, Andrews E, Jones J, et al. Pitfalls in meta-analyses on adverse events reported from clinical trials. Pharmacoepidemiology and Drug Safety 2011;20(10):1014-20
AC C
36. Haidich AB, Birtsou C, Dardavessis T, et al. The quality of safety reporting in trials is still suboptimal: survey of major general medical journals. Journal of Clinical Epidemiology
2011;64(2):124-35
37. de Vries TW, van Roon EN. Low quality of reporting adverse drug reactions in paediatric randomised controlled trials. Archives of Disease in Childhood 2010;95(12):1023-6
24
ACCEPTED MANUSCRIPT
38. Personal communication with Dr N. Balasubramanian from the Department of Statistics at Stanford University 39. United States of America, ex. rel. Juan N. Walterspiel MD, FAAP v. Bayer A.G., Quintiles False Claims Act action under 31 U.S.C. § 3729(a).
AC C
EP
TE D
M AN U
SC
https://www.pacer.gov/ Accessed 01/18/2015.
RI PT
Transnational Corp.
25
ACCEPTED MANUSCRIPT
Table 1: Harm-related endpoints
RI PT
Severe adverse events Title
Comparison
Description of Harmrelated endpoint
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant vs control
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet vs control acute
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent vs control
CD002130
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic
Platelet glycoprotein IIb/IIIa blockers during PCI vs control
Major extracranial haemorrhage during treatment period Major extracranial 6 (796) haemorrhage during treatment period Symptomatic (including 20 (6,244) fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding 20 (23,259) (all patients)
CD002309
TE D
M AN U
N of RCTs (N patients) from more developed countries 15 (2,506)
N of RCTs (N patients) from less developed countries 2 (344)
ROR (95% CIs)
1 (21,106)
0.67 (0.14-3.09)
2 (567)
1.97 (0.57-6.85)
3 (448)
2.39 (0.30-19.03)
2.48 (0.22-27.75)
EP
AC C
CD002147
SC
Review
Terlipressin vs vasopressin
Adverse events causing death
3 (204)
2 (101)
0.54 (0.02-19.03)
PDE4 inhibitors vs placebo
Non fatal serious adverse events
8 (2,965)
1 (1,018)
0.38 (0.18-0.83)
26
ACCEPTED MANUSCRIPT
CD003897b
CD008521a
CD008521b
IL2Ra vs placebo or no treatment RV1 vs placebo
Review
Title
CD000171a
Treatment of latent tuberculosis infection in
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.04)
Infection: CMV invasive
3 (881)
1 (118)
0.93 (0.12-7.26)
Infection: serious allcause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
Serious adverse events
7 (3,179)
18 (29,581)
0.88 (0.63-1.25)
3 (2,547)
7 (7,863)
0.81(0.45-1.46)
N of RCTs (N patients) from more developed countries 2 (838)
N of RCTs (N patients) from less developed countries 6 (4,695)
ROR (95% CIs)
RI PT
1 (29)
SC
Intraventricular haemorrhage (III-IV)
RV5 versus placebo All serious adverse events
AC C
Discontinuations due to adverse events
IL2Ra vs placebo or no treatment
5 (266)
M AN U
CD003897a
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interleukin 2 receptor antagonists for kidney transplant recipients Interleukin 2 receptor antagonists for kidney transplant recipients Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
Major infection
TE D
CD003481
Immunosuppressiv e agent plus steroids vs steroids alone Ibuprofen vs indomethacin
EP
CD002922
obstructive pulmonary disease Treatment for lupus nephritis
Comparison
Description of Harmrelated endpoint
Any TB drug vs placebo
Incidence of adverse events leading to
1.05 (0.04-26.24)
27
ACCEPTED MANUSCRIPT
CD002863
CD004386a
CD004386b
CD004773
Beta-blocker vs placebo/no betablocker Drug vs placebo/no intervention
Changed drugs due to maternal side effects
Side effects requiring discontinuation
Quinolone vs TMP- Side effects requiring SMZ discontinuation
Antifungal vs Placebo
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
3 (206)
2 (103)
0.22 (0.01-5.92)
7 (2,891)
1 (1,018)
0.61 (0.27-1.36)
10 (923)
1 (154)
1.22 (0.19-7.67)
16 (2,057)
1 (97)
2.11 (0.04-113.13)
6 (752)
1 (102)
4.03 (0.40-40.29)
2 (723)
1 (129)
0.74 (0.07-8.17)
RI PT
SC
PDE4 inhibitors vs placebo
M AN U
CD002309
TE D
CD002147
stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Adverse events causing withdrawal of treatment Withdrawals due to adverse effects
EP
CD000171c
Isoniazid vs isoniazid + rifampicin Isoniazid vs rifampicin + pyrazinimide Terlipressin versus vasopressin
AC C
CD000171b
HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Discontinuation secondary to adverse events
28
ACCEPTED MANUSCRIPT
CD008521b
Withdrawal due to adverse events
RV5 vs placebo
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Adverse events requiring discontinuation (end of follow-up) Adverse events requiring discontinuation (end of follow-up)
Description of Harmrelated endpoint
Any adverse event
8 (1,652)
3 (236)
0.61 (0.13-2.93)
5 (3,709)
4 (7,735)
1.00 (0.21-4.73)
6 (2,184)
15 (21,774)
1.92 (0.63-5.87)
N of RCTs (N patients) from more developed countries 8 (2,913)
N of RCTs (N patients) from less developed countries 1 (1,018)
ROR (95% CIs)
RI PT
Statins vs placebo/no treatment
SC
CD008521a
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use
M AN U
CD007784
Title
Comparison
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Interventions for preventing oral candidiasis for patients with cancer receiving treatment Antioxidant supplements for preventing
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
All drug types vs placebo/no treatment
Toxicity (adverse events 'probably due to drug')
2 (661)
1 (210)
1.47 (0.44-4.90)
Antioxidants (betacarotene) vs placebo
Adverse effects
3 (91,080)
1 (174)
0.88 (0.03-23.17)
CD004183
EP
AC C
CD003807
TE D
Review
1.21 (0.73-2.00)
29
ACCEPTED MANUSCRIPT
CD006536
CD008076
1.78 (0.03-92.13)
9 (929)
1 (102)
1.15 (0.36-3.63)
Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Adverse events
6(450)
1 (102)
0.75 (0.02-24.58)
RI PT
1 (97)
Antibiotic prophylaxis vs placebo/ no treatment
SC
Antifungal vs placebo
M AN U
CD004791
33 (5,002)
TE D
CD004773
Side effects
Quinolone vs TMP- Side effects SMZ
Stem cells vs no stem cells
Adverse effects
8 (669)
2 (213)
0.06 (0.02-0.20)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
Safety outcomes during follow-up
1 (3,202)
1 (787)
0.56 (0.31-1.00)
EP
CD004386b
Drug vs placebo/ no intervention
AC C
CD004386a
gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction Cilostazol versus aspirin for secondary prevention of vascular events after stroke of arterial origin
Combined Primary-Harms (Serious Adverse event, Discontinuations due to adverse event and Any adverse event)
30
ACCEPTED MANUSCRIPT
Anticoagulants for acute ischaemic stroke
CD000213
CD002130
N of RCTs (N patients) from more developed countries 15 (2,506)
N of RCTs (N patients) from less developed countries 2 (344)
ROR (95% CIs)
Major extracranial haemorrhage during treatment period
6 (796)
1 (21,106)
0.67 (0.14-3.09)
Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding
2 (838)
6 (4,695)
1.05 (0.04-26.24)
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
20 (5,622)
2 (567)
1.97 (0.57-6.85)
20 (23,259)
3 (448)
2.39 (0.30-19.03)
Major extracranial haemorrhage during treatment period
TE D
CD000171c
EP
CD000171b
AC C
CD000171a
Anticoagulant vs control in acute presumed ischaemic stroke Antiplatelet therapy for Antiplatelet vs acute ischaemic stroke control in acute presumed ischaemic stroke Treatment of latent Any TB drug vs tuberculosis infection in placebo HIV infected persons Treatment of latent Isoniazid vs tuberculosis infection in isoniazid + HIV infected persons rifampicin Treatment of latent Isoniazid vs tuberculosis infection in rifampicin + HIV infected persons pyrazinamide Thrombolysis for acute Any thrombolytic ischaemic stroke agent vs control
Description of Harmrelated endpoint
RI PT
CD000024
CD000029
Comparison
SC
Title
M AN U
Review
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST
Platelet glycoprotein IIb/IIIa blockers vs placebo/usual care (all patients)
2.48 (0.22-27.75)
31
ACCEPTED MANUSCRIPT
CD003481
CD003807
CD003897
CD004183
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interventions for preventing oral candidiasis for patients with cancer receiving treatment Interleukin 2 receptor antagonists for kidney transplant recipients Antioxidant supplements for preventing gastrointestinal cancers
1.21 (0.73-2.00)
10 (923)
1 (154)
1.22 (0.19-7.67)
5 (266)
1 (29)
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.03)
Toxicity (adverse events 'probably due to drug')
2 (661 )
1 (210)
1.47 (0.44-4.90)
IL2Ra vs placebo or no treatment
Infection: serious allcause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
Antioxidants vs placebo
Adverse effects - betacarotene
3 (91,080)
1 (174)
0.88 (0.03-23.17)
SC
Beta-blocker vs placebo/no betablocker
RI PT
1 (1,018)
Changed drugs due to maternal side effects
M AN U
CD002922
0.22 (0.01-5.88)
PDE4 inhibitors vs placebo
Adverse events causing 3 (204) withdrawal of treatment No of patients 8 (2,913) experiencing an adverse effect
Immunosuppressive Major infection agent plus steroids vs steroids alone Ibuprofen vs Intraventricular indomethacin haemorrhage (III-IV)
TE D
CD002863
2 (101)
Comparisons with placebo/no treatment for all drug types
EP
CD002309
Terlipressin vs vasopressin
AC C
CD002147
segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Treatment for lupus nephritis
32
ACCEPTED MANUSCRIPT
CD006536
CD007784
CD008076
Antifungal vs Placebo
Antibiotic prophylaxis vs placebo or no treatment
1 (97)
1.78 (0.03-92.13)
9 (929)
1 (102)
1.15 (0.36-3.63)
1 (295)
1 (129)
1.44 (0.60-3.46)
6 (450)
1 (102)
0.75 (0.02-24.58)
RI PT SC
Quinolone vs TMP- Side effects SMZ
Adverse events
M AN U
CD004791
33 (5,002)
Adverse events
TE D
CD004773
Drug vs placebo/ no Side effects intervention
Stem cells compared to no stem cells, Statins vs placebo/no treatment
Adverse effects
6 (617)
2 (213)
0.06 (0.02-0.18)
Withdrawal due to adverse events
8 (1,652)
3 (236)
0.61 (0.13-2.93)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
Other outcomes of safety during followup.
1 (3,202)
1 (787)
0.56 (0.31-1.00)
EP
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular after stroke of arterial
AC C
CD004386a
33
ACCEPTED MANUSCRIPT
RI PT
RV1 vs placebo
5 (3,708)
7 (3,179)
4 (7,7733)
1.00 (0.21-4.73)
18 (29,581)
0.88 (0.63-1.25)
TE D
M AN U
SC
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Adverse events requiring discontinuation (end of follow-up) Serious adverse events
EP
CD008521b
RV5 vs placebo
AC C
CD008521a
origin Vaccines for preventing rotavirus diarrhoea: vaccines in use
34
ACCEPTED MANUSCRIPT
Table 2: Individual meta-analyses with nominally significant relative differences between more developed and less developed
ROR (95% CIs)
sOR_MD (95% sOR_LD (95% CIs) CIs)
0.38 (0.18-0.83)
0.75 (0.55-1.03)
1.96 (0.96-3.98)
Stem cells compared to no stem cells,
Adverse effects
0.06 (0.02-0.20)
0.84 (0.48-1.45)
13.01 (4.90-34.57)
Any zidovudine vs placebo/no treatment
Neonatal haematological toxicity
2.77 (1.23-6.22)
2.09 (1.22-3.59)
0.76 (0.41-1.38)
Gastrointestinal complications
0.003 (0.001-0.06)
1.15 (0.34-3.81)
441.20 (24.93-7808)
Third generation cephalosporins versus conventional therapy
Diarrhoea
3.92 (1.35-11.34)
2.50 (1.54-4.05)
0.64 (0.25-1.64)
Rotarix vs placebo: after dose 2
Vomitingreactogenicity
0.66 (0.44-0.99)
0.68 (0.47-0.98)
1.03 (0.86-1.24)
M AN U
PDE4 inhibitors vs placebo
TE D
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease CD006536 Stem cell treatment for acute myocardial infarction CD000102 Interventions for reducing the risk of mother-to-child transmission of HIV infection CD001556c Preoperative chemotherapy for resectable thoracic esophageal cancer CD001832 Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis CD008521b Vaccines for preventing rotavirus diarrhoea: vaccines in use
Description of Harm-related endpoint Non fatal serious adverse events
SC
Comparison
Preoperative chemotherapy vs surgery alone
AC C
CD002309
Title
EP
Review
RI PT
countries
35
ACCEPTED MANUSCRIPT
Abbreviations: ROR: relative odds ratio; sOR_MD: summary odds ratio in more developed countries; sROR_LD: summary odds
AC C
EP
TE D
M AN U
SC
RI PT
ratio in less developed countries; CIs: confidence intervals; Footnote: ORs and RORs were calculated by random effects models
36
ACCEPTED MANUSCRIPT
Figure 1: Flow Chart
RI PT
Screened 131 Systematic Reviews
SC
Excluded 89 Systematic Reviews No adverse events (n=40) No pharmacologic interventions (n=30) No More developed and Less developed countries (n=18) No country information (n=1)
Discontinuations due to Adverse Events: 12 ma; 108 RCTs (70 from MD and 38 from LD countries)
TE D
Severe Adverse Events: 12 ma; 143 RCTs (99 from MD and 44 from LD countries)
M AN U
Included 42 eligible Systematic Reviews; 128 meta-analyses on Harm-related endpoints with 1186 study-level analyses from 474 independent RCTs (337 from MD and 137 from LD countries)a b
AC C
EP
Organ System Specific Adverse Events: 5 ma; 20 RCTs (11 from MD and 9 from LD countries)
Individual adverse events (n=60) 14 adverse events with ≥2 ma (43 ma) 46 adverse events with 1 ma (46 ma)
Any Adverse Event: 9 ma; 81 RCTs (71 from MD and 10 from LD countries)
All Discontinuations: 1 ma; 7 RCTs (6 from MD and 1 from LD countries)
89 ma, 767 RCTS corresponding to 311 independent RCTs (222 from MD and 89 from LD countries)
Footnotes: a
Each RCT could have contributed data to more than one study-level analyses, for the different harm-related endpoints.
b
We excluded from these 128 meta-analyses data from 47 RCTs that were from mixed-country settings (n=38); 37Eastern European countries (n=4) or did not report country (n=5).
*Conflict of Interest/Financial Disclosure
ACCEPTED MANUSCRIPT
Conflicts of Interest There are no conflicts of interest to declare for Contopoulos-Ioannidis D; Tseretopoulou X,
RI PT
Ancker M; Panagiotou OA; Maldonado Y and Ioannidis JP. Walerspiel JN worked as a consultant for US Bayer A.G. in 2002-3. “Whistleblower" on Bayer A.G. Currently does not hold any Bayer A.G or subsidiarie’s stock, nor does he give any lectures
AC C
EP
TE D
M AN U
SC
that are paid for directly or indirectly by Bayer A.G.
Supplemental Material (Online Only)
ACCEPTED MANUSCRIPT
SUPPLEMENTARY FILES Supplementary file 1: References of included Systematic Reviews
Cochrane Database Syst Rev. 2008;(4): CD000024
RI PT
1. Sandercock PAG, Counsell C, Kamal AK. Anticoagulants for acute ischaemic stroke.
2. Sandercock PAG, Counsell C, Gubitz GJ, Tseng MC. Antiplatelet therapy for acute ischaemic stroke. Cochrane Database Syst Rev. 2008;(3):CD000029.
SC
3. Brocklehurst P. Interventions for reducing the risk of mother-to-child transmission of HIV infection. Cochrane Database Syst Rev. 2002; (1): CD000102
M AN U
4. Akolo C, Adetifa I, Shepperd S, Volmink J. Treatment of latent tuberculosis infection in HIV infected persons. Cochrane Database Syst Rev. 2010;(1):CD000171 5.
Wardlaw JM, Murray V, Berge E, del Zoppo GJ. Thrombolysis for acute ischaemic stroke. Cochrane Database Syst Rev. 2009;(4): CD000213
TE D
6. Mahomed K, Bhutta ZA, Middleton P. Zinc supplementation for improving pregnancy and infant outcome. Cochrane Database Syst Rev. 2007;(2):CD000230 7. Dorhout Mees S, Rinkel GJE, Feigin VL, Algra A, van den Bergh WM, Vermeulen M,
EP
van Gijn J. Calcium antagonists for aneurysmal subarachnoid haemorrhage. Cochrane Database Syst Rev. 2007;(3): CD000277
AC C
8. Handoll HHG, Farrar MJ, McBirnie J, Tytherleigh-Strong GM, Milne AA, Gillespie WJ. Heparin, low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures.
Cochrane
Database Syst Rev. 2002;(4): CD000305.
9. Ohlsson A, Lacy JB. Intravenous immunoglobulin for preventing infection in preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2013;(7): CD000361
ACCEPTED MANUSCRIPT
10. Wong PF, Chong LY, Mikhailidis DP, Robless P, Stansby G. Antiplatelet agents for intermittent claudication. Cochrane Database Syst Rev. 2011;(11): CD001272. 11. Vogt K, Fenlon D, Rhodes S, Malthaner R. Preoperative chemotherapy for resectable
RI PT
thoracic esophageal cancer . Cochrane Database Syst Rev. 2006;(3): CD001556
12. Prasad K, Kumar A, Singhal T, Gupta PK. Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis. Cochrane Database Syst
SC
Rev. 2007;(4): CD001832
13. Bosch X, Marrugat J, Sanchis J. Platelet glycoprotein IIb/IIIa blockers during
M AN U
percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes. Cochrane Database Syst Rev. 2010;(9): : CD002130
14. Ioannou GN, Doust J, Rockey DC. Terlipressin for acute esophageal variceal
TE D
hemorrhage. Cochrane Database Syst Rev.2003;(1): CD002147 15. Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, Kupfer Y. Corticosteroids for treating severe sepsis and septic shock. Cochrane Database
EP
CD002243
Syst Rev.2004;(1):
16. Chong J, Poole P, Leung B, Black PN. Phosphodiesterase 4 inhibitors for chronic
AC C
obstructive pulmonary disease. Cochrane Database Syst Rev.2011;(5): CD002309
17. Magee L, Duley L. Oral beta-blockers for mild to moderate hypertension during pregnancy. Cochrane Database Syst Rev. 2003;(3): CD002863
18. Flanc RS, Roberts MA, Strippoli GFM, Chadban SJ, Kerr PG, Atkins RC. Treatment for lupus nephritis. Cochrane Database Syst Rev.2004;(1):CD002922
ACCEPTED MANUSCRIPT
19. Ohlsson A, Walia R, Shah SS. Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants. Cochrane Database
Syst Rev. 2010;(4):
CD003481
RI PT
20. Clarkson JE, Worthington HV, Eden TOB. Interventions for preventing oral candidiasis for patients with cancer receiving treatment. Cochrane Database Syst Rev. 2007;(1): CD003807
SC
21. Webster AC, Ruster LP, McGee R, Matheson SL, Higgins GY, Willis NS, Chapman JR, Craig JC. Interleukin 2 receptor antagonists for kidney transplant recipients. Cochrane
M AN U
Database Syst Rev. 2010;(1): CD003897
22. 22. Greb A, Bohlius J, Schiefer D, Schwarzer G, Schulz H, Engert A. High-dose chemotherapy with autologous stem cell transplantation in the first line treatment of aggressive Non-Hodgkin Lymphoma (NHL) in adults. Cochrane Database
TE D
Rev.2008;(1): CD004024
Syst
23. Rumbold A, Crowther CA. Vitamin E supplementation in pregnancy. Cochrane Database Syst Rev. 2005;(2): CD004069
EP
24. Bjelakovic G, Nikolova D, Simonetti RG, Gluud C. Antioxidant supplements for preventing gastrointestinal cancers. Cochrane Database Syst Rev. 2008;(3): CD004183
AC C
25. Ohlsson A, Shah SS. Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2006;(1): CD004213
26. Rumbold A, Duley L, Crowther CA, Haslam RR. Antioxidants for preventing preeclampsia. Cochrane Database Syst Rev.2008;(1):CD004227 27. Hunt R, Hey E. Ethamsylate for the prevention of morbidity and mortality in preterm or very low birth weight infants. Cochrane Database Syst Rev.2010;(1): CD004343
ACCEPTED MANUSCRIPT
28. Gafter-Gvili A, Fraser A, Paul M, Vidal L, Lawrie TA, van de Wetering MD, Kremer LCM,Leibovici L. Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy. Cochrane Database Syst Rev.2012;(1): CD004386
RI PT
29. Roberts D, Dalziel SR. Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth . Cochrane Database Syst Rev.2006;(3): CD004454 30. Duley L, Henderson-Smart DJ, Meher S, King JF. Antiplatelet agents for preventing
SC
pre-eclampsia and its complications. Cochrane Database Syst Rev. 2007;(2): CD004659 31. Chang LW, Phipps WT, Kennedy GE, Rutherford G. Antifungal interventions for the
Rev. 2005;(3): CD004773
M AN U
primary prevention of cryptococcal disease in adults with HIV. Cochrane Database Syst
32. Cohen MJ, Sahar T, Benenson S, Elinav E, Brezis M, Soares-Weiser K. Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites,
TE D
without gastro-intestinal bleeding. Cochrane Database Syst Rev.2009;(2): CD004791 33. Ohlsson A, Aher SM. Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2014;(4):
EP
CD004863
34. Aher SM, Ohlsson A. Late erythropoietin for preventing red blood cell transfusion in
AC C
preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2014;(4): CD004868
35. AlFaleh K, Anabrees J, Bassler D, Al-Kharfi T. Probiotics for prevention of necrotizing enterocolitis in preterm infants. Cochrane Database Syst Rev. 2011;(3): CD005496
ACCEPTED MANUSCRIPT
36. Raanani P, Gafter-Gvili A, Paul M, Ben-Bassat I, Leibovici L, Shpilberg O. Immunoglobulin prophylaxis in hematological malignancies and hematopoietic stem cell transplantation. Cochrane Database Syst Rev. 2008;(4): CD006501
RI PT
37. Clifford DM, Fisher SA, Brunskill SJ, Doree C, Mathur A, Watt S, M artin-Rendon E. Stem cell treatment for acute myocardial infarction. Cochrane Database Syst Rev. 2012;(2): CD006536
SC
38. Whitfield K, Rambaldi A, Wetterslev J, Gluud C. Pentoxifylline for alcoholic hepatitis. Cochrane Database Syst Rev. 2009; (4): CD007339
M AN U
39. Navaneethan SD, Pansini F, Perkovic V, Manno C, Pellegrini F, Johnson DW, Craig JC,Strippoli GFM. HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis. Cochrane Database Syst Rev. 2009;(2): CD007784 40. Kamal AK, Naqvi I, Husain MR, Khealani BA. Cilostazol versus aspirin for secondary Syst
TE D
prevention of vascular events after stroke of arterial origin. Cochrane Database Rev. 2011;(1): CD008076
41. Gurusamy KS, Koti R, Fusai G, Davidson BR. Somatostatin analogues for pancreatic
EP
surgery. Cochrane Database Syst Rev. 2012;(3): CD008370 42. Soares-Weiser K, MacLehose H, Bergman H, Ben-Aharon I, Nagpal S, Goldberg E,
AC C
Pitan F, Cunliffe N. Vaccines for preventing rotavirus diarrhoea: vaccines in use. Cochrane Database Syst Rev. 2012;(11): CD008521
ACCEPTED MANUSCRIPT
RI PT
Supplementary file 2: Randomized trials from more developed and less developed countries
337 RCTs from More Developed Countries
137 RCTs from Less Developed Countries Turkey
Belgium Canada
USA
Africa Argentina
Tunisia
SC
Austria
Thailand
Denmark
M AN U
Australia
Vietnam Zambia Uganda
Europe
South Korea
Finland
Asia
Bangladesh Barbados
Taiwan
Brazil
Singapore
France
Norway
New Zealand Netherlands
Japan
EP
Scandinavia
AC C
Spain Scotland
Panama Mexico
Israel
TE D
Italy
Sweden
Portugal
Philippines
Greece
Switzerland
China
Qatar
Germany UK
Chile S. Africa
Jordan
Costa Rica
Mali
Latin America Korea
Kenya Jamaica
Dom. Republic Egypt Haiti Iran Indonesia India Hong Kong
French caribbean Ghana
ACCEPTED MANUSCRIPT
RI PT
Supplementary file 3: List of meta-analyses per category of Harm; with ROR of harm per meta-analysis between trials from more developed vs less developed countries [and 95% CIs thereof] ; summary OR of harm in trials from more developed countries per meta-analysis [and 95% CIs thereof] and summary OR of harm in trials in less developed countries per meta-analysis [and 95% CIs thereof])
Severe adverse events
M AN U
SC
Abbreviations: MD: more developed countries, LD: less developed countries, ROR: relative odds ratio of harm ; LCI_ROR: lower 95% confidence interval [CI] of ROR; UCI_ROR: upper 95% CI of ROR; sOR_MD: summary odds ration from trials in MD countries, sOR_LD: summary odds ratio from trials in LD countries; LCI_OR_MD: lower 95% CI of summary OR from trials in MD countries; LCI_OR_LD: lower 95% CI of summary OR from trials in LD countries; UCI_OR_MD: upper 95% CI of summary OR from trials in MD countries; UCI_OR_LD: upper 95% CI of summary OR from trials in LD countries
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
2.48
0.22
27.75
1.57
0.71
3.47
0.63
0.06
6.18
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
0.67
0.14
3.09
0.99
0.22
4.43
1.49
1.06
2.08
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent versus control
1.97
0.57
6.85
3.63
2.52
5.24
1.84
0.56
6.06
CD002130
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of nonST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage
During PCI (all patients)
Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding
2.39
0.30
19.03
1.40
1.14
1.73
0.59
0.07
4.62
Adverse events causing death
0.54
0.02
19.03
0.60
0.07
5.42
1.11
0.07
18.24
EP
AC C
CD002147
TE D
title
review
Terlipressin versus vasopressin
ACCEPTED MANUSCRIPT
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Non fatal serious adverse events
0.38
0.18
CD002922
Treatment for lupus nephritis
Immunosuppressive agent plus steroids versus steroids alone
Major infection
0.33
0.02
4.53
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Intraventricular haemorrhage (III-IV)
1.75
0.51
6.03
CD003897a
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Infection: CMV invasive
0.93
0.12
CD003897b
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Infection: serious allcause total
1.08
CD008521a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Serious adverse events
CD008521b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV5 versus placebo
All serious adverse events
0.55
1.03
1.96
0.96
3.98
0.77
0.37
1.59
2.33
0.19
29.04
1.52
0.70
3.28
0.87
0.33
2.29
7.26
0.97
0.59
1.58
1.04
0.14
7.61
0.39
2.98
1.00
0.62
1.60
0.92
0.38
2.26
0.88
0.63
1.25
0.82
0.62
1.08
0.93
0.76
1.13
0.81
0.45
1.46
0.75
0.47
1.20
0.92
0.65
1.31
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
Incidence of adverse events leading to stopping treatment
1.06
0.04
26.24
3.73
0.16
85.90
3.53
1.76
7.09
SC
M AN U
TE D
Discontinuations due to adverse events
0.83
0.75
RI PT
CD002309
title
comparison
CD000171a
Treatment of latent tuberculosis infection in HIV infected persons
Any TB drug vs placebo
CD000171b
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs isoniazid + rifampicin
Incidence of adverse events leading to stopping treatment
4.59
0.70
29.98
1.08
0.27
4.33
0.24
0.07
0.83
CD000171c
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs rifampicin + pyrazinimide
Incidence of adverse events leading to stopping treatment
0.53
0.19
1.52
0.46
0.23
0.93
0.86
0.40
1.85
AC C
EP
review
ACCEPTED MANUSCRIPT
Terlipressin for acute esophageal variceal hemorrhage
Terlipressin versus vasopressin
Adverse events causing withdrawal of treatment
0.22
0.01
0.05
1.36
1.11
0.07
18.19
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Withdrawals due to adverse effects
0.61
0.27
1.36
1.96
1.50
2.58
3.23
1.51
6.91
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
Changed drugs due to maternal side effects
1.22
0.19
7.67
1.61
0.58
4.43
1.32
0.28
6.08
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/no intervention
Side effects requiring discontinuation
2.11
0.04
113.13
2.25
1.29
3.93
1.06
0.02
54.76
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Quinolone vs. TMP-SMZ
Side effects requiring discontinuation
4.03
0.40
40.29
0.39
0.16
1.00
0.10
0.01
0.80
CD004773
Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Antifungal vs. Placebo
Discontinuation secondary to adverse events
0.74
0.07
8.17
2.40
1.17
4.95
3.25
0.33
32.10
CD007784
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis
Statins versus placebo/no treatment
Withdrawal due to adverse events
0.61
0.13
2.93
1.32
0.72
2.42
2.17
0.51
9.27
CD008521a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Adverse events requiring discontinuation (end of follow-up) Adverse events requiring discontinuation (end of follow-up)
1.92
0.63
5.87
1.48
0.56
3.96
0.77
0.45
1.32
CD008521b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV5 versus placebo
1.00
0.21
4.73
0.64
0.16
2.45
0.63
0.29
1.36
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
1.21
0.73
2.00
1.51
1.00
2.28
1.26
0.93
1.69
0.25
SC
M AN U
EP
TE D
5.92
RI PT
CD002147
AC C
Any adverse event
review
title
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
ACCEPTED MANUSCRIPT
Interventions for preventing oral candidiasis for patients with cancer receiving treatment
Comparisons with placebo/no treatment for all drug types
Toxicity (adverse events 'probably due to drug')
1.47
0.44
CD004183
Antioxidant supplements for preventing gastrointestinal cancers
Antioxidants versus placebo
Adverse effects beta-carotene
0.88
0.03
23.17
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/ no intervention
Side effects
1.78
0.03
92.13
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Quinolone vs. TMP-SMZ
Side effects
1.15
0.36
CD004773
Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Antifungal vs. Placebo
Adverse events
1.44
CD004791
Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastro-intestinal bleeding Stem cell treatment for acute myocardial infarction
Antibiotic prophylaxis versus placebo or no treatment
Adverse events
Stem cells compared to no stem cells,
Cilostazol versus aspirin for secondary prevention of vascular
Cilostazol versus aspirin in patients with ischaemic stroke or TIA,
review
title
CD000102a
Interventions for reducing the risk of mother-to-child transmission of HIV infection
CD000102b
Interventions for reducing the risk of mother-to-child transmission of HIV infection
0.83
1.95
0.87
0.28
2.67
2.07
0.65
6.60
2.34
0.11
49.56
1.89
1.47
2.44
1.06
0.02
54.76
3.63
0.48
0.24
0.95
0.41
0.17
1.04
0.60
3.46
1.89
1.13
3.16
1.31
0.64
2.68
0.75
0.02
24.58
2.30
0.61
8.71
3.06
0.12
76.95
Adverse effects
0.06
0.02
0.20
0.84
0.48
1.45
13.01
4.90
34.57
Other outcomes of safety during followup.
0.56
0.31
1.00
1.44
1.19
1.74
2.57
1.48
4.46
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
comparison
SC
M AN U
TE D
Organ system specific adverse events
EP
CD008076
AC C
CD006536
4.90
1.27
RI PT
CD003807
description of harm
Any zidovudine vs placebo/no treatment
Neonatal haematological toxicity
2.77
1.23
6.22
2.09
1.22
3.59
0.76
0.41
1.38
Any zidovudine vs placebo/no treatment
Maternal haematological toxicity
0.74
0.23
2.36
1.06
0.53
2.11
1.43
0.56
3.64
ACCEPTED MANUSCRIPT
Preoperative chemotherapy for resectable thoracic esophageal cancer
Preoperative chemotherapy vs surgery alone
Pulmonary complications
1.94
0.83
4.54
CD001556b
Preoperative chemotherapy for resectable thoracic esophageal cancer
Preoperative chemotherapy vs surgery alone
Cardiac complications
0.74
0.27
2.03
CD001556c
Preoperative chemotherapy for resectable thoracic esophageal cancer
Preoperative chemotherapy vs surgery alone
Gastrointestinal complications
0.00
0.00
0.06
1.14
0.67
1.96
0.59
0.31
1.14
0.89
0.44
1.80
1.19
0.59
2.41
1.15
0.34
3.81
441.20
24.93
7808.00
Individual adverse events (Group 1; ≥2 ma/adverse event)
M AN U
SC
RI PT
CD001556a
title
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD004069
Vitamin E supplementation in pregnancy
Any vitamin E supplementation
Bleeding (pregnancy)
1.08
0.06
18.45
0.33
0.03
3.22
0.31
0.06
1.67
CD004227
Antioxidants for preventing pre-eclampsia
Any antioxidants versus control or placebo
Bleeding (pregnancy)
0.13
0.02
1.16
0.99
0.59
1.63
7.45
0.90
61.73
CD001832
Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis
Third generation cephalosporins versus conventional therapy
Diarrhoea
3.92
1.35
11.34
2.50
1.54
4.05
0.64
0.25
1.64
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Diarrhoea
0.77
0.27
2.20
1.99
0.79
4.99
2.58
1.56
4.25
CD008521_a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 1
Diarrhoeareactogenicity
1.33
0.90
1.97
1.22
0.90
1.65
0.92
0.71
1.19
CD008521_b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 2
Diarrhoeareactogenicity
1.10
0.61
1.97
0.99
0.58
1.67
0.90
0.70
1.15
CD008521_c
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: at the end of follow up
Diarrhoeareactogenicity
1.20
0.84
1.72
1.05
0.79
1.41
0.88
0.71
1.08
AC C
EP
TE D
review
ACCEPTED MANUSCRIPT
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/ no intervention
Drug resistant infection
0.81
0.17
1.05
2.15
1.87
0.42
8.31
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Quinolone vs. TMP-SMZ
Drug resistant infection
4.89
0.48
50.17
0.48
0.18
1.29
0.10
0.01
0.80
CD001272
Antiplatelet agents for intermittent claudication
Antiplatelet agent versus placebo
Dyspepsia
0.82
0.22
3.10
2.78
0.93
8.26
3.40
1.58
7.34
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Dyspepsia
0.59
0.19
1.86
3.12
1.92
5.08
5.27
1.87
14.84
CD008521_a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 1
Fever- reactogenicity
1.06
0.68
1.66
1.22
0.81
1.84
1.15
0.96
1.38
CD008521_b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 2
Fever-reactogenicity
0.89
0.69
1.15
0.92
0.74
1.13
1.03
0.89
1.20
CD008521_c
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: at the end of follow up
Fever- reactogenicity
0.87
0.69
1.10
0.89
0.75
1.05
1.02
0.87
1.19
CD008076
Cilostazol versus aspirin for secondary prevention of vascular
Analysis 1.9. Comparison 1 Cilostazol versus aspirin in patients with ischaemic stroke or TIA, Ibuprofen vs placebo or none
GI bleeding
0.38
0.11
1.32
0.38
0.17
0.86
1.00
0.39
2.54
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
GI bleeding
1.10
0.03
35.38
3.08
0.12
77.80
2.80
0.78
9.99
CD002243
Corticosteroids for treating severe sepsis and septic shock
Steroids versus control
GI bleeding
1.07
0.06
18.24
1.10
0.74
1.64
1.03
0.06
17.20
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
Hypertension
1.81
0.08
43.14
1.09
0.39
3.08
0.60
0.03
11.99
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
Hypertension
3.02
0.11
86.62
1.49
0.23
9.54
0.49
0.03
8.10
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
IVH
0.73
0.10
5.35
0.71
0.14
3.66
0.97
0.31
2.98
SC
M AN U
TE D
EP
AC C
3.73
1.51
RI PT
CD004386a
ACCEPTED MANUSCRIPT
Intravenous immunoglobulin for preventing infection in preterm and/or low birth weight infants
IVIG vs placebo or no treatment
IVH
0.84
0.22
0.77
1.32
1.20
0.32
4.45
CD004659
Antiplatelet agents for preventing preeclampsia and its complications
IVH
0.87
0.12
6.42
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Antiplatelet agents versus placebo/no antiplatelet for primary prevention (subgrouped by maternal risk) Erythropoietin vs. placebo or no treatment
0.90
0.64
1.26
1.03
0.14
7.30
IVH
1.44
0.39
5.31
1.00
0.64
1.55
0.69
0.20
2.37
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
IVH
0.79
45.79
0.79
0.39
1.61
1.00
0.02
54.47
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet versus control in acute presumed ischaemic stroke
Intracranial hemorrhage
1.37
0.44
4.27
1.70
0.56
5.11
1.24
0.94
1.63
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant versus control in acute presumed ischaemic stroke
Intracranial hemorrhage
7.28
0.53
99.58
2.47
1.14
5.34
0.34
0.03
4.12
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs placebo or none
NEC
0.69
0.12
3.92
1.00
0.32
3.13
1.44
0.39
5.27
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
NEC
0.58
0.21
1.61
0.93
0.58
1.51
1.60
0.65
3.95
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
NEC
0.85
0.30
2.42
0.44
0.24
0.82
0.52
0.22
1.22
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
NEC
1.69
0.63
4.51
0.81
0.40
1.66
0.48
0.24
0.94
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
NEC
1.74
0.12
24.72
0.88
0.38
2.06
0.51
0.04
6.25
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
Neutropenia
1.10
0.06
18.82
0.81
0.48
1.37
0.74
0.05
11.98
SC 0.01
M AN U
TE D
EP
AC C
3.20
1.01
RI PT
CD000361
ACCEPTED MANUSCRIPT
Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis
Third generation cephalosporins versus conventional therapy
Neutropenia
1.43
0.24
0.18
2.83
0.50
0.16
1.57
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
ROP
1.51
0.85
2.69
1.42
0.99
2.05
0.94
0.60
1.46
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
ROP
1.33
0.31
5.66
1.00
0.30
3.35
0.75
0.34
1.66
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs placebo or none
ROP
0.75
0.10
5.61
1.07
0.58
1.98
1.42
0.21
9.55
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Sepsis
1.59
0.49
5.20
1.55
0.63
3.79
0.97
0.45
2.10
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
Sepsis
2.19
0.19
24.96
0.69
0.32
1.49
0.32
0.03
3.18
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
Sepsis
1.05
0.51
2.18
0.88
0.62
1.24
0.83
0.44
1.59
CD005496
Probiotics for prevention of necrotizing enterocolitis in preterm infants
Probiotics vs. control
Sepsis
0.98
0.42
2.32
0.84
0.49
1.46
0.86
0.44
1.66
CD008521
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 1
Vomitingreactogenicity
0.99
0.74
1.34
1.07
0.83
1.39
1.08
0.93
1.26
CD008521_a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 2
Vomitingreactogenicity
0.66
0.44
0.99
0.68
0.47
0.98
1.03
0.86
1.24
CD008521_b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: ath the end of follow up
Vomitingreactogenicity
0.88
0.63
1.24
0.88
0.66
1.19
1.00
0.86
1.16
SC
M AN U
TE D
EP
AC C
8.49
0.72
RI PT
CD001832
ACCEPTED MANUSCRIPT
RI PT
Individual adverse events (Group 2; 1 ma per adverse event)
title
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Abdominal pain
1.23
0.48
3.14
2.28
1.25
4.18
1.85
0.91
3.78
CD001556
Preoperative chemotherapy for resectable thoracic esophageal cancer
Treatment morbidity and mortality
Anastomotic leaks
0.25
0.02
2.86
0.89
0.54
1.45
3.51
0.33
37.79
CD000230
Zinc supplementation for improving pregnancy and infant outcome
Zinc supplementation versus no zinc (with or without placebo)
Any maternal infection
1.62
0.68
3.89
1.26
0.76
2.10
0.78
0.38
1.58
CD008076
Cilostazol versus aspirin for secondary prevention of vascular
Analysis 1.8. Comparison 1 Cilostazol versus aspirin in patients with ischaemic stroke or TIA,
Bleeding
1.29
0.56
2.94
0.75
0.60
0.95
0.59
0.26
1.30
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
CLD
1.95
0.72
5.26
1.69
0.79
3.61
0.87
0.46
1.66
CD000277
Calcium antagonists for aneurysmal subarachnoid haemorrhage
Calcium antagonists
Cerebral infarction on CT/MR
3.56
0.92
13.77
0.65
0.52
0.81
0.18
0.05
0.69
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Chorioamnionitis
0.47
0.21
1.10
0.81
0.56
1.18
1.71
0.81
3.61
CD000277
Calcium antagonists for aneurysmal subarachnoid haemorrhage
Calcium antagonists
1.58
0.46
5.46
0.63
0.36
1.09
0.40
0.13
1.21
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Clinical signs of secondary ischaemia: studies with magnesium in addition of nimodipine Decreased urine output
0.52
0.10
2.68
0.22
0.07
0.70
0.43
0.13
1.40
CD008370
Somatostatin analogues for pancreatic surgery
Somatostatin analogues versus none
Delayed gastric emptying
0.55
0.15
2.04
0.65
0.32
1.31
1.17
0.39
3.49
M AN U
TE D
EP AC C
SC
review
ACCEPTED MANUSCRIPT
Statins versus placebo/no treatment
Elevated LFTs
1.16
0.05
Corticosteroids versus placebo or no treatment
Fever in women after trial entry requiring the use of antibiotics
1.28
0.24
6.65
CD004386
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
quinolone vs. TMP-SMZ
Fungal infection
1.20
0.09
15.41
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
GI Bleeding (LBWI)
3.14
CD002243
Corticosteroids for treating severe sepsis and septic shock
Steroids versus control
Hyperglycaemia
1.67
CD004343
Ethamsylate for the prevention of morbidity and mortality in preterm or very low birth weight infants
Ethamsylate versus placebo
Hypotension
CD001556
Preoperative chemotherapy for resectable thoracic esophageal cancer
Treatment morbidity and mortality
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
CD003481
1.26
0.30
5.30
1.08
0.07
17.94
0.95
0.23
3.95
0.75
0.32
1.73
0.61
0.29
1.32
0.51
0.04
5.81
89.66
3.27
0.13
84.37
1.04
0.46
2.35
0.03
88.94
1.67
1.26
2.20
1.00
0.02
52.85
1.31
0.36
4.81
1.03
0.51
2.06
0.78
0.26
2.35
Infection
1.13
0.30
4.24
0.65
0.38
1.09
0.57
0.17
1.92
IL2Ra versus placebo or no treatment
Infection: CMV all
0.83
0.23
3.07
0.84
0.63
1.13
1.01
0.28
3.62
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Intestinal perforation
0.28
0.01
7.52
0.19
0.01
4.21
0.67
0.23
1.99
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Intrapartum fever in woman requiring the use of antibiotics
0.12
0.00
3.23
0.24
0.03
2.22
1.98
0.18
22.18
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
Low APGAR scores
0.46
0.06
3.38
0.69
0.30
1.58
1.50
0.24
9.24
CD000102
Interventions for reducing the risk of mother-to-child transmission of HIV infection
Any zidovudine vs placebo/no treatment
Low birth weight
0.65
0.29
1.48
0.73
0.43
1.25
1.12
0.60
2.08
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Malignancy
1.34
0.21
8.70
0.68
0.38
1.20
0.51
0.09
3.02
AC C
EP
TE D
0.11
M AN U
CD004454
27.26
RI PT
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
SC
CD007784
ACCEPTED MANUSCRIPT
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Nausea
0.68
0.20
3.14
6.95
6.86
2.10
22.35
CD004659
Antiplatelet agents for preventing pre-eclampsia and its complications
Neonatal bleeding
0.82
0.39
1.70
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
Antiplatelet agents versus placebo/no antiplatelet for primary prevention (subgrouped by maternal risk) IL2Ra versus placebo or no treatment
1.07
0.76
1.51
1.32
0.69
2.52
PTLD/lymphoma
0.40
0.02
7.97
0.40
0.15
1.09
1.02
0.06
17.22
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
PVL
1.09
5.90
1.10
0.35
3.42
1.01
0.29
3.51
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Post-transplant diabetes mellitus
9.14
0.73
114.64
5.17
0.57
47.00
0.57
0.16
1.95
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Postnatal fever in woman
1.25
0.54
2.87
0.99
0.60
1.66
0.80
0.41
1.53
CD000230
Zinc supplementation for improving pregnancy and infant outcome
Zinc supplementation versus no zinc (with or without placebo)
Postpartum haemorrhage
2.93
0.60
14.44
1.62
0.86
3.06
0.55
0.13
2.39
CD000102
Interventions for reducing the risk of mother-to-child transmission of HIV infection
Any zidovudine vs placebo/no treatment
Premature delivery
2.75
0.83
9.10
1.27
0.59
2.71
0.46
0.18
1.16
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Proven Infection in NICU
1.23
0.48
3.16
0.81
0.43
1.53
0.66
0.33
1.31
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Puerperal sepsis
1.89
0.52
6.86
1.98
1.09
3.59
1.04
0.33
3.27
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs placebo or none
Pulmonary hypertension
3.94
0.04
401.58
3.59
0.33
38.70
0.91
0.02
48.05
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
RDS
0.63
0.10
4.04
0.23
0.06
0.93
0.36
0.11
1.21
CD001832
Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis
Third generation cephalosporins versus conventional therapy
Rash
0.22
0.01
3.54
0.29
0.04
2.24
1.31
0.20
8.48
SC 0.20
M AN U
TE D
EP
AC C
2.37
4.67
RI PT
CD002309
ACCEPTED MANUSCRIPT
Statins versus placebo/no treatment
Rhabdomyolysis
0.77
0.05
Late initiation of EPO (8-28 days) vs. placebo or no intervention
SIDS
0.61
0.03
12.78
Calcium antagonists
Secondary ischaemia
2.06
0.58
7.36
CD002243
Corticosteroids for treating severe sepsis and septic shock
Steroids versus control
Superinfections
3.10
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Systemic infection in the first 48 hours of life
0.83
CD004024
High-dose chemotherapy with autologous stem cell transplantation in the first line treatment of aggressive NonHodgkin Lymphoma (NHL) in adults Heparin low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures Immunoglobulin prophylaxis in hematological malignancies and hematopoietic stem cell transplantation Pentoxifylline for alcoholic hepatitis
High Dose Chemotherapy vs Control
Toxicities (secondary neoplasms)
Any heparin vs Control/Placebo
Polyvalent immunoglobulins vs. placebo / no intervention - HSCT
Heparin low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures
Any heparin vs Control/Placebo
CD000305
CD006501
CD007339
CD000305
0.27
6.08
1.65
0.19
14.16
0.83
0.15
4.71
1.38
0.11
16.97
0.53
0.44
0.64
0.26
0.07
0.90
31.89
1.01
0.75
1.35
0.33
0.03
3.28
0.24
2.84
0.47
0.21
1.09
0.57
0.23
1.40
0.22
0.01
6.89
0.76
0.25
2.28
3.40
0.13
87.73
Transfusion
0.99
0.02
55.27
0.82
0.48
1.42
0.83
0.02
44.40
VOD
0.76
0.02
29.17
2.65
0.66
10.56
3.46
0.12
100.52
Variceal bleeding
0.85
0.03
28.17
2.13
0.19
24.20
2.50
0.20
31.00
Wound haematoma
3.62
0.11
122.44
2.09
0.18
24.62
0.58
0.05
7.12
Hepatic-related morbidity pentoxifylline versus control
SC 0.30
M AN U
TE D
CD000277
EP
CD004868
11.00
1.28
RI PT
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants Calcium antagonists for aneurysmal subarachnoid haemorrhage
AC C
CD007784
Discontinuations for any reason
title
comparison
description of harm
ROR
CD001272
Antiplatelet agents for intermittent claudication
Antiplatelet agent versus placebo
Early cessation of treatment
1.66
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
3.68
2.48
1.70
3.62
1.49
0.74
3.02
SC
review
RI PT
ACCEPTED MANUSCRIPT
M AN U
0.75
Primary Harms (Severe adverse events, Discontinuations due to adverse events and any adverse event)
title
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
2.48
0.22
27.75
1.57
0.71
3.47
0.63
0.06
6.18
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
0.67
0.14
3.09
0.99
0.22
4.43
1.49
1.06
2.08
CD000171a
Treatment of latent tuberculosis infection in HIV infected persons
Any TB drug vs placebo
Incidence of adverse events leading to stopping treatment
1.05
0.04
26.24
3.73
0.16
85.90
3.53
1.76
7.09
CD000171b
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs isoniazid + rifampicin
Incidence of adverse events leading to stopping treatment
4.59
0.70
29.98
1.08
0.27
4.33
0.24
0.07
0.83
CD000171c
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs rifampicin + pyrazinimide
Incidence of adverse events leading to stopping treatment
0.53
0.19
1.52
0.46
0.23
0.93
0.86
0.40
1.85
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent versus control
Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days
1.97
0.57
6.85
3.63
2.52
5.24
1.84
0.56
6.06
AC C
EP
TE D
review
ACCEPTED MANUSCRIPT
Platelet glycoprotein IIb/IIIa blockers vs Placebo/usual care (all patients)
30-day major bleeding
2.39
0.30
Terlipressin versus vasopressin
Adverse events causing withdrawal of treatment
0.22
0.01
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
1.21
0.73
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
Changed drugs due to maternal side effects
1.22
0.19
CD002922
Treatment for lupus nephritis
Immunosuppressive agent plus steroids versus steroids alone
Major infection
0.33
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interventions for preventing oral candidiasis for patients with cancer receiving treatment Interleukin 2 receptor antagonists for kidney transplant recipients
Ibuprofen vs. indomethacin
Intraventricular haemorrhage (III-IV)
Comparisons with placebo/no treatment for all drug types
CD004183
Antioxidant supplements for preventing gastrointestinal cancers
Antioxidants versus placebo
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/ no intervention
CD003897
CD004386b
IL2Ra versus placebo or no treatment
1.73
0.59
0.07
4.62
0.25
0.05
1.35
1.11
0.07
18.24
2.00
1.51
1.00
2.28
1.26
0.93
1.69
7.67
1.61
0.58
4.43
1.32
0.28
6.08
0.02
4.53
0.77
0.37
1.59
2.33
0.19
29.04
1.75
0.51
6.03
1.52
0.70
3.28
0.87
0.33
2.29
Toxicity (adverse events 'probably due to drug')
1.47
0.44
4.90
1.27
0.83
1.95
0.87
0.28
2.67
Infection: serious all-cause total
1.08
0.39
2.98
1.00
0.62
1.60
0.92
0.38
2.26
Adverse effects - betacarotene
0.88
0.03
23.17
2.07
0.65
6.60
2.34
0.11
49.56
Side effects
1.78
0.03
92.13
1.89
1.47
2.44
1.06
0.02
54.76
Side effects
1.15
0.36
3.63
0.48
0.24
0.95
0.41
0.17
1.04
Quinolone vs. TMP-SMZ
5.88
SC
M AN U
TE D
CD003807
1.14
EP
CD002147
19.03
1.40
RI PT
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage
AC C
CD002130
ACCEPTED MANUSCRIPT
Antifungal vs. Placebo
Adverse events
1.44
0.60
Antibiotic prophylaxis versus placebo or no treatment
Adverse events
0.75
0.02
24.58
Stem cells compared to no stem cells,
Adverse effects
0.06
0.02
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular
Statins versus placebo/no treatment
Withdrawal due to adverse events
0.61
0.13
Cilostazol versus aspirin in patients with ischaemic stroke or TIA,
Other outcomes of safety during follow-up.
0.56
CD008521a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Serious adverse events
CD008521b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV5 versus placebo
Adverse events requiring discontinuation (end of follow-up)
CD007784
CD008076
3.16
1.31
0.64
2.68
2.30
0.61
8.71
3.06
0.12
76.95
0.18
0.72
0.39
1.36
13.01
4.90
34.57
2.93
1.32
0.72
2.42
2.17
0.51
9.27
0.31
1.00
1.44
1.19
1.74
2.57
1.48
4.46
0.88
0.63
1.25
0.82
0.62
1.08
0.93
0.76
1.13
1.00
0.21
4.73
0.64
0.16
2.45
0.63
0.29
1.36
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
SC
CD006536
1.13
M AN U
CD004791
3.46
1.89
RI PT
Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction
TE D
CD004773
Primary Harms
EP
USA vs Developing Countries
title
comparison
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
0.97
0.61
1.55
1.22
0.86
1.75
1.26
0.93
1.69
CD003807
Interventions for preventing oral candidiasis for patients with cancer receiving treatment
Comparisons with placebo/no treatment for all drug types
Toxicity (adverse events 'probably due to drug')
1.14
0.30
4.30
0.99
0.49
1.99
0.87
0.28
2.67
AC C
review
ACCEPTED MANUSCRIPT
Antioxidant supplements for preventing gastrointestinal cancers
Antioxidants vs placebo
Adverse effects - betacarotene
0.48
0.02
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction
Drug vs placebo/ no intervention
Side effects
2.29
0.04
121.40
Quinolone vs TMP-SMZ
Side effects
3.62
1.20
10.92
Antifungal vs Placebo
Adverse events
1.44
Antibiotic prophylaxis vs placebo or no treatment
Adverse events
1.01
Stem cells vs no stem cells
Adverse effects
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant vs control in acute presumed ischaemic stroke
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent vs control
CD002130
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
Platelet glycoprotein IIb/IIIa blockers vs Placebo/usual care (all patients)
PDE4 inhibitors vs placebo
Treatment for lupus nephritis
Immunosuppressive agent plus steroids vs steroids alone
CD002309
CD002922
1.19
2.34
0.11
49.56
2.44
1.51
3.95
1.06
0.02
54.76
1.50
0.81
2.77
0.41
0.17
1.04
3.46
1.89
1.13
3.16
1.31
0.64
2.68
0.01
97.98
3.10
0.12
79.23
3.06
0.12
76.95
0.10
0.03
0.40
1.35
0.53
3.46
13.01
4.90
34.57
Major extracranial haemorrhage during treatment period
3.42
0.28
42.14
2.16
0.76
6.18
0.63
0.06
6.18
Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days
2.45
0.54
11.05
4.52
1.80
11.36
1.84
0.56
6.06
30-day major bleeding
2.26
0.28
18.58
1.33
0.87
2.01
0.59
0.07
4.62
Non fatal serious adverse events
0.36
0.16
0.81
0.70
0.48
1.04
1.96
0.96
3.98
Major infection
0.33
0.02
4.53
0.77
0.37
1.59
2.33
0.19
29.04
SC 0.60
M AN U
CD006536
1.08
TE D
CD004791
1.13
EP
CD004773
AC C
CD004386b
10.25
RI PT
CD004183
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra vs placebo or no treatment
Infection: CMV invasive
0.72
0.07
CD003897b
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra vs placebo or no treatment
Infection: serious all-cause total
0.91
0.29
2.85
CD008521
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo
Serious adverse events
0.52
0.21
1.25
CD000171
Treatment of latent tuberculosis infection in HIV infected persons
Any TB drug vs placebo
Incidence of adverse events leading to stopping treatment
0.28
0.11
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Withdrawals due to adverse effects
0.74
CD004386
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use
Drug vs placebo/no intervention
Side effects requiring discontinuation
Quinolone vs TMP-SMZ
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
CD008521
CD008521
Statins vs placebo/no treatment
RV5 vs placebo
1.04
0.14
7.61
0.84
0.41
1.70
0.92
0.38
2.26
0.48
0.20
1.13
0.93
0.76
1.13
0.70
0.99
0.55
1.79
3.53
1.76
7.09
0.33
1.68
2.40
1.77
3.25
3.23
1.51
6.91
2.07
0.04
114.47
2.21
1.04
4.68
1.06
0.02
54.76
Side effects requiring discontinuation
5.17
0.52
51.44
0.50
0.20
1.26
0.10
0.01
0.80
Discontinuation secondary to adverse events
0.74
0.07
8.17
2.40
1.17
4.95
3.25
0.33
32.10
Withdrawal due to adverse events
0.51
0.01
35.21
1.11
0.02
59.20
2.17
0.51
9.27
Adverse events requiring discontinuation (end of follow-up)
1.46
0.06
37.19
0.93
0.04
21.46
0.63
0.29
1.36
Adverse events requiring discontinuation (end of follow-up)
2.15
0.33
14.06
1.66
0.28
10.06
0.77
0.45
1.32
SC
M AN U
EP
CD007784
Antifungal vs Placebo
2.18
AC C
CD004773
6.91
0.74
RI PT
CD003897a
CD004386
0.25
TE D
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Discontinuations due to adverse events Any adverse event
99 RCTS
Sample size per RCT: median(IQR; range) 175 (63-620; 15-4,010)
Adverse event rate per RCT: median (IQR; range) 3.5% (1.5%-7.6%; 0.02%-85.6%)
LD
44 RCTs
278 (101-965; 14-21,106)
3.4% (1.3%-8.0%; 0.02%-29%)
MD
70 RCTs
151 (61-306; 18-1,310)
4.8% (1.9%-11.9%; 0.3%-23.3%)
LD
38 RCTS
377 (120-752; 47-10,519)
1.0% (0.5%-2.3%; 0.01%-9.8%)
MD
71 RCTS
LD
10 RCTS
MD
337 RCTs (688 study-level analyses†) 137 RCTs (368 study-level analyses†)
LD
82 (61-167; 20-39,876)
10.1% (3.7%-39.1%; 0.9%-94.3%)
143 (102-210; 56-1,018)
10.1% (2.0%-26.5%; 1.0%-74.9%)
128 (60-299; 15-39,876)
6.3% (3.0%-13.6%; 0.2%-94.3%)
157 (72-400; 4-21,106)
6.9% (2.6%-14.1%; 0.01%-97.0%)
TE D
All harms* (474 independent RCTs included in 1056 study-level analyses†)
N of RCTs
SC
Severe adverse events
Country -Setting MD
M AN U
Harm
RI PT
Supplementary file 4: Sample size and adverse event rate per RCT for the primary harm endpoints, acording to country setting.
Abbreviations: LD: from less developed countries, m-a: meta-analysis, MD: from more developed countries, RCT: randomized trials Footnote:
EP
*All harms (Severe adverse events or Discontinuations due to adverse events or Any adverse event or Organ-system specific adverse events or Individual adverse events or Any discontinuations)
AC C
†Each RCT could have contributed data to more than one study-level analyses for the different types of harms/harm-related endpoints.
ACCEPTED MANUSCRIPT
Supplementary file 5a: Severe adverse events (summary OR [sOR] by countrysetting per meta-analysis)
Severe Adverse Events review
RI PT
sOR (95% CIs)
1 CD000024_LD CD000024_MD . 2 CD000029_LD CD000029_MD . 3 CD000213_LD CD000213_MD . 4 CD002130_LD CD002130_MD . 5 CD002147_LD CD002147_MD . 6 CD002309_LD CD002309_MD . 7 CD002922_LD CD002922_MD . 8 CD003481_LD CD003481_MD . 9 CD003897a_LD CD003897a_MD . 10 CD003897b_LD CD003897b_MD . 11 CD008521a_LD CD008521a_MD . 12 CD008521b_LD CD008521b_MD . NOTE: Weights are from random effects analysis
0.63 (0.06, 6.18) 1.57 (0.71, 3.47)
SC
1.49 (1.06, 2.08) 0.99 (0.22, 4.43)
1.84 (0.56, 6.06) 3.63 (2.52, 5.24)
M AN U
0.59 (0.07, 4.62) 1.40 (1.14, 1.73)
1.11 (0.07, 18.24) 0.60 (0.07, 5.42)
1.96 (0.96, 3.98) 0.75 (0.55, 1.03)
TE D
2.33 (0.19, 29.04) 0.77 (0.37, 1.59)
0.87 (0.33, 2.29) 1.52 (0.70, 3.28)
EP
1.04 (0.14, 7.61) 0.97 (0.59, 1.58)
AC C
0.92 (0.38, 2.26) 1.00 (0.62, 1.60)
.3
.5
0.93 (0.76, 1.13) 0.82 (0.62, 1.08)
0.92 (0.65, 1.31) 0.75 (0.47, 1.20)
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 5b: Severe adverse events (RORs per meta-analysis and summary ROR [sROR] across meta-analyses)
RI PT
Severe Adverse Events ROR (95% CI)
review
CD000024
2.48 (0.22, 27.75)
CD000029
0.67 (0.14, 3.09)
CD000213
1.97 (0.57, 6.85)
SC
CD002130 CD002147 CD002309
M AN U
CD002922 CD003481 CD003897a CD003897b CD008521a CD008521b
TE D
EP AC C
1
2
0.38 (0.18, 0.83) 0.33 (0.02, 4.53)
0.93 (0.12, 7.26) 1.08 (0.39, 2.98) 0.88 (0.63, 1.25) 0.81 (0.45, 1.46) 0.86 (0.68, 1.11)
NOTE: Weights are from random effects analysis .5
0.54 (0.02, 19.03)
1.75 (0.51, 6.03)
Overall (I-squared = 0.0%, p = 0.550)
.3
2.39 (0.30, 19.03)
3
ACCEPTED MANUSCRIPT
Supplementary file 6a: Discontinuations due to adverse events, by country setting (summary OR [sOR] by country-setting per meta-analysis)
RI PT
Discontinuations due to Adverse Events sOR (95% CI)
1 CD000171a_LD CD000171a_MD
3.53 (1.76, 7.09) 3.73 (0.16, 85.90)
2 CD000171b_LD CD000171b_MD
0.24 (0.07, 0.83) 1.08 (0.27, 4.33)
SC
review
3 CD000171c_LD CD000171c_MD
0.86 (0.40, 1.85) 0.46 (0.23, 0.93)
4 CD002147_LD CD002147_MD
M AN U
1.11 (0.07, 18.19) 0.25 (0.05, 1.36)
5 CD002309_LD CD002309_MD
3.23 (1.51, 6.91) 1.96 (1.50, 2.58)
6 CD002863_LD CD002863_MD
1.32 (0.28, 6.08) 1.61 (0.58, 4.43)
7 CD004386a_LD CD004386a_MD
9 CD004773_LD CD004773_MD
0.10 (0.01, 0.80) 0.39 (0.16, 1.00)
3.25 (0.33, 32.10) 2.40 (1.17, 4.95)
2.17 (0.51, 9.27) 1.32 (0.72, 2.42)
EP
10 CD007784_LD CD007784_MD
TE D
8 CD004386b_LD CD004386b_MD
1.06 (0.02, 54.75) 2.25 (1.29, 3.93)
AC C
11 CD008521a_LD CD008521a_MD
0.63 (0.29, 1.36) 0.64 (0.16, 2.45)
12 CD008521a_LD CD008521b_MD
0.77 (0.45, 1.32) 1.48 (0.56, 3.96)
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 6b: Discontinuations due to adverse events (RORs per metaanalysis and summary ROR [sROR] across meta-analyses)
RI PT
Discontinuations due to Adverse Events review
ROR (95% CI)
CD000171a
1.05 (0.04, 26.24)
CD000171b
4.59 (0.70, 29.97)
CD000171c
0.53 (0.19, 1.52)
0.22 (0.01, 5.92)
SC
CD002147 CD002309
0.61 (0.27, 1.36) 1.22 (0.19, 7.67)
CD004386a CD004386b CD004773 CD007784 CD008521a CD008521b Overall (I-squared = 0.0%, p = 0.585)
2.11 (0.04, 113.12)
M AN U
CD002863
4.03 (0.40, 40.29) 0.74 (0.07, 8.17) 0.61 (0.13, 2.93) 1.00 (0.21, 4.73) 1.92 (0.63, 5.87) 0.93 (0.60, 1.43)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 7a: Any adverse event per country setting (summary OR [sOR] by country-setting per meta-analysis)
Any Adverse Event sOR (95% CI)
RI PT
review 1 CD002309a_LD CD002309a_MD
1.26 (0.93, 1.69) 1.51 (1.01, 2.28)
2 CD003807a_LD CD003807a_MD
0.87 (0.28, 2.67) 1.27 (0.83, 1.95)
SC
3 CD004183a_LD CD004183a_MD
2.34 (0.11, 49.56) 2.07 (0.65, 6.60)
M AN U
4 CD004386a_LD CD004386a_MD 5 CD004386b_LD CD004386b_MD 6 CD004773a_LD CD004773a_MD
1.06 (0.02, 54.75) 1.89 (1.47, 2.44)
0.41 (0.17, 1.04) 0.48 (0.24, 0.95)
1.31 (0.64, 2.68) 1.89 (1.13, 3.16)
3.06 (0.12, 76.95) 2.30 (0.61, 8.71)
8 CD006536a_LD CD006536a_MD
13.01 (4.90, 34.57) 0.84 (0.48, 1.45)
9 CD008076a_LD CD008076a_MD
2.57 (1.48, 4.46) 1.44 (1.19, 1.74)
AC C
EP
TE D
7 CD004791a_LD CD004791a_MD
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 7b: Any adverse event (ROR per meta-analysis and summary ROR [sROR] across metaanalyses)
RI PT
Any Adverse Event review
ROR (95% CI)
CD002309
1.21 (0.73, 2.00)
CD003807
1.47 (0.44, 4.90)
0.88 (0.03, 23.16)
SC
CD004183 CD004386a
1.78 (0.03, 92.13)
1.15 (0.36, 3.63)
CD004773 CD004791 CD006536 CD008076 Overall (I-squared = 69.9%, p = 0.001)
NOTE: Weights are from random effects analysis
.5
AC C
EP
TE D
.3
M AN U
CD004386b
1
2
1.44 (0.60, 3.46) 0.75 (0.02, 24.58) 0.06 (0.02, 0.20) 0.56 (0.31, 1.00) 0.73 (0.37, 1.46)
3
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Supplementary file 8a: Combined primary harms (summary OR [sOR] by countrysetting per meta-analysis)
review
sOR (95% CI)
1 Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period
CD000024_LD CD000024_MD
2 Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period
CD000029_LD CD000029_MD
3 Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment
CD000171a_LD CD000171a_MD
4 Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment
CD000171b_LD CD000171b_MD
5 Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment
CD000171c_LD CD000171c_MD
0.63 (0.06, 6.18) 1.57 (0.71, 3.47) 1.49 (1.06, 2.08) 0.99 (0.22, 4.43)
3.53 (1.76, 7.09) 3.73 (0.16, 85.90) 0.24 (0.07, 0.83) 1.08 (0.27, 4.33) 0.86 (0.40, 1.85) 0.46 (0.23, 0.93)
SC
outcome
RI PT
Combined Primary Harms
6 Symptomatic (including fatal) intracranial haemorrhage within 7 to 10CdD ay0s00213_LD dD ay0s00213_MD Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 C
1.84 (0.56, 6.06) 3.63 (2.52, 5.24)
7 30-day major bleeding 30-day major bleeding
CD002130_LD CD002130_MD
8 Adverse events causing withdrawal of treatment Adverse events causing withdrawal of treatment
CD002147_LD CD002147_MD
9 No of patients experiencing an adverse eff ect No of patients experiencing an adverse eff ect
CD002309_LD CD002309_MD
1.26 (0.93, 1.69) 1.51 (1.01, 2.28)
10 Changed drugs due to maternal side effects Changed drugs due to maternal side effects
CD002863_LD CD002863_MD
1.32 (0.28, 6.08) 1.61 (0.58, 4.43)
11 Major infection Major infection
CD002922_LD CD002922_MD
2.33 (0.19, 29.04) 0.77 (0.37, 1.59)
12 Intraventricular haem orrhage (III-IV) Intraventricular haem orrhage (III-IV)
CD003481_LD CD003481_MD
0.87 (0.33, 2.29) 1.52 (0.70, 3.28)
13 Toxicity (adverse events 'probably due to drug') Toxicity (adverse events 'probably due to drug')
CD003807_LD CD003807_MD
0.87 (0.28, 2.67) 1.27 (0.83, 1.95)
14 Infection: serious all-cause total Infection: serious all-cause total
CD003897_LD CD003897_MD
0.92 (0.38, 2.26) 1.00 (0.62, 1.60)
15 Adverse effects - beta-carotene Adverse effects - beta-carotene
CD004183_LD CD004183_MD
2.34 (0.11, 49.56) 2.07 (0.65, 6.60)
16 Side effects Side effects
CD004386a_LD CD004386a_MD
1.06 (0.02, 54.75) 1.89 (1.47, 2.44)
18 Adverse events Adverse events 19 Adverse events Adverse events 20 Adverse effects Adverse effects
M AN U
1.11 (0.07, 18.24) 0.25 (0.05, 1.35)
TE D
17 Side effects Side effects
0.59 (0.07, 4.62) 1.40 (1.14, 1.73)
CD004386b_LD CD004386b_MD
0.41 (0.17, 1.04) 0.48 (0.24, 0.95)
CD004773_LD CD004773_MD
1.31 (0.64, 2.68) 1.89 (1.13, 3.16)
CD004791_LD CD004791_MD
3.06 (0.12, 76.95) 2.30 (0.61, 8.71) 13.01 (4.90, 34.57) 0.72 (0.39, 1.36)
CD007784_LD CD007784_MD
2.17 (0.51, 9.27) 1.32 (0.72, 2.42)
22 Other outcomes of saf ety during follow-up. Other outcomes of saf ety during follow-up.
CD008076_LD CD008076_MD
2.57 (1.48, 4.46) 1.44 (1.19, 1.74)
23 Adverse events requiring discontinuation (end of f ollow-up) Adverse events requiring discontinuation (end of follow-up)
CD008521b_LD CD008521b_MD
0.63 (0.29, 1.36) 0.64 (0.16, 2.45)
24 Serious adverse events Serious adverse events
CD008521a_LD CD008521a_MD
0.93 (0.76, 1.13) 0.82 (0.62, 1.08)
AC C
EP
CD006536_LD CD006536_MD
21 W ithdrawal due to adverse events W ithdrawal due to adverse events
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 8b: Combined Primary Harms (Severe adverse events OR discontinuations due to adverse events OR any adverse event) (RORs per metaanalysis and summary ROR [sROR] across metaanalyses)
RI PT
Combined Primary Harms review
ROR (95% CI)
2.48 (0.22, 27.75) 0.67 (0.14, 3.09) 1.05 (0.04, 26.24) 4.59 (0.70, 29.97) 0.53 (0.19, 1.52) 1.97 (0.57, 6.85) 2.39 (0.30, 19.03) 0.22 (0.01, 5.88) 1.21 (0.73, 2.00) 1.22 (0.19, 7.67) 0.33 (0.02, 4.53) 1.75 (0.51, 6.03) 1.47 (0.44, 4.90) 1.08 (0.39, 2.98) 0.88 (0.03, 23.16) 1.78 (0.03, 92.13) 1.15 (0.36, 3.63) 1.44 (0.60, 3.46) 0.75 (0.02, 24.58) 0.06 (0.02, 0.18) 0.61 (0.13, 2.93) 0.56 (0.31, 1.00) 1.00 (0.21, 4.73) 0.88 (0.63, 1.25) 0.90 (0.65, 1.24)
TE D
M AN U
SC
CD000024 CD000029 CD000171a CD000171b CD000171c CD000213 CD002130 CD002147 CD002309 CD002863 CD002922 CD003481 CD003807 CD003897 CD004183 CD004386a CD004386b CD004773 CD004791 CD006536 CD007784 CD008076 CD008521a CD008521b Overall (I-squared = 39.3%, p = 0.026)
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 9a: Organ system specific adverse events (summary OR [sOR] by country-setting per meta-
RI PT
analysis)
Organ System Specific Adverse Events review
sOR (95% CI)
Maternal haematological toxicity CD000102a_LD CD000102a_MD . Neonatal haematological toxicity CD000102b_LD CD000102b_MD . Pulmonary complications CD001556a_LD CD001556a_MD . Cardiac complications CD001556b_LD CD001556b_MD . Gastrointestinal complications CD001556c_LD CD001556c_MD .
SC
1.43 (0.56, 3.64) 1.06 (0.53, 2.11)
M AN U
0.76 (0.41, 1.38) 2.09 (1.22, 3.59)
0.59 (0.31, 1.14) 1.14 (0.67, 1.96)
TE D
1.19 (0.59, 2.41) 0.89 (0.44, 1.80)
.5
AC C
EP
.3
1
441.21 (24.93, 7807.45) 1.15 (0.34, 3.81)
2
3
ACCEPTED MANUSCRIPT
Supplementary file 9b: Organ system specific adverse events (ROR per metaanalysis)
RI PT
Organ System Specific Adverse Events review
ROR (95% CI)
SC
0.74 (0.23, 2.35)
2.77 (1.23, 6.22)
TE D
M AN U
Maternal haematological toxicity CD000102 . Neonatal haematological toxicity CD000102 . Pulmonary complications CD001556 . Cardiac complications CD001556 . Gastrointestinal complications CD001556 .
AC C
EP
.3
.5
1
1.94 (0.83, 4.54)
0.74 (0.27, 2.03)
0.00 (0.00, 0.06)
2
3
ACCEPTED MANUSCRIPT
Supplementary file 10a: Individual adverse events (RORs per meta-analyses and summary ROR [sROR] per individual adverse event)
Individual Adverse Events_Group 1 review
RORr (95% CI)
Bleeding (pregnancy) Bleeding (pregnancy) CD004227 Bleeding (pregnancy) CD004069 Subtotal (I-squared = 24.4%, p = 0.250) . Diarrhoea Diarrhoea CD008521 Diarrhoea CD002309 Diarrhoea CD008521 Diarrhoea CD001832 Diarrhoea CD008521 Subtotal (I-squared = 28.2%, p = 0.234) . Drug resistant infection CD004386 Drug resistant infection Drug resistant infection CD004386 Subtotal (I-squared = 37.7%, p = 0.205) . Dyspepsia Dyspepsia CD001272 Dyspepsia CD002309 Subtotal (I-squared = 0.0%, p = 0.722) . Fever Fever CD008521 Fever CD008521 Fever CD008521 Subtotal (I-squared = 0.0%, p = 0.737) . GI bleeding GI bleeding CD008076 GI bleeding CD002243 GI bleeding CD004213 Subtotal (I-squared = 0.0%, p = 0.722) . Hypertension Hypertension CD004863 Hypertension CD004868 Subtotal (I-squared = 0.0%, p = 0.829) . IVH IVH CD000361 IVH CD004863 IVH CD004868 IVH CD003481 IVH CD004659 Subtotal (I-squared = 0.0%, p = 0.975) . Intracranial hemorrhage Intracranial hemorrhage CD000029 Intracranial hemorrhage CD000024 Subtotal (I-squared = 24.2%, p = 0.251) . NEC NEC CD004863 NEC CD004868 NEC CD004454 NEC CD004213 NEC CD003481 Subtotal (I-squared = 0.0%, p = 0.633) . Neutropenia Neutropenia CD001832 Neutropenia CD004863 Subtotal (I-squared = 0.0%, p = 0.879) . ROP ROP CD003481 ROP CD004213 ROP CD004863 Subtotal (I-squared = 0.0%, p = 0.804) . Sepsis Sepsis CD005496 Sepsis CD004868 Sepsis CD004863 Sepsis CD003481 Subtotal (I-squared = 0.0%, p = 0.861) . Vomiting Vomiting CD008521 Vomiting CD008521 CD008521 Vomiting Subtotal (I-squared = 22.1%, p = 0.277) . NOTE: Weights are from random effects analysis
0.13 (0.02, 1.16) 1.08 (0.06, 18.45) 0.31 (0.04, 2.31) 1.20 0.77 1.33 3.92 1.10 1.28
(0.84, (0.27, (0.90, (1.35, (0.61, (0.96,
1.72) 2.20) 1.97) 11.34) 1.97) 1.72)
4.89 (0.48, 50.17) 0.81 (0.17, 3.73) 1.59 (0.29, 8.81)
SC
0.82 (0.22, 3.10) 0.59 (0.19, 1.86) 0.68 (0.28, 1.62) (0.69, (0.69, (0.68, (0.77,
1.15) 1.10) 1.66) 1.06)
0.38 1.07 1.10 0.49
(0.11, (0.06, (0.03, (0.17,
1.32) 18.24) 35.38) 1.45)
M AN U
0.89 0.87 1.06 0.90
3.02 (0.11, 86.62) 1.81 (0.08, 43.14) 2.31 (0.23, 23.10)
0.99 (0.46, 2.13)
1.37 (0.44, 4.27) 7.28 (0.53, 99.58) 2.05 (0.50, 8.33)
TE D
0.58 1.74 0.85 0.69 1.69 0.95
.5
1
(0.21, (0.12, (0.30, (0.12, (0.63, (0.55,
1.61) 24.72) 2.42) 3.92) 4.51) 1.64)
1.43 (0.24, 8.49) 1.10 (0.06, 18.82) 1.33 (0.29, 6.00)
EP
AC C
.3
RI PT
AdverseEvent
2
3
1.33 0.75 1.51 1.42
(0.31, (0.10, (0.85, (0.85,
5.66) 5.61) 2.69) 2.38)
0.98 2.19 1.05 1.59 1.14
(0.42, (0.19, (0.51, (0.49, (0.70,
2.32) 24.96) 2.18) 5.20) 1.86)
0.66 0.99 0.88 0.86
(0.44, (0.74, (0.63, (0.69,
0.99) 1.34) 1.24) 1.08)
ACCEPTED MANUSCRIPT
Supplementary file 10b: Individual adverse events in Group 2 (1 ma per adverse event) (ROR per meta-analysis)
Individual Adverse Events_Group 2 AdverseEvent
RORr (95% CI)
1.23 (0.48, 3.14)
Anastomotic leaks
0.25 (0.02, 2.86)
Any maternal infection
1.62 (0.68, 3.89)
RI PT
Abdominal pain
Bleeding
1.29 (0.56, 2.94)
CLD
1.95 (0.72, 5.26)
Cerebral infarction on CT/MR
3.56 (0.92, 13.77) 0.47 (0.21, 1.10)
Chorioamnionitis Clinical signs of secondary ischaemia: studies with magnesium in addition of nimodipine
1.58 (0.46, 5.46) 0.52 (0.10, 2.68)
Decreased urine output Delayed gastric emptying
0.55 (0.15, 2.04)
1.16 (0.05, 27.26)
Elevated LFTs
1.28 (0.24, 6.65)
Fever in women after trial entry requiring the use of antibiotics Fungal infection
1.20 (0.09, 15.41) 3.14 (0.11, 89.65)
GI Bleeding (LBWI)
1.67 (0.03, 88.94)
Hyperglycaemia
1.31 (0.36, 4.81)
Hypotension
1.13 (0.30, 4.24)
Infection
0.83 (0.23, 3.07)
Infection: CMV all
0.28 (0.01, 7.52)
SC
Intestinal perforation Intrapartum fever in woman requiring the use of antibiotics Low APGAR scores Low birth weight Malignancy Nausea Neonatal bleeding PTLD/lymphoma
Post-transplant diabetes mellitus Postnatal fever in woman Postpartum haemorrhage Premature delivery Proven Infection in NICU Puerperal sepsis Pulmonary hypertension RDS Rash Rhabdomyolysis SIDS Secondary ischaemia Superinfections Systemic infection in the first 48 hours of life Toxicities (secondary neoplasms) Transfusion VOD Variceal bleeding W ound haematoma NOTE: W eights are from random effects analysis
EP
TE D
.3
AC C
0.46 (0.06, 3.38) 0.65 (0.29, 1.48) 1.34 (0.21, 8.70) 0.68 (0.20, 2.37) 0.82 (0.39, 1.70) 0.40 (0.02, 7.97) 1.09 (0.20, 5.90) 9.14 (0.73, 114.64)
M AN U
PVL
0.12 (0.00, 3.23)
.5
1
1.25 (0.54, 2.87) 2.93 (0.60, 14.44) 2.75 (0.83, 9.10) 1.23 (0.48, 3.16) 1.89 (0.52, 6.86) 3.94 (0.04, 401.55) 0.63 (0.10, 4.04) 0.22 (0.01, 3.54) 0.77 (0.05, 11.00) 0.61 (0.03, 12.78) 2.06 (0.58, 7.36) 3.10 (0.30, 31.89) 0.83 (0.24, 2.84) 0.22 (0.01, 6.89) 0.99 (0.02, 55.27) 0.76 (0.02, 29.17) 0.85 (0.03, 28.17) 3.62 (0.11, 122.43)
2
3
ACCEPTED MANUSCRIPT
Supplementary file 11: All Harm-related endpoints (ROR per meta-analysis) ALL Harm-related endpoints RORr (95% CII)
M AN U 1.66 (0.75,, 3.68))
0.06 (0.02,, 0.20)) 0.56 (0.31,, 1.00)) 0.75 (0.02,, 24.58)) 0.88 (0.03,, 23.16)) 1.15 (0.36,, 3.63)) 1.21 (0.73,, 2.00)) 1.44 (0.60,, 3.46)) 1.47 (0.44,, 4.90)) 1.78 (0.03,, 92.13))
0.22 (0.01,, 5.92)) 0.53 (0.19,, 1.52)) 0.61 (0.13,, 2.93)) 0.61 (0.27,, 1.36)) 0.74 (0.07,, 8.17)) 1.00 (0.21,, 4.73)) 1.05 (0.04,, 26.24)) 1.22 (0.19,, 7.67)) 1.92 (0.63,, 5.87)) 2.11 (0.04,, 113.12)) 4.03 (0.40,, 40.29)) 4.59 (0.70,, 29.97))
EP
TE D
0.13 (0.02,, 1.16)) 0.38 (0.11,, 1.32)) 0.58 (0.21,, 1.61)) 0.59 (0.19,, 1.86)) 0.66 (0.44,, 0.99)) 0.69 (0.12,, 3.92)) 0.73 (0.10,, 5.35)) 0.75 (0.10,, 5.61)) 0.77 (0.27,, 2.20)) 0.79 (0.01,, 45.79)) 0.81 (0.17,, 3.73)) 0.82 (0.22,, 3.10)) 0.84 (0.22,, 3.20)) 0.85 (0.30,, 2.42)) 0.87 (0.69,, 1.10)) 0.87 (0.12,, 6.42)) 0.88 (0.63,, 1.24)) 0.89 (0.69,, 1.15)) 0.98 (0.42,, 2.32)) 0.99 (0.74,, 1.34)) 1.05 (0.51,, 2.18)) 1.06 (0.68,, 1.66)) 1.07 (0.06,, 18.24)) 1.08 (0.06,, 18.45)) 1.10 (0.03,, 35.38)) 1.10 (0.61,, 1.97)) 1.10 (0.06,, 18.82)) 1.20 (0.84,, 1.72)) 1.33 (0.90,, 1.97)) 1.33 (0.31,, 5.66)) 1.37 (0.44,, 4.27)) 1.43 (0.24,, 8.49)) 1.44 (0.39,, 5.31)) 1.51 (0.85,, 2.69)) 1.59 (0.49,, 5.20)) 1.69 (0.63,, 4.51)) 1.74 (0.12,, 24.72)) 1.81 (0.08,, 43.14)) 2.19 (0.19,, 24.96)) 3.02 (0.11,, 86.62)) 3.92 (1.35,, 11.34)) 4.89 (0.48,, 50.17)) 7.28 (0.53,, 99.58))
AC C .3
3.23)) 6.89)) 3.54)) 2.86)) 7.52)) 4.53)) 7.97)) 3.38)) 1.10)) 2.68)) 2.04)) 12.78)) 4.04)) 1.48)) 2.37)) 29.17)) 11.00)) 1.70)) 3.07)) 2.84)) 28.17)) 55.27)) 5.90)) 4.24)) 27.26)) 15.41)) 3.16)) 3.14)) 2.87)) 6.65)) 2.94)) 4.81)) 8.70)) 5.46)) 3.89)) 88.94)) 6.86)) 5.26)) 7.36)) 9.10)) 14.44)) 31.89)) 89.65)) 13.77)) 122.43)) 401.55)) 114.64))
SC
0.12 (0.00,, 0.22 (0.01,, 0.22 (0.01,, 0.25 (0.02,, 0.28 (0.01,, 0.33 (0.02,, 0.40 (0.02,, 0.46 (0.06,, 0.47 (0.21,, 0.52 (0.10,, 0.55 (0.15,, 0.61 (0.03,, 0.63 (0.10,, 0.65 (0.29,, 0.68 (0.20,, 0.76 (0.02,, 0.77 (0.05,, 0.82 (0.39,, 0.83 (0.23,, 0.83 (0.24,, 0.85 (0.03,, 0.99 (0.02,, 1.09 (0.20,, 1.13 (0.30,, 1.16 (0.05,, 1.20 (0.09,, 1.23 (0.48,, 1.23 (0.48,, 1.25 (0.54,, 1.28 (0.24,, 1.29 (0.56,, 1.31 (0.36,, 1.34 (0.21,, 1.58 (0.46,, 1.62 (0.68,, 1.67 (0.03,, 1.89 (0.52,, 1.95 (0.72,, 2.06 (0.58,, 2.75 (0.83,, 2.93 (0.60,, 3.10 (0.30,, 3.14 (0.11,, 3.56 (0.92,, 3.62 (0.11,, 3.94 (0.04,, 9.14 (0.73,,
RI PT
review Individual ae_grp2 CD004454 CD004024 CD001832 CD001556 CD003481 CD002922 CD003897 CD002863 CD004454 CD003481 CD008370 CD004868 CD002863 CD000102 CD002309 CD006501 CD007784 CD004659 CD003897 CD004454 CD007339 CD000305 CD003481 CD001556 CD007784 CD004386 CD004454 CD002309 CD004454 CD004454 CD008076 CD004343 CD003897 CD000277 CD000230 CD002243 CD004454 CD003481 CD000277 CD000102 CD000230 CD002243 CD003481 CD000277 CD000305 CD004213 CD003897 . Any Discontinuation CD001272 . Any ae CD006536 CD008076 CD004791 CD004183 CD004386 CD002309 CD004773 CD003807 CD004386 . Discontinuations due to ae CD002147 CD000171 CD007784 CD002309 CD004773 CD008521 CD000171 CD002863 CD008521 CD004386 CD004386 CD000171 . Individual ae_grp1 CD004227 CD008076 CD004863 CD002309 CD008521 CD004213 CD003481 CD004213 CD002309 CD004868 CD004386 CD001272 CD000361 CD004454 CD008521 CD004659 CD008521 CD008521 CD005496 CD008521 CD004863 CD008521 CD002243 CD004069 CD004213 CD008521 CD004863 CD008521 CD008521 CD003481 CD000029 CD001832 CD004863 CD004863 CD003481 CD003481 CD004868 CD004868 CD004868 CD004863 CD001832 CD004386 CD000024 . Organ specific ae CD001556 CD000102 CD001556 CD001556 CD000102 . Severe ae CD002922 CD002309 CD002147 CD000029 CD008521b CD008521 CD003897 CD003897 CD003481 CD000213 CD002130 CD000024 . NOTE: W eights are from random effects analysis
.5
1
2
3
0.00 (0.00,, 0.74 (0.23,, 0.74 (0.27,, 1.94 (0.83,, 2.77 (1.23,,
0.06)) 2.35)) 2.03)) 4.54)) 6.22))
0.33 (0.02,, 0.38 (0.18,, 0.54 (0.02,, 0.67 (0.14,, 0.81 (0.45,, 0.88 (0.63,, 0.93 (0.12,, 1.08 (0.39,, 1.75 (0.51,, 1.97 (0.57,, 2.39 (0.30,, 2.48 (0.22,,
4.53)) 0.83)) 19.03)) 3.09)) 1.46)) 1.25)) 7.26)) 2.98)) 6.03)) 6.85)) 19.03)) 27.75))
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Footnote: Individual adverse events-group 1: adverse events with ≥2 meta-analyses per adverse event; Individual adverse events-group 2: adverse events with 1 meta-analysis per adverse event
ACCEPTED MANUSCRIPT
Supplementary file 12: Summary estimates (sRORs [Summary Relative Odds Ratios]: when there were ≥2 meta-analyses per harm/harm-related endpoint or RORs: when there was only 1 meta-analysis per harm/harm-related endpoint). Summary Estimates
Primary Endpoints
sROR (95% CIs)
I2 (95% CI), pvalue
Severe adverse events (12 m-a) Discontinuations due to adverse events (12 m-a) Any adverse event (9 m-a)
0.86 (0.68-1.11) 0.93 (0.60-1.43) 0.73 (0.37-1.46)
0 (0-58%); p=0.550 0 (0-58%); p=0.585 70% (40%-85%); p<0.001
Composite Harm Endpoint
sROR (95% CIs)
Combined Primary-Harms
0.90 (0.65-1.24)
SC
TE D
Gastrointestinal complications
ROR (95% CIs)
0.74 (0.23-2.35) 2.77 (1.23-6.22) 1.94 (0.83-4.54) 0.74 (0.27-2.03)
0.00 (0.00-0.06)
sROR (95% CIs) 0.31 (0.04-2.31) 1.28 (0.96-1.72) 1.59 (0.29-8.81) 0.68 (0.28-1.62) 0.90 (0.77-1.06) 0.49 (0.17-1.45) 2.31 (0.23-23.10) 0.99 (0.46-2.13) 2.05 (0.50-8.33) 0.95 (0.55-1.64) 1.33 (0.29-6.00) 1.42 (0.85-2.38) 1.14 (0.70-1.86)
Vomiting
0.86 (0.69-1.08)
Individual Harms –group 2 (N=46 harms; 1 m-a/harm) Abdominal pain Anastomotic leaks Any maternal infection
ROR (95% CIs)
EP
Individual Harms –group 1 (N=14 harms; ≥2 m-a/harm) Bleeding (during pregnancy) Diarrhoea Drug resistant infection Dyspepsia Fever GI bleeding Hypertension IVH Intracranial hemorrhage NEC Neutropenia ROP Sepsis
AC C
39% (1%-63%); p=0.026
M AN U
Organ system specific harms (N=5 harms; 1 m-a/ harm) Maternal haematological toxicity Neonatal haematological toxicity Pulmonary complications Cardiac complications
RI PT
Harm-related endpoints
1.23 (0.48-3.14) 0.25 (0.02-2.86) 1.62 (0.68-3.89)
1.20 (0.09-15.41) 3.14 (0.11-89.66) 1.67 (0.03-88.94) 1.31 (0.36-4.81) 1.13 (0.30-4.24) 0.83 (0.23-3.07) 0.28 (0.01-7.52) 0.12 (0.00-3.23)
TE D
M AN U
Intestinal perforation Intrapartum fever in woman requiring the use of antibiotics Low APGAR scores Low birth weight Malignancy Nausea Neonatal bleeding PTLD/lymphoma PVL Post-transplant diabetes mellitus
1.29 (0.56-2.94) 1.95 (0.72-5.26) 3.56 (0.92-13.77) 0.47 (0.21-1.10) 1.58 (0.46-5.46) 0.52 (0.10-2.68) 0.55 (0.15-2.04) 1.16 (0.05-27.26) 1.28 (0.24-6.65)
SC
Bleeding CLD Cerebral infarction on CT/MR Chorioamnionitis Clinical signs of secondary ischaemia Decreased urine output Delayed gastric emptying Elevated LFTs Fever in women after trial entry requiring the use of antibiotics Fungal infection GI Bleeding (LBWI) Hyperglycaemia Hypotension Infection Infection: CMV all
RI PT
ACCEPTED MANUSCRIPT
EP
Postnatal fever in woman Postpartum haemorrhage Premature delivery Proven Infection in NICU Puerperal sepsis Pulmonary hypertension
AC C
RDS Rash Rhabdomyolysis SIDS Secondary ischaemia Superinfections Systemic infection in the first 48 hours of life Toxicities (Secondary Neoplasms) Transfusion VOD Variceal bleeding Wound haematoma Any Discontinuation (1 m-a) Early cessation of treatment
0.46 (0.06-3.38) 0.65 (0.29-1.48) 1.34 (0.21-8.70) 0.68 (0.20-2.37) 0.82 (0.39-1.70) 0.40 (0.02-7.97) 1.09 (0.20-5.90) 9.14 (0.73114.64) 1.25 (0.54-2.87) 2.93 (0.60-14.44) 2.75 (0.83-9.10) 1.23 (0.48-3.16) 1.89 (0.52-6.86) 3.94 (0.04401.58) 0.63 (0.10-4.04) 0.22 (0.01-3.54) 0.77 (0.05-11.00) 0.61 (0.03-12.78) 2.06 (0.58-7.36) 3.10 (0.30-31.89) 0.83 (0.24-2.84) 0.22 (0.001-6.89) 0.99 (0.02-55.27) 0.76 (0.02-29.17) 0.85 (0.03-28.17) 3.62 (0.11122.43) ROR (95% CIs) 1.66 (0.75-3.68)
ACCEPTED MANUSCRIPT
Supplementary file 13: Severe adverse events (ROR per meta-analyses, calculated by FEM and summary ROR [sROR] across meta-analyses calculated by REM).
RI PT
Severe Adverse Events review
RORf (95% CI)
2.48 (0.22, 27.75)
CD000024
0.67 (0.14, 3.09)
SC
CD000029
1.94 (0.57, 6.63)
CD000213 CD002130
2.40 (0.30, 19.00) 0.54 (0.02, 19.03)
M AN U
CD002147 CD002309 CD002922 CD003481 CD003897 CD003897 CD008521
TE D
Overall (I-squared = 0.0%, p = 0.468)
0.38 (0.18, 0.83) 0.38 (0.03, 4.69) 1.75 (0.51, 6.03) 0.93 (0.12, 7.26) 1.07 (0.41, 2.78) 0.94 (0.69, 1.26) 0.91 (0.71, 1.17)
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3 3.5
ACCEPTED MANUSCRIPT
Supplementary file 14: Discontinuations due to adverse events (ROR per metaanalyses, calculated by FEM and summary ROR [sROR] across meta-analyses
RI PT
calculated by REM)
Discontinuations due to Adverse Events review
RORf (95% CI)
CD000171a
4.38 (1.08, 17.76)
CD000171b
0.53 (0.19, 1.52)
SC
CD000171c CD000171d CD000171e
M AN U
CD000171f CD002147 CD002309 CD002863 CD004386a CD004386b CD004773 CD007784 CD008521
TE D
Overall (I-squared = 39.6%, p = 0.063)
0.42 (0.16, 1.07) 3.21 (0.10, 105.42) 6.75 (0.30, 150.93) 0.32 (0.13, 0.80) 0.22 (0.01, 5.88) 0.60 (0.27, 1.33) 1.22 (0.19, 7.67) 2.11 (0.04, 112.90) 5.49 (0.63, 47.56) 0.74 (0.07, 8.17) 0.51 (0.11, 2.26) 1.86 (0.63, 5.51) 0.88 (0.52, 1.47)
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3 3.5
ACCEPTED MANUSCRIPT
Supplementary file 15: Any adverse event (ROR per meta-analyses, calculated by FEM and summary ROR [sROR] across meta-analyses calculated by REM)
RI PT
Any Adverse Event review
RORf (95% CI)
CD002309
0.99 (0.70, 1.39)
CD003807
1.47 (0.44, 4.90) 0.22 (0.01, 6.89)
SC
CD004024 CD004183 CD004386
CD004773 CD004791 CD006536 CD008076 Overall (I-squared = 67.1%, p = 0.001)
M AN U
CD004386
0.85 (0.04, 18.02) 1.79 (0.03, 92.44) 1.55 (0.58, 4.12) 1.44 (0.60, 3.46) 0.75 (0.02, 24.58) 0.06 (0.02, 0.20) 0.56 (0.31, 1.00) 0.72 (0.39, 1.33)
NOTE: Weights are from random effects analysis .5
AC C
EP
TE D
.3
1
2
3 3.5
ACCEPTED MANUSCRIPT
Supplementary file 16: Severe adverse events: USA trials vs trials from Less Developed (DC) Countries (ROR per meta-analysis and summary ROR [sROR]
RI PT
across meta-analyses; calculations by REMs)
Severe Adverse Events_USA vs DC review
RORr (95% CI)
3.42 (0.28, 42.14)
SC
CD000024 CD000213 CD002130
2.26 (0.28, 18.57) 0.36 (0.16, 0.81)
M AN U
CD002309
2.45 (0.54, 11.05)
CD002922 CD003897 CD003897 CD008521
0.33 (0.02, 4.53) 0.72 (0.07, 6.91) 0.91 (0.29, 2.85) 0.52 (0.21, 1.25)
Overall (I-squared = 21.6%, p = 0.258)
0.74 (0.42, 1.29)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 17: Discontinuations due to adverse events: USA trials vs trials from Less Developed (DC) countries. (ROR per meta-analysis and summary ROR
RI PT
[sROR] across meta-analyses; calculations by REM)
Discontinuations due to Adverse Events_USA vs DC review
RORr (95% CI)
0.28 (0.11, 0.70)
SC
CD000171 CD002309
CD004386 CD004773 CD007784 CD008521 CD008521 Overall (I-squared = 18.9%, p = 0.280)
2.07 (0.04, 114.46) 5.17 (0.52, 51.44)
M AN U
CD004386
0.74 (0.33, 1.68)
0.74 (0.07, 8.17) 0.51 (0.01, 35.21) 1.46 (0.06, 37.19) 2.15 (0.33, 14.06) 0.75 (0.38, 1.48)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 18: Any adverse event: USA trials vs trials from Less Developed (DC) Countries. (ROR per meta-analysis and summary ROR [sROR] across meta-
RI PT
analyses; calculations by REM)
Any Adverse Event_USA vs DC review
RORr (95% CI)
0.97 (0.61, 1.55)
SC
CD002309 CD003807
CD004386 CD004386 CD004773 CD004791 CD006536 Overall (I-squared = 58.8%, p = 0.017)
0.48 (0.02, 10.25) 2.29 (0.04, 121.39)
M AN U
CD004183
1.14 (0.30, 4.30)
3.62 (1.20, 10.92) 1.44 (0.60, 3.46) 1.01 (0.01, 97.98) 0.10 (0.03, 0.40) 0.97 (0.47, 2.03)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
RI PT
Supplementary file 19: Baseline risks of harm/harm-related endpoints in the control group in meta-analyses from more developed and less developed countries in topics with statistically significant differences between trials from the two country settings (ROR and 95% CIs >1 or <1) Title
Comparison
Outcome
CD000102
Interventions for reducing the risk of mother-to-child transmission of HIV infection Preoperative chemotherapy for resectable thoracic esophageal cancer Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
Any ZDV vs placebo/no treatment
Neonatal haematological toxicity GI complications Diarrhoea
Stem cell treatment for acute myocardial infarction Vaccines for preventing rotavirus diarrhoea: vaccines in use
Stem cells vs no stem cells Rotarix vs placebo: after dose 2
CD002309
CD006536
Baseline risk (Control Rate) in Less Developed Countries (95% CIs)
P of Difference in Baseline Risk
0.05 (0.01-0.11)
0.061
0.80
0.02 (0.01-0.04)
0.02 (0.00-0.07)
0.873
0.80
0.14 (0.08-0.21)
0.10 (0.04-0.19)
0.497
Non fatal serious adverse events Adverse effects
0.37
0.06 (0.04-0.08)
0.03 (0.02-0.05)
0.05
0.32
0.15 (0.08-0.25)
0.05 (0.02-0.10)
0.023
Reactogenicity: vomiting
0.50
0.12 (0.08-0.15)
0.11 (0.08-0.15)
0.922
EP
CD008521a
Baseline risk (Control Rat)e in More Developed Countries (95% CIs) 0.12 (0.08-0.16)
M AN U
CD001832
Treatment morbidity and mortality Third generation cephalosporins versus conventional therapy PDE4 inhibitors vs placebo
TE D
CD001556
Small study effect (pvalue) 0.34
SC
Review
Abbreviations: CI: confidence intervalsFootnote: p-value considered significant if <0.05
AC C
Footnote: Baseline harm/harm-related endpoint risks per country-setting were synthesized by random effects using the Freeman-Tukey transformation method and results between the two country settings by chi-square test.*p-value for Haber test: For the analysis for small study effect per meta-analysis, all studies included in each meta-analysis from all country settings were consider
Marked Revision
ACCEPTED MANUSCRIPT
Comparative Rates of Harms in Randomized Trials from More-Developed Versus LessDeveloped Countries May be Different
1 2 3
, Xanthippi Tseretopoulou,
RI PT
Despina Contopoulos-Ioannidis, clinical associate professor
registar in paediatrics 4, Megan Ancker, field physician 5, Juan N. Walterspiel, pediatric infectious diseases consultant
6
, Orestis A Panagiotou, post-doctoral fellow
Department of Pediatrics, Division of Infectious Diseases, Stanford University School of
Medicine, Stanford, CA, USA
M AN U
1
2
Palo Alto Medical Foundation Research Institute, Palo Alto, CA, USA
3
Meta-Research Innovation Center at Stanford, Stanford, CA, USA
6
7
TE D
5
Leeds Teaching Hospital, NHS Trust, Leeds, UK Medecins San Frontieres, Paris, France
Mendocino Coast District Hospital, Fort Bragg CA, USA Division of Cancer Epidemiology & Genetics, National Cancer Institute, National Institute of
EP
4
, Yvonne
SC
Maldonado, professor 1 8 9, John PA Ioannidis, professor 3 8 10 11
7
Health, Bethesda, MD, USA
Department of Health Research and Policy, Stanford University School of Medicine, Stanford,
AC C
8
CA, USA 9
Senior Associate Dean for Faculty Development and Diversity, Stanford University School of
Medicine, Stanford, CA, USA 10
Stanford Prevention Research Center, Department of Medicine, Stanford University School of
Medicine, Stanford, CA, USA
1
ACCEPTED MANUSCRIPT
11
Department of Statistics, Stanford University School of Humanities and Sciences, Stanford,
CA, USA Address correspondence to: Despina Contopoulos-Ioannidis, MD; Department of Pediatrics,
RI PT
Division of Infectious Diseases, Stanford University School of Medicine, 300 Pasteur Drive,
AC C
EP
TE D
M AN U
SC
Room G312, Stanford CA, 94305, USA; e-mail: [email protected]
2
ACCEPTED MANUSCRIPT
Funding: There was no funding for this study. Conflicts of interest: There are no conflicts of interest to declare for DCI, XT, MA, OAP, YM and JPAI. JNW worked as a consultant for US Bayer A.G. in 2002-3. “Whistleblower" on Bayer
that are paid for directly or indirectly by Bayer A.G.
RI PT
A.G. Currently does not hold any Bayer A.G or subsidiarie’s stock, nor does he give any lectures
The views expressed in this paper are solely those of the authors and do not necessarily represent
SC
the official views of the National Institutes of Health or the US Department of Health and
AC C
EP
TE D
M AN U
Human Services.
3
ACCEPTED MANUSCRIPT
Abstract Objectives: We set up to evaluate the relative risk of harms in trials performed in less-developed versus more-developed countries.
RI PT
Study Design: Meta-epidemiologic evaluation using the Cochrane Database of Systematic Reviews. We considered meta-analyses with at least one RCT in a less-developed country and one RCT in a more-developed country. We targeted severe adverse events (AEs),
SC
discontinuations due to AEs, any AE, organ system-specific AEs, individual AEs and all discontinuations due to any reason. We estimated the relative odds ratio (ROR) of harms
under each category of harms.
M AN U
between more and less-developed countries for each topic and the summary sROR across topics
Results: We identified 42 systematic-reviews (128 meta-analyses, 521 independent RCTs). Summary sRORs did not differ significantly from 1.00 for any harm category. Nominally
TE D
significant RORs were found in only 6/128 meta-analyses. However, in 27% (35/128) of metaanalyses the ROR point-estimates indicated relative differences between country-settings >2fold. Considering also ROR 95% confidence-intervals, in 92% (118/128) of meta-analyses one
EP
could not exclude a 2-fold difference in both directions. Conclusions: We identified limited comparative evidence on harms in trials from these two
AC C
country-settings. Substantial differences in the risk point-estimates were common; the potential for modest differences could rarely be excluded with confidence.
Key words: Comparative Safety, Comparative Harms, More developed countries, Less developed countries, Randomized trials, Meta-analyses.
4
ACCEPTED MANUSCRIPT
Running title: Comparative Harms in RCTs from More vs Less Developed Countries
What is new? Randomized trials on important clinical questions are frequently performed in less-developed
RI PT
•
countries, without long-standing clinical research tradition. We have previously shown that trials in such country-settings tend to report more favorable mortality outcomes. It remains
SC
unknown whether similar differences exist in the reporting of harms. We studied the comparative safety of medical interventions in randomized trials from more- versus less
•
M AN U
developed countries.
The identified comparative evidence on harms in trials from these two country-settings was limited.
•
Not many nominally-significant discrepancies in relative-risks of harms between country
TE D
settings were detected. However, substantial differences in the point-estimates of the risks of harms were common and the potential for modest differences between the two country settings could rarely be excluded with confidence.
•
EP
What this adds to what was known?
Due to the limited comparative evidence, it remains unclear whether safety data from trials in
AC C
less developed countries could be reliably used to guide policy development and clinical decision making in more developed countries, but also less developed countries. What is the implication, what should change now? •
Evaluation of the comparative safety of medical interventions in trials from these two country settings should be further pursued and additional comparative evidence needs to be accumulated.
5
ACCEPTED MANUSCRIPT
Alternative methodological approaches, i.e. through analysis of comparative safety data reported within international multisite trials, between trial sites in more developed countries
EP
TE D
M AN U
SC
RI PT
and trial sites in less developed countries should be also considered.
AC C
•
6
ACCEPTED MANUSCRIPT
Introduction As the budget for and participation rates in randomized clinical trials (RCTs) in more developed countries are limited, it is becoming increasingly more common that clinical
RI PT
guidelines and clinical decision-making about important questions of health interventions and health care will depend on evidence from trials performed in less developed countries.[1] With the globalization of clinical research, emerging countries are increasingly more actively involved
SC
in clinical trials.[2-4] Asian and Latin America regions have recently shown the largest annual increase in the number of registered clinical trials [5]. Trials done exclusively in less developed
M AN U
countries often have low methodological quality. [6] Only 56% of 670 surveyed researchers from developing countries reported that their research had been reviewed by a local institutional review board [7] and only 11% of published clinical trials conducted in China in 2004 report that their study protocol was reviewed by an ethical review committee.[8] Some reports are raising concerns for underreporting of adverse events from studies
TE D
performed in developing countries.[9] For example, large differences in the reported rates of ciprofloxacin associated arthropathy were seen in children from North America compared to
EP
children from Latin America.[9] However, the frequency of this phenomenon has not been systematically studied. If results from trials performed in countries without a longstanding
AC C
tradition in clinical research will be used to guide clinical decision-making, an empirical largescale evaluation of these trends is needed. While there are potential benefits from the globalization of clinical research, it is important to evaluate whether results are similar and possible to extrapolate across different settings. In a previous evaluation [10] we assessed differences in mortality and primary efficacy outcomes in RCTs performed on the same topic in more versus less developed countries. We found that on average, trials in less developed countries tended to report more favorable results for the 7
ACCEPTED MANUSCRIPT
experimental intervention [10]. Sometimes genuine differences between country settings could explain differences in results; however, selective outcome reporting, publication, language and other biases [11-13] in the literature coming from less developed countries was considered more It would be important to assess whether major harms
RI PT
likely to explain these discrepancies.
outcomes are also similar or different in RCTs from more versus less developed countries.
Very often the number of patients studied in pre-licensure trials is small to allow the
SC
robust evaluation of both safety and efficacy outcomes. [14] I Individual studies and even individual meta-analyses are on average underpowered to detect differences in reported rates of
M AN U
clinically important adverse events from different country settings.[15] Therefore, we performed a large-scale meta-epidemiological evaluation of safety outcomes in trials from more vs. less developed countries. Methods
TE D
Definitions of countries
The categorization of countries into more developed and less developed countries was done as previously described in our earlier paper on comparative results for mortality and
EP
primary efficacy outcomes.[10] In brief we considered more developed countries to be those with both longstanding established marker economies and longstanding tradition in clinical
AC C
research as previously suggested.[16] Such countries included the United States, Canada, Australia, New Zealand, Israel, Japan and Western European countries. All other countries except for those in Eastern Europe were considered as less developed. We excluded RCTs from Eastern European countries as these represent another unique type of countries in transition. [17] Harm-related endpoints
8
ACCEPTED MANUSCRIPT
We targeted 6 main categories of harm-related endpoints [18], three of which were considered as primary endpoints because they combine adverse events of all types and they only include harms. We did not focus on mortality, as this was the focus in our previous paper [10].
RI PT
The primary endpoint categories were: severe adverse events; discontinuations due to adverse events; and any adverse event. The secondary endpoint categories were: organ system-specific adverse events; individual adverse events; and all discontinuations-due to any reason (in some
SC
studies, discontinuations due to harm might not have been separately reported, but they could have been included under such a broader study endpoint; this endpoint would then be relevant
M AN U
for harms, even though not fully specific). Under the categories of organ system-specific and individual adverse events we considered several subcategories (e.g. gastrointestinal adverse events, hematologic adverse events etc.; and headache, neutropenia etc., respectively). These endpoints were considered secondary because they either do not include all adverse events or
TE D
they count also some events that are not due to adverse events.
We also considered a composite primary-harms endpoint (combined primary-harms), where all three primary harm endpoints (severe adverse events, discontinuations due to adverse
EP
events and any adverse event) were considered together. Eligible meta-analyses
AC C
We included meta-analyses that quantitatively synthesized evidence on harm-related
endpoints and included at least one RCT from a less developed country and at least one RCT from a more developed country for the same compared interventions and the same type of harmrelated endpoint. We focused on RCTs performed exclusively in more developed countries or exclusively in less developed countries; international multicenter trials were excluded unless all participating sites were from the same country-setting.
9
ACCEPTED MANUSCRIPT
Search We screened a list of 131 systematic reviews from the Cochrane Database of Systematic reviews (last update August 27, 2012) that we had previously identified [10] as having RCTs
RI PT
from more developed and less developed countries for mortality. The reasons for this choice were that this selected group of systematic reviews would have been more likely to have included also evidence on harms from both country settings; and the questions addressed were
SC
likely to pertain to serious clinical conditions and diseases given that death was also an outcome of interest. Meta-analyses within the same systematic review for different compared
M AN U
interventions or for different types of harm-related-endpoints were considered as separate metaanalyses. Thus, each systematic review could have contributed data to more than one category of harm-related endpoints. Data extraction
TE D
From each eligible meta-analysis we extracted information on the compared interventions, the harm-related endpoint, the study (author, year), the number of patients with harm-related endpoints and the total number of patients in the experimental group and the control
EP
group respectively per study. No continues harms were encountered. Statistical synthesis of data
AC C
We synthesized the data in the following sequential steps as previously described [10 18
19]: First, for each RCT we calculated the odds ratio (OR) of harm-related endpoints for the experimental vs. the control intervention. When needed we coined estimates so that an OR>1 always corresponds to an increased risk of the harm-related outcome. If in the 2×2 table for each study (events/total number of subjects in the experimental arm and the control arm respectively) there were one or more cells with zero events, then a standard correction of 0.5 was added to all
10
ACCEPTED MANUSCRIPT
cells to allow for a meaningful calculation of the OR, as previously described.[20] Second, for each individual meta-analysis, under each category of harm-related endpoint, we calculated the summary OR (sOR) from all RCTs from the more developed countries and the sOR from all
RI PT
RCTs from the less developed countries using random-effect model for the meta-analysis; we also performed sensitivity analyses using fixed-effects model. [21 22] Random effects model assumes that the combined studies aim at identifying an average treatment effect, while fixed
SC
effects model assumes a common effect across the combined studies. Third, for each individual meta-analysis, and for each category of harm-related endpoint, we calculated the relative OR
M AN U
(ROR) between more developed and less developed countries by dividing the sOR from trials in more developed countries by the sOR from trials in less developed countries. Fourth, for each category of harm-related endpoint, we generated a summary ROR (sROR) between more and less developed countries by synthesizing the RORs from all the individual meta-analyses, under
TE D
each category of harm-related endpoint using random-effects models. A ROR>1 (or sROR>1) means that less developed countries reported less unfavorable results for the experimental intervention relative to more developed countries.
EP
For the combined primary-harms endpoint, we synthesized available RORs from all 3 primary endpoints; whenever RORs existed for more than one primary endpoint for the same
AC C
meta-analysis, we used the ROR that had the smaller variance. For the meta-analyses for organ system-specific adverse events and individual adverse events, we combined RORs only when they were targeting exactly the same types of organ system-specific or individual adverse events. We calculated the between meta-analyses heterogeneity under each category of harm-
related endpoint by using the I2 metric with its corresponding 95% CIs.[23 24]
11
ACCEPTED MANUSCRIPT
We calculated the number of meta-analyses where there was a statistically significant difference in the sOR between the two country settings (95% CIs of ROR excluding 1.00) and
the two country settings (ROR≥ 2.0 or ROR ≤0.50).
RI PT
the number of meta-analyses where the ROR suggested difference exceeding 2.0-fold between
In topics with statistically significant differences in harm-related endpoints between the two country settings, we also explored whether there were statistically significant differences in
SC
the baseline risks of these endpoints in the control group between meta-analyses from more developed and less developed countries. We synthesized the baseline risks per country setting by
M AN U
random effects using the Freeman-Tukey transformation method [25] for the meta-analysis of proportions and compared the results in the two country settings by chi-square test. We also examined whether there was evidence for small study effects using the Harbord’s test (considered to be significant for P<0.10). [26]
TE D
In sensitivity analyses we compared meta-analyses of U.S. trials only vs. trials from less developed countries. We also performed sensitivity analyses for the primary endpoints where we excluded trials from four Asian “tigers” (Hong Kong, Taiwan, Singapore, and South Korea) that
EP
evolved from less developed countries into advanced economies according to the International Monetary Fund in 1997, although their tradition of clinical research is still not as long-standing
AC C
as in the main more developed countries. All analyses were done in Stata SE12 (College Station, TX: StataCorp LP). We followed
the PRISMA [27] (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines for reporting. Results Eligible topics 12
ACCEPTED MANUSCRIPT
Of the previously identified 131 systematic reviews [10], 42 were considered eligible (Supplementary file 1) with a total of 128 separate meta-analyses: 12 meta-analyses for severe adverse events, 12 for discontinuations due to adverse events, 9 for any adverse event, 5 for
RI PT
organ system-specific adverse events, 89 for individual adverse events and 1 for all discontinuations due to any reason (Figure 1). From a total of 521 independent RCTs -after excluding 38 from mixed country settings, 5 with no reported country setting and 4 from Eastern
SC
European countries- we eventually analyzed 474 RCTs; 337 from more developed countries and 137 from less developed countries (Supplementary file 2). The topics of the 42 systematic
M AN U
reviews were related to infectious diseases (n=10), neonatal diseases (n=6), maternal-fetal conditions (n=6), cardiovascular diseases (n=5), neurologic diseases (n=5), gastrointestinal diseases (n=3), oncologic conditions (n=3), transplantations (n=2), rheumatologic diseases (n=1) and pulmonary diseases (n=1). The individual topics appear in Supplementary file 3.
TE D
The median sample size per RCT from more developed and less developed countries for the three primary harm endpoints was 175 and 278 respectively for severe adverse events; 151 and 377 respectively for discontinuations due to adverse events and 82 and 143 respectively for
EP
any adverse event (Supplementary file 4). The median adverse event rate per RCT in more developed and less developed countries was 3.5% and 3.4% respectively for severe adverse
AC C
events; 4.8% and 1.0% for discontinuations due to adverse events and 10.1% and 10.1% respectively for any adverse event (Supplementary file 4). Relative differences in individual meta-analyses (more-developed vs. less-developed) The clinical topics and the ROR estimates between the two country settings for the three primary endpoints and the combined primary-harms endpoint are shown in Table 1 and Supplementary file 3; the respective figures for the summary OR per meta-analysis according to 13
ACCEPTED MANUSCRIPT
country-setting and the ROR per meta-analysis, for each category of these harms are shown in Supplementary files 5-8. Respectively, the summary ORs per country-setting per meta-analysis and the ROR per meta-analysis for each of the secondary endpoint categories are shown in
RI PT
Supplementary files 9-11. The summary estimates (sRORs) for the different primary and secondary endpoints and the combined primary-harms endpoint are shown in Supplementary file
SC
12.
The summary sRORs did not differ significantly from 1.00 for any category of harm.
M AN U
When all harm-related endpoints were considered, nominally statistically significant relative differences between more developed and less developed countries were identified in 5% (6/128) of all meta-analyses (ROR>1 in two and ROR<1 in four) (Supplementary file 3). However, in 27% (35/128) of meta-analyses the relative differences in the summary ORs between more and less developed countries, based on the ROR point estimates, were larger than 2-fold. More
TE D
specifically, RORs≥ 2.0 were seen on 20 meta-analyses while ROR ≤ 0.50 in 15 meta-analyses (p=0.36) (Supplementary files 3 and 11).
EP
There was also large uncertainty left around the estimated RORs and in 92% (118/128) of meta-analyses, when the ROR 95% confidence intervals were also considered, one could not
AC C
exclude a 2-fold difference in harms in both directions (lower ROR CI≥0.5 and upper ROR CI≤2.0) (Supplementary file 11). Moreover, in 67% (86/128) of the meta-analyses, there were more than 10-fold differences between the upper and lower CIs of the RORs. More specifically this occurred in 67% of meta-analyses for severe adverse events; 75% of meta-analyses for discontinuations due to adverse events; 44% of meta-analyses for any adverse events; 40% of meta-analyses for organ system specific adverse events; and 69% of meta-analyses for individual adverse events. 14
ACCEPTED MANUSCRIPT
Results were similar when fixed effects were used to synthesize the sORs for the primary harm-related endpoints per country-setting within each meta-analysis. (Supplementary files 1315). In sensitivity analyses comparing only USA trials against trials from less developed
RI PT
countries 4/24 (17%) of meta-analyses for the primary endpoints showed nominally differences beyond chance for the comparison by the country setting (ROR>1 in one and ROR<1 in three) (Supplementary files 3, 16-18). Further sensitivity analyses excluding trials from the four Asian
SC
“tigers” also gave similar results for the primary endpoints and the combined primary-harms
M AN U
endpoint. RORs
Table 2 shows in more detail the 6 topics for which a nominally statistically significant difference was found between studies from less developed and more developed countries. For 2
TE D
topics, less developed countries reported less unfavorable results (RORs>1); while the opposite occurred in 4 topics (RORs<1). Studies from less developed countries had reported relatively less neonatal hematologic toxicities with zidovudine (vs. placebo) for prevention of HIV mother
EP
to child transmission; and relatively less diarrhea with third generation cephalosporins (vs. conventional antibiotics) for acute bacterial meningitis. While studies from less developed
AC C
countries had reported relatively more severe adverse events with the use of phosphodiesterase inhibitors (vs. placebo) for chronic obstructive pulmonary disease; relatively more adverse events of any type with stem cell therapy for acute myocardial infarction; relatively more gastrointestinal complications with preoperative chemotherapy (vs. surgery alone) for resected thoracic and esophageal cancer; and relatively more vomiting after the second dose of rotavirus vaccines (vs placebo).
15
ACCEPTED MANUSCRIPT
For the above 6 topics there were no statistically significant difference in the baseline risks of harms in the control groups (control rates) between trials from less and more developed
RI PT
countries; nor was there evidence of small study effects (Supplementary file 19). Discussion
Our study provided a bird’s eye view of the pattern of reporting of harms and harm-
SC
related endpoints in trials from the two country settings. We explored the relative differences in rates of harms across different country-settings using diverse approaches in order to obtain a
M AN U
more comprehensive appraisal. Our inferences need to be cautious given the large uncertainty in the relative incidence of harms for many of the examined topics. Due to the dearth of harm data in individual randomized trials but also in individual meta-analyses, it would not have been possible to identify significant differences between country-sett ings and onl y a bird’s e ye view
TE D
and large scale evaluation across diverse topics was able to probe such country-setting specific harm differences. Overall, across all analyzed meta-analyses for all types of harm-related endpoints, we did not detect nominally significant discrepancies in the reported rates of harm-
EP
related endpoints between trials from less developed and more developed countries when results were summarized for each harm category. However, when the point estimates of the relative
AC C
differences of harms between more and less developed countries were considered, in 27% of the meta-analyses the relative differences were larger than 2-fold. Many meta-analyses had limited evidence and substantial uncertainty and when the uncertainty around the relative differences were also considered, differences in the relative odds of harms exceeding 2-fold in both directions could not be excluded in 92% of meta-analyses.
16
ACCEPTED MANUSCRIPT
We should acknowledge some study limitations. First, we did not screen the whole Cochrane Database for Systematic Reviews with harm-related endpoints. However, the 131 systematic reviews we screened was a group of meta-analyses that was highly likely to have
RI PT
included harm-related endpoints both from trials in more developed and less developed countries as they had already included trials from both country settings for mortality. By focusing also on mortality, our pool of these 131 screened systematic reviews was about clinically important
SC
medical conditions and interventions. Trials from more developed countries were the preponderance, as we had also observed in our assessments of mortality previously. [10]
M AN U
However, trials from less developed countries are expected to become more common in the future. Second, in the 128 individual meta-analyses, included in the 42 eligible systematic reviews, with comparative evidence on harm-related endpoints from the two country settings only a tenth had data on the primary harm endpoints; the majority had reported individual
TE D
specific adverse events. The clinical significance of these individual harms, without any additional information on their severity grading, is unclear.[28] This probably reflects poor reporting for harms and poor compliance with CONSORT standards of reporting, which is a
37].
EP
universal problem in randomized trials independently of the trial’s country setting [29-31] [32Third, we focused on trials performed exclusively in more-developed or less-developed
AC C
countries and did not analyze international multisite trials. It remains unknown whether the reported rates of harms in trials from other countries in economic and social transition and from study-sites from different country settings included in the same international multisite trials are systematically different. However, the vast majority of published reports of international studies included in the CDSR do not report results separately per country setting. In our database, less than 10% of the RCTs were international studies from mixed country settings. In Bayer A.G.’s
17
ACCEPTED MANUSCRIPT
ciprofloxacin patent extension studies the FDA noted an arthropathy rate of 21 % reported by investigators in the United States compared to none in a less developed country [9]. This Bayer/Quintiles database was further analyzed and divided into countries [38], and ultimately
conclusions are still pending.
RI PT
handed over to the US Department of Justice for a false claims investigation [39]; final Fraud must be ruled out whenever significant discrepancies
findings, rather than following a patient for adverse events.
SC
between investigator groups are found, as it is easy for an untrained investigator to report no
In summary, although we did not identify consistent over- or under-estimation of risks of
M AN U
harms between trials done in more versus less developed countries, substantial differences in the estimates of risks were relatively common and the potential for modestly large differences often could not be excluded with confidence. Collection and reporting of harms should be emphasized and standardized in all randomized trials, both from more developed and less developed
TE D
countries. More emphasis should be given to the standardized reporting of clinically important harms, such as severe adverse events or discontinuations due to adverse events. It is very important to have available the information required for the prompt identification of systematic
EP
differences in harms across different country settings. Differences may reflect genuine variability in the incidence of harms or differences in the reporting of harms. Reported rates of harms
AC C
may be influenced by the genetic profiles of populations, but may also depend upon ethical practices and monitoring systems in place in different countries. Disentangling these reasons would require that harms/harm-related endpoints are collected and reported transparently and in a harmonized way across randomized trials [28]. Country-setting discrepancies in the reported rates of harms should be routinely incorporated and explored in meta-analyses and systematic reviews. Moreover, regulatory agencies, such as the FDA or the EMEA, should enforce that such
18
ACCEPTED MANUSCRIPT
transparent harm-related information is provided per country and site for all randomized trials, including also multisite international trials. This may allow understanding how often and why some countries and sites have different results. Until such transparent comparative evidence on
RI PT
harms is available from trials in different country-settings, guideline developers and policy makers should cautiously extrapolate harm data from trials from one country setting to guide
AC C
EP
TE D
M AN U
SC
clinical decision making in very different country settings.
19
ACCEPTED MANUSCRIPT
Contributorship: DCI had the original idea and developed the study analysis plan; JPAI, XT, MA, OAP, JNW and YM critically commented on the study analysis plan; DCI, XT, MA performed data extraction; DCI and OAP performed the statistical analyses; DCI drafted the first
RI PT
manuscript draft and all authors critically reviewed it and approved the final version; DCI is the guarantor for this paper. Ethical approval: Not applicable.
SC
Transparency declaration: Dr. Contopoulos-Ioannidis (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that
as planned have been explained.
M AN U
no important aspects of the study have been omitted; and that any discrepancies from the study
Copyright/license for publication: The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, “a worldwide license to the Publishers and
TE D
its licensees in perpetuity, in all forms, formats and media (whether known now or created in the future), to: i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include within collections and
EP
create summaries, extracts and/or, abstracts of the Contribution, iii) create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the
AC C
inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) license any third party to do any or all of the above.”
20
ACCEPTED MANUSCRIPT
References 1. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials: race-, sex-, and age-based disparities. JAMA 2004;291(22):2720-6
RI PT
2. George M, Selvarajan S, S SK, et al. Globalization of clinical trials - where are we heading? Current Clinical Pharmacology 2013;8(2):115-23
3. Duke T, Fuller D. Randomised controlled trials in child health in developing countries: trends
SC
and lessons over 11 years. Archives of Disease in Childhood 2014;99(7):615-20.
4. Dandona L, Katoch VM, Dandona R. Research to achieve health care for all in India. Lancet
M AN U
2011;377(9771):1055-7
5. Drain PK, Robine M, Holmes KK, et al. Trail watch: global migration of clinical trials. Nature reviews. Drug Discovery 2014;13(3):166-7
6. Zhang D, Freemantle N, Cheng KK. Are randomized trials conducted in China or India
2011;64(1):90-5
TE D
biased? A comparative empirical analysis. Journal of Clinical Epidemiology
7. Hyder AA, Wali SA, Khan AN, et al. Ethical review of health research: a perspective from
EP
developing country researchers. Journal of Medical Ethics 2004;30(1):68-72 8. Zhang D, Yin P, Freemantle N, et al. An assessment of the quality of randomised controlled
AC C
trials conducted in China. Trials 2008;9:22
9. Division of special pathogen and immunologic drug products. Summary of clinical review of studies
submitted
in
response
to
a
pediatric
written
request.
http://www.fda.gov/ohrms/dockets/ac/05/briefing/20054152b1_03_03_Cipro%20medical.pdf Accessed 1/21/2015
21
ACCEPTED MANUSCRIPT
10. Panagiotou OA, Contopoulos-Ioannidis DG, Ioannidis JP. Comparative effect sizes in randomised trials from less developed and more developed countries: metaepidemiological assessment. BMJ 2013;346:f707
RI PT
11. Lorenzo C, Garrafa V, Solbakk JH, et al. Hidden risks associated with clinical trials in developing countries. Journal of Medical Ethics 2010;36(2):111-5
2002;21(19):2869-77
SC
12. Smith R. Publishing research from developing countries. Statistics in Medicine
13. Juni P, Holenstein F, Sterne J, et al. Direction and impact of language bias in meta-analyses trials:
empirical
study.
International
M AN U
of controlled
2002;31(1):115-23
Journal
of Epidemiology
14. Duijnhoven RG, Straus SM, Raine JM, et al. Number of patients studied prior to approval of new medicines: a database analysis. PLoS Medicine 2013;10(3):e1001407
TE D
15. Tsang R, Colley L, Lynd LD. Inadequate statistical power to detect clinically significant differences in adverse event rates in randomized controlled trials. Journal of Clinical Epidemiology 2009;62(6):609-16
gs-r,
EP
16. United Nations Statistics Division. Composition of macro geographical (continental) regions and
selected
economic
and
other
groupings.
AC C
2013.http://unstats.un.org/unsd/methods/m49/m49regin.htm.
17. Platonov P. Clinical trials in Russia and Eastern Europe: recruitment and quality. International Journal of Clinical Pharmacology and Therapeutics 2003;41(7):277-80
18. Lathyris D, Panagiotou OA, Baltogianni M, et al. Safety of medical interventions in children versus adults. Pediatrics 2014;133(3):e666-73
22
ACCEPTED MANUSCRIPT
19. Contopoulos-Ioannidis DG, Baltogianni MS, Ioannidis JP. Comparative effectiveness of medical
interventions
in
adults
versus
children.
The Journal
of Pediatrics
20. http://handbook.cochrane.org/ ChfsroiTCCVuM.
RI PT
2010;157(2):322-30
21. Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews. Annals of Internal Medicine 1997;127(9):820-6
SC
22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials 1986;7(3):177-88
M AN U
23. Ioannidis JP, Patsopoulos NA, Evangelou E. Uncertainty in heterogeneity estimates in metaanalyses. BMJ 2007;335(7626):914-6
24. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine 2002;21(11):1539-58
1950;21:607-11
TE D
25. Freeman MF TJ. Transformations related to the angular and the square root. Ann Math Statist
26. Harbord RM EM, Sterne JA. . A modified test for small-study effects in meta-analyses of
EP
controlled trials with binary endpoints. Statistics in Medicine 2006;25:3443-57. 27. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic
AC C
reviews and meta-analyses of studies that evaluate health care interventions: explanation
and elaboration. Annals of Internal Medicine 2009;151(4):W65-94
28. Ioannidis JP, Evans SJ, Gotzsche PC, et al. Better reporting of harms in randomized trials: an extension of the CONSORT statement. Annals of Internal Medicine 2004;141(10):781-8
23
ACCEPTED MANUSCRIPT
29. Papanikolaou PN, Ioannidis JP. Availability of large-scale evidence on specific harms from systematic reviews of randomized trials. The American Journal of Medicine 2004;117(8):582-9
medical areas. JAMA 2001;285(4):437-43
RI PT
30. Ioannidis JP, Lau J. Completeness of safety reporting in randomized trials: an evaluation of 7
31. Ioannidis JP. Adverse events in randomized trials: neglected, restricted, distorted, and
SC
silenced. Archives of Internal Medicine 2009;169(19):1737-9
32. Mukherjee SD, Coombes ME, Levine M, et al. A qualitative study evaluating causality
M AN U
attribution for serious adverse events during early phase oncology clinical trials. Investigational New Drugs 2011;29(5):1013-20 .
33. Wahab IA, Pratt NL, Kalisch LM, et al. The detection of adverse events in randomized clinical trials: can we really say new medicines are safe? Current Drug Safety
TE D
2013;8(2):104-13
34. Schroll JB, Maund E, Gotzsche PC. Challenges in coding adverse events in clinical trials: a systematic review. PloS One 2012;7(7):e41174.
EP
35. Huang HY, Andrews E, Jones J, et al. Pitfalls in meta-analyses on adverse events reported from clinical trials. Pharmacoepidemiology and Drug Safety 2011;20(10):1014-20
AC C
36. Haidich AB, Birtsou C, Dardavessis T, et al. The quality of safety reporting in trials is still suboptimal: survey of major general medical journals. Journal of Clinical Epidemiology 2011;64(2):124-35
37. de Vries TW, van Roon EN. Low quality of reporting adverse drug reactions in paediatric randomised controlled trials. Archives of Disease in Childhood 2010;95(12):1023-6
24
ACCEPTED MANUSCRIPT
38. Personal communication with Dr N. Balasubramanian from the Department of Statistics at Stanford University 39. United States of America, ex. rel. Juan N. Walterspiel MD, FAAP v. Bayer A.G., Quintiles False Claims Act action under 31 U.S.C. § 3729(a).
AC C
EP
TE D
M AN U
SC
https://www.pacer.gov/ Accessed 01/18/2015.
RI PT
Transnational Corp.
25
ACCEPTED MANUSCRIPT
Table 1: Harm-related endpoints
CD000029
CD000213
SC
Description of Harm- N of RCTs related endpoint (N patients) from more developed countries Anticoagulants for acute Anticoagulant vs Major extracranial 15 (2,506) ischaemic stroke control haemorrhage during treatment period Antiplatelet therapy for Antiplatelet vs Major extracranial 6 (796) acute ischaemic stroke control acute haemorrhage during treatment period Thrombolysis for acute Any thrombolytic Symptomatic (including 20 (6,244) ischaemic stroke agent vs control fatal) intracranial haemorrhage within 7 to 10 days Platelet glycoprotein Platelet 30-day major bleeding 20 (23,259) IIb/IIIa blockers during glycoprotein (all patients) percutaneous coronary IIb/IIIa blockers intervention and as the during PCI vs initial medical treatment control of non-ST segment elevation acute coronary syndromes Terlipressin for acute Terlipressin vs Adverse events causing 3 (204) esophageal variceal vasopressin death hemorrhage Phosphodiesterase 4 PDE4 inhibitors vs Non fatal serious 8 (2,965) inhibitors for chronic placebo adverse events
N of RCTs ROR (95% CIs) (N patients) from less developed countries 2 (344) 2.48 (0.22-27.75)
1 (21,106)
0.67 (0.14-3.09)
2 (567)
1.97 (0.57-6.85)
3 (448)
2.39 (0.30-19.03)
2 (101)
0.54 (0.02-19.03)
1 (1,018)
0.38 (0.18-0.83)
CD002147
CD002309
AC C
EP
CD002130
Comparison
M AN U
CD000024
Title
TE D
Review
RI PT
Severe adverse events
26
ACCEPTED MANUSCRIPT
CD003897b
CD008521a
CD008521b
AC C
Discontinuations due to adverse events
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.04)
CMV 3 (881)
1 (118)
0.93 (0.12-7.26)
Infection: serious all- 6 (1,009) cause total
2 (132)
1.08 (0.39-2.98)
Serious adverse events
18 (29,581)
0.88 (0.63-1.25)
7 (7,863)
0.81(0.45-1.46)
RI PT
5 (266)
Infection: invasive
SC
Intraventricular haemorrhage (III-IV)
M AN U
CD003897a
1 (29)
TE D
CD003481
Major infection
All serious events
7 (3,179)
adverse 3 (2,547)
EP
CD002922
obstructive pulmonary disease Treatment for lupus Immunosuppressiv nephritis e agent plus steroids vs steroids alone Ibuprofen for the Ibuprofen vs treatment of patent indomethacin ductus arteriosus in preterm and/or low birth weight infants Interleukin 2 receptor IL2Ra vs placebo antagonists for kidney or no treatment transplant recipients Interleukin 2 receptor IL2Ra vs placebo antagonists for kidney or no treatment transplant recipients Vaccines for preventing RV1 vs placebo rotavirus diarrhoea: vaccines in use Vaccines for preventing RV5 versus placebo rotavirus diarrhoea: vaccines in use
Review
Title
Comparison
Description of Harmrelated endpoint
CD000171a
Treatment of latent Any TB drug vs Incidence of adverse tuberculosis infection in placebo events leading to
N of RCTs (N patients) from more developed countries 2 (838)
N of RCTs ROR (95% CIs) (N patients) from less developed countries 6 (4,695) 1.05 (0.04-26.24)
27
ACCEPTED MANUSCRIPT
CD002863
CD004386a
CD004386b
CD004773
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
3 (206)
2 (103)
0.22 (0.01-5.92)
7 (2,891)
1 (1,018)
0.61 (0.27-1.36)
1 (154)
1.22 (0.19-7.67)
1 (97)
2.11 (0.04-113.13)
1 (102)
4.03 (0.40-40.29)
1 (129)
0.74 (0.07-8.17)
RI PT
SC
PDE4 inhibitors vs placebo
1 (1,092)
M AN U
CD002309
3 (509)
Beta-blocker vs Changed drugs due to 10 (923) placebo/no beta- maternal side effects blocker
TE D
CD002147
stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Adverse events causing withdrawal of treatment Withdrawals due to adverse effects
Drug vs placebo/no Side effects requiring 16 (2,057) intervention discontinuation
EP
CD000171c
Isoniazid vs isoniazid + rifampicin Isoniazid vs rifampicin + pyrazinimide Terlipressin versus vasopressin
Quinolone vs TMP- Side effects requiring 6 (752) SMZ discontinuation
AC C
CD000171b
HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Antifungal Placebo
vs Discontinuation 2 (723) secondary to adverse events
28
ACCEPTED MANUSCRIPT
CD008521b
RV5 vs placebo
Vaccines for preventing RV1 vs placebo rotavirus diarrhoea: vaccines in use
Any adverse event
CD003807
CD004183
Adverse requiring discontinuation follow-up) Adverse requiring discontinuation follow-up)
to 8 (1,652)
events 5 (3,709)
3 (236)
0.61 (0.13-2.93)
4 (7,735)
1.00 (0.21-4.73)
15 (21,774)
1.92 (0.63-5.87)
(end of
events 6 (2,184)
(end of
Description of Harm- N of RCTs related endpoint (N patients) from more developed countries Phosphodiesterase 4 PDE4 inhibitors vs No of patients 8 (2,913) inhibitors for chronic placebo experiencing an obstructive pulmonary adverse effect disease Interventions for All drug types vs Toxicity (adverse 2 (661) preventing oral placebo/no events 'probably due to candidiasis for patients treatment drug') with cancer receiving treatment Antioxidant Antioxidants (beta- Adverse effects 3 (91,080) supplements for carotene) vs preventing placebo
TE D
Comparison
EP
CD002309
Title
AC C
Review
vs Withdrawal due adverse events
RI PT
Statins placebo/no treatment
SC
CD008521a
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use
M AN U
CD007784
N of RCTs ROR (95% CIs) (N patients) from less developed countries 1 (1,018) 1.21 (0.73-2.00)
1 (210)
1.47 (0.44-4.90)
1 (174)
0.88 (0.03-23.17)
29
ACCEPTED MANUSCRIPT
CD006536
CD008076
9 (929)
1 (102)
1.15 (0.36-3.63)
vs Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Adverse events
6(450)
1 (102)
0.75 (0.02-24.58)
8 (669)
2 (213)
0.06 (0.02-0.20)
1 (787)
0.56 (0.31-1.00)
RI PT
1.78 (0.03-92.13)
Antibiotic prophylaxis placebo/ treatment
SC
Antifungal placebo
vs no
M AN U
CD004791
1 (97)
TE D
CD004773
33 (5,002)
Quinolone vs TMP- Side effects SMZ
Stem cells vs no Adverse effects stem cells
EP
CD004386b
Drug vs placebo/ Side effects no intervention
Cilostazol vs Safety outcomes 1 (3,202) aspirin in patients during follow-up with ischaemic stroke or TIA
AC C
CD004386a
gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction Cilostazol versus aspirin for secondary prevention of vascular events after stroke of arterial origin
Combined Primary-Harms (Serious Adverse event, Discontinuations due to adverse event and Any adverse event)
30
ACCEPTED MANUSCRIPT
CD000171a
Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Thrombolysis for acute ischaemic stroke
CD000171c
CD000213
CD002130
RI PT
Antiplatelet therapy for acute ischaemic stroke
SC
CD000029
M AN U
Anticoagulants for acute ischaemic stroke
Description of Harm- N of RCTs (N patients) related endpoint from more developed countries Anticoagulant vs Major extracranial 15 (2,506) control in acute haemorrhage during presumed treatment period ischaemic stroke Antiplatelet vs Major extracranial 6 (796) control in acute haemorrhage during presumed treatment period ischaemic stroke Any TB drug vs Incidence of adverse 2 (838) placebo events leading to stopping treatment Isoniazid vs Incidence of adverse 3 (509) isoniazid + events leading to rifampicin stopping treatment Isoniazid vs Incidence of adverse 2 (373) rifampicin + events leading to pyrazinamide stopping treatment Any thrombolytic Symptomatic 20 (5,622) agent vs control (including fatal) intracranial haemorrhage within 7 to 10 days Platelet 30-day major bleeding 20 (23,259) glycoprotein IIb/IIIa blockers vs placebo/usual care (all patients)
TE D
CD000024
CD000171b
Comparison
EP
Title
AC C
Review
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST
N of RCTs ROR (95% CIs) (N patients) from less developed countries 2 (344) 2.48 (0.22-27.75)
1 (21,106)
0.67 (0.14-3.09)
6 (4,695)
1.05 (0.04-26.24)
1 (1,092)
4.59 (0.70-29.98)
2 (1,455)
0.53 (0.19-1.52)
2 (567)
1.97 (0.57-6.85)
3 (448)
2.39 (0.30-19.03)
31
ACCEPTED MANUSCRIPT
CD002922
CD003481
1 (1,018)
1.21 (0.73-2.00)
1 (154)
1.22 (0.19-7.67)
5 (266)
1 (29)
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.03)
2 (661 )
1 (210)
1.47 (0.44-4.90)
6 (1,009)
2 (132)
1.08 (0.39-2.98)
3 (91,080)
1 (174)
0.88 (0.03-23.17)
RI PT
0.22 (0.01-5.88)
Beta-blocker vs Changed drugs due to 10 (923) placebo/no beta- maternal side effects blocker
Immunosuppressive Major infection agent plus steroids vs steroids alone Ibuprofen for the Ibuprofen vs Intraventricular treatment of patent indomethacin haemorrhage (III-IV) ductus arteriosus in preterm and/or low birth weight infants Interventions for Comparisons with Toxicity (adverse preventing oral placebo/no events 'probably due to candidiasis for patients treatment for all drug') with cancer receiving drug types treatment Interleukin 2 receptor IL2Ra vs placebo Infection: serious allantagonists for kidney or no treatment cause total transplant recipients Antioxidant Antioxidants vs Adverse effects - betasupplements for placebo carotene preventing gastrointestinal cancers
CD003897
CD004183
AC C
EP
CD003807
2 (101)
SC
CD002863
vs Adverse events causing 3 (204) withdrawal of treatment PDE4 inhibitors vs No of patients 8(2,913) placebo experiencing an adverse effect
M AN U
CD002309
Terlipressin vasopressin
TE D
CD002147
segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Treatment for lupus nephritis
32
ACCEPTED MANUSCRIPT
CD006536
CD007784
CD008076
1.78 (0.03-92.13)
9 (929)
1 (102)
1.15 (0.36-3.63)
1 (295)
1 (129)
1.44 (0.60-3.46)
6 (450)
1 (102)
0.75 (0.02-24.58)
6 (617)
2 (213)
0.06 (0.02-0.18)
to 8 (1,652)
3 (236)
0.61 (0.13-2.93)
Cilostazol vs Other outcomes of 1 (3,202) aspirin in patients safety during followwith ischaemic up. stroke or TIA
1 (787)
0.56 (0.31-1.00)
RI PT
33 (5,002)
Antifungal Placebo
Antibiotic prophylaxis placebo or treatment
SC
Quinolone vs TMP- Side effects SMZ
vs Adverse events
M AN U
CD004791
1 (97)
Adverse events
vs no
TE D
CD004773
Drug vs placebo/ no Side effects intervention
Stem cells Adverse effects compared to no stem cells, Statins vs Withdrawal due placebo/no adverse events treatment
EP
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular after stroke of arterial
AC C
CD004386a
33
ACCEPTED MANUSCRIPT
RI PT
1.00 (0.21-4.73)
18 (29,581)
0.88 (0.63-1.25)
TE D
M AN U
SC
Vaccines for preventing RV1 vs placebo rotavirus diarrhoea: vaccines in use
4 (7,7733)
EP
CD008521b
Adverse events 5 (3,708) requiring discontinuation (end of follow-up) Serious adverse events 7 (3,179)
AC C
CD008521a
origin Vaccines for preventing RV5 vs placebo rotavirus diarrhoea: vaccines in use
34
ACCEPTED MANUSCRIPT
Table 2: Individual meta-analyses with nominally significant relative differences between more developed and less developed
PDE4 inhibitors placebo
sOR_MD (95% sOR_LD CIs) CIs)
(95%
0.75 (0.55-1.03)
1.96 (0.96-3.98)
0.06 (0.02-0.20)
0.84 (0.48-1.45)
13.01 (4.90-34.57)
Any zidovudine vs Neonatal placebo/no treatment haematological toxicity
2.77 (1.23-6.22)
2.09 (1.22-3.59)
0.76 (0.41-1.38)
0.003 (0.001-0.06)
1.15 (0.34-3.81)
441.20 (24.93-7808)
Third generation Diarrhoea cephalosporins versus conventional therapy
3.92 (1.35-11.34)
2.50 (1.54-4.05)
0.64 (0.25-1.64)
Rotarix vs after dose 2
0.66 (0.44-0.99)
0.68 (0.47-0.98)
1.03 (0.86-1.24)
M AN U
Stem cells compared to Adverse effects no stem cells,
TE D
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease CD006536 Stem cell treatment for acute myocardial infarction CD000102 Interventions for reducing the risk of mother-to-child transmission of HIV infection CD001556c Preoperative chemotherapy for resectable thoracic esophageal cancer CD001832 Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis CD008521b Vaccines for preventing rotavirus diarrhoea: vaccines in use
Description of ROR (95% CIs) Harm-related endpoint vs Non fatal 0.38 (0.18-0.83) serious adverse events
SC
Comparison
Preoperative chemotherapy surgery alone
AC C
CD002309
Title
Gastrointestinal vs complications
EP
Review
RI PT
countries
placebo: Vomitingreactogenicity
35
ACCEPTED MANUSCRIPT
Abbreviations: ROR: relative odds ratio; sOR_MD: summary odds ratio in more developed countries; sROR_LD: summary odds
AC C
EP
TE D
M AN U
SC
RI PT
ratio in less developed countries; CIs: confidence intervals; Footnote: ORs and RORs were calculated by random effects models
36
ACCEPTED MANUSCRIPT
Figure 1: Flow Chart
RI PT
Screened 131 Systematic Reviews
SC
Excluded 89 Systematic Reviews No adverse events (n=40) No pharmacologic interventions (n=30) No More developed and Less developed countries (n=18) No country information (n=1)
Discontinuations due to Adverse Events: 12 ma; 108 RCTs (70 from MD and 38 from LD countries)
TE D
Severe Adverse Events: 12 ma; 143 RCTs (99 from MD and 44 from LD countries)
M AN U
Included 42 eligible Systematic Reviews; 128 meta-analyses on Harm-related endpoints with 1186 study-level analyses from 474 independent RCTs (337 from MD and 137 from LD countries)a b
AC C
EP
Organ System Specific Adverse Events: 5 ma; 20 RCTs (11 from MD and 9 from LD countries)
Individual adverse events (n=60) 14 adverse events with ≥2 ma (43 ma) 46 adverse events with 1 ma (46 ma)
Any Adverse Event: 9 ma; 81 RCTs (71 from MD and 10 from LD countries)
All Discontinuations: 1 ma; 7 RCTs (6 from MD and 1 from LD countries)
89 ma, 767 RCTS corresponding to 311 independent RCTs (222 from MD and 89 from LD countries)
Footnotes: a
Each RCT could have contributed data to more than one study-level analyses, for the different harm-related endpoints.
b
We excluded from these 128 meta-analyses data from 47 RCTs that were from mixed-country settings (n=38); 37Eastern European countries (n=4) or did not report country (n=5).
ACCEPTED MANUSCRIPT
Table 1a: Characteristics of Meta-analyses and Endpoints for Severe Adverse Events Severe adverse events Review Title
Comparison
Description endpoint
CD000024
Anticoagulants for acute ischaemic stroke Antiplatelet therapy for acute ischaemic stroke Thrombolysis for acute ischaemic stroke
Anticoagulant vs control
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 (PDE4) inhibitors for chronic obstructive pulmonary disease Treatment for lupus nephritis
Platelet glycoprotein IIb/IIIa blockers during PCI vs control
Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding (all patients)
CD002309
CD002922 CD003481
CD003897a CD003897b CD008521a CD008521b
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interleukin 2receptor antagonists (IL2Ra) for kidney transplant recipients Interleukin 2 receptor antagonists (IL2Ra) for kidney transplant recipients Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
1 (21,106)
0.67 (0.14-3.09)
20 (6,244)
2 (567)
1.97 (0.57-6.85)
20 (23,259)
3 (448)
2.39 (0.30-19.03)
Adverse events causing death
3 (204)
2 (101)
0.54 (0.02-19.03)
Non fatal serious adverse events
8 (2,965)
1 (1,018)
0.38 (0.18-0.83)
Major infection
5 (266)
1 (29)
0.33 (0.02-4.53)
Intraventricular haemorrhage (III-IV)
3 (285)
4 (261)
1.75 (0.51-6.04)
IL2Ra vs placebo or no treatment
Infection: CMV invasive
3 (881)
1 (118)
0.93 (0.12-7.26)
IL2Ra vs placebo or no treatment
Infection: serious all-cause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
RV1 vs placebo
Serious adverse events
7 (3,179)
18 (29,581)
0.88 (0.63-1.25)
RV5 versus placebo
All serious adverse events
3 (2,547)
7 (7,863)
0.81(0.45-1.46)
Terlipressin vs vasopressin PDE4 inhibitors vs placebo
SC
Any thrombolytic agent vs control
RI PT
6 (796)
Antiplatelet vs control acute
Harm-related
M AN U
CD002147
ROR (95% CIs)
TE D
CD002130
N of RCTs (N patients) from less developed countries 2 (344)
Immunosuppressive agent steroids vs steroids alone Ibuprofen vs indomethacin
EP
CD000213
N of RCTs (N patients) from more developed countries 15 (2,506)
AC C
CD000029
of
plus
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); RV5: pentavalent rotavirus vaccine; RV1: monovalent rotavirus vaccine; TB: tuberculosis;vs: versus
2.48 (0.22-27.75)
ACCEPTED MANUSCRIPT
Table 1b: Characteristics of Meta-analyses and Endpoints for Discontinuations due to Adverse Events Discontinuations due to adverse events Review Title
Comparison
Description endpoint
CD000171a
Any TB drug vs placebo
Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Adverse events causing withdrawal of treatment Withdrawals due to adverse effects
CD002863 CD004386a
CD004386b
CD004773
CD007784
CD008521a CD008521b
PDE4 inhibitors vs placebo
SC
RI PT
Harm-related
M AN U
CD002309
+
N of RCTs (N patients) from more developed countries 2 (838)
N of RCTs (N patients) from less developed countries 6 (4,695)
ROR (95% CIs)
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
3 (206)
2 (103)
0.22 (0.01-5.92)
7 (2,891)
1 (1,018)
0.61 (0.27-1.36)
1.05 (0.04-26.24)
Beta-blocker vs placebo/no betablocker Drug vs placebo/no intervention
Changed drugs due to maternal side effects Side effects requiring discontinuation
10 (923)
1 (154)
1.22 (0.19-7.67)
16 (2,057)
1 (97)
2.11 (0.04-113.13)
Quinolone vs TMP-SMZ
Side effects requiring discontinuation
6 (752)
1 (102)
4.03 (0.40-40.29)
Antifungal vs Placebo
Discontinuation secondary to adverse events
2 (723)
1 (129)
0.74 (0.07-8.17)
Statins vs placebo/no treatment
Withdrawal due to adverse events
8 (1,652)
3 (236)
0.61 (0.13-2.93)
RV5 vs placebo
Adverse events requiring discontinuation (end of follow-up) Adverse events requiring discontinuation (end of follow-up)
5 (3,709)
4 (7,735)
1.00 (0.21-4.73)
6 (2,184)
15 (21,774)
1.92 (0.63-5.87)
TE D
CD002147
Isoniazid vs rifampicin pyrazinimide Terlipressin versus vasopressin
EP
CD000171c
Isoniazid vs isoniazid + rifampicin
AC C
CD000171b
Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4(PDE4) inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
of
RV1 vs placebo
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); RV5: pentavalent rotavirus vaccine; RV1: monovalent rotavirus vaccine; TB: tuberculosis; TMP-SMX: trimethoprim-sulfamethoxazole; vs: versus
ACCEPTED MANUSCRIPT
Table 1c: Characteristics of Meta-analyses and Endpoints for Any adverse event Any adverse event Review Title
Comparison
Description endpoint
CD002309
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
All drug types vs placebo/no treatment
CD004773
CD004791
CD006536 CD008076
2 (661)
1 (210)
1.47 (0.44-4.90)
Antioxidants (beta-carotene) vs placebo Drug vs placebo/ no intervention
Adverse effects
3 (91,080)
1 (174)
0.88 (0.03-23.17)
Side effects
33 (5,002)
1 (97)
1.78 (0.03-92.13)
Quinolone vs TMP-SMZ
Side effects
9 (929)
1 (102)
1.15 (0.36-3.63)
Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Antibiotic prophylaxis vs placebo/ no treatment
Adverse events
6(450)
1 (102)
0.75 (0.02-24.58)
Stem cells vs no stem cells
Adverse effects
8 (669)
2 (213)
0.06 (0.02-0.20)
Safety outcomes during follow-up
1 (3,202)
1 (787)
0.56 (0.31-1.00)
SC
RI PT
Toxicity (adverse events 'probably due to drug')
Antifungal vs placebo
Harm-related
M AN U
CD004386b
ROR (95% CIs)
TE D
CD004386a
N of RCTs (N patients) from less developed countries 1 (1,018)
EP
CD004183
N of RCTs (N patients) from more developed countries 8 (2,913)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
AC C
CD003807
Phosphodiesterase 4 (PDE4) inhibitors for chronic obstructive pulmonary disease Interventions for preventing oral candidiasis for patients with cancer receiving treatment Antioxidant supplements for preventing gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastro-intestinal bleeding Stem cell treatment for acute myocardial infarction Cilostazol versus aspirin for secondary prevention of vascular events after stroke of arterial origin
of
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); TIA: transient ischemic attack; TMP-SMX: trimethoprim-sulfamethoxazole; vs: versus
1.21 (0.73-2.00)
ACCEPTED MANUSCRIPT
Table 1d: Characteristics of Meta-analyses and Endpoints for Combined Harms (Severe adverse events; Discontinuations due to adverse events or Any adverse event) Combined Primary-Harms (Serious Adverse event, Discontinuations due to adverse event and Any adverse event) Review Title Comparison Description of endpoint
CD000171c CD000213
CD002130
CD002147 CD002309
CD002863 CD002922 CD003481
CD003807
CD003897
1 (21,106)
0.67 (0.14-3.09)
2 (838)
6 (4,695)
1.05 (0.04-26.24)
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
20 (5,622)
2 (567)
1.97 (0.57-6.85)
20 (23,259)
3 (448)
2.39 (0.30-19.03)
Adverse events causing withdrawal of treatment No of patients experiencing an adverse effect
3 (204)
2 (101)
0.22 (0.01-5.88)
8(2,913)
1 (1,018)
1.21 (0.73-2.00)
Beta-blocker vs placebo/no betablocker Immunosuppressive agent plus steroids vs steroids alone Ibuprofen vs indomethacin
Changed drugs due to maternal side effects Major infection
10 (923)
1 (154)
1.22 (0.19-7.67)
5 (266)
1 (29)
0.33 (0.02-4.53)
Intraventricular haemorrhage (III-IV)
3 (285)
4 (261)
1.75 (0.51-6.03)
Comparisons with placebo/no treatment for all drug types
Toxicity (adverse events 'probably due to drug')
2 (661 )
1 (210)
1.47 (0.44-4.90)
IL2Ra vs placebo or no treatment
Infection: serious all-cause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 (PDE4) inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Treatment for lupus nephritis
Platelet glycoprotein IIb/IIIa blockers vs placebo/usual care (all patients)
Isoniazid vs isoniazid + rifampicin
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interventions for preventing oral candidiasis for patients with cancer receiving treatment Interleukin 2 (IL 2) receptor antagonists for kidney transplant recipients
Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding
SC
Anticoagulant vs control in acute presumed ischaemic stroke Antiplatelet vs control in acute presumed ischaemic stroke Any TB drug vs placebo
RI PT
6 (796)
Anticoagulants for acute ischaemic stroke Antiplatelet therapy for acute ischaemic stroke Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Thrombolysis for acute ischaemic stroke
M AN U
CD000171b
ROR (95% CIs)
Isoniazid vs rifampicin + pyrazinamide Any thrombolytic agent vs control
TE D
CD000171a
N of RCTs (N patients) from less developed countries 2 (344)
Terlipressin vs vasopressin
PDE4 inhibitors vs placebo
EP
CD000029
N of RCTs (N patients) from more developed countries 15 (2,506)
AC C
CD000024
Harm-related
2.48 (0.22-27.75)
ACCEPTED MANUSCRIPT
CD004791
CD006536 CD007784
CD008076
CD008521a CD008521b
3 (91,080)
1 (174)
0.88 (0.03-23.17)
Drug vs placebo/ no intervention
Side effects
33 (5,002)
1 (97)
1.78 (0.03-92.13)
Quinolone vs TMP-SMZ
Side effects
9 (929)
1 (102)
1.15 (0.36-3.63)
Antifungal vs Placebo
Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Antibiotic prophylaxis vs placebo or no treatment
Adverse events
6 (450)
1 (102)
0.75 (0.02-24.58)
Stem cells compared to no stem cells, Statins vs placebo/no treatment
Adverse effects
6 (617)
2 (213)
0.06 (0.02-0.18)
Withdrawal due to adverse events
8 (1,652)
3 (236)
0.61 (0.13-2.93)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
Other outcomes of safety during follow-up.
1 (3,202)
1 (787)
0.56 (0.31-1.00)
RV5 vs placebo
Adverse events requiring discontinuation (end of follow-up) Serious adverse events
5 (3,708)
4 (7,7733)
1.00 (0.21-4.73)
7 (3,179)
18 (29,581)
0.88 (0.63-1.25)
RV1 vs placebo
RI PT
CD004773
Adverse effects - beta-carotene
SC
CD004386b
Antioxidants vs placebo
M AN U
CD004386a
Antioxidant supplements for preventing gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastro-intestinal bleeding Stem cell treatment for acute myocardial infarction HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular after stroke of arterial origin Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
TE D
CD004183
AC C
EP
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); RV5: pentavalent rotavirus vaccine; RV1: monovalent rotavirus vaccine; TB: tuberculosis; TIA: transient ischemic attack; TMP-SMX: trimethoprim-sulfamethoxazole; vs: versus
ACCEPTED MANUSCRIPT
Table 2: Individual meta-analyses with nominally significant relative differences between more developed and less developed
ROR (95% CIs)
sOR_MD (95% sOR_LD (95% CIs) CIs)
0.38 (0.18-0.83)
0.75 (0.55-1.03)
1.96 (0.96-3.98)
Stem cells compared to no stem cells,
Adverse effects
0.06 (0.02-0.20)
0.84 (0.48-1.45)
13.01 (4.90-34.57)
Any zidovudine vs placebo/no treatment
Neonatal haematological toxicity
2.77 (1.23-6.22)
2.09 (1.22-3.59)
0.76 (0.41-1.38)
Gastrointestinal complications
0.003 (0.001-0.06)
1.15 (0.34-3.81)
441.20 (24.93-7808)
Third generation cephalosporins versus conventional therapy
Diarrhoea
3.92 (1.35-11.34)
2.50 (1.54-4.05)
0.64 (0.25-1.64)
Rotarix vs placebo: after dose 2
Vomitingreactogenicity
0.66 (0.44-0.99)
0.68 (0.47-0.98)
1.03 (0.86-1.24)
M AN U
PDE4 inhibitors vs placebo
TE D
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease CD006536 Stem cell treatment for acute myocardial infarction CD000102 Interventions for reducing the risk of mother-to-child transmission of HIV infection CD001556c Preoperative chemotherapy for resectable thoracic esophageal cancer CD001832 Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis CD008521b Vaccines for preventing rotavirus diarrhoea: vaccines in use
Description of Harm-related endpoint Non fatal serious adverse events
SC
Comparison
Preoperative chemotherapy vs surgery alone
AC C
CD002309
Title
EP
Review
RI PT
countries
ACCEPTED MANUSCRIPT
Abbreviations: ROR: relative odds ratio; sOR_MD: summary odds ratio in more developed countries; sROR_LD: summary odds
AC C
EP
TE D
M AN U
SC
RI PT
ratio in less developed countries; CIs: confidence intervals; Footnote: ORs and RORs were calculated by random effects models
ACCEPTED MANUSCRIPT
Figure 1: Flow Chart
RI PT
Screened 131 Systematic Reviews
SC
Excluded 89 Systematic Reviews No adverse events (n=40) No pharmacologic interventions (n=30) No More developed and Less developed countries (n=18) No country information (n=1)
Discontinuations due to Adverse Events: 12 ma; 108 RCTs (70 from MD and 38 from LD countries)
TE D
Severe Adverse Events: 12 ma; 143 RCTs (99 from MD and 44 from LD countries)
M AN U
Included 42 eligible Systematic Reviews; 128 meta-analyses on Harm-related endpoints with 1186 study-level analyses from 474 independent RCTs (337 from MD and 137 from LD countries)a b
AC C
EP
Organ System Specific Adverse Events: 5 ma; 20 RCTs (11 from MD and 9 from LD countries)
Individual adverse events (n=60) 14 adverse events with ≥2 ma (43 ma) 46 adverse events with 1 ma (46 ma)
Any Adverse Event: 9 ma; 81 RCTs (71 from MD and 10 from LD countries)
All Discontinuations: 1 ma; 7 RCTs (6 from MD and 1 from LD countries)
89 ma, 767 RCTS corresponding to 311 independent RCTs (222 from MD and 89 from LD countries)
Footnotes: a
Each RCT could have contributed data to more than one study-level analyses, for the different harm-related endpoints.
b
We excluded from these 128 meta-analyses data from 47 RCTs that were from mixed-country settings (n=38); Eastern European countries (n=4) or did not report country (n=5).
S0895-4356(16)30028-2
DOI:
10.1016/j.jclinepi.2016.02.032
Reference:
JCE 9143
To appear in:
Journal of Clinical Epidemiology
Received Date: 23 April 2015 Revised Date:
4 January 2016
Accepted Date: 4 February 2016
Please cite this article as: Contopoulos-Ioannidis D, Tseretopoulou X, Ancker M, Walterspiel JN, Panagiotou OA, Maldonado Y, Ioannidis J, Comparative Rates of Harms in Randomized Trials from More-Developed Versus Less- Developed Countries May be Different, Journal of Clinical Epidemiology (2016), doi: 10.1016/j.jclinepi.2016.02.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
*Manuscript (with PAGE numbers)
ACCEPTED MANUSCRIPT
Comparative Rates of Harms in Randomized Trials from More-Developed Versus LessDeveloped Countries May be Different
1 2 3
, Xanthippi Tseretopoulou,
RI PT
Despina Contopoulos-Ioannidis, clinical associate professor
registar in paediatrics 4, Megan Ancker, field physician 5, Juan N. Walterspiel, pediatric infectious diseases consultant
6
, Orestis A Panagiotou, post-doctoral fellow
Department of Pediatrics, Division of Infectious Diseases, Stanford University School of
Medicine, Stanford, CA, USA
M AN U
1
2
Palo Alto Medical Foundation Research Institute, Palo Alto, CA, USA
3
Meta-Research Innovation Center at Stanford, Stanford, CA, USA
6
7
TE D
5
Leeds Teaching Hospital, NHS Trust, Leeds, UK Medecins San Frontieres, Paris, France
Mendocino Coast District Hospital, Fort Bragg CA, USA Division of Cancer Epidemiology & Genetics, National Cancer Institute, National Institute of
EP
4
, Yvonne
SC
Maldonado, professor 1 8 9, John PA Ioannidis, professor 3 8 10 11
7
Health, Bethesda, MD, USA
Department of Health Research and Policy, Stanford University School of Medicine, Stanford,
AC C
8
CA, USA 9
Senior Associate Dean for Faculty Development and Diversity, Stanford University School of
Medicine, Stanford, CA, USA 10
Stanford Prevention Research Center, Department of Medicine, Stanford University School of
Medicine, Stanford, CA, USA
1
ACCEPTED MANUSCRIPT
11
Department of Statistics, Stanford University School of Humanities and Sciences, Stanford,
CA, USA Address correspondence to: Despina Contopoulos-Ioannidis, MD; Department of Pediatrics,
RI PT
Division of Infectious Diseases, Stanford University School of Medicine, 300 Pasteur Drive,
AC C
EP
TE D
M AN U
SC
Room G312, Stanford CA, 94305, USA; e-mail: [email protected]
2
ACCEPTED MANUSCRIPT
Funding: There was no funding for this study. Conflicts of interest: There are no conflicts of interest to declare for DCI, XT, MA, OAP, YM and JPAI. JNW worked as a consultant for US Bayer A.G. in 2002-3. “Whistleblower" on Bayer
that are paid for directly or indirectly by Bayer A.G.
RI PT
A.G. Currently does not hold any Bayer A.G or subsidiarie’s stock, nor does he give any lectures
The views expressed in this paper are solely those of the authors and do not necessarily represent
SC
the official views of the National Institutes of Health or the US Department of Health and
AC C
EP
TE D
M AN U
Human Services.
3
ACCEPTED MANUSCRIPT
Abstract Objectives: We set up to evaluate the relative risk of harms in trials performed in less-developed versus more-developed countries.
RI PT
Study Design: Meta-epidemiologic evaluation using the Cochrane Database of Systematic Reviews. We considered meta-analyses with at least one RCT in a less-developed country and one RCT in a more-developed country. We targeted severe adverse events (AEs),
SC
discontinuations due to AEs, any AE, organ system-specific AEs, individual AEs and all discontinuations due to any reason. We estimated the relative odds ratio (ROR) of harms
under each category of harms.
M AN U
between more and less-developed countries for each topic and the summary sROR across topics
Results: We identified 42 systematic-reviews (128 meta-analyses, 521 independent RCTs). Summary sRORs did not differ significantly from 1.00 for any harm category. Nominally
TE D
significant RORs were found in only 6/128 meta-analyses. However, in 27% (35/128) of metaanalyses the ROR point-estimates indicated relative differences between country-settings >2fold. Considering also ROR 95% confidence-intervals, in 92% (118/128) of meta-analyses one
EP
could not exclude a 2-fold difference in both directions. Conclusions: We identified limited comparative evidence on harms in trials from these two
AC C
country-settings. Substantial differences in the risk point-estimates were common; the potential for modest differences could rarely be excluded with confidence.
Key words: Comparative Safety, Comparative Harms, More developed countries, Less developed countries, Randomized trials, Meta-analyses.
4
ACCEPTED MANUSCRIPT
Running title: Comparative Harms in RCTs from More vs Less Developed Countries What is new? •
Randomized trials on important clinical questions are frequently performed in less-developed
RI PT
countries, without long-standing clinical research tradition. We have previously shown that trials in such country-settings tend to report more favorable mortality outcomes. It remains unknown whether similar differences exist in the reporting of harms. We studied the
SC
comparative safety of medical interventions in randomized trials from more- versus less developed countries.
The identified comparative evidence on harms in trials from these two country-settings was
M AN U
•
limited. •
Not many nominally-significant discrepancies in relative-risks of harms between country settings were detected. However, substantial differences in the point-estimates of the risks of
TE D
harms were common and the potential for modest differences between the two country settings could rarely be excluded with confidence. What this adds to what was known?
Due to the limited comparative evidence, it remains unclear whether safety data from trials in
EP
•
less developed countries could be reliably used to guide policy development and clinical
AC C
decision making in more developed countries, but also less developed countries. What is the implication, what should change now? •
Evaluation of the comparative safety of medical interventions in trials from these two country settings should be further pursued and additional comparative evidence needs to be accumulated.
5
ACCEPTED MANUSCRIPT
Alternative methodological approaches, i.e. through analysis of comparative safety data reported within international multisite trials, between trial sites in more developed countries
EP
TE D
M AN U
SC
RI PT
and trial sites in less developed countries should be also considered.
AC C
•
6
ACCEPTED MANUSCRIPT
Introduction As the budget for and participation rates in randomized clinical trials (RCTs) in more developed countries are limited, it is becoming increasingly more common that clinical
RI PT
guidelines and clinical decision-making about important questions of health interventions and health care will depend on evidence from trials performed in less developed countries.[1] With the globalization of clinical research, emerging countries are increasingly more actively involved
SC
in clinical trials.[2-4] Asian and Latin America regions have recently shown the largest annual increase in the number of registered clinical trials [5]. Trials done exclusively in less developed
M AN U
countries often have low methodological quality. [6] Only 56% of 670 surveyed researchers from developing countries reported that their research had been reviewed by a local institutional review board [7] and only 11% of published clinical trials conducted in China in 2004 report that their study protocol was reviewed by an ethical review committee.[8] Some reports are raising concerns for underreporting of adverse events from studies
TE D
performed in developing countries.[9] For example, large differences in the reported rates of ciprofloxacin associated arthropathy were seen in children from North America compared to
EP
children from Latin America.[9] However, the frequency of this phenomenon has not been systematically studied. If results from trials performed in countries without a longstanding
AC C
tradition in clinical research will be used to guide clinical decision-making, an empirical largescale evaluation of these trends is needed. While there are potential benefits from the globalization of clinical research, it is important to evaluate whether results are similar and possible to extrapolate across different settings. In a previous evaluation [10] we assessed differences in mortality and primary efficacy outcomes in RCTs performed on the same topic in more versus less developed countries. We found that on average, trials in less developed countries tended to report more favorable results for the 7
ACCEPTED MANUSCRIPT
experimental intervention [10]. Sometimes genuine differences between country settings could explain differences in results; however, selective outcome reporting, publication, language and other biases [11-13] in the literature coming from less developed countries was considered more It would be important to assess whether major harms
RI PT
likely to explain these discrepancies.
outcomes are also similar or different in RCTs from more versus less developed countries.
Very often the number of patients studied in pre-licensure trials is small to allow the
SC
robust evaluation of both safety and efficacy outcomes. [14] I Individual studies and even individual meta-analyses are on average underpowered to detect differences in reported rates of
M AN U
clinically important adverse events from different country settings.[15] Therefore, we performed a large-scale meta-epidemiological evaluation of safety outcomes in trials from more vs. less developed countries. Methods
TE D
Definitions of countries
The categorization of countries into more developed and less developed countries was done as previously described in our earlier paper on comparative results for mortality and
EP
primary efficacy outcomes.[10] In brief we considered more developed countries to be those with both longstanding established marker economies and longstanding tradition in clinical
AC C
research as previously suggested.[16] Such countries included the United States, Canada, Australia, New Zealand, Israel, Japan and Western European countries. All other countries except for those in Eastern Europe were considered as less developed. We excluded RCTs from Eastern European countries as these represent another unique type of countries in transition. [17] Harm-related endpoints
8
ACCEPTED MANUSCRIPT
We targeted 6 main categories of harm-related endpoints [18], three of which were considered as primary endpoints because they combine adverse events of all types and they only include harms. We did not focus on mortality, as this was the focus in our previous paper [10].
RI PT
The primary endpoint categories were: severe adverse events; discontinuations due to adverse events; and any adverse event. The secondary endpoint categories were: organ system-specific adverse events; individual adverse events; and all discontinuations-due to any reason (in some
SC
studies, discontinuations due to harm might not have been separately reported, but they could have been included under such a broader study endpoint; this endpoint would then be relevant
M AN U
for harms, even though not fully specific). Under the categories of organ system-specific and individual adverse events we considered several subcategories (e.g. gastrointestinal adverse events, hematologic adverse events etc.; and headache, neutropenia etc., respectively). These endpoints were considered secondary because they either do not include all adverse events or
TE D
they count also some events that are not due to adverse events.
We also considered a composite primary-harms endpoint (combined primary-harms), where all three primary harm endpoints (severe adverse events, discontinuations due to adverse
EP
events and any adverse event) were considered together. Eligible meta-analyses
AC C
We included meta-analyses that quantitatively synthesized evidence on harm-related
endpoints and included at least one RCT from a less developed country and at least one RCT from a more developed country for the same compared interventions and the same type of harmrelated endpoint. We focused on RCTs performed exclusively in more developed countries or exclusively in less developed countries; international multicenter trials were excluded unless all participating sites were from the same country-setting.
9
ACCEPTED MANUSCRIPT
Search We screened a list of 131 systematic reviews from the Cochrane Database of Systematic reviews (last update August 27, 2012) that we had previously identified [10] as having RCTs
RI PT
from more developed and less developed countries for mortality. The reasons for this choice were that this selected group of systematic reviews would have been more likely to have included also evidence on harms from both country settings; and the questions addressed were
SC
likely to pertain to serious clinical conditions and diseases given that death was also an outcome of interest. Meta-analyses within the same systematic review for different compared
M AN U
interventions or for different types of harm-related-endpoints were considered as separate metaanalyses. Thus, each systematic review could have contributed data to more than one category of harm-related endpoints. Data extraction
TE D
From each eligible meta-analysis we extracted information on the compared interventions, the harm-related endpoint, the study (author, year), the number of patients with harm-related endpoints and the total number of patients in the experimental group and the control
EP
group respectively per study. No continues harms were encountered. Statistical synthesis of data
AC C
We synthesized the data in the following sequential steps as previously described [10 18
19]: First, for each RCT we calculated the odds ratio (OR) of harm-related endpoints for the experimental vs. the control intervention. When needed we coined estimates so that an OR>1 always corresponds to an increased risk of the harm-related outcome. If in the 2×2 table for each study (events/total number of subjects in the experimental arm and the control arm respectively) there were one or more cells with zero events, then a standard correction of 0.5 was added to all
10
ACCEPTED MANUSCRIPT
cells to allow for a meaningful calculation of the OR, as previously described.[20] Second, for each individual meta-analysis, under each category of harm-related endpoint, we calculated the summary OR (sOR) from all RCTs from the more developed countries and the sOR from all
RI PT
RCTs from the less developed countries using random-effect model for the meta-analysis; we also performed sensitivity analyses using fixed-effects model. [21 22] Random effects model assumes that the combined studies aim at identifying an average treatment effect, while fixed
SC
effects model assumes a common effect across the combined studies. Third, for each individual meta-analysis, and for each category of harm-related endpoint, we calculated the relative OR
M AN U
(ROR) between more developed and less developed countries by dividing the sOR from trials in more developed countries by the sOR from trials in less developed countries. Fourth, for each category of harm-related endpoint, we generated a summary ROR (sROR) between more and less developed countries by synthesizing the RORs from all the individual meta-analyses, under
TE D
each category of harm-related endpoint using random-effects models. A ROR>1 (or sROR>1) means that less developed countries reported less unfavorable results for the experimental intervention relative to more developed countries.
EP
For the combined primary-harms endpoint, we synthesized available RORs from all 3 primary endpoints; whenever RORs existed for more than one primary endpoint for the same
AC C
meta-analysis, we used the ROR that had the smaller variance. For the meta-analyses for organ system-specific adverse events and individual adverse events, we combined RORs only when they were targeting exactly the same types of organ system-specific or individual adverse events. We calculated the between meta-analyses heterogeneity under each category of harm-
related endpoint by using the I2 metric with its corresponding 95% CIs.[23 24]
11
ACCEPTED MANUSCRIPT
We calculated the number of meta-analyses where there was a statistically significant difference in the sOR between the two country settings (95% CIs of ROR excluding 1.00) and
the two country settings (ROR≥ 2.0 or ROR ≤0.50).
RI PT
the number of meta-analyses where the ROR suggested difference exceeding 2.0-fold between
In topics with statistically significant differences in harm-related endpoints between the two country settings, we also explored whether there were statistically significant differences in
SC
the baseline risks of these endpoints in the control group between meta-analyses from more developed and less developed countries. We synthesized the baseline risks per country setting by
M AN U
random effects using the Freeman-Tukey transformation method [25] for the meta-analysis of proportions and compared the results in the two country settings by chi-square test. We also examined whether there was evidence for small study effects using the Harbord’s test (considered to be significant for P<0.10). [26]
TE D
In sensitivity analyses we compared meta-analyses of U.S. trials only vs. trials from less developed countries. We also performed sensitivity analyses for the primary endpoints where we excluded trials from four Asian “tigers” (Hong Kong, Taiwan, Singapore, and South Korea) that
EP
evolved from less developed countries into advanced economies according to the International Monetary Fund in 1997, although their tradition of clinical research is still not as long-standing
AC C
as in the main more developed countries. All analyses were done in Stata SE12 (College Station, TX: StataCorp LP). We followed
the PRISMA [27] (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines for reporting. Results Eligible topics 12
ACCEPTED MANUSCRIPT
Of the previously identified 131 systematic reviews [10], 42 were considered eligible (Supplementary file 1) with a total of 128 separate meta-analyses: 12 meta-analyses for severe adverse events, 12 for discontinuations due to adverse events, 9 for any adverse event, 5 for
RI PT
organ system-specific adverse events, 89 for individual adverse events and 1 for all discontinuations due to any reason (Figure 1). From a total of 521 independent RCTs -after excluding 38 from mixed country settings, 5 with no reported country setting and 4 from Eastern
SC
European countries- we eventually analyzed 474 RCTs; 337 from more developed countries and 137 from less developed countries (Supplementary file 2). The topics of the 42 systematic
M AN U
reviews were related to infectious diseases (n=10), neonatal diseases (n=6), maternal-fetal conditions (n=6), cardiovascular diseases (n=5), neurologic diseases (n=5), gastrointestinal diseases (n=3), oncologic conditions (n=3), transplantations (n=2), rheumatologic diseases (n=1) and pulmonary diseases (n=1). The individual topics appear in Supplementary file 3.
TE D
The median sample size per RCT from more developed and less developed countries for the three primary harm endpoints was 175 and 278 respectively for severe adverse events; 151 and 377 respectively for discontinuations due to adverse events and 82 and 143 respectively for
EP
any adverse event (Supplementary file 4). The median adverse event rate per RCT in more developed and less developed countries was 3.5% and 3.4% respectively for severe adverse
AC C
events; 4.8% and 1.0% for discontinuations due to adverse events and 10.1% and 10.1% respectively for any adverse event (Supplementary file 4). Relative differences in individual meta-analyses (more-developed vs. less-developed) The clinical topics and the ROR estimates between the two country settings for the three primary endpoints and the combined primary-harms endpoint are shown in Table 1 and Supplementary file 3; the respective figures for the summary OR per meta-analysis according to 13
ACCEPTED MANUSCRIPT
country-setting and the ROR per meta-analysis, for each category of these harms are shown in Supplementary files 5-8. Respectively, the summary ORs per country-setting per meta-analysis and the ROR per meta-analysis for each of the secondary endpoint categories are shown in
RI PT
Supplementary files 9-11. The summary estimates (sRORs) for the different primary and secondary endpoints and the combined primary-harms endpoint are shown in Supplementary file
SC
12.
The summary sRORs did not differ significantly from 1.00 for any category of harm.
M AN U
When all harm-related endpoints were considered, nominally statistically significant relative differences between more developed and less developed countries were identified in 5% (6/128) of all meta-analyses (ROR>1 in two and ROR<1 in four) (Supplementary file 3). However, in 27% (35/128) of meta-analyses the relative differences in the summary ORs between more and less developed countries, based on the ROR point estimates, were larger than 2-fold. More
TE D
specifically, RORs≥ 2.0 were seen on 20 meta-analyses while ROR ≤ 0.50 in 15 meta-analyses (p=0.36) (Supplementary files 3 and 11).
EP
There was also large uncertainty left around the estimated RORs and in 92% (118/128) of meta-analyses, when the ROR 95% confidence intervals were also considered, one could not
AC C
exclude a 2-fold difference in harms in both directions (lower ROR CI≥0.5 and upper ROR CI≤2.0) (Supplementary file 11). Moreover, in 67% (86/128) of the meta-analyses, there were more than 10-fold differences between the upper and lower CIs of the RORs. More specifically this occurred in 67% of meta-analyses for severe adverse events; 75% of meta-analyses for discontinuations due to adverse events; 44% of meta-analyses for any adverse events; 40% of meta-analyses for organ system specific adverse events; and 69% of meta-analyses for individual adverse events. 14
ACCEPTED MANUSCRIPT
Results were similar when fixed effects were used to synthesize the sORs for the primary harm-related endpoints per country-setting within each meta-analysis. (Supplementary files 1315). In sensitivity analyses comparing only USA trials against trials from less developed
RI PT
countries 4/24 (17%) of meta-analyses for the primary endpoints showed nominally differences beyond chance for the comparison by the country setting (ROR>1 in one and ROR<1 in three) (Supplementary files 3, 16-18). Further sensitivity analyses excluding trials from the four Asian
SC
“tigers” also gave similar results for the primary endpoints and the combined primary-harms
M AN U
endpoint. RORs
Table 2 shows in more detail the 6 topics for which a nominally statistically significant difference was found between studies from less developed and more developed countries. For 2
TE D
topics, less developed countries reported less unfavorable results (RORs>1); while the opposite occurred in 4 topics (RORs<1). Studies from less developed countries had reported relatively less neonatal hematologic toxicities with zidovudine (vs. placebo) for prevention of HIV mother
EP
to child transmission; and relatively less diarrhea with third generation cephalosporins (vs. conventional antibiotics) for acute bacterial meningitis. While studies from less developed
AC C
countries had reported relatively more severe adverse events with the use of phosphodiesterase inhibitors (vs. placebo) for chronic obstructive pulmonary disease; relatively more adverse events of any type with stem cell therapy for acute myocardial infarction; relatively more gastrointestinal complications with preoperative chemotherapy (vs. surgery alone) for resected thoracic and esophageal cancer; and relatively more vomiting after the second dose of rotavirus vaccines (vs placebo).
15
ACCEPTED MANUSCRIPT
For the above 6 topics there were no statistically significant difference in the baseline risks of harms in the control groups (control rates) between trials from less and more developed
RI PT
countries; nor was there evidence of small study effects (Supplementary file 19). Discussion
Our study provided a bird’s eye view of the pattern of reporting of harms and harm-
SC
related endpoints in trials from the two country settings. We explored the relative differences in rates of harms across different country-settings using diverse approaches in order to obtain a
M AN U
more comprehensive appraisal. Our inferences need to be cautious given the large uncertainty in the relative incidence of harms for many of the examined topics. Due to the dearth of harm data in individual randomized trials but also in individual meta-analyses, it would not have been possible to identify significant differences between country-settings and only a bird’s eye view
TE D
and large scale evaluation across diverse topics was able to probe such country-setting specific harm differences. Overall, across all analyzed meta-analyses for all types of harm-related endpoints, we did not detect nominally significant discrepancies in the reported rates of harm-
EP
related endpoints between trials from less developed and more developed countries when results were summarized for each harm category. However, when the point estimates of the relative
AC C
differences of harms between more and less developed countries were considered, in 27% of the meta-analyses the relative differences were larger than 2-fold. Many meta-analyses had limited evidence and substantial uncertainty and when the uncertainty around the relative differences were also considered, differences in the relative odds of harms exceeding 2-fold in both directions could not be excluded in 92% of meta-analyses.
16
ACCEPTED MANUSCRIPT
We should acknowledge some study limitations. First, we did not screen the whole Cochrane Database for Systematic Reviews with harm-related endpoints. However, the 131 systematic reviews we screened was a group of meta-analyses that was highly likely to have
RI PT
included harm-related endpoints both from trials in more developed and less developed countries as they had already included trials from both country settings for mortality. By focusing also on mortality, our pool of these 131 screened systematic reviews was about clinically important
SC
medical conditions and interventions. Trials from more developed countries were the preponderance, as we had also observed in our assessments of mortality previously. [10]
M AN U
However, trials from less developed countries are expected to become more common in the future. Second, in the 128 individual meta-analyses, included in the 42 eligible systematic reviews, with comparative evidence on harm-related endpoints from the two country settings only a tenth had data on the primary harm endpoints; the majority had reported individual
TE D
specific adverse events. The clinical significance of these individual harms, without any additional information on their severity grading, is unclear.[28] This probably reflects poor reporting for harms and poor compliance with CONSORT standards of reporting, which is a
37].
EP
universal problem in randomized trials independently of the trial’s country setting [29-31] [32Third, we focused on trials performed exclusively in more-developed or less-developed
AC C
countries and did not analyze international multisite trials. It remains unknown whether the reported rates of harms in trials from other countries in economic and social transition and from study-sites from different country settings included in the same international multisite trials are systematically different. However, the vast majority of published reports of international studies included in the CDSR do not report results separately per country setting. In our database, less than 10% of the RCTs were international studies from mixed country settings. In Bayer A.G.’s
17
ACCEPTED MANUSCRIPT
ciprofloxacin patent extension studies the FDA noted an arthropathy rate of 21 % reported by investigators in the United States compared to none in a less developed country [9]. This Bayer/Quintiles database was further analyzed and divided into countries [38], and ultimately
conclusions are still pending.
RI PT
handed over to the US Department of Justice for a false claims investigation [39]; final Fraud must be ruled out whenever significant discrepancies
findings, rather than following a patient for adverse events.
SC
between investigator groups are found, as it is easy for an untrained investigator to report no
In summary, although we did not identify consistent over- or under-estimation of risks of
M AN U
harms between trials done in more versus less developed countries, substantial differences in the estimates of risks were relatively common and the potential for modestly large differences often could not be excluded with confidence. Collection and reporting of harms should be emphasized and standardized in all randomized trials, both from more developed and less developed
TE D
countries. More emphasis should be given to the standardized reporting of clinically important harms, such as severe adverse events or discontinuations due to adverse events. It is very important to have available the information required for the prompt identification of systematic
EP
differences in harms across different country settings. Differences may reflect genuine variability in the incidence of harms or differences in the reporting of harms. Reported rates of harms
AC C
may be influenced by the genetic profiles of populations, but may also depend upon ethical practices and monitoring systems in place in different countries. Disentangling these reasons would require that harms/harm-related endpoints are collected and reported transparently and in a harmonized way across randomized trials [28]. Country-setting discrepancies in the reported rates of harms should be routinely incorporated and explored in meta-analyses and systematic reviews. Moreover, regulatory agencies, such as the FDA or the EMEA, should enforce that such
18
ACCEPTED MANUSCRIPT
transparent harm-related information is provided per country and site for all randomized trials, including also multisite international trials. This may allow understanding how often and why some countries and sites have different results. Until such transparent comparative evidence on
RI PT
harms is available from trials in different country-settings, guideline developers and policy makers should cautiously extrapolate harm data from trials from one country setting to guide
AC C
EP
TE D
M AN U
SC
clinical decision making in very different country settings.
19
ACCEPTED MANUSCRIPT
Contributorship: DCI had the original idea and developed the study analysis plan; JPAI, XT, MA, OAP, JNW and YM critically commented on the study analysis plan; DCI, XT, MA performed data extraction; DCI and OAP performed the statistical analyses; DCI drafted the first
RI PT
manuscript draft and all authors critically reviewed it and approved the final version; DCI is the guarantor for this paper. Ethical approval: Not applicable.
SC
Transparency declaration: Dr. Contopoulos-Ioannidis (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that
as planned have been explained.
M AN U
no important aspects of the study have been omitted; and that any discrepancies from the study
Copyright/license for publication: The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, “a worldwide license to the Publishers and
TE D
its licensees in perpetuity, in all forms, formats and media (whether known now or created in the future), to: i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include within collections and
EP
create summaries, extracts and/or, abstracts of the Contribution, iii) create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the
AC C
inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) license any third party to do any or all of the above.”
20
ACCEPTED MANUSCRIPT
References 1. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials: race-, sex-, and age-based disparities. JAMA 2004;291(22):2720-6
RI PT
2. George M, Selvarajan S, S SK, et al. Globalization of clinical trials - where are we heading? Current Clinical Pharmacology 2013;8(2):115-23
3. Duke T, Fuller D. Randomised controlled trials in child health in developing countries: trends
SC
and lessons over 11 years. Archives of Disease in Childhood 2014;99(7):615-20.
4. Dandona L, Katoch VM, Dandona R. Research to achieve health care for all in India. Lancet
M AN U
2011;377(9771):1055-7
5. Drain PK, Robine M, Holmes KK, et al. Trail watch: global migration of clinical trials. Nature reviews. Drug Discovery 2014;13(3):166-7
6. Zhang D, Freemantle N, Cheng KK. Are randomized trials conducted in China or India
2011;64(1):90-5
TE D
biased? A comparative empirical analysis. Journal of Clinical Epidemiology
7. Hyder AA, Wali SA, Khan AN, et al. Ethical review of health research: a perspective from
EP
developing country researchers. Journal of Medical Ethics 2004;30(1):68-72 8. Zhang D, Yin P, Freemantle N, et al. An assessment of the quality of randomised controlled
AC C
trials conducted in China. Trials 2008;9:22
9. Division of special pathogen and immunologic drug products. Summary of clinical review of studies
submitted
in
response
to
a
pediatric
written
request.
http://www.fda.gov/ohrms/dockets/ac/05/briefing/20054152b1_03_03_Cipro%20medical.pdf Accessed 1/21/2015
21
ACCEPTED MANUSCRIPT
10. Panagiotou OA, Contopoulos-Ioannidis DG, Ioannidis JP. Comparative effect sizes in randomised trials from less developed and more developed countries: metaepidemiological assessment. BMJ 2013;346:f707
RI PT
11. Lorenzo C, Garrafa V, Solbakk JH, et al. Hidden risks associated with clinical trials in developing countries. Journal of Medical Ethics 2010;36(2):111-5
2002;21(19):2869-77
SC
12. Smith R. Publishing research from developing countries. Statistics in Medicine
13. Juni P, Holenstein F, Sterne J, et al. Direction and impact of language bias in meta-analyses trials:
empirical
study.
International
M AN U
of controlled
2002;31(1):115-23
Journal
of Epidemiology
14. Duijnhoven RG, Straus SM, Raine JM, et al. Number of patients studied prior to approval of new medicines: a database analysis. PLoS Medicine 2013;10(3):e1001407
TE D
15. Tsang R, Colley L, Lynd LD. Inadequate statistical power to detect clinically significant differences in adverse event rates in randomized controlled trials. Journal of Clinical Epidemiology 2009;62(6):609-16
gs-r,
EP
16. United Nations Statistics Division. Composition of macro geographical (continental) regions and
selected
economic
and
other
groupings.
AC C
2013.http://unstats.un.org/unsd/methods/m49/m49regin.htm.
17. Platonov P. Clinical trials in Russia and Eastern Europe: recruitment and quality. International Journal of Clinical Pharmacology and Therapeutics 2003;41(7):277-80
18. Lathyris D, Panagiotou OA, Baltogianni M, et al. Safety of medical interventions in children versus adults. Pediatrics 2014;133(3):e666-73
22
ACCEPTED MANUSCRIPT
19. Contopoulos-Ioannidis DG, Baltogianni MS, Ioannidis JP. Comparative effectiveness of medical
interventions
in
adults
versus
children.
The Journal
of Pediatrics
20. http://handbook.cochrane.org/ ChfsroiTCCVuM.
RI PT
2010;157(2):322-30
21. Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews. Annals of Internal Medicine 1997;127(9):820-6
SC
22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials 1986;7(3):177-88
M AN U
23. Ioannidis JP, Patsopoulos NA, Evangelou E. Uncertainty in heterogeneity estimates in metaanalyses. BMJ 2007;335(7626):914-6
24. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine 2002;21(11):1539-58
1950;21:607-11
TE D
25. Freeman MF TJ. Transformations related to the angular and the square root. Ann Math Statist
26. Harbord RM EM, Sterne JA. . A modified test for small-study effects in meta-analyses of
EP
controlled trials with binary endpoints. Statistics in Medicine 2006;25:3443-57. 27. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic
AC C
reviews and meta-analyses of studies that evaluate health care interventions: explanation
and elaboration. Annals of Internal Medicine 2009;151(4):W65-94
28. Ioannidis JP, Evans SJ, Gotzsche PC, et al. Better reporting of harms in randomized trials: an extension of the CONSORT statement. Annals of Internal Medicine 2004;141(10):781-8
23
ACCEPTED MANUSCRIPT
29. Papanikolaou PN, Ioannidis JP. Availability of large-scale evidence on specific harms from systematic reviews of randomized trials. The American Journal of Medicine 2004;117(8):582-9
medical areas. JAMA 2001;285(4):437-43
RI PT
30. Ioannidis JP, Lau J. Completeness of safety reporting in randomized trials: an evaluation of 7
31. Ioannidis JP. Adverse events in randomized trials: neglected, restricted, distorted, and
SC
silenced. Archives of Internal Medicine 2009;169(19):1737-9
32. Mukherjee SD, Coombes ME, Levine M, et al. A qualitative study evaluating causality
M AN U
attribution for serious adverse events during early phase oncology clinical trials. Investigational New Drugs 2011;29(5):1013-20 .
33. Wahab IA, Pratt NL, Kalisch LM, et al. The detection of adverse events in randomized clinical trials: can we really say new medicines are safe? Current Drug Safety
TE D
2013;8(2):104-13
34. Schroll JB, Maund E, Gotzsche PC. Challenges in coding adverse events in clinical trials: a systematic review. PloS One 2012;7(7):e41174.
EP
35. Huang HY, Andrews E, Jones J, et al. Pitfalls in meta-analyses on adverse events reported from clinical trials. Pharmacoepidemiology and Drug Safety 2011;20(10):1014-20
AC C
36. Haidich AB, Birtsou C, Dardavessis T, et al. The quality of safety reporting in trials is still suboptimal: survey of major general medical journals. Journal of Clinical Epidemiology
2011;64(2):124-35
37. de Vries TW, van Roon EN. Low quality of reporting adverse drug reactions in paediatric randomised controlled trials. Archives of Disease in Childhood 2010;95(12):1023-6
24
ACCEPTED MANUSCRIPT
38. Personal communication with Dr N. Balasubramanian from the Department of Statistics at Stanford University 39. United States of America, ex. rel. Juan N. Walterspiel MD, FAAP v. Bayer A.G., Quintiles False Claims Act action under 31 U.S.C. § 3729(a).
AC C
EP
TE D
M AN U
SC
https://www.pacer.gov/ Accessed 01/18/2015.
RI PT
Transnational Corp.
25
ACCEPTED MANUSCRIPT
Table 1: Harm-related endpoints
RI PT
Severe adverse events Title
Comparison
Description of Harmrelated endpoint
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant vs control
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet vs control acute
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent vs control
CD002130
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic
Platelet glycoprotein IIb/IIIa blockers during PCI vs control
Major extracranial haemorrhage during treatment period Major extracranial 6 (796) haemorrhage during treatment period Symptomatic (including 20 (6,244) fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding 20 (23,259) (all patients)
CD002309
TE D
M AN U
N of RCTs (N patients) from more developed countries 15 (2,506)
N of RCTs (N patients) from less developed countries 2 (344)
ROR (95% CIs)
1 (21,106)
0.67 (0.14-3.09)
2 (567)
1.97 (0.57-6.85)
3 (448)
2.39 (0.30-19.03)
2.48 (0.22-27.75)
EP
AC C
CD002147
SC
Review
Terlipressin vs vasopressin
Adverse events causing death
3 (204)
2 (101)
0.54 (0.02-19.03)
PDE4 inhibitors vs placebo
Non fatal serious adverse events
8 (2,965)
1 (1,018)
0.38 (0.18-0.83)
26
ACCEPTED MANUSCRIPT
CD003897b
CD008521a
CD008521b
IL2Ra vs placebo or no treatment RV1 vs placebo
Review
Title
CD000171a
Treatment of latent tuberculosis infection in
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.04)
Infection: CMV invasive
3 (881)
1 (118)
0.93 (0.12-7.26)
Infection: serious allcause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
Serious adverse events
7 (3,179)
18 (29,581)
0.88 (0.63-1.25)
3 (2,547)
7 (7,863)
0.81(0.45-1.46)
N of RCTs (N patients) from more developed countries 2 (838)
N of RCTs (N patients) from less developed countries 6 (4,695)
ROR (95% CIs)
RI PT
1 (29)
SC
Intraventricular haemorrhage (III-IV)
RV5 versus placebo All serious adverse events
AC C
Discontinuations due to adverse events
IL2Ra vs placebo or no treatment
5 (266)
M AN U
CD003897a
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interleukin 2 receptor antagonists for kidney transplant recipients Interleukin 2 receptor antagonists for kidney transplant recipients Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
Major infection
TE D
CD003481
Immunosuppressiv e agent plus steroids vs steroids alone Ibuprofen vs indomethacin
EP
CD002922
obstructive pulmonary disease Treatment for lupus nephritis
Comparison
Description of Harmrelated endpoint
Any TB drug vs placebo
Incidence of adverse events leading to
1.05 (0.04-26.24)
27
ACCEPTED MANUSCRIPT
CD002863
CD004386a
CD004386b
CD004773
Beta-blocker vs placebo/no betablocker Drug vs placebo/no intervention
Changed drugs due to maternal side effects
Side effects requiring discontinuation
Quinolone vs TMP- Side effects requiring SMZ discontinuation
Antifungal vs Placebo
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
3 (206)
2 (103)
0.22 (0.01-5.92)
7 (2,891)
1 (1,018)
0.61 (0.27-1.36)
10 (923)
1 (154)
1.22 (0.19-7.67)
16 (2,057)
1 (97)
2.11 (0.04-113.13)
6 (752)
1 (102)
4.03 (0.40-40.29)
2 (723)
1 (129)
0.74 (0.07-8.17)
RI PT
SC
PDE4 inhibitors vs placebo
M AN U
CD002309
TE D
CD002147
stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Adverse events causing withdrawal of treatment Withdrawals due to adverse effects
EP
CD000171c
Isoniazid vs isoniazid + rifampicin Isoniazid vs rifampicin + pyrazinimide Terlipressin versus vasopressin
AC C
CD000171b
HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Discontinuation secondary to adverse events
28
ACCEPTED MANUSCRIPT
CD008521b
Withdrawal due to adverse events
RV5 vs placebo
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Adverse events requiring discontinuation (end of follow-up) Adverse events requiring discontinuation (end of follow-up)
Description of Harmrelated endpoint
Any adverse event
8 (1,652)
3 (236)
0.61 (0.13-2.93)
5 (3,709)
4 (7,735)
1.00 (0.21-4.73)
6 (2,184)
15 (21,774)
1.92 (0.63-5.87)
N of RCTs (N patients) from more developed countries 8 (2,913)
N of RCTs (N patients) from less developed countries 1 (1,018)
ROR (95% CIs)
RI PT
Statins vs placebo/no treatment
SC
CD008521a
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use
M AN U
CD007784
Title
Comparison
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Interventions for preventing oral candidiasis for patients with cancer receiving treatment Antioxidant supplements for preventing
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
All drug types vs placebo/no treatment
Toxicity (adverse events 'probably due to drug')
2 (661)
1 (210)
1.47 (0.44-4.90)
Antioxidants (betacarotene) vs placebo
Adverse effects
3 (91,080)
1 (174)
0.88 (0.03-23.17)
CD004183
EP
AC C
CD003807
TE D
Review
1.21 (0.73-2.00)
29
ACCEPTED MANUSCRIPT
CD006536
CD008076
1.78 (0.03-92.13)
9 (929)
1 (102)
1.15 (0.36-3.63)
Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Adverse events
6(450)
1 (102)
0.75 (0.02-24.58)
RI PT
1 (97)
Antibiotic prophylaxis vs placebo/ no treatment
SC
Antifungal vs placebo
M AN U
CD004791
33 (5,002)
TE D
CD004773
Side effects
Quinolone vs TMP- Side effects SMZ
Stem cells vs no stem cells
Adverse effects
8 (669)
2 (213)
0.06 (0.02-0.20)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
Safety outcomes during follow-up
1 (3,202)
1 (787)
0.56 (0.31-1.00)
EP
CD004386b
Drug vs placebo/ no intervention
AC C
CD004386a
gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction Cilostazol versus aspirin for secondary prevention of vascular events after stroke of arterial origin
Combined Primary-Harms (Serious Adverse event, Discontinuations due to adverse event and Any adverse event)
30
ACCEPTED MANUSCRIPT
Anticoagulants for acute ischaemic stroke
CD000213
CD002130
N of RCTs (N patients) from more developed countries 15 (2,506)
N of RCTs (N patients) from less developed countries 2 (344)
ROR (95% CIs)
Major extracranial haemorrhage during treatment period
6 (796)
1 (21,106)
0.67 (0.14-3.09)
Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding
2 (838)
6 (4,695)
1.05 (0.04-26.24)
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
20 (5,622)
2 (567)
1.97 (0.57-6.85)
20 (23,259)
3 (448)
2.39 (0.30-19.03)
Major extracranial haemorrhage during treatment period
TE D
CD000171c
EP
CD000171b
AC C
CD000171a
Anticoagulant vs control in acute presumed ischaemic stroke Antiplatelet therapy for Antiplatelet vs acute ischaemic stroke control in acute presumed ischaemic stroke Treatment of latent Any TB drug vs tuberculosis infection in placebo HIV infected persons Treatment of latent Isoniazid vs tuberculosis infection in isoniazid + HIV infected persons rifampicin Treatment of latent Isoniazid vs tuberculosis infection in rifampicin + HIV infected persons pyrazinamide Thrombolysis for acute Any thrombolytic ischaemic stroke agent vs control
Description of Harmrelated endpoint
RI PT
CD000024
CD000029
Comparison
SC
Title
M AN U
Review
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST
Platelet glycoprotein IIb/IIIa blockers vs placebo/usual care (all patients)
2.48 (0.22-27.75)
31
ACCEPTED MANUSCRIPT
CD003481
CD003807
CD003897
CD004183
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interventions for preventing oral candidiasis for patients with cancer receiving treatment Interleukin 2 receptor antagonists for kidney transplant recipients Antioxidant supplements for preventing gastrointestinal cancers
1.21 (0.73-2.00)
10 (923)
1 (154)
1.22 (0.19-7.67)
5 (266)
1 (29)
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.03)
Toxicity (adverse events 'probably due to drug')
2 (661 )
1 (210)
1.47 (0.44-4.90)
IL2Ra vs placebo or no treatment
Infection: serious allcause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
Antioxidants vs placebo
Adverse effects - betacarotene
3 (91,080)
1 (174)
0.88 (0.03-23.17)
SC
Beta-blocker vs placebo/no betablocker
RI PT
1 (1,018)
Changed drugs due to maternal side effects
M AN U
CD002922
0.22 (0.01-5.88)
PDE4 inhibitors vs placebo
Adverse events causing 3 (204) withdrawal of treatment No of patients 8 (2,913) experiencing an adverse effect
Immunosuppressive Major infection agent plus steroids vs steroids alone Ibuprofen vs Intraventricular indomethacin haemorrhage (III-IV)
TE D
CD002863
2 (101)
Comparisons with placebo/no treatment for all drug types
EP
CD002309
Terlipressin vs vasopressin
AC C
CD002147
segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Treatment for lupus nephritis
32
ACCEPTED MANUSCRIPT
CD006536
CD007784
CD008076
Antifungal vs Placebo
Antibiotic prophylaxis vs placebo or no treatment
1 (97)
1.78 (0.03-92.13)
9 (929)
1 (102)
1.15 (0.36-3.63)
1 (295)
1 (129)
1.44 (0.60-3.46)
6 (450)
1 (102)
0.75 (0.02-24.58)
RI PT SC
Quinolone vs TMP- Side effects SMZ
Adverse events
M AN U
CD004791
33 (5,002)
Adverse events
TE D
CD004773
Drug vs placebo/ no Side effects intervention
Stem cells compared to no stem cells, Statins vs placebo/no treatment
Adverse effects
6 (617)
2 (213)
0.06 (0.02-0.18)
Withdrawal due to adverse events
8 (1,652)
3 (236)
0.61 (0.13-2.93)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
Other outcomes of safety during followup.
1 (3,202)
1 (787)
0.56 (0.31-1.00)
EP
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular after stroke of arterial
AC C
CD004386a
33
ACCEPTED MANUSCRIPT
RI PT
RV1 vs placebo
5 (3,708)
7 (3,179)
4 (7,7733)
1.00 (0.21-4.73)
18 (29,581)
0.88 (0.63-1.25)
TE D
M AN U
SC
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Adverse events requiring discontinuation (end of follow-up) Serious adverse events
EP
CD008521b
RV5 vs placebo
AC C
CD008521a
origin Vaccines for preventing rotavirus diarrhoea: vaccines in use
34
ACCEPTED MANUSCRIPT
Table 2: Individual meta-analyses with nominally significant relative differences between more developed and less developed
ROR (95% CIs)
sOR_MD (95% sOR_LD (95% CIs) CIs)
0.38 (0.18-0.83)
0.75 (0.55-1.03)
1.96 (0.96-3.98)
Stem cells compared to no stem cells,
Adverse effects
0.06 (0.02-0.20)
0.84 (0.48-1.45)
13.01 (4.90-34.57)
Any zidovudine vs placebo/no treatment
Neonatal haematological toxicity
2.77 (1.23-6.22)
2.09 (1.22-3.59)
0.76 (0.41-1.38)
Gastrointestinal complications
0.003 (0.001-0.06)
1.15 (0.34-3.81)
441.20 (24.93-7808)
Third generation cephalosporins versus conventional therapy
Diarrhoea
3.92 (1.35-11.34)
2.50 (1.54-4.05)
0.64 (0.25-1.64)
Rotarix vs placebo: after dose 2
Vomitingreactogenicity
0.66 (0.44-0.99)
0.68 (0.47-0.98)
1.03 (0.86-1.24)
M AN U
PDE4 inhibitors vs placebo
TE D
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease CD006536 Stem cell treatment for acute myocardial infarction CD000102 Interventions for reducing the risk of mother-to-child transmission of HIV infection CD001556c Preoperative chemotherapy for resectable thoracic esophageal cancer CD001832 Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis CD008521b Vaccines for preventing rotavirus diarrhoea: vaccines in use
Description of Harm-related endpoint Non fatal serious adverse events
SC
Comparison
Preoperative chemotherapy vs surgery alone
AC C
CD002309
Title
EP
Review
RI PT
countries
35
ACCEPTED MANUSCRIPT
Abbreviations: ROR: relative odds ratio; sOR_MD: summary odds ratio in more developed countries; sROR_LD: summary odds
AC C
EP
TE D
M AN U
SC
RI PT
ratio in less developed countries; CIs: confidence intervals; Footnote: ORs and RORs were calculated by random effects models
36
ACCEPTED MANUSCRIPT
Figure 1: Flow Chart
RI PT
Screened 131 Systematic Reviews
SC
Excluded 89 Systematic Reviews No adverse events (n=40) No pharmacologic interventions (n=30) No More developed and Less developed countries (n=18) No country information (n=1)
Discontinuations due to Adverse Events: 12 ma; 108 RCTs (70 from MD and 38 from LD countries)
TE D
Severe Adverse Events: 12 ma; 143 RCTs (99 from MD and 44 from LD countries)
M AN U
Included 42 eligible Systematic Reviews; 128 meta-analyses on Harm-related endpoints with 1186 study-level analyses from 474 independent RCTs (337 from MD and 137 from LD countries)a b
AC C
EP
Organ System Specific Adverse Events: 5 ma; 20 RCTs (11 from MD and 9 from LD countries)
Individual adverse events (n=60) 14 adverse events with ≥2 ma (43 ma) 46 adverse events with 1 ma (46 ma)
Any Adverse Event: 9 ma; 81 RCTs (71 from MD and 10 from LD countries)
All Discontinuations: 1 ma; 7 RCTs (6 from MD and 1 from LD countries)
89 ma, 767 RCTS corresponding to 311 independent RCTs (222 from MD and 89 from LD countries)
Footnotes: a
Each RCT could have contributed data to more than one study-level analyses, for the different harm-related endpoints.
b
We excluded from these 128 meta-analyses data from 47 RCTs that were from mixed-country settings (n=38); 37Eastern European countries (n=4) or did not report country (n=5).
*Conflict of Interest/Financial Disclosure
ACCEPTED MANUSCRIPT
Conflicts of Interest There are no conflicts of interest to declare for Contopoulos-Ioannidis D; Tseretopoulou X,
RI PT
Ancker M; Panagiotou OA; Maldonado Y and Ioannidis JP. Walerspiel JN worked as a consultant for US Bayer A.G. in 2002-3. “Whistleblower" on Bayer A.G. Currently does not hold any Bayer A.G or subsidiarie’s stock, nor does he give any lectures
AC C
EP
TE D
M AN U
SC
that are paid for directly or indirectly by Bayer A.G.
Supplemental Material (Online Only)
ACCEPTED MANUSCRIPT
SUPPLEMENTARY FILES Supplementary file 1: References of included Systematic Reviews
Cochrane Database Syst Rev. 2008;(4): CD000024
RI PT
1. Sandercock PAG, Counsell C, Kamal AK. Anticoagulants for acute ischaemic stroke.
2. Sandercock PAG, Counsell C, Gubitz GJ, Tseng MC. Antiplatelet therapy for acute ischaemic stroke. Cochrane Database Syst Rev. 2008;(3):CD000029.
SC
3. Brocklehurst P. Interventions for reducing the risk of mother-to-child transmission of HIV infection. Cochrane Database Syst Rev. 2002; (1): CD000102
M AN U
4. Akolo C, Adetifa I, Shepperd S, Volmink J. Treatment of latent tuberculosis infection in HIV infected persons. Cochrane Database Syst Rev. 2010;(1):CD000171 5.
Wardlaw JM, Murray V, Berge E, del Zoppo GJ. Thrombolysis for acute ischaemic stroke. Cochrane Database Syst Rev. 2009;(4): CD000213
TE D
6. Mahomed K, Bhutta ZA, Middleton P. Zinc supplementation for improving pregnancy and infant outcome. Cochrane Database Syst Rev. 2007;(2):CD000230 7. Dorhout Mees S, Rinkel GJE, Feigin VL, Algra A, van den Bergh WM, Vermeulen M,
EP
van Gijn J. Calcium antagonists for aneurysmal subarachnoid haemorrhage. Cochrane Database Syst Rev. 2007;(3): CD000277
AC C
8. Handoll HHG, Farrar MJ, McBirnie J, Tytherleigh-Strong GM, Milne AA, Gillespie WJ. Heparin, low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures.
Cochrane
Database Syst Rev. 2002;(4): CD000305.
9. Ohlsson A, Lacy JB. Intravenous immunoglobulin for preventing infection in preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2013;(7): CD000361
ACCEPTED MANUSCRIPT
10. Wong PF, Chong LY, Mikhailidis DP, Robless P, Stansby G. Antiplatelet agents for intermittent claudication. Cochrane Database Syst Rev. 2011;(11): CD001272. 11. Vogt K, Fenlon D, Rhodes S, Malthaner R. Preoperative chemotherapy for resectable
RI PT
thoracic esophageal cancer . Cochrane Database Syst Rev. 2006;(3): CD001556
12. Prasad K, Kumar A, Singhal T, Gupta PK. Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis. Cochrane Database Syst
SC
Rev. 2007;(4): CD001832
13. Bosch X, Marrugat J, Sanchis J. Platelet glycoprotein IIb/IIIa blockers during
M AN U
percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes. Cochrane Database Syst Rev. 2010;(9): : CD002130
14. Ioannou GN, Doust J, Rockey DC. Terlipressin for acute esophageal variceal
TE D
hemorrhage. Cochrane Database Syst Rev.2003;(1): CD002147 15. Annane D, Bellissant E, Bollaert PE, Briegel J, Keh D, Kupfer Y. Corticosteroids for treating severe sepsis and septic shock. Cochrane Database
EP
CD002243
Syst Rev.2004;(1):
16. Chong J, Poole P, Leung B, Black PN. Phosphodiesterase 4 inhibitors for chronic
AC C
obstructive pulmonary disease. Cochrane Database Syst Rev.2011;(5): CD002309
17. Magee L, Duley L. Oral beta-blockers for mild to moderate hypertension during pregnancy. Cochrane Database Syst Rev. 2003;(3): CD002863
18. Flanc RS, Roberts MA, Strippoli GFM, Chadban SJ, Kerr PG, Atkins RC. Treatment for lupus nephritis. Cochrane Database Syst Rev.2004;(1):CD002922
ACCEPTED MANUSCRIPT
19. Ohlsson A, Walia R, Shah SS. Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants. Cochrane Database
Syst Rev. 2010;(4):
CD003481
RI PT
20. Clarkson JE, Worthington HV, Eden TOB. Interventions for preventing oral candidiasis for patients with cancer receiving treatment. Cochrane Database Syst Rev. 2007;(1): CD003807
SC
21. Webster AC, Ruster LP, McGee R, Matheson SL, Higgins GY, Willis NS, Chapman JR, Craig JC. Interleukin 2 receptor antagonists for kidney transplant recipients. Cochrane
M AN U
Database Syst Rev. 2010;(1): CD003897
22. 22. Greb A, Bohlius J, Schiefer D, Schwarzer G, Schulz H, Engert A. High-dose chemotherapy with autologous stem cell transplantation in the first line treatment of aggressive Non-Hodgkin Lymphoma (NHL) in adults. Cochrane Database
TE D
Rev.2008;(1): CD004024
Syst
23. Rumbold A, Crowther CA. Vitamin E supplementation in pregnancy. Cochrane Database Syst Rev. 2005;(2): CD004069
EP
24. Bjelakovic G, Nikolova D, Simonetti RG, Gluud C. Antioxidant supplements for preventing gastrointestinal cancers. Cochrane Database Syst Rev. 2008;(3): CD004183
AC C
25. Ohlsson A, Shah SS. Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2006;(1): CD004213
26. Rumbold A, Duley L, Crowther CA, Haslam RR. Antioxidants for preventing preeclampsia. Cochrane Database Syst Rev.2008;(1):CD004227 27. Hunt R, Hey E. Ethamsylate for the prevention of morbidity and mortality in preterm or very low birth weight infants. Cochrane Database Syst Rev.2010;(1): CD004343
ACCEPTED MANUSCRIPT
28. Gafter-Gvili A, Fraser A, Paul M, Vidal L, Lawrie TA, van de Wetering MD, Kremer LCM,Leibovici L. Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy. Cochrane Database Syst Rev.2012;(1): CD004386
RI PT
29. Roberts D, Dalziel SR. Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth . Cochrane Database Syst Rev.2006;(3): CD004454 30. Duley L, Henderson-Smart DJ, Meher S, King JF. Antiplatelet agents for preventing
SC
pre-eclampsia and its complications. Cochrane Database Syst Rev. 2007;(2): CD004659 31. Chang LW, Phipps WT, Kennedy GE, Rutherford G. Antifungal interventions for the
Rev. 2005;(3): CD004773
M AN U
primary prevention of cryptococcal disease in adults with HIV. Cochrane Database Syst
32. Cohen MJ, Sahar T, Benenson S, Elinav E, Brezis M, Soares-Weiser K. Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites,
TE D
without gastro-intestinal bleeding. Cochrane Database Syst Rev.2009;(2): CD004791 33. Ohlsson A, Aher SM. Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2014;(4):
EP
CD004863
34. Aher SM, Ohlsson A. Late erythropoietin for preventing red blood cell transfusion in
AC C
preterm and/or low birth weight infants. Cochrane Database Syst Rev. 2014;(4): CD004868
35. AlFaleh K, Anabrees J, Bassler D, Al-Kharfi T. Probiotics for prevention of necrotizing enterocolitis in preterm infants. Cochrane Database Syst Rev. 2011;(3): CD005496
ACCEPTED MANUSCRIPT
36. Raanani P, Gafter-Gvili A, Paul M, Ben-Bassat I, Leibovici L, Shpilberg O. Immunoglobulin prophylaxis in hematological malignancies and hematopoietic stem cell transplantation. Cochrane Database Syst Rev. 2008;(4): CD006501
RI PT
37. Clifford DM, Fisher SA, Brunskill SJ, Doree C, Mathur A, Watt S, M artin-Rendon E. Stem cell treatment for acute myocardial infarction. Cochrane Database Syst Rev. 2012;(2): CD006536
SC
38. Whitfield K, Rambaldi A, Wetterslev J, Gluud C. Pentoxifylline for alcoholic hepatitis. Cochrane Database Syst Rev. 2009; (4): CD007339
M AN U
39. Navaneethan SD, Pansini F, Perkovic V, Manno C, Pellegrini F, Johnson DW, Craig JC,Strippoli GFM. HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis. Cochrane Database Syst Rev. 2009;(2): CD007784 40. Kamal AK, Naqvi I, Husain MR, Khealani BA. Cilostazol versus aspirin for secondary Syst
TE D
prevention of vascular events after stroke of arterial origin. Cochrane Database Rev. 2011;(1): CD008076
41. Gurusamy KS, Koti R, Fusai G, Davidson BR. Somatostatin analogues for pancreatic
EP
surgery. Cochrane Database Syst Rev. 2012;(3): CD008370 42. Soares-Weiser K, MacLehose H, Bergman H, Ben-Aharon I, Nagpal S, Goldberg E,
AC C
Pitan F, Cunliffe N. Vaccines for preventing rotavirus diarrhoea: vaccines in use. Cochrane Database Syst Rev. 2012;(11): CD008521
ACCEPTED MANUSCRIPT
RI PT
Supplementary file 2: Randomized trials from more developed and less developed countries
337 RCTs from More Developed Countries
137 RCTs from Less Developed Countries Turkey
Belgium Canada
USA
Africa Argentina
Tunisia
SC
Austria
Thailand
Denmark
M AN U
Australia
Vietnam Zambia Uganda
Europe
South Korea
Finland
Asia
Bangladesh Barbados
Taiwan
Brazil
Singapore
France
Norway
New Zealand Netherlands
Japan
EP
Scandinavia
AC C
Spain Scotland
Panama Mexico
Israel
TE D
Italy
Sweden
Portugal
Philippines
Greece
Switzerland
China
Qatar
Germany UK
Chile S. Africa
Jordan
Costa Rica
Mali
Latin America Korea
Kenya Jamaica
Dom. Republic Egypt Haiti Iran Indonesia India Hong Kong
French caribbean Ghana
ACCEPTED MANUSCRIPT
RI PT
Supplementary file 3: List of meta-analyses per category of Harm; with ROR of harm per meta-analysis between trials from more developed vs less developed countries [and 95% CIs thereof] ; summary OR of harm in trials from more developed countries per meta-analysis [and 95% CIs thereof] and summary OR of harm in trials in less developed countries per meta-analysis [and 95% CIs thereof])
Severe adverse events
M AN U
SC
Abbreviations: MD: more developed countries, LD: less developed countries, ROR: relative odds ratio of harm ; LCI_ROR: lower 95% confidence interval [CI] of ROR; UCI_ROR: upper 95% CI of ROR; sOR_MD: summary odds ration from trials in MD countries, sOR_LD: summary odds ratio from trials in LD countries; LCI_OR_MD: lower 95% CI of summary OR from trials in MD countries; LCI_OR_LD: lower 95% CI of summary OR from trials in LD countries; UCI_OR_MD: upper 95% CI of summary OR from trials in MD countries; UCI_OR_LD: upper 95% CI of summary OR from trials in LD countries
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
2.48
0.22
27.75
1.57
0.71
3.47
0.63
0.06
6.18
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
0.67
0.14
3.09
0.99
0.22
4.43
1.49
1.06
2.08
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent versus control
1.97
0.57
6.85
3.63
2.52
5.24
1.84
0.56
6.06
CD002130
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of nonST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage
During PCI (all patients)
Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding
2.39
0.30
19.03
1.40
1.14
1.73
0.59
0.07
4.62
Adverse events causing death
0.54
0.02
19.03
0.60
0.07
5.42
1.11
0.07
18.24
EP
AC C
CD002147
TE D
title
review
Terlipressin versus vasopressin
ACCEPTED MANUSCRIPT
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Non fatal serious adverse events
0.38
0.18
CD002922
Treatment for lupus nephritis
Immunosuppressive agent plus steroids versus steroids alone
Major infection
0.33
0.02
4.53
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Intraventricular haemorrhage (III-IV)
1.75
0.51
6.03
CD003897a
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Infection: CMV invasive
0.93
0.12
CD003897b
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Infection: serious allcause total
1.08
CD008521a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Serious adverse events
CD008521b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV5 versus placebo
All serious adverse events
0.55
1.03
1.96
0.96
3.98
0.77
0.37
1.59
2.33
0.19
29.04
1.52
0.70
3.28
0.87
0.33
2.29
7.26
0.97
0.59
1.58
1.04
0.14
7.61
0.39
2.98
1.00
0.62
1.60
0.92
0.38
2.26
0.88
0.63
1.25
0.82
0.62
1.08
0.93
0.76
1.13
0.81
0.45
1.46
0.75
0.47
1.20
0.92
0.65
1.31
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
Incidence of adverse events leading to stopping treatment
1.06
0.04
26.24
3.73
0.16
85.90
3.53
1.76
7.09
SC
M AN U
TE D
Discontinuations due to adverse events
0.83
0.75
RI PT
CD002309
title
comparison
CD000171a
Treatment of latent tuberculosis infection in HIV infected persons
Any TB drug vs placebo
CD000171b
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs isoniazid + rifampicin
Incidence of adverse events leading to stopping treatment
4.59
0.70
29.98
1.08
0.27
4.33
0.24
0.07
0.83
CD000171c
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs rifampicin + pyrazinimide
Incidence of adverse events leading to stopping treatment
0.53
0.19
1.52
0.46
0.23
0.93
0.86
0.40
1.85
AC C
EP
review
ACCEPTED MANUSCRIPT
Terlipressin for acute esophageal variceal hemorrhage
Terlipressin versus vasopressin
Adverse events causing withdrawal of treatment
0.22
0.01
0.05
1.36
1.11
0.07
18.19
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Withdrawals due to adverse effects
0.61
0.27
1.36
1.96
1.50
2.58
3.23
1.51
6.91
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
Changed drugs due to maternal side effects
1.22
0.19
7.67
1.61
0.58
4.43
1.32
0.28
6.08
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/no intervention
Side effects requiring discontinuation
2.11
0.04
113.13
2.25
1.29
3.93
1.06
0.02
54.76
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Quinolone vs. TMP-SMZ
Side effects requiring discontinuation
4.03
0.40
40.29
0.39
0.16
1.00
0.10
0.01
0.80
CD004773
Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Antifungal vs. Placebo
Discontinuation secondary to adverse events
0.74
0.07
8.17
2.40
1.17
4.95
3.25
0.33
32.10
CD007784
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis
Statins versus placebo/no treatment
Withdrawal due to adverse events
0.61
0.13
2.93
1.32
0.72
2.42
2.17
0.51
9.27
CD008521a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Adverse events requiring discontinuation (end of follow-up) Adverse events requiring discontinuation (end of follow-up)
1.92
0.63
5.87
1.48
0.56
3.96
0.77
0.45
1.32
CD008521b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV5 versus placebo
1.00
0.21
4.73
0.64
0.16
2.45
0.63
0.29
1.36
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
1.21
0.73
2.00
1.51
1.00
2.28
1.26
0.93
1.69
0.25
SC
M AN U
EP
TE D
5.92
RI PT
CD002147
AC C
Any adverse event
review
title
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
ACCEPTED MANUSCRIPT
Interventions for preventing oral candidiasis for patients with cancer receiving treatment
Comparisons with placebo/no treatment for all drug types
Toxicity (adverse events 'probably due to drug')
1.47
0.44
CD004183
Antioxidant supplements for preventing gastrointestinal cancers
Antioxidants versus placebo
Adverse effects beta-carotene
0.88
0.03
23.17
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/ no intervention
Side effects
1.78
0.03
92.13
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Quinolone vs. TMP-SMZ
Side effects
1.15
0.36
CD004773
Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Antifungal vs. Placebo
Adverse events
1.44
CD004791
Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastro-intestinal bleeding Stem cell treatment for acute myocardial infarction
Antibiotic prophylaxis versus placebo or no treatment
Adverse events
Stem cells compared to no stem cells,
Cilostazol versus aspirin for secondary prevention of vascular
Cilostazol versus aspirin in patients with ischaemic stroke or TIA,
review
title
CD000102a
Interventions for reducing the risk of mother-to-child transmission of HIV infection
CD000102b
Interventions for reducing the risk of mother-to-child transmission of HIV infection
0.83
1.95
0.87
0.28
2.67
2.07
0.65
6.60
2.34
0.11
49.56
1.89
1.47
2.44
1.06
0.02
54.76
3.63
0.48
0.24
0.95
0.41
0.17
1.04
0.60
3.46
1.89
1.13
3.16
1.31
0.64
2.68
0.75
0.02
24.58
2.30
0.61
8.71
3.06
0.12
76.95
Adverse effects
0.06
0.02
0.20
0.84
0.48
1.45
13.01
4.90
34.57
Other outcomes of safety during followup.
0.56
0.31
1.00
1.44
1.19
1.74
2.57
1.48
4.46
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
comparison
SC
M AN U
TE D
Organ system specific adverse events
EP
CD008076
AC C
CD006536
4.90
1.27
RI PT
CD003807
description of harm
Any zidovudine vs placebo/no treatment
Neonatal haematological toxicity
2.77
1.23
6.22
2.09
1.22
3.59
0.76
0.41
1.38
Any zidovudine vs placebo/no treatment
Maternal haematological toxicity
0.74
0.23
2.36
1.06
0.53
2.11
1.43
0.56
3.64
ACCEPTED MANUSCRIPT
Preoperative chemotherapy for resectable thoracic esophageal cancer
Preoperative chemotherapy vs surgery alone
Pulmonary complications
1.94
0.83
4.54
CD001556b
Preoperative chemotherapy for resectable thoracic esophageal cancer
Preoperative chemotherapy vs surgery alone
Cardiac complications
0.74
0.27
2.03
CD001556c
Preoperative chemotherapy for resectable thoracic esophageal cancer
Preoperative chemotherapy vs surgery alone
Gastrointestinal complications
0.00
0.00
0.06
1.14
0.67
1.96
0.59
0.31
1.14
0.89
0.44
1.80
1.19
0.59
2.41
1.15
0.34
3.81
441.20
24.93
7808.00
Individual adverse events (Group 1; ≥2 ma/adverse event)
M AN U
SC
RI PT
CD001556a
title
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD004069
Vitamin E supplementation in pregnancy
Any vitamin E supplementation
Bleeding (pregnancy)
1.08
0.06
18.45
0.33
0.03
3.22
0.31
0.06
1.67
CD004227
Antioxidants for preventing pre-eclampsia
Any antioxidants versus control or placebo
Bleeding (pregnancy)
0.13
0.02
1.16
0.99
0.59
1.63
7.45
0.90
61.73
CD001832
Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis
Third generation cephalosporins versus conventional therapy
Diarrhoea
3.92
1.35
11.34
2.50
1.54
4.05
0.64
0.25
1.64
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Diarrhoea
0.77
0.27
2.20
1.99
0.79
4.99
2.58
1.56
4.25
CD008521_a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 1
Diarrhoeareactogenicity
1.33
0.90
1.97
1.22
0.90
1.65
0.92
0.71
1.19
CD008521_b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 2
Diarrhoeareactogenicity
1.10
0.61
1.97
0.99
0.58
1.67
0.90
0.70
1.15
CD008521_c
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: at the end of follow up
Diarrhoeareactogenicity
1.20
0.84
1.72
1.05
0.79
1.41
0.88
0.71
1.08
AC C
EP
TE D
review
ACCEPTED MANUSCRIPT
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/ no intervention
Drug resistant infection
0.81
0.17
1.05
2.15
1.87
0.42
8.31
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Quinolone vs. TMP-SMZ
Drug resistant infection
4.89
0.48
50.17
0.48
0.18
1.29
0.10
0.01
0.80
CD001272
Antiplatelet agents for intermittent claudication
Antiplatelet agent versus placebo
Dyspepsia
0.82
0.22
3.10
2.78
0.93
8.26
3.40
1.58
7.34
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Dyspepsia
0.59
0.19
1.86
3.12
1.92
5.08
5.27
1.87
14.84
CD008521_a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 1
Fever- reactogenicity
1.06
0.68
1.66
1.22
0.81
1.84
1.15
0.96
1.38
CD008521_b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 2
Fever-reactogenicity
0.89
0.69
1.15
0.92
0.74
1.13
1.03
0.89
1.20
CD008521_c
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: at the end of follow up
Fever- reactogenicity
0.87
0.69
1.10
0.89
0.75
1.05
1.02
0.87
1.19
CD008076
Cilostazol versus aspirin for secondary prevention of vascular
Analysis 1.9. Comparison 1 Cilostazol versus aspirin in patients with ischaemic stroke or TIA, Ibuprofen vs placebo or none
GI bleeding
0.38
0.11
1.32
0.38
0.17
0.86
1.00
0.39
2.54
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
GI bleeding
1.10
0.03
35.38
3.08
0.12
77.80
2.80
0.78
9.99
CD002243
Corticosteroids for treating severe sepsis and septic shock
Steroids versus control
GI bleeding
1.07
0.06
18.24
1.10
0.74
1.64
1.03
0.06
17.20
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
Hypertension
1.81
0.08
43.14
1.09
0.39
3.08
0.60
0.03
11.99
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
Hypertension
3.02
0.11
86.62
1.49
0.23
9.54
0.49
0.03
8.10
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
IVH
0.73
0.10
5.35
0.71
0.14
3.66
0.97
0.31
2.98
SC
M AN U
TE D
EP
AC C
3.73
1.51
RI PT
CD004386a
ACCEPTED MANUSCRIPT
Intravenous immunoglobulin for preventing infection in preterm and/or low birth weight infants
IVIG vs placebo or no treatment
IVH
0.84
0.22
0.77
1.32
1.20
0.32
4.45
CD004659
Antiplatelet agents for preventing preeclampsia and its complications
IVH
0.87
0.12
6.42
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Antiplatelet agents versus placebo/no antiplatelet for primary prevention (subgrouped by maternal risk) Erythropoietin vs. placebo or no treatment
0.90
0.64
1.26
1.03
0.14
7.30
IVH
1.44
0.39
5.31
1.00
0.64
1.55
0.69
0.20
2.37
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
IVH
0.79
45.79
0.79
0.39
1.61
1.00
0.02
54.47
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet versus control in acute presumed ischaemic stroke
Intracranial hemorrhage
1.37
0.44
4.27
1.70
0.56
5.11
1.24
0.94
1.63
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant versus control in acute presumed ischaemic stroke
Intracranial hemorrhage
7.28
0.53
99.58
2.47
1.14
5.34
0.34
0.03
4.12
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs placebo or none
NEC
0.69
0.12
3.92
1.00
0.32
3.13
1.44
0.39
5.27
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
NEC
0.58
0.21
1.61
0.93
0.58
1.51
1.60
0.65
3.95
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
NEC
0.85
0.30
2.42
0.44
0.24
0.82
0.52
0.22
1.22
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
NEC
1.69
0.63
4.51
0.81
0.40
1.66
0.48
0.24
0.94
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
NEC
1.74
0.12
24.72
0.88
0.38
2.06
0.51
0.04
6.25
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
Neutropenia
1.10
0.06
18.82
0.81
0.48
1.37
0.74
0.05
11.98
SC 0.01
M AN U
TE D
EP
AC C
3.20
1.01
RI PT
CD000361
ACCEPTED MANUSCRIPT
Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis
Third generation cephalosporins versus conventional therapy
Neutropenia
1.43
0.24
0.18
2.83
0.50
0.16
1.57
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
ROP
1.51
0.85
2.69
1.42
0.99
2.05
0.94
0.60
1.46
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
ROP
1.33
0.31
5.66
1.00
0.30
3.35
0.75
0.34
1.66
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs placebo or none
ROP
0.75
0.10
5.61
1.07
0.58
1.98
1.42
0.21
9.55
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Sepsis
1.59
0.49
5.20
1.55
0.63
3.79
0.97
0.45
2.10
CD004868
Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Late initiation of EPO (8-28 days) vs. placebo or no intervention
Sepsis
2.19
0.19
24.96
0.69
0.32
1.49
0.32
0.03
3.18
CD004863
Early erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants
Erythropoietin vs. placebo or no treatment
Sepsis
1.05
0.51
2.18
0.88
0.62
1.24
0.83
0.44
1.59
CD005496
Probiotics for prevention of necrotizing enterocolitis in preterm infants
Probiotics vs. control
Sepsis
0.98
0.42
2.32
0.84
0.49
1.46
0.86
0.44
1.66
CD008521
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 1
Vomitingreactogenicity
0.99
0.74
1.34
1.07
0.83
1.39
1.08
0.93
1.26
CD008521_a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: after dose 2
Vomitingreactogenicity
0.66
0.44
0.99
0.68
0.47
0.98
1.03
0.86
1.24
CD008521_b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo: ath the end of follow up
Vomitingreactogenicity
0.88
0.63
1.24
0.88
0.66
1.19
1.00
0.86
1.16
SC
M AN U
TE D
EP
AC C
8.49
0.72
RI PT
CD001832
ACCEPTED MANUSCRIPT
RI PT
Individual adverse events (Group 2; 1 ma per adverse event)
title
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Abdominal pain
1.23
0.48
3.14
2.28
1.25
4.18
1.85
0.91
3.78
CD001556
Preoperative chemotherapy for resectable thoracic esophageal cancer
Treatment morbidity and mortality
Anastomotic leaks
0.25
0.02
2.86
0.89
0.54
1.45
3.51
0.33
37.79
CD000230
Zinc supplementation for improving pregnancy and infant outcome
Zinc supplementation versus no zinc (with or without placebo)
Any maternal infection
1.62
0.68
3.89
1.26
0.76
2.10
0.78
0.38
1.58
CD008076
Cilostazol versus aspirin for secondary prevention of vascular
Analysis 1.8. Comparison 1 Cilostazol versus aspirin in patients with ischaemic stroke or TIA,
Bleeding
1.29
0.56
2.94
0.75
0.60
0.95
0.59
0.26
1.30
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
CLD
1.95
0.72
5.26
1.69
0.79
3.61
0.87
0.46
1.66
CD000277
Calcium antagonists for aneurysmal subarachnoid haemorrhage
Calcium antagonists
Cerebral infarction on CT/MR
3.56
0.92
13.77
0.65
0.52
0.81
0.18
0.05
0.69
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Chorioamnionitis
0.47
0.21
1.10
0.81
0.56
1.18
1.71
0.81
3.61
CD000277
Calcium antagonists for aneurysmal subarachnoid haemorrhage
Calcium antagonists
1.58
0.46
5.46
0.63
0.36
1.09
0.40
0.13
1.21
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Clinical signs of secondary ischaemia: studies with magnesium in addition of nimodipine Decreased urine output
0.52
0.10
2.68
0.22
0.07
0.70
0.43
0.13
1.40
CD008370
Somatostatin analogues for pancreatic surgery
Somatostatin analogues versus none
Delayed gastric emptying
0.55
0.15
2.04
0.65
0.32
1.31
1.17
0.39
3.49
M AN U
TE D
EP AC C
SC
review
ACCEPTED MANUSCRIPT
Statins versus placebo/no treatment
Elevated LFTs
1.16
0.05
Corticosteroids versus placebo or no treatment
Fever in women after trial entry requiring the use of antibiotics
1.28
0.24
6.65
CD004386
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
quinolone vs. TMP-SMZ
Fungal infection
1.20
0.09
15.41
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
GI Bleeding (LBWI)
3.14
CD002243
Corticosteroids for treating severe sepsis and septic shock
Steroids versus control
Hyperglycaemia
1.67
CD004343
Ethamsylate for the prevention of morbidity and mortality in preterm or very low birth weight infants
Ethamsylate versus placebo
Hypotension
CD001556
Preoperative chemotherapy for resectable thoracic esophageal cancer
Treatment morbidity and mortality
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
CD003481
1.26
0.30
5.30
1.08
0.07
17.94
0.95
0.23
3.95
0.75
0.32
1.73
0.61
0.29
1.32
0.51
0.04
5.81
89.66
3.27
0.13
84.37
1.04
0.46
2.35
0.03
88.94
1.67
1.26
2.20
1.00
0.02
52.85
1.31
0.36
4.81
1.03
0.51
2.06
0.78
0.26
2.35
Infection
1.13
0.30
4.24
0.65
0.38
1.09
0.57
0.17
1.92
IL2Ra versus placebo or no treatment
Infection: CMV all
0.83
0.23
3.07
0.84
0.63
1.13
1.01
0.28
3.62
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
Intestinal perforation
0.28
0.01
7.52
0.19
0.01
4.21
0.67
0.23
1.99
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Intrapartum fever in woman requiring the use of antibiotics
0.12
0.00
3.23
0.24
0.03
2.22
1.98
0.18
22.18
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
Low APGAR scores
0.46
0.06
3.38
0.69
0.30
1.58
1.50
0.24
9.24
CD000102
Interventions for reducing the risk of mother-to-child transmission of HIV infection
Any zidovudine vs placebo/no treatment
Low birth weight
0.65
0.29
1.48
0.73
0.43
1.25
1.12
0.60
2.08
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Malignancy
1.34
0.21
8.70
0.68
0.38
1.20
0.51
0.09
3.02
AC C
EP
TE D
0.11
M AN U
CD004454
27.26
RI PT
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
SC
CD007784
ACCEPTED MANUSCRIPT
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Nausea
0.68
0.20
3.14
6.95
6.86
2.10
22.35
CD004659
Antiplatelet agents for preventing pre-eclampsia and its complications
Neonatal bleeding
0.82
0.39
1.70
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
Antiplatelet agents versus placebo/no antiplatelet for primary prevention (subgrouped by maternal risk) IL2Ra versus placebo or no treatment
1.07
0.76
1.51
1.32
0.69
2.52
PTLD/lymphoma
0.40
0.02
7.97
0.40
0.15
1.09
1.02
0.06
17.22
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs. indomethacin
PVL
1.09
5.90
1.10
0.35
3.42
1.01
0.29
3.51
CD003897
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra versus placebo or no treatment
Post-transplant diabetes mellitus
9.14
0.73
114.64
5.17
0.57
47.00
0.57
0.16
1.95
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Postnatal fever in woman
1.25
0.54
2.87
0.99
0.60
1.66
0.80
0.41
1.53
CD000230
Zinc supplementation for improving pregnancy and infant outcome
Zinc supplementation versus no zinc (with or without placebo)
Postpartum haemorrhage
2.93
0.60
14.44
1.62
0.86
3.06
0.55
0.13
2.39
CD000102
Interventions for reducing the risk of mother-to-child transmission of HIV infection
Any zidovudine vs placebo/no treatment
Premature delivery
2.75
0.83
9.10
1.27
0.59
2.71
0.46
0.18
1.16
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Proven Infection in NICU
1.23
0.48
3.16
0.81
0.43
1.53
0.66
0.33
1.31
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Puerperal sepsis
1.89
0.52
6.86
1.98
1.09
3.59
1.04
0.33
3.27
CD004213
Ibuprofen for the prevention of patent ductus arteriosus in preterm and/or low birth weight infants
Ibuprofen vs placebo or none
Pulmonary hypertension
3.94
0.04
401.58
3.59
0.33
38.70
0.91
0.02
48.05
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
RDS
0.63
0.10
4.04
0.23
0.06
0.93
0.36
0.11
1.21
CD001832
Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis
Third generation cephalosporins versus conventional therapy
Rash
0.22
0.01
3.54
0.29
0.04
2.24
1.31
0.20
8.48
SC 0.20
M AN U
TE D
EP
AC C
2.37
4.67
RI PT
CD002309
ACCEPTED MANUSCRIPT
Statins versus placebo/no treatment
Rhabdomyolysis
0.77
0.05
Late initiation of EPO (8-28 days) vs. placebo or no intervention
SIDS
0.61
0.03
12.78
Calcium antagonists
Secondary ischaemia
2.06
0.58
7.36
CD002243
Corticosteroids for treating severe sepsis and septic shock
Steroids versus control
Superinfections
3.10
CD004454
Antenatal corticosteroids for accelerating fetal lung maturation for women at risk of preterm birth
Corticosteroids versus placebo or no treatment
Systemic infection in the first 48 hours of life
0.83
CD004024
High-dose chemotherapy with autologous stem cell transplantation in the first line treatment of aggressive NonHodgkin Lymphoma (NHL) in adults Heparin low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures Immunoglobulin prophylaxis in hematological malignancies and hematopoietic stem cell transplantation Pentoxifylline for alcoholic hepatitis
High Dose Chemotherapy vs Control
Toxicities (secondary neoplasms)
Any heparin vs Control/Placebo
Polyvalent immunoglobulins vs. placebo / no intervention - HSCT
Heparin low molecular weight heparin and physical methods for preventing deep vein thrombosis and pulmonary embolism following surgery for hip fractures
Any heparin vs Control/Placebo
CD000305
CD006501
CD007339
CD000305
0.27
6.08
1.65
0.19
14.16
0.83
0.15
4.71
1.38
0.11
16.97
0.53
0.44
0.64
0.26
0.07
0.90
31.89
1.01
0.75
1.35
0.33
0.03
3.28
0.24
2.84
0.47
0.21
1.09
0.57
0.23
1.40
0.22
0.01
6.89
0.76
0.25
2.28
3.40
0.13
87.73
Transfusion
0.99
0.02
55.27
0.82
0.48
1.42
0.83
0.02
44.40
VOD
0.76
0.02
29.17
2.65
0.66
10.56
3.46
0.12
100.52
Variceal bleeding
0.85
0.03
28.17
2.13
0.19
24.20
2.50
0.20
31.00
Wound haematoma
3.62
0.11
122.44
2.09
0.18
24.62
0.58
0.05
7.12
Hepatic-related morbidity pentoxifylline versus control
SC 0.30
M AN U
TE D
CD000277
EP
CD004868
11.00
1.28
RI PT
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Late erythropoietin for preventing red blood cell transfusion in preterm and/or low birth weight infants Calcium antagonists for aneurysmal subarachnoid haemorrhage
AC C
CD007784
Discontinuations for any reason
title
comparison
description of harm
ROR
CD001272
Antiplatelet agents for intermittent claudication
Antiplatelet agent versus placebo
Early cessation of treatment
1.66
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
3.68
2.48
1.70
3.62
1.49
0.74
3.02
SC
review
RI PT
ACCEPTED MANUSCRIPT
M AN U
0.75
Primary Harms (Severe adverse events, Discontinuations due to adverse events and any adverse event)
title
comparison
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
2.48
0.22
27.75
1.57
0.71
3.47
0.63
0.06
6.18
CD000029
Antiplatelet therapy for acute ischaemic stroke
Antiplatelet versus control in acute presumed ischaemic stroke
Major extracranial haemorrhage during treatment period
0.67
0.14
3.09
0.99
0.22
4.43
1.49
1.06
2.08
CD000171a
Treatment of latent tuberculosis infection in HIV infected persons
Any TB drug vs placebo
Incidence of adverse events leading to stopping treatment
1.05
0.04
26.24
3.73
0.16
85.90
3.53
1.76
7.09
CD000171b
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs isoniazid + rifampicin
Incidence of adverse events leading to stopping treatment
4.59
0.70
29.98
1.08
0.27
4.33
0.24
0.07
0.83
CD000171c
Treatment of latent tuberculosis infection in HIV infected persons
Isoniazid vs rifampicin + pyrazinimide
Incidence of adverse events leading to stopping treatment
0.53
0.19
1.52
0.46
0.23
0.93
0.86
0.40
1.85
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent versus control
Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days
1.97
0.57
6.85
3.63
2.52
5.24
1.84
0.56
6.06
AC C
EP
TE D
review
ACCEPTED MANUSCRIPT
Platelet glycoprotein IIb/IIIa blockers vs Placebo/usual care (all patients)
30-day major bleeding
2.39
0.30
Terlipressin versus vasopressin
Adverse events causing withdrawal of treatment
0.22
0.01
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
1.21
0.73
CD002863
Oral beta-blockers for mild to moderate hypertension during pregnancy
Beta-blocker versus placebo/no beta-blocker
Changed drugs due to maternal side effects
1.22
0.19
CD002922
Treatment for lupus nephritis
Immunosuppressive agent plus steroids versus steroids alone
Major infection
0.33
CD003481
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interventions for preventing oral candidiasis for patients with cancer receiving treatment Interleukin 2 receptor antagonists for kidney transplant recipients
Ibuprofen vs. indomethacin
Intraventricular haemorrhage (III-IV)
Comparisons with placebo/no treatment for all drug types
CD004183
Antioxidant supplements for preventing gastrointestinal cancers
Antioxidants versus placebo
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy
Drug vs. placebo/ no intervention
CD003897
CD004386b
IL2Ra versus placebo or no treatment
1.73
0.59
0.07
4.62
0.25
0.05
1.35
1.11
0.07
18.24
2.00
1.51
1.00
2.28
1.26
0.93
1.69
7.67
1.61
0.58
4.43
1.32
0.28
6.08
0.02
4.53
0.77
0.37
1.59
2.33
0.19
29.04
1.75
0.51
6.03
1.52
0.70
3.28
0.87
0.33
2.29
Toxicity (adverse events 'probably due to drug')
1.47
0.44
4.90
1.27
0.83
1.95
0.87
0.28
2.67
Infection: serious all-cause total
1.08
0.39
2.98
1.00
0.62
1.60
0.92
0.38
2.26
Adverse effects - betacarotene
0.88
0.03
23.17
2.07
0.65
6.60
2.34
0.11
49.56
Side effects
1.78
0.03
92.13
1.89
1.47
2.44
1.06
0.02
54.76
Side effects
1.15
0.36
3.63
0.48
0.24
0.95
0.41
0.17
1.04
Quinolone vs. TMP-SMZ
5.88
SC
M AN U
TE D
CD003807
1.14
EP
CD002147
19.03
1.40
RI PT
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage
AC C
CD002130
ACCEPTED MANUSCRIPT
Antifungal vs. Placebo
Adverse events
1.44
0.60
Antibiotic prophylaxis versus placebo or no treatment
Adverse events
0.75
0.02
24.58
Stem cells compared to no stem cells,
Adverse effects
0.06
0.02
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular
Statins versus placebo/no treatment
Withdrawal due to adverse events
0.61
0.13
Cilostazol versus aspirin in patients with ischaemic stroke or TIA,
Other outcomes of safety during follow-up.
0.56
CD008521a
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
Serious adverse events
CD008521b
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV5 versus placebo
Adverse events requiring discontinuation (end of follow-up)
CD007784
CD008076
3.16
1.31
0.64
2.68
2.30
0.61
8.71
3.06
0.12
76.95
0.18
0.72
0.39
1.36
13.01
4.90
34.57
2.93
1.32
0.72
2.42
2.17
0.51
9.27
0.31
1.00
1.44
1.19
1.74
2.57
1.48
4.46
0.88
0.63
1.25
0.82
0.62
1.08
0.93
0.76
1.13
1.00
0.21
4.73
0.64
0.16
2.45
0.63
0.29
1.36
description of harm
ROR
LCI_ROR
UCI_ROR
sOR_MD
LCI_OR_MD
UCI_OR_MD
sOR_LD
LCI_OR_LD
UCI_OR_LD
SC
CD006536
1.13
M AN U
CD004791
3.46
1.89
RI PT
Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction
TE D
CD004773
Primary Harms
EP
USA vs Developing Countries
title
comparison
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
0.97
0.61
1.55
1.22
0.86
1.75
1.26
0.93
1.69
CD003807
Interventions for preventing oral candidiasis for patients with cancer receiving treatment
Comparisons with placebo/no treatment for all drug types
Toxicity (adverse events 'probably due to drug')
1.14
0.30
4.30
0.99
0.49
1.99
0.87
0.28
2.67
AC C
review
ACCEPTED MANUSCRIPT
Antioxidant supplements for preventing gastrointestinal cancers
Antioxidants vs placebo
Adverse effects - betacarotene
0.48
0.02
CD004386a
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction
Drug vs placebo/ no intervention
Side effects
2.29
0.04
121.40
Quinolone vs TMP-SMZ
Side effects
3.62
1.20
10.92
Antifungal vs Placebo
Adverse events
1.44
Antibiotic prophylaxis vs placebo or no treatment
Adverse events
1.01
Stem cells vs no stem cells
Adverse effects
CD000024
Anticoagulants for acute ischaemic stroke
Anticoagulant vs control in acute presumed ischaemic stroke
CD000213
Thrombolysis for acute ischaemic stroke
Any thrombolytic agent vs control
CD002130
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
Platelet glycoprotein IIb/IIIa blockers vs Placebo/usual care (all patients)
PDE4 inhibitors vs placebo
Treatment for lupus nephritis
Immunosuppressive agent plus steroids vs steroids alone
CD002309
CD002922
1.19
2.34
0.11
49.56
2.44
1.51
3.95
1.06
0.02
54.76
1.50
0.81
2.77
0.41
0.17
1.04
3.46
1.89
1.13
3.16
1.31
0.64
2.68
0.01
97.98
3.10
0.12
79.23
3.06
0.12
76.95
0.10
0.03
0.40
1.35
0.53
3.46
13.01
4.90
34.57
Major extracranial haemorrhage during treatment period
3.42
0.28
42.14
2.16
0.76
6.18
0.63
0.06
6.18
Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days
2.45
0.54
11.05
4.52
1.80
11.36
1.84
0.56
6.06
30-day major bleeding
2.26
0.28
18.58
1.33
0.87
2.01
0.59
0.07
4.62
Non fatal serious adverse events
0.36
0.16
0.81
0.70
0.48
1.04
1.96
0.96
3.98
Major infection
0.33
0.02
4.53
0.77
0.37
1.59
2.33
0.19
29.04
SC 0.60
M AN U
CD006536
1.08
TE D
CD004791
1.13
EP
CD004773
AC C
CD004386b
10.25
RI PT
CD004183
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra vs placebo or no treatment
Infection: CMV invasive
0.72
0.07
CD003897b
Interleukin 2 receptor antagonists for kidney transplant recipients
IL2Ra vs placebo or no treatment
Infection: serious all-cause total
0.91
0.29
2.85
CD008521
Vaccines for preventing rotavirus diarrhoea: vaccines in use
Rotarix vs placebo
Serious adverse events
0.52
0.21
1.25
CD000171
Treatment of latent tuberculosis infection in HIV infected persons
Any TB drug vs placebo
Incidence of adverse events leading to stopping treatment
0.28
0.11
CD002309
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
PDE4 inhibitors vs placebo
Withdrawals due to adverse effects
0.74
CD004386
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use
Drug vs placebo/no intervention
Side effects requiring discontinuation
Quinolone vs TMP-SMZ
Vaccines for preventing rotavirus diarrhoea: vaccines in use
RV1 vs placebo
CD008521
CD008521
Statins vs placebo/no treatment
RV5 vs placebo
1.04
0.14
7.61
0.84
0.41
1.70
0.92
0.38
2.26
0.48
0.20
1.13
0.93
0.76
1.13
0.70
0.99
0.55
1.79
3.53
1.76
7.09
0.33
1.68
2.40
1.77
3.25
3.23
1.51
6.91
2.07
0.04
114.47
2.21
1.04
4.68
1.06
0.02
54.76
Side effects requiring discontinuation
5.17
0.52
51.44
0.50
0.20
1.26
0.10
0.01
0.80
Discontinuation secondary to adverse events
0.74
0.07
8.17
2.40
1.17
4.95
3.25
0.33
32.10
Withdrawal due to adverse events
0.51
0.01
35.21
1.11
0.02
59.20
2.17
0.51
9.27
Adverse events requiring discontinuation (end of follow-up)
1.46
0.06
37.19
0.93
0.04
21.46
0.63
0.29
1.36
Adverse events requiring discontinuation (end of follow-up)
2.15
0.33
14.06
1.66
0.28
10.06
0.77
0.45
1.32
SC
M AN U
EP
CD007784
Antifungal vs Placebo
2.18
AC C
CD004773
6.91
0.74
RI PT
CD003897a
CD004386
0.25
TE D
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Discontinuations due to adverse events Any adverse event
99 RCTS
Sample size per RCT: median(IQR; range) 175 (63-620; 15-4,010)
Adverse event rate per RCT: median (IQR; range) 3.5% (1.5%-7.6%; 0.02%-85.6%)
LD
44 RCTs
278 (101-965; 14-21,106)
3.4% (1.3%-8.0%; 0.02%-29%)
MD
70 RCTs
151 (61-306; 18-1,310)
4.8% (1.9%-11.9%; 0.3%-23.3%)
LD
38 RCTS
377 (120-752; 47-10,519)
1.0% (0.5%-2.3%; 0.01%-9.8%)
MD
71 RCTS
LD
10 RCTS
MD
337 RCTs (688 study-level analyses†) 137 RCTs (368 study-level analyses†)
LD
82 (61-167; 20-39,876)
10.1% (3.7%-39.1%; 0.9%-94.3%)
143 (102-210; 56-1,018)
10.1% (2.0%-26.5%; 1.0%-74.9%)
128 (60-299; 15-39,876)
6.3% (3.0%-13.6%; 0.2%-94.3%)
157 (72-400; 4-21,106)
6.9% (2.6%-14.1%; 0.01%-97.0%)
TE D
All harms* (474 independent RCTs included in 1056 study-level analyses†)
N of RCTs
SC
Severe adverse events
Country -Setting MD
M AN U
Harm
RI PT
Supplementary file 4: Sample size and adverse event rate per RCT for the primary harm endpoints, acording to country setting.
Abbreviations: LD: from less developed countries, m-a: meta-analysis, MD: from more developed countries, RCT: randomized trials Footnote:
EP
*All harms (Severe adverse events or Discontinuations due to adverse events or Any adverse event or Organ-system specific adverse events or Individual adverse events or Any discontinuations)
AC C
†Each RCT could have contributed data to more than one study-level analyses for the different types of harms/harm-related endpoints.
ACCEPTED MANUSCRIPT
Supplementary file 5a: Severe adverse events (summary OR [sOR] by countrysetting per meta-analysis)
Severe Adverse Events review
RI PT
sOR (95% CIs)
1 CD000024_LD CD000024_MD . 2 CD000029_LD CD000029_MD . 3 CD000213_LD CD000213_MD . 4 CD002130_LD CD002130_MD . 5 CD002147_LD CD002147_MD . 6 CD002309_LD CD002309_MD . 7 CD002922_LD CD002922_MD . 8 CD003481_LD CD003481_MD . 9 CD003897a_LD CD003897a_MD . 10 CD003897b_LD CD003897b_MD . 11 CD008521a_LD CD008521a_MD . 12 CD008521b_LD CD008521b_MD . NOTE: Weights are from random effects analysis
0.63 (0.06, 6.18) 1.57 (0.71, 3.47)
SC
1.49 (1.06, 2.08) 0.99 (0.22, 4.43)
1.84 (0.56, 6.06) 3.63 (2.52, 5.24)
M AN U
0.59 (0.07, 4.62) 1.40 (1.14, 1.73)
1.11 (0.07, 18.24) 0.60 (0.07, 5.42)
1.96 (0.96, 3.98) 0.75 (0.55, 1.03)
TE D
2.33 (0.19, 29.04) 0.77 (0.37, 1.59)
0.87 (0.33, 2.29) 1.52 (0.70, 3.28)
EP
1.04 (0.14, 7.61) 0.97 (0.59, 1.58)
AC C
0.92 (0.38, 2.26) 1.00 (0.62, 1.60)
.3
.5
0.93 (0.76, 1.13) 0.82 (0.62, 1.08)
0.92 (0.65, 1.31) 0.75 (0.47, 1.20)
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 5b: Severe adverse events (RORs per meta-analysis and summary ROR [sROR] across meta-analyses)
RI PT
Severe Adverse Events ROR (95% CI)
review
CD000024
2.48 (0.22, 27.75)
CD000029
0.67 (0.14, 3.09)
CD000213
1.97 (0.57, 6.85)
SC
CD002130 CD002147 CD002309
M AN U
CD002922 CD003481 CD003897a CD003897b CD008521a CD008521b
TE D
EP AC C
1
2
0.38 (0.18, 0.83) 0.33 (0.02, 4.53)
0.93 (0.12, 7.26) 1.08 (0.39, 2.98) 0.88 (0.63, 1.25) 0.81 (0.45, 1.46) 0.86 (0.68, 1.11)
NOTE: Weights are from random effects analysis .5
0.54 (0.02, 19.03)
1.75 (0.51, 6.03)
Overall (I-squared = 0.0%, p = 0.550)
.3
2.39 (0.30, 19.03)
3
ACCEPTED MANUSCRIPT
Supplementary file 6a: Discontinuations due to adverse events, by country setting (summary OR [sOR] by country-setting per meta-analysis)
RI PT
Discontinuations due to Adverse Events sOR (95% CI)
1 CD000171a_LD CD000171a_MD
3.53 (1.76, 7.09) 3.73 (0.16, 85.90)
2 CD000171b_LD CD000171b_MD
0.24 (0.07, 0.83) 1.08 (0.27, 4.33)
SC
review
3 CD000171c_LD CD000171c_MD
0.86 (0.40, 1.85) 0.46 (0.23, 0.93)
4 CD002147_LD CD002147_MD
M AN U
1.11 (0.07, 18.19) 0.25 (0.05, 1.36)
5 CD002309_LD CD002309_MD
3.23 (1.51, 6.91) 1.96 (1.50, 2.58)
6 CD002863_LD CD002863_MD
1.32 (0.28, 6.08) 1.61 (0.58, 4.43)
7 CD004386a_LD CD004386a_MD
9 CD004773_LD CD004773_MD
0.10 (0.01, 0.80) 0.39 (0.16, 1.00)
3.25 (0.33, 32.10) 2.40 (1.17, 4.95)
2.17 (0.51, 9.27) 1.32 (0.72, 2.42)
EP
10 CD007784_LD CD007784_MD
TE D
8 CD004386b_LD CD004386b_MD
1.06 (0.02, 54.75) 2.25 (1.29, 3.93)
AC C
11 CD008521a_LD CD008521a_MD
0.63 (0.29, 1.36) 0.64 (0.16, 2.45)
12 CD008521a_LD CD008521b_MD
0.77 (0.45, 1.32) 1.48 (0.56, 3.96)
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 6b: Discontinuations due to adverse events (RORs per metaanalysis and summary ROR [sROR] across meta-analyses)
RI PT
Discontinuations due to Adverse Events review
ROR (95% CI)
CD000171a
1.05 (0.04, 26.24)
CD000171b
4.59 (0.70, 29.97)
CD000171c
0.53 (0.19, 1.52)
0.22 (0.01, 5.92)
SC
CD002147 CD002309
0.61 (0.27, 1.36) 1.22 (0.19, 7.67)
CD004386a CD004386b CD004773 CD007784 CD008521a CD008521b Overall (I-squared = 0.0%, p = 0.585)
2.11 (0.04, 113.12)
M AN U
CD002863
4.03 (0.40, 40.29) 0.74 (0.07, 8.17) 0.61 (0.13, 2.93) 1.00 (0.21, 4.73) 1.92 (0.63, 5.87) 0.93 (0.60, 1.43)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 7a: Any adverse event per country setting (summary OR [sOR] by country-setting per meta-analysis)
Any Adverse Event sOR (95% CI)
RI PT
review 1 CD002309a_LD CD002309a_MD
1.26 (0.93, 1.69) 1.51 (1.01, 2.28)
2 CD003807a_LD CD003807a_MD
0.87 (0.28, 2.67) 1.27 (0.83, 1.95)
SC
3 CD004183a_LD CD004183a_MD
2.34 (0.11, 49.56) 2.07 (0.65, 6.60)
M AN U
4 CD004386a_LD CD004386a_MD 5 CD004386b_LD CD004386b_MD 6 CD004773a_LD CD004773a_MD
1.06 (0.02, 54.75) 1.89 (1.47, 2.44)
0.41 (0.17, 1.04) 0.48 (0.24, 0.95)
1.31 (0.64, 2.68) 1.89 (1.13, 3.16)
3.06 (0.12, 76.95) 2.30 (0.61, 8.71)
8 CD006536a_LD CD006536a_MD
13.01 (4.90, 34.57) 0.84 (0.48, 1.45)
9 CD008076a_LD CD008076a_MD
2.57 (1.48, 4.46) 1.44 (1.19, 1.74)
AC C
EP
TE D
7 CD004791a_LD CD004791a_MD
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 7b: Any adverse event (ROR per meta-analysis and summary ROR [sROR] across metaanalyses)
RI PT
Any Adverse Event review
ROR (95% CI)
CD002309
1.21 (0.73, 2.00)
CD003807
1.47 (0.44, 4.90)
0.88 (0.03, 23.16)
SC
CD004183 CD004386a
1.78 (0.03, 92.13)
1.15 (0.36, 3.63)
CD004773 CD004791 CD006536 CD008076 Overall (I-squared = 69.9%, p = 0.001)
NOTE: Weights are from random effects analysis
.5
AC C
EP
TE D
.3
M AN U
CD004386b
1
2
1.44 (0.60, 3.46) 0.75 (0.02, 24.58) 0.06 (0.02, 0.20) 0.56 (0.31, 1.00) 0.73 (0.37, 1.46)
3
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
ACCEPTED MANUSCRIPT
Supplementary file 8a: Combined primary harms (summary OR [sOR] by countrysetting per meta-analysis)
review
sOR (95% CI)
1 Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period
CD000024_LD CD000024_MD
2 Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period
CD000029_LD CD000029_MD
3 Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment
CD000171a_LD CD000171a_MD
4 Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment
CD000171b_LD CD000171b_MD
5 Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment
CD000171c_LD CD000171c_MD
0.63 (0.06, 6.18) 1.57 (0.71, 3.47) 1.49 (1.06, 2.08) 0.99 (0.22, 4.43)
3.53 (1.76, 7.09) 3.73 (0.16, 85.90) 0.24 (0.07, 0.83) 1.08 (0.27, 4.33) 0.86 (0.40, 1.85) 0.46 (0.23, 0.93)
SC
outcome
RI PT
Combined Primary Harms
6 Symptomatic (including fatal) intracranial haemorrhage within 7 to 10CdD ay0s00213_LD dD ay0s00213_MD Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 C
1.84 (0.56, 6.06) 3.63 (2.52, 5.24)
7 30-day major bleeding 30-day major bleeding
CD002130_LD CD002130_MD
8 Adverse events causing withdrawal of treatment Adverse events causing withdrawal of treatment
CD002147_LD CD002147_MD
9 No of patients experiencing an adverse eff ect No of patients experiencing an adverse eff ect
CD002309_LD CD002309_MD
1.26 (0.93, 1.69) 1.51 (1.01, 2.28)
10 Changed drugs due to maternal side effects Changed drugs due to maternal side effects
CD002863_LD CD002863_MD
1.32 (0.28, 6.08) 1.61 (0.58, 4.43)
11 Major infection Major infection
CD002922_LD CD002922_MD
2.33 (0.19, 29.04) 0.77 (0.37, 1.59)
12 Intraventricular haem orrhage (III-IV) Intraventricular haem orrhage (III-IV)
CD003481_LD CD003481_MD
0.87 (0.33, 2.29) 1.52 (0.70, 3.28)
13 Toxicity (adverse events 'probably due to drug') Toxicity (adverse events 'probably due to drug')
CD003807_LD CD003807_MD
0.87 (0.28, 2.67) 1.27 (0.83, 1.95)
14 Infection: serious all-cause total Infection: serious all-cause total
CD003897_LD CD003897_MD
0.92 (0.38, 2.26) 1.00 (0.62, 1.60)
15 Adverse effects - beta-carotene Adverse effects - beta-carotene
CD004183_LD CD004183_MD
2.34 (0.11, 49.56) 2.07 (0.65, 6.60)
16 Side effects Side effects
CD004386a_LD CD004386a_MD
1.06 (0.02, 54.75) 1.89 (1.47, 2.44)
18 Adverse events Adverse events 19 Adverse events Adverse events 20 Adverse effects Adverse effects
M AN U
1.11 (0.07, 18.24) 0.25 (0.05, 1.35)
TE D
17 Side effects Side effects
0.59 (0.07, 4.62) 1.40 (1.14, 1.73)
CD004386b_LD CD004386b_MD
0.41 (0.17, 1.04) 0.48 (0.24, 0.95)
CD004773_LD CD004773_MD
1.31 (0.64, 2.68) 1.89 (1.13, 3.16)
CD004791_LD CD004791_MD
3.06 (0.12, 76.95) 2.30 (0.61, 8.71) 13.01 (4.90, 34.57) 0.72 (0.39, 1.36)
CD007784_LD CD007784_MD
2.17 (0.51, 9.27) 1.32 (0.72, 2.42)
22 Other outcomes of saf ety during follow-up. Other outcomes of saf ety during follow-up.
CD008076_LD CD008076_MD
2.57 (1.48, 4.46) 1.44 (1.19, 1.74)
23 Adverse events requiring discontinuation (end of f ollow-up) Adverse events requiring discontinuation (end of follow-up)
CD008521b_LD CD008521b_MD
0.63 (0.29, 1.36) 0.64 (0.16, 2.45)
24 Serious adverse events Serious adverse events
CD008521a_LD CD008521a_MD
0.93 (0.76, 1.13) 0.82 (0.62, 1.08)
AC C
EP
CD006536_LD CD006536_MD
21 W ithdrawal due to adverse events W ithdrawal due to adverse events
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 8b: Combined Primary Harms (Severe adverse events OR discontinuations due to adverse events OR any adverse event) (RORs per metaanalysis and summary ROR [sROR] across metaanalyses)
RI PT
Combined Primary Harms review
ROR (95% CI)
2.48 (0.22, 27.75) 0.67 (0.14, 3.09) 1.05 (0.04, 26.24) 4.59 (0.70, 29.97) 0.53 (0.19, 1.52) 1.97 (0.57, 6.85) 2.39 (0.30, 19.03) 0.22 (0.01, 5.88) 1.21 (0.73, 2.00) 1.22 (0.19, 7.67) 0.33 (0.02, 4.53) 1.75 (0.51, 6.03) 1.47 (0.44, 4.90) 1.08 (0.39, 2.98) 0.88 (0.03, 23.16) 1.78 (0.03, 92.13) 1.15 (0.36, 3.63) 1.44 (0.60, 3.46) 0.75 (0.02, 24.58) 0.06 (0.02, 0.18) 0.61 (0.13, 2.93) 0.56 (0.31, 1.00) 1.00 (0.21, 4.73) 0.88 (0.63, 1.25) 0.90 (0.65, 1.24)
TE D
M AN U
SC
CD000024 CD000029 CD000171a CD000171b CD000171c CD000213 CD002130 CD002147 CD002309 CD002863 CD002922 CD003481 CD003807 CD003897 CD004183 CD004386a CD004386b CD004773 CD004791 CD006536 CD007784 CD008076 CD008521a CD008521b Overall (I-squared = 39.3%, p = 0.026)
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 9a: Organ system specific adverse events (summary OR [sOR] by country-setting per meta-
RI PT
analysis)
Organ System Specific Adverse Events review
sOR (95% CI)
Maternal haematological toxicity CD000102a_LD CD000102a_MD . Neonatal haematological toxicity CD000102b_LD CD000102b_MD . Pulmonary complications CD001556a_LD CD001556a_MD . Cardiac complications CD001556b_LD CD001556b_MD . Gastrointestinal complications CD001556c_LD CD001556c_MD .
SC
1.43 (0.56, 3.64) 1.06 (0.53, 2.11)
M AN U
0.76 (0.41, 1.38) 2.09 (1.22, 3.59)
0.59 (0.31, 1.14) 1.14 (0.67, 1.96)
TE D
1.19 (0.59, 2.41) 0.89 (0.44, 1.80)
.5
AC C
EP
.3
1
441.21 (24.93, 7807.45) 1.15 (0.34, 3.81)
2
3
ACCEPTED MANUSCRIPT
Supplementary file 9b: Organ system specific adverse events (ROR per metaanalysis)
RI PT
Organ System Specific Adverse Events review
ROR (95% CI)
SC
0.74 (0.23, 2.35)
2.77 (1.23, 6.22)
TE D
M AN U
Maternal haematological toxicity CD000102 . Neonatal haematological toxicity CD000102 . Pulmonary complications CD001556 . Cardiac complications CD001556 . Gastrointestinal complications CD001556 .
AC C
EP
.3
.5
1
1.94 (0.83, 4.54)
0.74 (0.27, 2.03)
0.00 (0.00, 0.06)
2
3
ACCEPTED MANUSCRIPT
Supplementary file 10a: Individual adverse events (RORs per meta-analyses and summary ROR [sROR] per individual adverse event)
Individual Adverse Events_Group 1 review
RORr (95% CI)
Bleeding (pregnancy) Bleeding (pregnancy) CD004227 Bleeding (pregnancy) CD004069 Subtotal (I-squared = 24.4%, p = 0.250) . Diarrhoea Diarrhoea CD008521 Diarrhoea CD002309 Diarrhoea CD008521 Diarrhoea CD001832 Diarrhoea CD008521 Subtotal (I-squared = 28.2%, p = 0.234) . Drug resistant infection CD004386 Drug resistant infection Drug resistant infection CD004386 Subtotal (I-squared = 37.7%, p = 0.205) . Dyspepsia Dyspepsia CD001272 Dyspepsia CD002309 Subtotal (I-squared = 0.0%, p = 0.722) . Fever Fever CD008521 Fever CD008521 Fever CD008521 Subtotal (I-squared = 0.0%, p = 0.737) . GI bleeding GI bleeding CD008076 GI bleeding CD002243 GI bleeding CD004213 Subtotal (I-squared = 0.0%, p = 0.722) . Hypertension Hypertension CD004863 Hypertension CD004868 Subtotal (I-squared = 0.0%, p = 0.829) . IVH IVH CD000361 IVH CD004863 IVH CD004868 IVH CD003481 IVH CD004659 Subtotal (I-squared = 0.0%, p = 0.975) . Intracranial hemorrhage Intracranial hemorrhage CD000029 Intracranial hemorrhage CD000024 Subtotal (I-squared = 24.2%, p = 0.251) . NEC NEC CD004863 NEC CD004868 NEC CD004454 NEC CD004213 NEC CD003481 Subtotal (I-squared = 0.0%, p = 0.633) . Neutropenia Neutropenia CD001832 Neutropenia CD004863 Subtotal (I-squared = 0.0%, p = 0.879) . ROP ROP CD003481 ROP CD004213 ROP CD004863 Subtotal (I-squared = 0.0%, p = 0.804) . Sepsis Sepsis CD005496 Sepsis CD004868 Sepsis CD004863 Sepsis CD003481 Subtotal (I-squared = 0.0%, p = 0.861) . Vomiting Vomiting CD008521 Vomiting CD008521 CD008521 Vomiting Subtotal (I-squared = 22.1%, p = 0.277) . NOTE: Weights are from random effects analysis
0.13 (0.02, 1.16) 1.08 (0.06, 18.45) 0.31 (0.04, 2.31) 1.20 0.77 1.33 3.92 1.10 1.28
(0.84, (0.27, (0.90, (1.35, (0.61, (0.96,
1.72) 2.20) 1.97) 11.34) 1.97) 1.72)
4.89 (0.48, 50.17) 0.81 (0.17, 3.73) 1.59 (0.29, 8.81)
SC
0.82 (0.22, 3.10) 0.59 (0.19, 1.86) 0.68 (0.28, 1.62) (0.69, (0.69, (0.68, (0.77,
1.15) 1.10) 1.66) 1.06)
0.38 1.07 1.10 0.49
(0.11, (0.06, (0.03, (0.17,
1.32) 18.24) 35.38) 1.45)
M AN U
0.89 0.87 1.06 0.90
3.02 (0.11, 86.62) 1.81 (0.08, 43.14) 2.31 (0.23, 23.10)
0.99 (0.46, 2.13)
1.37 (0.44, 4.27) 7.28 (0.53, 99.58) 2.05 (0.50, 8.33)
TE D
0.58 1.74 0.85 0.69 1.69 0.95
.5
1
(0.21, (0.12, (0.30, (0.12, (0.63, (0.55,
1.61) 24.72) 2.42) 3.92) 4.51) 1.64)
1.43 (0.24, 8.49) 1.10 (0.06, 18.82) 1.33 (0.29, 6.00)
EP
AC C
.3
RI PT
AdverseEvent
2
3
1.33 0.75 1.51 1.42
(0.31, (0.10, (0.85, (0.85,
5.66) 5.61) 2.69) 2.38)
0.98 2.19 1.05 1.59 1.14
(0.42, (0.19, (0.51, (0.49, (0.70,
2.32) 24.96) 2.18) 5.20) 1.86)
0.66 0.99 0.88 0.86
(0.44, (0.74, (0.63, (0.69,
0.99) 1.34) 1.24) 1.08)
ACCEPTED MANUSCRIPT
Supplementary file 10b: Individual adverse events in Group 2 (1 ma per adverse event) (ROR per meta-analysis)
Individual Adverse Events_Group 2 AdverseEvent
RORr (95% CI)
1.23 (0.48, 3.14)
Anastomotic leaks
0.25 (0.02, 2.86)
Any maternal infection
1.62 (0.68, 3.89)
RI PT
Abdominal pain
Bleeding
1.29 (0.56, 2.94)
CLD
1.95 (0.72, 5.26)
Cerebral infarction on CT/MR
3.56 (0.92, 13.77) 0.47 (0.21, 1.10)
Chorioamnionitis Clinical signs of secondary ischaemia: studies with magnesium in addition of nimodipine
1.58 (0.46, 5.46) 0.52 (0.10, 2.68)
Decreased urine output Delayed gastric emptying
0.55 (0.15, 2.04)
1.16 (0.05, 27.26)
Elevated LFTs
1.28 (0.24, 6.65)
Fever in women after trial entry requiring the use of antibiotics Fungal infection
1.20 (0.09, 15.41) 3.14 (0.11, 89.65)
GI Bleeding (LBWI)
1.67 (0.03, 88.94)
Hyperglycaemia
1.31 (0.36, 4.81)
Hypotension
1.13 (0.30, 4.24)
Infection
0.83 (0.23, 3.07)
Infection: CMV all
0.28 (0.01, 7.52)
SC
Intestinal perforation Intrapartum fever in woman requiring the use of antibiotics Low APGAR scores Low birth weight Malignancy Nausea Neonatal bleeding PTLD/lymphoma
Post-transplant diabetes mellitus Postnatal fever in woman Postpartum haemorrhage Premature delivery Proven Infection in NICU Puerperal sepsis Pulmonary hypertension RDS Rash Rhabdomyolysis SIDS Secondary ischaemia Superinfections Systemic infection in the first 48 hours of life Toxicities (secondary neoplasms) Transfusion VOD Variceal bleeding W ound haematoma NOTE: W eights are from random effects analysis
EP
TE D
.3
AC C
0.46 (0.06, 3.38) 0.65 (0.29, 1.48) 1.34 (0.21, 8.70) 0.68 (0.20, 2.37) 0.82 (0.39, 1.70) 0.40 (0.02, 7.97) 1.09 (0.20, 5.90) 9.14 (0.73, 114.64)
M AN U
PVL
0.12 (0.00, 3.23)
.5
1
1.25 (0.54, 2.87) 2.93 (0.60, 14.44) 2.75 (0.83, 9.10) 1.23 (0.48, 3.16) 1.89 (0.52, 6.86) 3.94 (0.04, 401.55) 0.63 (0.10, 4.04) 0.22 (0.01, 3.54) 0.77 (0.05, 11.00) 0.61 (0.03, 12.78) 2.06 (0.58, 7.36) 3.10 (0.30, 31.89) 0.83 (0.24, 2.84) 0.22 (0.01, 6.89) 0.99 (0.02, 55.27) 0.76 (0.02, 29.17) 0.85 (0.03, 28.17) 3.62 (0.11, 122.43)
2
3
ACCEPTED MANUSCRIPT
Supplementary file 11: All Harm-related endpoints (ROR per meta-analysis) ALL Harm-related endpoints RORr (95% CII)
M AN U 1.66 (0.75,, 3.68))
0.06 (0.02,, 0.20)) 0.56 (0.31,, 1.00)) 0.75 (0.02,, 24.58)) 0.88 (0.03,, 23.16)) 1.15 (0.36,, 3.63)) 1.21 (0.73,, 2.00)) 1.44 (0.60,, 3.46)) 1.47 (0.44,, 4.90)) 1.78 (0.03,, 92.13))
0.22 (0.01,, 5.92)) 0.53 (0.19,, 1.52)) 0.61 (0.13,, 2.93)) 0.61 (0.27,, 1.36)) 0.74 (0.07,, 8.17)) 1.00 (0.21,, 4.73)) 1.05 (0.04,, 26.24)) 1.22 (0.19,, 7.67)) 1.92 (0.63,, 5.87)) 2.11 (0.04,, 113.12)) 4.03 (0.40,, 40.29)) 4.59 (0.70,, 29.97))
EP
TE D
0.13 (0.02,, 1.16)) 0.38 (0.11,, 1.32)) 0.58 (0.21,, 1.61)) 0.59 (0.19,, 1.86)) 0.66 (0.44,, 0.99)) 0.69 (0.12,, 3.92)) 0.73 (0.10,, 5.35)) 0.75 (0.10,, 5.61)) 0.77 (0.27,, 2.20)) 0.79 (0.01,, 45.79)) 0.81 (0.17,, 3.73)) 0.82 (0.22,, 3.10)) 0.84 (0.22,, 3.20)) 0.85 (0.30,, 2.42)) 0.87 (0.69,, 1.10)) 0.87 (0.12,, 6.42)) 0.88 (0.63,, 1.24)) 0.89 (0.69,, 1.15)) 0.98 (0.42,, 2.32)) 0.99 (0.74,, 1.34)) 1.05 (0.51,, 2.18)) 1.06 (0.68,, 1.66)) 1.07 (0.06,, 18.24)) 1.08 (0.06,, 18.45)) 1.10 (0.03,, 35.38)) 1.10 (0.61,, 1.97)) 1.10 (0.06,, 18.82)) 1.20 (0.84,, 1.72)) 1.33 (0.90,, 1.97)) 1.33 (0.31,, 5.66)) 1.37 (0.44,, 4.27)) 1.43 (0.24,, 8.49)) 1.44 (0.39,, 5.31)) 1.51 (0.85,, 2.69)) 1.59 (0.49,, 5.20)) 1.69 (0.63,, 4.51)) 1.74 (0.12,, 24.72)) 1.81 (0.08,, 43.14)) 2.19 (0.19,, 24.96)) 3.02 (0.11,, 86.62)) 3.92 (1.35,, 11.34)) 4.89 (0.48,, 50.17)) 7.28 (0.53,, 99.58))
AC C .3
3.23)) 6.89)) 3.54)) 2.86)) 7.52)) 4.53)) 7.97)) 3.38)) 1.10)) 2.68)) 2.04)) 12.78)) 4.04)) 1.48)) 2.37)) 29.17)) 11.00)) 1.70)) 3.07)) 2.84)) 28.17)) 55.27)) 5.90)) 4.24)) 27.26)) 15.41)) 3.16)) 3.14)) 2.87)) 6.65)) 2.94)) 4.81)) 8.70)) 5.46)) 3.89)) 88.94)) 6.86)) 5.26)) 7.36)) 9.10)) 14.44)) 31.89)) 89.65)) 13.77)) 122.43)) 401.55)) 114.64))
SC
0.12 (0.00,, 0.22 (0.01,, 0.22 (0.01,, 0.25 (0.02,, 0.28 (0.01,, 0.33 (0.02,, 0.40 (0.02,, 0.46 (0.06,, 0.47 (0.21,, 0.52 (0.10,, 0.55 (0.15,, 0.61 (0.03,, 0.63 (0.10,, 0.65 (0.29,, 0.68 (0.20,, 0.76 (0.02,, 0.77 (0.05,, 0.82 (0.39,, 0.83 (0.23,, 0.83 (0.24,, 0.85 (0.03,, 0.99 (0.02,, 1.09 (0.20,, 1.13 (0.30,, 1.16 (0.05,, 1.20 (0.09,, 1.23 (0.48,, 1.23 (0.48,, 1.25 (0.54,, 1.28 (0.24,, 1.29 (0.56,, 1.31 (0.36,, 1.34 (0.21,, 1.58 (0.46,, 1.62 (0.68,, 1.67 (0.03,, 1.89 (0.52,, 1.95 (0.72,, 2.06 (0.58,, 2.75 (0.83,, 2.93 (0.60,, 3.10 (0.30,, 3.14 (0.11,, 3.56 (0.92,, 3.62 (0.11,, 3.94 (0.04,, 9.14 (0.73,,
RI PT
review Individual ae_grp2 CD004454 CD004024 CD001832 CD001556 CD003481 CD002922 CD003897 CD002863 CD004454 CD003481 CD008370 CD004868 CD002863 CD000102 CD002309 CD006501 CD007784 CD004659 CD003897 CD004454 CD007339 CD000305 CD003481 CD001556 CD007784 CD004386 CD004454 CD002309 CD004454 CD004454 CD008076 CD004343 CD003897 CD000277 CD000230 CD002243 CD004454 CD003481 CD000277 CD000102 CD000230 CD002243 CD003481 CD000277 CD000305 CD004213 CD003897 . Any Discontinuation CD001272 . Any ae CD006536 CD008076 CD004791 CD004183 CD004386 CD002309 CD004773 CD003807 CD004386 . Discontinuations due to ae CD002147 CD000171 CD007784 CD002309 CD004773 CD008521 CD000171 CD002863 CD008521 CD004386 CD004386 CD000171 . Individual ae_grp1 CD004227 CD008076 CD004863 CD002309 CD008521 CD004213 CD003481 CD004213 CD002309 CD004868 CD004386 CD001272 CD000361 CD004454 CD008521 CD004659 CD008521 CD008521 CD005496 CD008521 CD004863 CD008521 CD002243 CD004069 CD004213 CD008521 CD004863 CD008521 CD008521 CD003481 CD000029 CD001832 CD004863 CD004863 CD003481 CD003481 CD004868 CD004868 CD004868 CD004863 CD001832 CD004386 CD000024 . Organ specific ae CD001556 CD000102 CD001556 CD001556 CD000102 . Severe ae CD002922 CD002309 CD002147 CD000029 CD008521b CD008521 CD003897 CD003897 CD003481 CD000213 CD002130 CD000024 . NOTE: W eights are from random effects analysis
.5
1
2
3
0.00 (0.00,, 0.74 (0.23,, 0.74 (0.27,, 1.94 (0.83,, 2.77 (1.23,,
0.06)) 2.35)) 2.03)) 4.54)) 6.22))
0.33 (0.02,, 0.38 (0.18,, 0.54 (0.02,, 0.67 (0.14,, 0.81 (0.45,, 0.88 (0.63,, 0.93 (0.12,, 1.08 (0.39,, 1.75 (0.51,, 1.97 (0.57,, 2.39 (0.30,, 2.48 (0.22,,
4.53)) 0.83)) 19.03)) 3.09)) 1.46)) 1.25)) 7.26)) 2.98)) 6.03)) 6.85)) 19.03)) 27.75))
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Footnote: Individual adverse events-group 1: adverse events with ≥2 meta-analyses per adverse event; Individual adverse events-group 2: adverse events with 1 meta-analysis per adverse event
ACCEPTED MANUSCRIPT
Supplementary file 12: Summary estimates (sRORs [Summary Relative Odds Ratios]: when there were ≥2 meta-analyses per harm/harm-related endpoint or RORs: when there was only 1 meta-analysis per harm/harm-related endpoint). Summary Estimates
Primary Endpoints
sROR (95% CIs)
I2 (95% CI), pvalue
Severe adverse events (12 m-a) Discontinuations due to adverse events (12 m-a) Any adverse event (9 m-a)
0.86 (0.68-1.11) 0.93 (0.60-1.43) 0.73 (0.37-1.46)
0 (0-58%); p=0.550 0 (0-58%); p=0.585 70% (40%-85%); p<0.001
Composite Harm Endpoint
sROR (95% CIs)
Combined Primary-Harms
0.90 (0.65-1.24)
SC
TE D
Gastrointestinal complications
ROR (95% CIs)
0.74 (0.23-2.35) 2.77 (1.23-6.22) 1.94 (0.83-4.54) 0.74 (0.27-2.03)
0.00 (0.00-0.06)
sROR (95% CIs) 0.31 (0.04-2.31) 1.28 (0.96-1.72) 1.59 (0.29-8.81) 0.68 (0.28-1.62) 0.90 (0.77-1.06) 0.49 (0.17-1.45) 2.31 (0.23-23.10) 0.99 (0.46-2.13) 2.05 (0.50-8.33) 0.95 (0.55-1.64) 1.33 (0.29-6.00) 1.42 (0.85-2.38) 1.14 (0.70-1.86)
Vomiting
0.86 (0.69-1.08)
Individual Harms –group 2 (N=46 harms; 1 m-a/harm) Abdominal pain Anastomotic leaks Any maternal infection
ROR (95% CIs)
EP
Individual Harms –group 1 (N=14 harms; ≥2 m-a/harm) Bleeding (during pregnancy) Diarrhoea Drug resistant infection Dyspepsia Fever GI bleeding Hypertension IVH Intracranial hemorrhage NEC Neutropenia ROP Sepsis
AC C
39% (1%-63%); p=0.026
M AN U
Organ system specific harms (N=5 harms; 1 m-a/ harm) Maternal haematological toxicity Neonatal haematological toxicity Pulmonary complications Cardiac complications
RI PT
Harm-related endpoints
1.23 (0.48-3.14) 0.25 (0.02-2.86) 1.62 (0.68-3.89)
1.20 (0.09-15.41) 3.14 (0.11-89.66) 1.67 (0.03-88.94) 1.31 (0.36-4.81) 1.13 (0.30-4.24) 0.83 (0.23-3.07) 0.28 (0.01-7.52) 0.12 (0.00-3.23)
TE D
M AN U
Intestinal perforation Intrapartum fever in woman requiring the use of antibiotics Low APGAR scores Low birth weight Malignancy Nausea Neonatal bleeding PTLD/lymphoma PVL Post-transplant diabetes mellitus
1.29 (0.56-2.94) 1.95 (0.72-5.26) 3.56 (0.92-13.77) 0.47 (0.21-1.10) 1.58 (0.46-5.46) 0.52 (0.10-2.68) 0.55 (0.15-2.04) 1.16 (0.05-27.26) 1.28 (0.24-6.65)
SC
Bleeding CLD Cerebral infarction on CT/MR Chorioamnionitis Clinical signs of secondary ischaemia Decreased urine output Delayed gastric emptying Elevated LFTs Fever in women after trial entry requiring the use of antibiotics Fungal infection GI Bleeding (LBWI) Hyperglycaemia Hypotension Infection Infection: CMV all
RI PT
ACCEPTED MANUSCRIPT
EP
Postnatal fever in woman Postpartum haemorrhage Premature delivery Proven Infection in NICU Puerperal sepsis Pulmonary hypertension
AC C
RDS Rash Rhabdomyolysis SIDS Secondary ischaemia Superinfections Systemic infection in the first 48 hours of life Toxicities (Secondary Neoplasms) Transfusion VOD Variceal bleeding Wound haematoma Any Discontinuation (1 m-a) Early cessation of treatment
0.46 (0.06-3.38) 0.65 (0.29-1.48) 1.34 (0.21-8.70) 0.68 (0.20-2.37) 0.82 (0.39-1.70) 0.40 (0.02-7.97) 1.09 (0.20-5.90) 9.14 (0.73114.64) 1.25 (0.54-2.87) 2.93 (0.60-14.44) 2.75 (0.83-9.10) 1.23 (0.48-3.16) 1.89 (0.52-6.86) 3.94 (0.04401.58) 0.63 (0.10-4.04) 0.22 (0.01-3.54) 0.77 (0.05-11.00) 0.61 (0.03-12.78) 2.06 (0.58-7.36) 3.10 (0.30-31.89) 0.83 (0.24-2.84) 0.22 (0.001-6.89) 0.99 (0.02-55.27) 0.76 (0.02-29.17) 0.85 (0.03-28.17) 3.62 (0.11122.43) ROR (95% CIs) 1.66 (0.75-3.68)
ACCEPTED MANUSCRIPT
Supplementary file 13: Severe adverse events (ROR per meta-analyses, calculated by FEM and summary ROR [sROR] across meta-analyses calculated by REM).
RI PT
Severe Adverse Events review
RORf (95% CI)
2.48 (0.22, 27.75)
CD000024
0.67 (0.14, 3.09)
SC
CD000029
1.94 (0.57, 6.63)
CD000213 CD002130
2.40 (0.30, 19.00) 0.54 (0.02, 19.03)
M AN U
CD002147 CD002309 CD002922 CD003481 CD003897 CD003897 CD008521
TE D
Overall (I-squared = 0.0%, p = 0.468)
0.38 (0.18, 0.83) 0.38 (0.03, 4.69) 1.75 (0.51, 6.03) 0.93 (0.12, 7.26) 1.07 (0.41, 2.78) 0.94 (0.69, 1.26) 0.91 (0.71, 1.17)
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3 3.5
ACCEPTED MANUSCRIPT
Supplementary file 14: Discontinuations due to adverse events (ROR per metaanalyses, calculated by FEM and summary ROR [sROR] across meta-analyses
RI PT
calculated by REM)
Discontinuations due to Adverse Events review
RORf (95% CI)
CD000171a
4.38 (1.08, 17.76)
CD000171b
0.53 (0.19, 1.52)
SC
CD000171c CD000171d CD000171e
M AN U
CD000171f CD002147 CD002309 CD002863 CD004386a CD004386b CD004773 CD007784 CD008521
TE D
Overall (I-squared = 39.6%, p = 0.063)
0.42 (0.16, 1.07) 3.21 (0.10, 105.42) 6.75 (0.30, 150.93) 0.32 (0.13, 0.80) 0.22 (0.01, 5.88) 0.60 (0.27, 1.33) 1.22 (0.19, 7.67) 2.11 (0.04, 112.90) 5.49 (0.63, 47.56) 0.74 (0.07, 8.17) 0.51 (0.11, 2.26) 1.86 (0.63, 5.51) 0.88 (0.52, 1.47)
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3 3.5
ACCEPTED MANUSCRIPT
Supplementary file 15: Any adverse event (ROR per meta-analyses, calculated by FEM and summary ROR [sROR] across meta-analyses calculated by REM)
RI PT
Any Adverse Event review
RORf (95% CI)
CD002309
0.99 (0.70, 1.39)
CD003807
1.47 (0.44, 4.90) 0.22 (0.01, 6.89)
SC
CD004024 CD004183 CD004386
CD004773 CD004791 CD006536 CD008076 Overall (I-squared = 67.1%, p = 0.001)
M AN U
CD004386
0.85 (0.04, 18.02) 1.79 (0.03, 92.44) 1.55 (0.58, 4.12) 1.44 (0.60, 3.46) 0.75 (0.02, 24.58) 0.06 (0.02, 0.20) 0.56 (0.31, 1.00) 0.72 (0.39, 1.33)
NOTE: Weights are from random effects analysis .5
AC C
EP
TE D
.3
1
2
3 3.5
ACCEPTED MANUSCRIPT
Supplementary file 16: Severe adverse events: USA trials vs trials from Less Developed (DC) Countries (ROR per meta-analysis and summary ROR [sROR]
RI PT
across meta-analyses; calculations by REMs)
Severe Adverse Events_USA vs DC review
RORr (95% CI)
3.42 (0.28, 42.14)
SC
CD000024 CD000213 CD002130
2.26 (0.28, 18.57) 0.36 (0.16, 0.81)
M AN U
CD002309
2.45 (0.54, 11.05)
CD002922 CD003897 CD003897 CD008521
0.33 (0.02, 4.53) 0.72 (0.07, 6.91) 0.91 (0.29, 2.85) 0.52 (0.21, 1.25)
Overall (I-squared = 21.6%, p = 0.258)
0.74 (0.42, 1.29)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 17: Discontinuations due to adverse events: USA trials vs trials from Less Developed (DC) countries. (ROR per meta-analysis and summary ROR
RI PT
[sROR] across meta-analyses; calculations by REM)
Discontinuations due to Adverse Events_USA vs DC review
RORr (95% CI)
0.28 (0.11, 0.70)
SC
CD000171 CD002309
CD004386 CD004773 CD007784 CD008521 CD008521 Overall (I-squared = 18.9%, p = 0.280)
2.07 (0.04, 114.46) 5.17 (0.52, 51.44)
M AN U
CD004386
0.74 (0.33, 1.68)
0.74 (0.07, 8.17) 0.51 (0.01, 35.21) 1.46 (0.06, 37.19) 2.15 (0.33, 14.06) 0.75 (0.38, 1.48)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
Supplementary file 18: Any adverse event: USA trials vs trials from Less Developed (DC) Countries. (ROR per meta-analysis and summary ROR [sROR] across meta-
RI PT
analyses; calculations by REM)
Any Adverse Event_USA vs DC review
RORr (95% CI)
0.97 (0.61, 1.55)
SC
CD002309 CD003807
CD004386 CD004386 CD004773 CD004791 CD006536 Overall (I-squared = 58.8%, p = 0.017)
0.48 (0.02, 10.25) 2.29 (0.04, 121.39)
M AN U
CD004183
1.14 (0.30, 4.30)
3.62 (1.20, 10.92) 1.44 (0.60, 3.46) 1.01 (0.01, 97.98) 0.10 (0.03, 0.40) 0.97 (0.47, 2.03)
TE D
NOTE: Weights are from random effects analysis
AC C
EP
.3
.5
1
2
3
ACCEPTED MANUSCRIPT
RI PT
Supplementary file 19: Baseline risks of harm/harm-related endpoints in the control group in meta-analyses from more developed and less developed countries in topics with statistically significant differences between trials from the two country settings (ROR and 95% CIs >1 or <1) Title
Comparison
Outcome
CD000102
Interventions for reducing the risk of mother-to-child transmission of HIV infection Preoperative chemotherapy for resectable thoracic esophageal cancer Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease
Any ZDV vs placebo/no treatment
Neonatal haematological toxicity GI complications Diarrhoea
Stem cell treatment for acute myocardial infarction Vaccines for preventing rotavirus diarrhoea: vaccines in use
Stem cells vs no stem cells Rotarix vs placebo: after dose 2
CD002309
CD006536
Baseline risk (Control Rate) in Less Developed Countries (95% CIs)
P of Difference in Baseline Risk
0.05 (0.01-0.11)
0.061
0.80
0.02 (0.01-0.04)
0.02 (0.00-0.07)
0.873
0.80
0.14 (0.08-0.21)
0.10 (0.04-0.19)
0.497
Non fatal serious adverse events Adverse effects
0.37
0.06 (0.04-0.08)
0.03 (0.02-0.05)
0.05
0.32
0.15 (0.08-0.25)
0.05 (0.02-0.10)
0.023
Reactogenicity: vomiting
0.50
0.12 (0.08-0.15)
0.11 (0.08-0.15)
0.922
EP
CD008521a
Baseline risk (Control Rat)e in More Developed Countries (95% CIs) 0.12 (0.08-0.16)
M AN U
CD001832
Treatment morbidity and mortality Third generation cephalosporins versus conventional therapy PDE4 inhibitors vs placebo
TE D
CD001556
Small study effect (pvalue) 0.34
SC
Review
Abbreviations: CI: confidence intervalsFootnote: p-value considered significant if <0.05
AC C
Footnote: Baseline harm/harm-related endpoint risks per country-setting were synthesized by random effects using the Freeman-Tukey transformation method and results between the two country settings by chi-square test.*p-value for Haber test: For the analysis for small study effect per meta-analysis, all studies included in each meta-analysis from all country settings were consider
Marked Revision
ACCEPTED MANUSCRIPT
Comparative Rates of Harms in Randomized Trials from More-Developed Versus LessDeveloped Countries May be Different
1 2 3
, Xanthippi Tseretopoulou,
RI PT
Despina Contopoulos-Ioannidis, clinical associate professor
registar in paediatrics 4, Megan Ancker, field physician 5, Juan N. Walterspiel, pediatric infectious diseases consultant
6
, Orestis A Panagiotou, post-doctoral fellow
Department of Pediatrics, Division of Infectious Diseases, Stanford University School of
Medicine, Stanford, CA, USA
M AN U
1
2
Palo Alto Medical Foundation Research Institute, Palo Alto, CA, USA
3
Meta-Research Innovation Center at Stanford, Stanford, CA, USA
6
7
TE D
5
Leeds Teaching Hospital, NHS Trust, Leeds, UK Medecins San Frontieres, Paris, France
Mendocino Coast District Hospital, Fort Bragg CA, USA Division of Cancer Epidemiology & Genetics, National Cancer Institute, National Institute of
EP
4
, Yvonne
SC
Maldonado, professor 1 8 9, John PA Ioannidis, professor 3 8 10 11
7
Health, Bethesda, MD, USA
Department of Health Research and Policy, Stanford University School of Medicine, Stanford,
AC C
8
CA, USA 9
Senior Associate Dean for Faculty Development and Diversity, Stanford University School of
Medicine, Stanford, CA, USA 10
Stanford Prevention Research Center, Department of Medicine, Stanford University School of
Medicine, Stanford, CA, USA
1
ACCEPTED MANUSCRIPT
11
Department of Statistics, Stanford University School of Humanities and Sciences, Stanford,
CA, USA Address correspondence to: Despina Contopoulos-Ioannidis, MD; Department of Pediatrics,
RI PT
Division of Infectious Diseases, Stanford University School of Medicine, 300 Pasteur Drive,
AC C
EP
TE D
M AN U
SC
Room G312, Stanford CA, 94305, USA; e-mail: [email protected]
2
ACCEPTED MANUSCRIPT
Funding: There was no funding for this study. Conflicts of interest: There are no conflicts of interest to declare for DCI, XT, MA, OAP, YM and JPAI. JNW worked as a consultant for US Bayer A.G. in 2002-3. “Whistleblower" on Bayer
that are paid for directly or indirectly by Bayer A.G.
RI PT
A.G. Currently does not hold any Bayer A.G or subsidiarie’s stock, nor does he give any lectures
The views expressed in this paper are solely those of the authors and do not necessarily represent
SC
the official views of the National Institutes of Health or the US Department of Health and
AC C
EP
TE D
M AN U
Human Services.
3
ACCEPTED MANUSCRIPT
Abstract Objectives: We set up to evaluate the relative risk of harms in trials performed in less-developed versus more-developed countries.
RI PT
Study Design: Meta-epidemiologic evaluation using the Cochrane Database of Systematic Reviews. We considered meta-analyses with at least one RCT in a less-developed country and one RCT in a more-developed country. We targeted severe adverse events (AEs),
SC
discontinuations due to AEs, any AE, organ system-specific AEs, individual AEs and all discontinuations due to any reason. We estimated the relative odds ratio (ROR) of harms
under each category of harms.
M AN U
between more and less-developed countries for each topic and the summary sROR across topics
Results: We identified 42 systematic-reviews (128 meta-analyses, 521 independent RCTs). Summary sRORs did not differ significantly from 1.00 for any harm category. Nominally
TE D
significant RORs were found in only 6/128 meta-analyses. However, in 27% (35/128) of metaanalyses the ROR point-estimates indicated relative differences between country-settings >2fold. Considering also ROR 95% confidence-intervals, in 92% (118/128) of meta-analyses one
EP
could not exclude a 2-fold difference in both directions. Conclusions: We identified limited comparative evidence on harms in trials from these two
AC C
country-settings. Substantial differences in the risk point-estimates were common; the potential for modest differences could rarely be excluded with confidence.
Key words: Comparative Safety, Comparative Harms, More developed countries, Less developed countries, Randomized trials, Meta-analyses.
4
ACCEPTED MANUSCRIPT
Running title: Comparative Harms in RCTs from More vs Less Developed Countries
What is new? Randomized trials on important clinical questions are frequently performed in less-developed
RI PT
•
countries, without long-standing clinical research tradition. We have previously shown that trials in such country-settings tend to report more favorable mortality outcomes. It remains
SC
unknown whether similar differences exist in the reporting of harms. We studied the comparative safety of medical interventions in randomized trials from more- versus less
•
M AN U
developed countries.
The identified comparative evidence on harms in trials from these two country-settings was limited.
•
Not many nominally-significant discrepancies in relative-risks of harms between country
TE D
settings were detected. However, substantial differences in the point-estimates of the risks of harms were common and the potential for modest differences between the two country settings could rarely be excluded with confidence.
•
EP
What this adds to what was known?
Due to the limited comparative evidence, it remains unclear whether safety data from trials in
AC C
less developed countries could be reliably used to guide policy development and clinical decision making in more developed countries, but also less developed countries. What is the implication, what should change now? •
Evaluation of the comparative safety of medical interventions in trials from these two country settings should be further pursued and additional comparative evidence needs to be accumulated.
5
ACCEPTED MANUSCRIPT
Alternative methodological approaches, i.e. through analysis of comparative safety data reported within international multisite trials, between trial sites in more developed countries
EP
TE D
M AN U
SC
RI PT
and trial sites in less developed countries should be also considered.
AC C
•
6
ACCEPTED MANUSCRIPT
Introduction As the budget for and participation rates in randomized clinical trials (RCTs) in more developed countries are limited, it is becoming increasingly more common that clinical
RI PT
guidelines and clinical decision-making about important questions of health interventions and health care will depend on evidence from trials performed in less developed countries.[1] With the globalization of clinical research, emerging countries are increasingly more actively involved
SC
in clinical trials.[2-4] Asian and Latin America regions have recently shown the largest annual increase in the number of registered clinical trials [5]. Trials done exclusively in less developed
M AN U
countries often have low methodological quality. [6] Only 56% of 670 surveyed researchers from developing countries reported that their research had been reviewed by a local institutional review board [7] and only 11% of published clinical trials conducted in China in 2004 report that their study protocol was reviewed by an ethical review committee.[8] Some reports are raising concerns for underreporting of adverse events from studies
TE D
performed in developing countries.[9] For example, large differences in the reported rates of ciprofloxacin associated arthropathy were seen in children from North America compared to
EP
children from Latin America.[9] However, the frequency of this phenomenon has not been systematically studied. If results from trials performed in countries without a longstanding
AC C
tradition in clinical research will be used to guide clinical decision-making, an empirical largescale evaluation of these trends is needed. While there are potential benefits from the globalization of clinical research, it is important to evaluate whether results are similar and possible to extrapolate across different settings. In a previous evaluation [10] we assessed differences in mortality and primary efficacy outcomes in RCTs performed on the same topic in more versus less developed countries. We found that on average, trials in less developed countries tended to report more favorable results for the 7
ACCEPTED MANUSCRIPT
experimental intervention [10]. Sometimes genuine differences between country settings could explain differences in results; however, selective outcome reporting, publication, language and other biases [11-13] in the literature coming from less developed countries was considered more It would be important to assess whether major harms
RI PT
likely to explain these discrepancies.
outcomes are also similar or different in RCTs from more versus less developed countries.
Very often the number of patients studied in pre-licensure trials is small to allow the
SC
robust evaluation of both safety and efficacy outcomes. [14] I Individual studies and even individual meta-analyses are on average underpowered to detect differences in reported rates of
M AN U
clinically important adverse events from different country settings.[15] Therefore, we performed a large-scale meta-epidemiological evaluation of safety outcomes in trials from more vs. less developed countries. Methods
TE D
Definitions of countries
The categorization of countries into more developed and less developed countries was done as previously described in our earlier paper on comparative results for mortality and
EP
primary efficacy outcomes.[10] In brief we considered more developed countries to be those with both longstanding established marker economies and longstanding tradition in clinical
AC C
research as previously suggested.[16] Such countries included the United States, Canada, Australia, New Zealand, Israel, Japan and Western European countries. All other countries except for those in Eastern Europe were considered as less developed. We excluded RCTs from Eastern European countries as these represent another unique type of countries in transition. [17] Harm-related endpoints
8
ACCEPTED MANUSCRIPT
We targeted 6 main categories of harm-related endpoints [18], three of which were considered as primary endpoints because they combine adverse events of all types and they only include harms. We did not focus on mortality, as this was the focus in our previous paper [10].
RI PT
The primary endpoint categories were: severe adverse events; discontinuations due to adverse events; and any adverse event. The secondary endpoint categories were: organ system-specific adverse events; individual adverse events; and all discontinuations-due to any reason (in some
SC
studies, discontinuations due to harm might not have been separately reported, but they could have been included under such a broader study endpoint; this endpoint would then be relevant
M AN U
for harms, even though not fully specific). Under the categories of organ system-specific and individual adverse events we considered several subcategories (e.g. gastrointestinal adverse events, hematologic adverse events etc.; and headache, neutropenia etc., respectively). These endpoints were considered secondary because they either do not include all adverse events or
TE D
they count also some events that are not due to adverse events.
We also considered a composite primary-harms endpoint (combined primary-harms), where all three primary harm endpoints (severe adverse events, discontinuations due to adverse
EP
events and any adverse event) were considered together. Eligible meta-analyses
AC C
We included meta-analyses that quantitatively synthesized evidence on harm-related
endpoints and included at least one RCT from a less developed country and at least one RCT from a more developed country for the same compared interventions and the same type of harmrelated endpoint. We focused on RCTs performed exclusively in more developed countries or exclusively in less developed countries; international multicenter trials were excluded unless all participating sites were from the same country-setting.
9
ACCEPTED MANUSCRIPT
Search We screened a list of 131 systematic reviews from the Cochrane Database of Systematic reviews (last update August 27, 2012) that we had previously identified [10] as having RCTs
RI PT
from more developed and less developed countries for mortality. The reasons for this choice were that this selected group of systematic reviews would have been more likely to have included also evidence on harms from both country settings; and the questions addressed were
SC
likely to pertain to serious clinical conditions and diseases given that death was also an outcome of interest. Meta-analyses within the same systematic review for different compared
M AN U
interventions or for different types of harm-related-endpoints were considered as separate metaanalyses. Thus, each systematic review could have contributed data to more than one category of harm-related endpoints. Data extraction
TE D
From each eligible meta-analysis we extracted information on the compared interventions, the harm-related endpoint, the study (author, year), the number of patients with harm-related endpoints and the total number of patients in the experimental group and the control
EP
group respectively per study. No continues harms were encountered. Statistical synthesis of data
AC C
We synthesized the data in the following sequential steps as previously described [10 18
19]: First, for each RCT we calculated the odds ratio (OR) of harm-related endpoints for the experimental vs. the control intervention. When needed we coined estimates so that an OR>1 always corresponds to an increased risk of the harm-related outcome. If in the 2×2 table for each study (events/total number of subjects in the experimental arm and the control arm respectively) there were one or more cells with zero events, then a standard correction of 0.5 was added to all
10
ACCEPTED MANUSCRIPT
cells to allow for a meaningful calculation of the OR, as previously described.[20] Second, for each individual meta-analysis, under each category of harm-related endpoint, we calculated the summary OR (sOR) from all RCTs from the more developed countries and the sOR from all
RI PT
RCTs from the less developed countries using random-effect model for the meta-analysis; we also performed sensitivity analyses using fixed-effects model. [21 22] Random effects model assumes that the combined studies aim at identifying an average treatment effect, while fixed
SC
effects model assumes a common effect across the combined studies. Third, for each individual meta-analysis, and for each category of harm-related endpoint, we calculated the relative OR
M AN U
(ROR) between more developed and less developed countries by dividing the sOR from trials in more developed countries by the sOR from trials in less developed countries. Fourth, for each category of harm-related endpoint, we generated a summary ROR (sROR) between more and less developed countries by synthesizing the RORs from all the individual meta-analyses, under
TE D
each category of harm-related endpoint using random-effects models. A ROR>1 (or sROR>1) means that less developed countries reported less unfavorable results for the experimental intervention relative to more developed countries.
EP
For the combined primary-harms endpoint, we synthesized available RORs from all 3 primary endpoints; whenever RORs existed for more than one primary endpoint for the same
AC C
meta-analysis, we used the ROR that had the smaller variance. For the meta-analyses for organ system-specific adverse events and individual adverse events, we combined RORs only when they were targeting exactly the same types of organ system-specific or individual adverse events. We calculated the between meta-analyses heterogeneity under each category of harm-
related endpoint by using the I2 metric with its corresponding 95% CIs.[23 24]
11
ACCEPTED MANUSCRIPT
We calculated the number of meta-analyses where there was a statistically significant difference in the sOR between the two country settings (95% CIs of ROR excluding 1.00) and
the two country settings (ROR≥ 2.0 or ROR ≤0.50).
RI PT
the number of meta-analyses where the ROR suggested difference exceeding 2.0-fold between
In topics with statistically significant differences in harm-related endpoints between the two country settings, we also explored whether there were statistically significant differences in
SC
the baseline risks of these endpoints in the control group between meta-analyses from more developed and less developed countries. We synthesized the baseline risks per country setting by
M AN U
random effects using the Freeman-Tukey transformation method [25] for the meta-analysis of proportions and compared the results in the two country settings by chi-square test. We also examined whether there was evidence for small study effects using the Harbord’s test (considered to be significant for P<0.10). [26]
TE D
In sensitivity analyses we compared meta-analyses of U.S. trials only vs. trials from less developed countries. We also performed sensitivity analyses for the primary endpoints where we excluded trials from four Asian “tigers” (Hong Kong, Taiwan, Singapore, and South Korea) that
EP
evolved from less developed countries into advanced economies according to the International Monetary Fund in 1997, although their tradition of clinical research is still not as long-standing
AC C
as in the main more developed countries. All analyses were done in Stata SE12 (College Station, TX: StataCorp LP). We followed
the PRISMA [27] (Preferred Reporting Items for Systematic Reviews and Meta-analyses) guidelines for reporting. Results Eligible topics 12
ACCEPTED MANUSCRIPT
Of the previously identified 131 systematic reviews [10], 42 were considered eligible (Supplementary file 1) with a total of 128 separate meta-analyses: 12 meta-analyses for severe adverse events, 12 for discontinuations due to adverse events, 9 for any adverse event, 5 for
RI PT
organ system-specific adverse events, 89 for individual adverse events and 1 for all discontinuations due to any reason (Figure 1). From a total of 521 independent RCTs -after excluding 38 from mixed country settings, 5 with no reported country setting and 4 from Eastern
SC
European countries- we eventually analyzed 474 RCTs; 337 from more developed countries and 137 from less developed countries (Supplementary file 2). The topics of the 42 systematic
M AN U
reviews were related to infectious diseases (n=10), neonatal diseases (n=6), maternal-fetal conditions (n=6), cardiovascular diseases (n=5), neurologic diseases (n=5), gastrointestinal diseases (n=3), oncologic conditions (n=3), transplantations (n=2), rheumatologic diseases (n=1) and pulmonary diseases (n=1). The individual topics appear in Supplementary file 3.
TE D
The median sample size per RCT from more developed and less developed countries for the three primary harm endpoints was 175 and 278 respectively for severe adverse events; 151 and 377 respectively for discontinuations due to adverse events and 82 and 143 respectively for
EP
any adverse event (Supplementary file 4). The median adverse event rate per RCT in more developed and less developed countries was 3.5% and 3.4% respectively for severe adverse
AC C
events; 4.8% and 1.0% for discontinuations due to adverse events and 10.1% and 10.1% respectively for any adverse event (Supplementary file 4). Relative differences in individual meta-analyses (more-developed vs. less-developed) The clinical topics and the ROR estimates between the two country settings for the three primary endpoints and the combined primary-harms endpoint are shown in Table 1 and Supplementary file 3; the respective figures for the summary OR per meta-analysis according to 13
ACCEPTED MANUSCRIPT
country-setting and the ROR per meta-analysis, for each category of these harms are shown in Supplementary files 5-8. Respectively, the summary ORs per country-setting per meta-analysis and the ROR per meta-analysis for each of the secondary endpoint categories are shown in
RI PT
Supplementary files 9-11. The summary estimates (sRORs) for the different primary and secondary endpoints and the combined primary-harms endpoint are shown in Supplementary file
SC
12.
The summary sRORs did not differ significantly from 1.00 for any category of harm.
M AN U
When all harm-related endpoints were considered, nominally statistically significant relative differences between more developed and less developed countries were identified in 5% (6/128) of all meta-analyses (ROR>1 in two and ROR<1 in four) (Supplementary file 3). However, in 27% (35/128) of meta-analyses the relative differences in the summary ORs between more and less developed countries, based on the ROR point estimates, were larger than 2-fold. More
TE D
specifically, RORs≥ 2.0 were seen on 20 meta-analyses while ROR ≤ 0.50 in 15 meta-analyses (p=0.36) (Supplementary files 3 and 11).
EP
There was also large uncertainty left around the estimated RORs and in 92% (118/128) of meta-analyses, when the ROR 95% confidence intervals were also considered, one could not
AC C
exclude a 2-fold difference in harms in both directions (lower ROR CI≥0.5 and upper ROR CI≤2.0) (Supplementary file 11). Moreover, in 67% (86/128) of the meta-analyses, there were more than 10-fold differences between the upper and lower CIs of the RORs. More specifically this occurred in 67% of meta-analyses for severe adverse events; 75% of meta-analyses for discontinuations due to adverse events; 44% of meta-analyses for any adverse events; 40% of meta-analyses for organ system specific adverse events; and 69% of meta-analyses for individual adverse events. 14
ACCEPTED MANUSCRIPT
Results were similar when fixed effects were used to synthesize the sORs for the primary harm-related endpoints per country-setting within each meta-analysis. (Supplementary files 1315). In sensitivity analyses comparing only USA trials against trials from less developed
RI PT
countries 4/24 (17%) of meta-analyses for the primary endpoints showed nominally differences beyond chance for the comparison by the country setting (ROR>1 in one and ROR<1 in three) (Supplementary files 3, 16-18). Further sensitivity analyses excluding trials from the four Asian
SC
“tigers” also gave similar results for the primary endpoints and the combined primary-harms
M AN U
endpoint. RORs
Table 2 shows in more detail the 6 topics for which a nominally statistically significant difference was found between studies from less developed and more developed countries. For 2
TE D
topics, less developed countries reported less unfavorable results (RORs>1); while the opposite occurred in 4 topics (RORs<1). Studies from less developed countries had reported relatively less neonatal hematologic toxicities with zidovudine (vs. placebo) for prevention of HIV mother
EP
to child transmission; and relatively less diarrhea with third generation cephalosporins (vs. conventional antibiotics) for acute bacterial meningitis. While studies from less developed
AC C
countries had reported relatively more severe adverse events with the use of phosphodiesterase inhibitors (vs. placebo) for chronic obstructive pulmonary disease; relatively more adverse events of any type with stem cell therapy for acute myocardial infarction; relatively more gastrointestinal complications with preoperative chemotherapy (vs. surgery alone) for resected thoracic and esophageal cancer; and relatively more vomiting after the second dose of rotavirus vaccines (vs placebo).
15
ACCEPTED MANUSCRIPT
For the above 6 topics there were no statistically significant difference in the baseline risks of harms in the control groups (control rates) between trials from less and more developed
RI PT
countries; nor was there evidence of small study effects (Supplementary file 19). Discussion
Our study provided a bird’s eye view of the pattern of reporting of harms and harm-
SC
related endpoints in trials from the two country settings. We explored the relative differences in rates of harms across different country-settings using diverse approaches in order to obtain a
M AN U
more comprehensive appraisal. Our inferences need to be cautious given the large uncertainty in the relative incidence of harms for many of the examined topics. Due to the dearth of harm data in individual randomized trials but also in individual meta-analyses, it would not have been possible to identify significant differences between country-sett ings and onl y a bird’s e ye view
TE D
and large scale evaluation across diverse topics was able to probe such country-setting specific harm differences. Overall, across all analyzed meta-analyses for all types of harm-related endpoints, we did not detect nominally significant discrepancies in the reported rates of harm-
EP
related endpoints between trials from less developed and more developed countries when results were summarized for each harm category. However, when the point estimates of the relative
AC C
differences of harms between more and less developed countries were considered, in 27% of the meta-analyses the relative differences were larger than 2-fold. Many meta-analyses had limited evidence and substantial uncertainty and when the uncertainty around the relative differences were also considered, differences in the relative odds of harms exceeding 2-fold in both directions could not be excluded in 92% of meta-analyses.
16
ACCEPTED MANUSCRIPT
We should acknowledge some study limitations. First, we did not screen the whole Cochrane Database for Systematic Reviews with harm-related endpoints. However, the 131 systematic reviews we screened was a group of meta-analyses that was highly likely to have
RI PT
included harm-related endpoints both from trials in more developed and less developed countries as they had already included trials from both country settings for mortality. By focusing also on mortality, our pool of these 131 screened systematic reviews was about clinically important
SC
medical conditions and interventions. Trials from more developed countries were the preponderance, as we had also observed in our assessments of mortality previously. [10]
M AN U
However, trials from less developed countries are expected to become more common in the future. Second, in the 128 individual meta-analyses, included in the 42 eligible systematic reviews, with comparative evidence on harm-related endpoints from the two country settings only a tenth had data on the primary harm endpoints; the majority had reported individual
TE D
specific adverse events. The clinical significance of these individual harms, without any additional information on their severity grading, is unclear.[28] This probably reflects poor reporting for harms and poor compliance with CONSORT standards of reporting, which is a
37].
EP
universal problem in randomized trials independently of the trial’s country setting [29-31] [32Third, we focused on trials performed exclusively in more-developed or less-developed
AC C
countries and did not analyze international multisite trials. It remains unknown whether the reported rates of harms in trials from other countries in economic and social transition and from study-sites from different country settings included in the same international multisite trials are systematically different. However, the vast majority of published reports of international studies included in the CDSR do not report results separately per country setting. In our database, less than 10% of the RCTs were international studies from mixed country settings. In Bayer A.G.’s
17
ACCEPTED MANUSCRIPT
ciprofloxacin patent extension studies the FDA noted an arthropathy rate of 21 % reported by investigators in the United States compared to none in a less developed country [9]. This Bayer/Quintiles database was further analyzed and divided into countries [38], and ultimately
conclusions are still pending.
RI PT
handed over to the US Department of Justice for a false claims investigation [39]; final Fraud must be ruled out whenever significant discrepancies
findings, rather than following a patient for adverse events.
SC
between investigator groups are found, as it is easy for an untrained investigator to report no
In summary, although we did not identify consistent over- or under-estimation of risks of
M AN U
harms between trials done in more versus less developed countries, substantial differences in the estimates of risks were relatively common and the potential for modestly large differences often could not be excluded with confidence. Collection and reporting of harms should be emphasized and standardized in all randomized trials, both from more developed and less developed
TE D
countries. More emphasis should be given to the standardized reporting of clinically important harms, such as severe adverse events or discontinuations due to adverse events. It is very important to have available the information required for the prompt identification of systematic
EP
differences in harms across different country settings. Differences may reflect genuine variability in the incidence of harms or differences in the reporting of harms. Reported rates of harms
AC C
may be influenced by the genetic profiles of populations, but may also depend upon ethical practices and monitoring systems in place in different countries. Disentangling these reasons would require that harms/harm-related endpoints are collected and reported transparently and in a harmonized way across randomized trials [28]. Country-setting discrepancies in the reported rates of harms should be routinely incorporated and explored in meta-analyses and systematic reviews. Moreover, regulatory agencies, such as the FDA or the EMEA, should enforce that such
18
ACCEPTED MANUSCRIPT
transparent harm-related information is provided per country and site for all randomized trials, including also multisite international trials. This may allow understanding how often and why some countries and sites have different results. Until such transparent comparative evidence on
RI PT
harms is available from trials in different country-settings, guideline developers and policy makers should cautiously extrapolate harm data from trials from one country setting to guide
AC C
EP
TE D
M AN U
SC
clinical decision making in very different country settings.
19
ACCEPTED MANUSCRIPT
Contributorship: DCI had the original idea and developed the study analysis plan; JPAI, XT, MA, OAP, JNW and YM critically commented on the study analysis plan; DCI, XT, MA performed data extraction; DCI and OAP performed the statistical analyses; DCI drafted the first
RI PT
manuscript draft and all authors critically reviewed it and approved the final version; DCI is the guarantor for this paper. Ethical approval: Not applicable.
SC
Transparency declaration: Dr. Contopoulos-Ioannidis (the manuscript’s guarantor) affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that
as planned have been explained.
M AN U
no important aspects of the study have been omitted; and that any discrepancies from the study
Copyright/license for publication: The Corresponding Author has the right to grant on behalf of all authors and does grant on behalf of all authors, “a worldwide license to the Publishers and
TE D
its licensees in perpetuity, in all forms, formats and media (whether known now or created in the future), to: i) publish, reproduce, distribute, display and store the Contribution, ii) translate the Contribution into other languages, create adaptations, reprints, include within collections and
EP
create summaries, extracts and/or, abstracts of the Contribution, iii) create any other derivative work(s) based on the Contribution, iv) to exploit all subsidiary rights in the Contribution, v) the
AC C
inclusion of electronic links from the Contribution to third party material where-ever it may be located; and, vi) license any third party to do any or all of the above.”
20
ACCEPTED MANUSCRIPT
References 1. Murthy VH, Krumholz HM, Gross CP. Participation in cancer clinical trials: race-, sex-, and age-based disparities. JAMA 2004;291(22):2720-6
RI PT
2. George M, Selvarajan S, S SK, et al. Globalization of clinical trials - where are we heading? Current Clinical Pharmacology 2013;8(2):115-23
3. Duke T, Fuller D. Randomised controlled trials in child health in developing countries: trends
SC
and lessons over 11 years. Archives of Disease in Childhood 2014;99(7):615-20.
4. Dandona L, Katoch VM, Dandona R. Research to achieve health care for all in India. Lancet
M AN U
2011;377(9771):1055-7
5. Drain PK, Robine M, Holmes KK, et al. Trail watch: global migration of clinical trials. Nature reviews. Drug Discovery 2014;13(3):166-7
6. Zhang D, Freemantle N, Cheng KK. Are randomized trials conducted in China or India
2011;64(1):90-5
TE D
biased? A comparative empirical analysis. Journal of Clinical Epidemiology
7. Hyder AA, Wali SA, Khan AN, et al. Ethical review of health research: a perspective from
EP
developing country researchers. Journal of Medical Ethics 2004;30(1):68-72 8. Zhang D, Yin P, Freemantle N, et al. An assessment of the quality of randomised controlled
AC C
trials conducted in China. Trials 2008;9:22
9. Division of special pathogen and immunologic drug products. Summary of clinical review of studies
submitted
in
response
to
a
pediatric
written
request.
http://www.fda.gov/ohrms/dockets/ac/05/briefing/20054152b1_03_03_Cipro%20medical.pdf Accessed 1/21/2015
21
ACCEPTED MANUSCRIPT
10. Panagiotou OA, Contopoulos-Ioannidis DG, Ioannidis JP. Comparative effect sizes in randomised trials from less developed and more developed countries: metaepidemiological assessment. BMJ 2013;346:f707
RI PT
11. Lorenzo C, Garrafa V, Solbakk JH, et al. Hidden risks associated with clinical trials in developing countries. Journal of Medical Ethics 2010;36(2):111-5
2002;21(19):2869-77
SC
12. Smith R. Publishing research from developing countries. Statistics in Medicine
13. Juni P, Holenstein F, Sterne J, et al. Direction and impact of language bias in meta-analyses trials:
empirical
study.
International
M AN U
of controlled
2002;31(1):115-23
Journal
of Epidemiology
14. Duijnhoven RG, Straus SM, Raine JM, et al. Number of patients studied prior to approval of new medicines: a database analysis. PLoS Medicine 2013;10(3):e1001407
TE D
15. Tsang R, Colley L, Lynd LD. Inadequate statistical power to detect clinically significant differences in adverse event rates in randomized controlled trials. Journal of Clinical Epidemiology 2009;62(6):609-16
gs-r,
EP
16. United Nations Statistics Division. Composition of macro geographical (continental) regions and
selected
economic
and
other
groupings.
AC C
2013.http://unstats.un.org/unsd/methods/m49/m49regin.htm.
17. Platonov P. Clinical trials in Russia and Eastern Europe: recruitment and quality. International Journal of Clinical Pharmacology and Therapeutics 2003;41(7):277-80
18. Lathyris D, Panagiotou OA, Baltogianni M, et al. Safety of medical interventions in children versus adults. Pediatrics 2014;133(3):e666-73
22
ACCEPTED MANUSCRIPT
19. Contopoulos-Ioannidis DG, Baltogianni MS, Ioannidis JP. Comparative effectiveness of medical
interventions
in
adults
versus
children.
The Journal
of Pediatrics
20. http://handbook.cochrane.org/ ChfsroiTCCVuM.
RI PT
2010;157(2):322-30
21. Lau J, Ioannidis JP, Schmid CH. Quantitative synthesis in systematic reviews. Annals of Internal Medicine 1997;127(9):820-6
SC
22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Controlled Clinical Trials 1986;7(3):177-88
M AN U
23. Ioannidis JP, Patsopoulos NA, Evangelou E. Uncertainty in heterogeneity estimates in metaanalyses. BMJ 2007;335(7626):914-6
24. Higgins JP, Thompson SG. Quantifying heterogeneity in a meta-analysis. Statistics in Medicine 2002;21(11):1539-58
1950;21:607-11
TE D
25. Freeman MF TJ. Transformations related to the angular and the square root. Ann Math Statist
26. Harbord RM EM, Sterne JA. . A modified test for small-study effects in meta-analyses of
EP
controlled trials with binary endpoints. Statistics in Medicine 2006;25:3443-57. 27. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic
AC C
reviews and meta-analyses of studies that evaluate health care interventions: explanation
and elaboration. Annals of Internal Medicine 2009;151(4):W65-94
28. Ioannidis JP, Evans SJ, Gotzsche PC, et al. Better reporting of harms in randomized trials: an extension of the CONSORT statement. Annals of Internal Medicine 2004;141(10):781-8
23
ACCEPTED MANUSCRIPT
29. Papanikolaou PN, Ioannidis JP. Availability of large-scale evidence on specific harms from systematic reviews of randomized trials. The American Journal of Medicine 2004;117(8):582-9
medical areas. JAMA 2001;285(4):437-43
RI PT
30. Ioannidis JP, Lau J. Completeness of safety reporting in randomized trials: an evaluation of 7
31. Ioannidis JP. Adverse events in randomized trials: neglected, restricted, distorted, and
SC
silenced. Archives of Internal Medicine 2009;169(19):1737-9
32. Mukherjee SD, Coombes ME, Levine M, et al. A qualitative study evaluating causality
M AN U
attribution for serious adverse events during early phase oncology clinical trials. Investigational New Drugs 2011;29(5):1013-20 .
33. Wahab IA, Pratt NL, Kalisch LM, et al. The detection of adverse events in randomized clinical trials: can we really say new medicines are safe? Current Drug Safety
TE D
2013;8(2):104-13
34. Schroll JB, Maund E, Gotzsche PC. Challenges in coding adverse events in clinical trials: a systematic review. PloS One 2012;7(7):e41174.
EP
35. Huang HY, Andrews E, Jones J, et al. Pitfalls in meta-analyses on adverse events reported from clinical trials. Pharmacoepidemiology and Drug Safety 2011;20(10):1014-20
AC C
36. Haidich AB, Birtsou C, Dardavessis T, et al. The quality of safety reporting in trials is still suboptimal: survey of major general medical journals. Journal of Clinical Epidemiology 2011;64(2):124-35
37. de Vries TW, van Roon EN. Low quality of reporting adverse drug reactions in paediatric randomised controlled trials. Archives of Disease in Childhood 2010;95(12):1023-6
24
ACCEPTED MANUSCRIPT
38. Personal communication with Dr N. Balasubramanian from the Department of Statistics at Stanford University 39. United States of America, ex. rel. Juan N. Walterspiel MD, FAAP v. Bayer A.G., Quintiles False Claims Act action under 31 U.S.C. § 3729(a).
AC C
EP
TE D
M AN U
SC
https://www.pacer.gov/ Accessed 01/18/2015.
RI PT
Transnational Corp.
25
ACCEPTED MANUSCRIPT
Table 1: Harm-related endpoints
CD000029
CD000213
SC
Description of Harm- N of RCTs related endpoint (N patients) from more developed countries Anticoagulants for acute Anticoagulant vs Major extracranial 15 (2,506) ischaemic stroke control haemorrhage during treatment period Antiplatelet therapy for Antiplatelet vs Major extracranial 6 (796) acute ischaemic stroke control acute haemorrhage during treatment period Thrombolysis for acute Any thrombolytic Symptomatic (including 20 (6,244) ischaemic stroke agent vs control fatal) intracranial haemorrhage within 7 to 10 days Platelet glycoprotein Platelet 30-day major bleeding 20 (23,259) IIb/IIIa blockers during glycoprotein (all patients) percutaneous coronary IIb/IIIa blockers intervention and as the during PCI vs initial medical treatment control of non-ST segment elevation acute coronary syndromes Terlipressin for acute Terlipressin vs Adverse events causing 3 (204) esophageal variceal vasopressin death hemorrhage Phosphodiesterase 4 PDE4 inhibitors vs Non fatal serious 8 (2,965) inhibitors for chronic placebo adverse events
N of RCTs ROR (95% CIs) (N patients) from less developed countries 2 (344) 2.48 (0.22-27.75)
1 (21,106)
0.67 (0.14-3.09)
2 (567)
1.97 (0.57-6.85)
3 (448)
2.39 (0.30-19.03)
2 (101)
0.54 (0.02-19.03)
1 (1,018)
0.38 (0.18-0.83)
CD002147
CD002309
AC C
EP
CD002130
Comparison
M AN U
CD000024
Title
TE D
Review
RI PT
Severe adverse events
26
ACCEPTED MANUSCRIPT
CD003897b
CD008521a
CD008521b
AC C
Discontinuations due to adverse events
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.04)
CMV 3 (881)
1 (118)
0.93 (0.12-7.26)
Infection: serious all- 6 (1,009) cause total
2 (132)
1.08 (0.39-2.98)
Serious adverse events
18 (29,581)
0.88 (0.63-1.25)
7 (7,863)
0.81(0.45-1.46)
RI PT
5 (266)
Infection: invasive
SC
Intraventricular haemorrhage (III-IV)
M AN U
CD003897a
1 (29)
TE D
CD003481
Major infection
All serious events
7 (3,179)
adverse 3 (2,547)
EP
CD002922
obstructive pulmonary disease Treatment for lupus Immunosuppressiv nephritis e agent plus steroids vs steroids alone Ibuprofen for the Ibuprofen vs treatment of patent indomethacin ductus arteriosus in preterm and/or low birth weight infants Interleukin 2 receptor IL2Ra vs placebo antagonists for kidney or no treatment transplant recipients Interleukin 2 receptor IL2Ra vs placebo antagonists for kidney or no treatment transplant recipients Vaccines for preventing RV1 vs placebo rotavirus diarrhoea: vaccines in use Vaccines for preventing RV5 versus placebo rotavirus diarrhoea: vaccines in use
Review
Title
Comparison
Description of Harmrelated endpoint
CD000171a
Treatment of latent Any TB drug vs Incidence of adverse tuberculosis infection in placebo events leading to
N of RCTs (N patients) from more developed countries 2 (838)
N of RCTs ROR (95% CIs) (N patients) from less developed countries 6 (4,695) 1.05 (0.04-26.24)
27
ACCEPTED MANUSCRIPT
CD002863
CD004386a
CD004386b
CD004773
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
3 (206)
2 (103)
0.22 (0.01-5.92)
7 (2,891)
1 (1,018)
0.61 (0.27-1.36)
1 (154)
1.22 (0.19-7.67)
1 (97)
2.11 (0.04-113.13)
1 (102)
4.03 (0.40-40.29)
1 (129)
0.74 (0.07-8.17)
RI PT
SC
PDE4 inhibitors vs placebo
1 (1,092)
M AN U
CD002309
3 (509)
Beta-blocker vs Changed drugs due to 10 (923) placebo/no beta- maternal side effects blocker
TE D
CD002147
stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Adverse events causing withdrawal of treatment Withdrawals due to adverse effects
Drug vs placebo/no Side effects requiring 16 (2,057) intervention discontinuation
EP
CD000171c
Isoniazid vs isoniazid + rifampicin Isoniazid vs rifampicin + pyrazinimide Terlipressin versus vasopressin
Quinolone vs TMP- Side effects requiring 6 (752) SMZ discontinuation
AC C
CD000171b
HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV
Antifungal Placebo
vs Discontinuation 2 (723) secondary to adverse events
28
ACCEPTED MANUSCRIPT
CD008521b
RV5 vs placebo
Vaccines for preventing RV1 vs placebo rotavirus diarrhoea: vaccines in use
Any adverse event
CD003807
CD004183
Adverse requiring discontinuation follow-up) Adverse requiring discontinuation follow-up)
to 8 (1,652)
events 5 (3,709)
3 (236)
0.61 (0.13-2.93)
4 (7,735)
1.00 (0.21-4.73)
15 (21,774)
1.92 (0.63-5.87)
(end of
events 6 (2,184)
(end of
Description of Harm- N of RCTs related endpoint (N patients) from more developed countries Phosphodiesterase 4 PDE4 inhibitors vs No of patients 8 (2,913) inhibitors for chronic placebo experiencing an obstructive pulmonary adverse effect disease Interventions for All drug types vs Toxicity (adverse 2 (661) preventing oral placebo/no events 'probably due to candidiasis for patients treatment drug') with cancer receiving treatment Antioxidant Antioxidants (beta- Adverse effects 3 (91,080) supplements for carotene) vs preventing placebo
TE D
Comparison
EP
CD002309
Title
AC C
Review
vs Withdrawal due adverse events
RI PT
Statins placebo/no treatment
SC
CD008521a
HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use
M AN U
CD007784
N of RCTs ROR (95% CIs) (N patients) from less developed countries 1 (1,018) 1.21 (0.73-2.00)
1 (210)
1.47 (0.44-4.90)
1 (174)
0.88 (0.03-23.17)
29
ACCEPTED MANUSCRIPT
CD006536
CD008076
9 (929)
1 (102)
1.15 (0.36-3.63)
vs Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Adverse events
6(450)
1 (102)
0.75 (0.02-24.58)
8 (669)
2 (213)
0.06 (0.02-0.20)
1 (787)
0.56 (0.31-1.00)
RI PT
1.78 (0.03-92.13)
Antibiotic prophylaxis placebo/ treatment
SC
Antifungal placebo
vs no
M AN U
CD004791
1 (97)
TE D
CD004773
33 (5,002)
Quinolone vs TMP- Side effects SMZ
Stem cells vs no Adverse effects stem cells
EP
CD004386b
Drug vs placebo/ Side effects no intervention
Cilostazol vs Safety outcomes 1 (3,202) aspirin in patients during follow-up with ischaemic stroke or TIA
AC C
CD004386a
gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction Cilostazol versus aspirin for secondary prevention of vascular events after stroke of arterial origin
Combined Primary-Harms (Serious Adverse event, Discontinuations due to adverse event and Any adverse event)
30
ACCEPTED MANUSCRIPT
CD000171a
Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Thrombolysis for acute ischaemic stroke
CD000171c
CD000213
CD002130
RI PT
Antiplatelet therapy for acute ischaemic stroke
SC
CD000029
M AN U
Anticoagulants for acute ischaemic stroke
Description of Harm- N of RCTs (N patients) related endpoint from more developed countries Anticoagulant vs Major extracranial 15 (2,506) control in acute haemorrhage during presumed treatment period ischaemic stroke Antiplatelet vs Major extracranial 6 (796) control in acute haemorrhage during presumed treatment period ischaemic stroke Any TB drug vs Incidence of adverse 2 (838) placebo events leading to stopping treatment Isoniazid vs Incidence of adverse 3 (509) isoniazid + events leading to rifampicin stopping treatment Isoniazid vs Incidence of adverse 2 (373) rifampicin + events leading to pyrazinamide stopping treatment Any thrombolytic Symptomatic 20 (5,622) agent vs control (including fatal) intracranial haemorrhage within 7 to 10 days Platelet 30-day major bleeding 20 (23,259) glycoprotein IIb/IIIa blockers vs placebo/usual care (all patients)
TE D
CD000024
CD000171b
Comparison
EP
Title
AC C
Review
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST
N of RCTs ROR (95% CIs) (N patients) from less developed countries 2 (344) 2.48 (0.22-27.75)
1 (21,106)
0.67 (0.14-3.09)
6 (4,695)
1.05 (0.04-26.24)
1 (1,092)
4.59 (0.70-29.98)
2 (1,455)
0.53 (0.19-1.52)
2 (567)
1.97 (0.57-6.85)
3 (448)
2.39 (0.30-19.03)
31
ACCEPTED MANUSCRIPT
CD002922
CD003481
1 (1,018)
1.21 (0.73-2.00)
1 (154)
1.22 (0.19-7.67)
5 (266)
1 (29)
0.33 (0.02-4.53)
3 (285)
4 (261)
1.75 (0.51-6.03)
2 (661 )
1 (210)
1.47 (0.44-4.90)
6 (1,009)
2 (132)
1.08 (0.39-2.98)
3 (91,080)
1 (174)
0.88 (0.03-23.17)
RI PT
0.22 (0.01-5.88)
Beta-blocker vs Changed drugs due to 10 (923) placebo/no beta- maternal side effects blocker
Immunosuppressive Major infection agent plus steroids vs steroids alone Ibuprofen for the Ibuprofen vs Intraventricular treatment of patent indomethacin haemorrhage (III-IV) ductus arteriosus in preterm and/or low birth weight infants Interventions for Comparisons with Toxicity (adverse preventing oral placebo/no events 'probably due to candidiasis for patients treatment for all drug') with cancer receiving drug types treatment Interleukin 2 receptor IL2Ra vs placebo Infection: serious allantagonists for kidney or no treatment cause total transplant recipients Antioxidant Antioxidants vs Adverse effects - betasupplements for placebo carotene preventing gastrointestinal cancers
CD003897
CD004183
AC C
EP
CD003807
2 (101)
SC
CD002863
vs Adverse events causing 3 (204) withdrawal of treatment PDE4 inhibitors vs No of patients 8(2,913) placebo experiencing an adverse effect
M AN U
CD002309
Terlipressin vasopressin
TE D
CD002147
segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Treatment for lupus nephritis
32
ACCEPTED MANUSCRIPT
CD006536
CD007784
CD008076
1.78 (0.03-92.13)
9 (929)
1 (102)
1.15 (0.36-3.63)
1 (295)
1 (129)
1.44 (0.60-3.46)
6 (450)
1 (102)
0.75 (0.02-24.58)
6 (617)
2 (213)
0.06 (0.02-0.18)
to 8 (1,652)
3 (236)
0.61 (0.13-2.93)
Cilostazol vs Other outcomes of 1 (3,202) aspirin in patients safety during followwith ischaemic up. stroke or TIA
1 (787)
0.56 (0.31-1.00)
RI PT
33 (5,002)
Antifungal Placebo
Antibiotic prophylaxis placebo or treatment
SC
Quinolone vs TMP- Side effects SMZ
vs Adverse events
M AN U
CD004791
1 (97)
Adverse events
vs no
TE D
CD004773
Drug vs placebo/ no Side effects intervention
Stem cells Adverse effects compared to no stem cells, Statins vs Withdrawal due placebo/no adverse events treatment
EP
CD004386b
Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastrointestinal bleeding Stem cell treatment for acute myocardial infarction HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular after stroke of arterial
AC C
CD004386a
33
ACCEPTED MANUSCRIPT
RI PT
1.00 (0.21-4.73)
18 (29,581)
0.88 (0.63-1.25)
TE D
M AN U
SC
Vaccines for preventing RV1 vs placebo rotavirus diarrhoea: vaccines in use
4 (7,7733)
EP
CD008521b
Adverse events 5 (3,708) requiring discontinuation (end of follow-up) Serious adverse events 7 (3,179)
AC C
CD008521a
origin Vaccines for preventing RV5 vs placebo rotavirus diarrhoea: vaccines in use
34
ACCEPTED MANUSCRIPT
Table 2: Individual meta-analyses with nominally significant relative differences between more developed and less developed
PDE4 inhibitors placebo
sOR_MD (95% sOR_LD CIs) CIs)
(95%
0.75 (0.55-1.03)
1.96 (0.96-3.98)
0.06 (0.02-0.20)
0.84 (0.48-1.45)
13.01 (4.90-34.57)
Any zidovudine vs Neonatal placebo/no treatment haematological toxicity
2.77 (1.23-6.22)
2.09 (1.22-3.59)
0.76 (0.41-1.38)
0.003 (0.001-0.06)
1.15 (0.34-3.81)
441.20 (24.93-7808)
Third generation Diarrhoea cephalosporins versus conventional therapy
3.92 (1.35-11.34)
2.50 (1.54-4.05)
0.64 (0.25-1.64)
Rotarix vs after dose 2
0.66 (0.44-0.99)
0.68 (0.47-0.98)
1.03 (0.86-1.24)
M AN U
Stem cells compared to Adverse effects no stem cells,
TE D
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease CD006536 Stem cell treatment for acute myocardial infarction CD000102 Interventions for reducing the risk of mother-to-child transmission of HIV infection CD001556c Preoperative chemotherapy for resectable thoracic esophageal cancer CD001832 Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis CD008521b Vaccines for preventing rotavirus diarrhoea: vaccines in use
Description of ROR (95% CIs) Harm-related endpoint vs Non fatal 0.38 (0.18-0.83) serious adverse events
SC
Comparison
Preoperative chemotherapy surgery alone
AC C
CD002309
Title
Gastrointestinal vs complications
EP
Review
RI PT
countries
placebo: Vomitingreactogenicity
35
ACCEPTED MANUSCRIPT
Abbreviations: ROR: relative odds ratio; sOR_MD: summary odds ratio in more developed countries; sROR_LD: summary odds
AC C
EP
TE D
M AN U
SC
RI PT
ratio in less developed countries; CIs: confidence intervals; Footnote: ORs and RORs were calculated by random effects models
36
ACCEPTED MANUSCRIPT
Figure 1: Flow Chart
RI PT
Screened 131 Systematic Reviews
SC
Excluded 89 Systematic Reviews No adverse events (n=40) No pharmacologic interventions (n=30) No More developed and Less developed countries (n=18) No country information (n=1)
Discontinuations due to Adverse Events: 12 ma; 108 RCTs (70 from MD and 38 from LD countries)
TE D
Severe Adverse Events: 12 ma; 143 RCTs (99 from MD and 44 from LD countries)
M AN U
Included 42 eligible Systematic Reviews; 128 meta-analyses on Harm-related endpoints with 1186 study-level analyses from 474 independent RCTs (337 from MD and 137 from LD countries)a b
AC C
EP
Organ System Specific Adverse Events: 5 ma; 20 RCTs (11 from MD and 9 from LD countries)
Individual adverse events (n=60) 14 adverse events with ≥2 ma (43 ma) 46 adverse events with 1 ma (46 ma)
Any Adverse Event: 9 ma; 81 RCTs (71 from MD and 10 from LD countries)
All Discontinuations: 1 ma; 7 RCTs (6 from MD and 1 from LD countries)
89 ma, 767 RCTS corresponding to 311 independent RCTs (222 from MD and 89 from LD countries)
Footnotes: a
Each RCT could have contributed data to more than one study-level analyses, for the different harm-related endpoints.
b
We excluded from these 128 meta-analyses data from 47 RCTs that were from mixed-country settings (n=38); 37Eastern European countries (n=4) or did not report country (n=5).
ACCEPTED MANUSCRIPT
Table 1a: Characteristics of Meta-analyses and Endpoints for Severe Adverse Events Severe adverse events Review Title
Comparison
Description endpoint
CD000024
Anticoagulants for acute ischaemic stroke Antiplatelet therapy for acute ischaemic stroke Thrombolysis for acute ischaemic stroke
Anticoagulant vs control
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 (PDE4) inhibitors for chronic obstructive pulmonary disease Treatment for lupus nephritis
Platelet glycoprotein IIb/IIIa blockers during PCI vs control
Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding (all patients)
CD002309
CD002922 CD003481
CD003897a CD003897b CD008521a CD008521b
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interleukin 2receptor antagonists (IL2Ra) for kidney transplant recipients Interleukin 2 receptor antagonists (IL2Ra) for kidney transplant recipients Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
1 (21,106)
0.67 (0.14-3.09)
20 (6,244)
2 (567)
1.97 (0.57-6.85)
20 (23,259)
3 (448)
2.39 (0.30-19.03)
Adverse events causing death
3 (204)
2 (101)
0.54 (0.02-19.03)
Non fatal serious adverse events
8 (2,965)
1 (1,018)
0.38 (0.18-0.83)
Major infection
5 (266)
1 (29)
0.33 (0.02-4.53)
Intraventricular haemorrhage (III-IV)
3 (285)
4 (261)
1.75 (0.51-6.04)
IL2Ra vs placebo or no treatment
Infection: CMV invasive
3 (881)
1 (118)
0.93 (0.12-7.26)
IL2Ra vs placebo or no treatment
Infection: serious all-cause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
RV1 vs placebo
Serious adverse events
7 (3,179)
18 (29,581)
0.88 (0.63-1.25)
RV5 versus placebo
All serious adverse events
3 (2,547)
7 (7,863)
0.81(0.45-1.46)
Terlipressin vs vasopressin PDE4 inhibitors vs placebo
SC
Any thrombolytic agent vs control
RI PT
6 (796)
Antiplatelet vs control acute
Harm-related
M AN U
CD002147
ROR (95% CIs)
TE D
CD002130
N of RCTs (N patients) from less developed countries 2 (344)
Immunosuppressive agent steroids vs steroids alone Ibuprofen vs indomethacin
EP
CD000213
N of RCTs (N patients) from more developed countries 15 (2,506)
AC C
CD000029
of
plus
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); RV5: pentavalent rotavirus vaccine; RV1: monovalent rotavirus vaccine; TB: tuberculosis;vs: versus
2.48 (0.22-27.75)
ACCEPTED MANUSCRIPT
Table 1b: Characteristics of Meta-analyses and Endpoints for Discontinuations due to Adverse Events Discontinuations due to adverse events Review Title
Comparison
Description endpoint
CD000171a
Any TB drug vs placebo
Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Adverse events causing withdrawal of treatment Withdrawals due to adverse effects
CD002863 CD004386a
CD004386b
CD004773
CD007784
CD008521a CD008521b
PDE4 inhibitors vs placebo
SC
RI PT
Harm-related
M AN U
CD002309
+
N of RCTs (N patients) from more developed countries 2 (838)
N of RCTs (N patients) from less developed countries 6 (4,695)
ROR (95% CIs)
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
3 (206)
2 (103)
0.22 (0.01-5.92)
7 (2,891)
1 (1,018)
0.61 (0.27-1.36)
1.05 (0.04-26.24)
Beta-blocker vs placebo/no betablocker Drug vs placebo/no intervention
Changed drugs due to maternal side effects Side effects requiring discontinuation
10 (923)
1 (154)
1.22 (0.19-7.67)
16 (2,057)
1 (97)
2.11 (0.04-113.13)
Quinolone vs TMP-SMZ
Side effects requiring discontinuation
6 (752)
1 (102)
4.03 (0.40-40.29)
Antifungal vs Placebo
Discontinuation secondary to adverse events
2 (723)
1 (129)
0.74 (0.07-8.17)
Statins vs placebo/no treatment
Withdrawal due to adverse events
8 (1,652)
3 (236)
0.61 (0.13-2.93)
RV5 vs placebo
Adverse events requiring discontinuation (end of follow-up) Adverse events requiring discontinuation (end of follow-up)
5 (3,709)
4 (7,735)
1.00 (0.21-4.73)
6 (2,184)
15 (21,774)
1.92 (0.63-5.87)
TE D
CD002147
Isoniazid vs rifampicin pyrazinimide Terlipressin versus vasopressin
EP
CD000171c
Isoniazid vs isoniazid + rifampicin
AC C
CD000171b
Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4(PDE4) inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
of
RV1 vs placebo
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); RV5: pentavalent rotavirus vaccine; RV1: monovalent rotavirus vaccine; TB: tuberculosis; TMP-SMX: trimethoprim-sulfamethoxazole; vs: versus
ACCEPTED MANUSCRIPT
Table 1c: Characteristics of Meta-analyses and Endpoints for Any adverse event Any adverse event Review Title
Comparison
Description endpoint
CD002309
PDE4 inhibitors vs placebo
No of patients experiencing an adverse effect
All drug types vs placebo/no treatment
CD004773
CD004791
CD006536 CD008076
2 (661)
1 (210)
1.47 (0.44-4.90)
Antioxidants (beta-carotene) vs placebo Drug vs placebo/ no intervention
Adverse effects
3 (91,080)
1 (174)
0.88 (0.03-23.17)
Side effects
33 (5,002)
1 (97)
1.78 (0.03-92.13)
Quinolone vs TMP-SMZ
Side effects
9 (929)
1 (102)
1.15 (0.36-3.63)
Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Antibiotic prophylaxis vs placebo/ no treatment
Adverse events
6(450)
1 (102)
0.75 (0.02-24.58)
Stem cells vs no stem cells
Adverse effects
8 (669)
2 (213)
0.06 (0.02-0.20)
Safety outcomes during follow-up
1 (3,202)
1 (787)
0.56 (0.31-1.00)
SC
RI PT
Toxicity (adverse events 'probably due to drug')
Antifungal vs placebo
Harm-related
M AN U
CD004386b
ROR (95% CIs)
TE D
CD004386a
N of RCTs (N patients) from less developed countries 1 (1,018)
EP
CD004183
N of RCTs (N patients) from more developed countries 8 (2,913)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
AC C
CD003807
Phosphodiesterase 4 (PDE4) inhibitors for chronic obstructive pulmonary disease Interventions for preventing oral candidiasis for patients with cancer receiving treatment Antioxidant supplements for preventing gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastro-intestinal bleeding Stem cell treatment for acute myocardial infarction Cilostazol versus aspirin for secondary prevention of vascular events after stroke of arterial origin
of
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); TIA: transient ischemic attack; TMP-SMX: trimethoprim-sulfamethoxazole; vs: versus
1.21 (0.73-2.00)
ACCEPTED MANUSCRIPT
Table 1d: Characteristics of Meta-analyses and Endpoints for Combined Harms (Severe adverse events; Discontinuations due to adverse events or Any adverse event) Combined Primary-Harms (Serious Adverse event, Discontinuations due to adverse event and Any adverse event) Review Title Comparison Description of endpoint
CD000171c CD000213
CD002130
CD002147 CD002309
CD002863 CD002922 CD003481
CD003807
CD003897
1 (21,106)
0.67 (0.14-3.09)
2 (838)
6 (4,695)
1.05 (0.04-26.24)
3 (509)
1 (1,092)
4.59 (0.70-29.98)
2 (373)
2 (1,455)
0.53 (0.19-1.52)
20 (5,622)
2 (567)
1.97 (0.57-6.85)
20 (23,259)
3 (448)
2.39 (0.30-19.03)
Adverse events causing withdrawal of treatment No of patients experiencing an adverse effect
3 (204)
2 (101)
0.22 (0.01-5.88)
8(2,913)
1 (1,018)
1.21 (0.73-2.00)
Beta-blocker vs placebo/no betablocker Immunosuppressive agent plus steroids vs steroids alone Ibuprofen vs indomethacin
Changed drugs due to maternal side effects Major infection
10 (923)
1 (154)
1.22 (0.19-7.67)
5 (266)
1 (29)
0.33 (0.02-4.53)
Intraventricular haemorrhage (III-IV)
3 (285)
4 (261)
1.75 (0.51-6.03)
Comparisons with placebo/no treatment for all drug types
Toxicity (adverse events 'probably due to drug')
2 (661 )
1 (210)
1.47 (0.44-4.90)
IL2Ra vs placebo or no treatment
Infection: serious all-cause total
6 (1,009)
2 (132)
1.08 (0.39-2.98)
Platelet glycoprotein IIb/IIIa blockers during percutaneous coronary intervention and as the initial medical treatment of non-ST segment elevation acute coronary syndromes Terlipressin for acute esophageal variceal hemorrhage Phosphodiesterase 4 (PDE4) inhibitors for chronic obstructive pulmonary disease Oral beta-blockers for mild to moderate hypertension during pregnancy Treatment for lupus nephritis
Platelet glycoprotein IIb/IIIa blockers vs placebo/usual care (all patients)
Isoniazid vs isoniazid + rifampicin
Ibuprofen for the treatment of patent ductus arteriosus in preterm and/or low birth weight infants Interventions for preventing oral candidiasis for patients with cancer receiving treatment Interleukin 2 (IL 2) receptor antagonists for kidney transplant recipients
Major extracranial haemorrhage during treatment period Major extracranial haemorrhage during treatment period Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Incidence of adverse events leading to stopping treatment Symptomatic (including fatal) intracranial haemorrhage within 7 to 10 days 30-day major bleeding
SC
Anticoagulant vs control in acute presumed ischaemic stroke Antiplatelet vs control in acute presumed ischaemic stroke Any TB drug vs placebo
RI PT
6 (796)
Anticoagulants for acute ischaemic stroke Antiplatelet therapy for acute ischaemic stroke Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Treatment of latent tuberculosis infection in HIV infected persons Thrombolysis for acute ischaemic stroke
M AN U
CD000171b
ROR (95% CIs)
Isoniazid vs rifampicin + pyrazinamide Any thrombolytic agent vs control
TE D
CD000171a
N of RCTs (N patients) from less developed countries 2 (344)
Terlipressin vs vasopressin
PDE4 inhibitors vs placebo
EP
CD000029
N of RCTs (N patients) from more developed countries 15 (2,506)
AC C
CD000024
Harm-related
2.48 (0.22-27.75)
ACCEPTED MANUSCRIPT
CD004791
CD006536 CD007784
CD008076
CD008521a CD008521b
3 (91,080)
1 (174)
0.88 (0.03-23.17)
Drug vs placebo/ no intervention
Side effects
33 (5,002)
1 (97)
1.78 (0.03-92.13)
Quinolone vs TMP-SMZ
Side effects
9 (929)
1 (102)
1.15 (0.36-3.63)
Antifungal vs Placebo
Adverse events
1 (295)
1 (129)
1.44 (0.60-3.46)
Antibiotic prophylaxis vs placebo or no treatment
Adverse events
6 (450)
1 (102)
0.75 (0.02-24.58)
Stem cells compared to no stem cells, Statins vs placebo/no treatment
Adverse effects
6 (617)
2 (213)
0.06 (0.02-0.18)
Withdrawal due to adverse events
8 (1,652)
3 (236)
0.61 (0.13-2.93)
Cilostazol vs aspirin in patients with ischaemic stroke or TIA
Other outcomes of safety during follow-up.
1 (3,202)
1 (787)
0.56 (0.31-1.00)
RV5 vs placebo
Adverse events requiring discontinuation (end of follow-up) Serious adverse events
5 (3,708)
4 (7,7733)
1.00 (0.21-4.73)
7 (3,179)
18 (29,581)
0.88 (0.63-1.25)
RV1 vs placebo
RI PT
CD004773
Adverse effects - beta-carotene
SC
CD004386b
Antioxidants vs placebo
M AN U
CD004386a
Antioxidant supplements for preventing gastrointestinal cancers Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antibiotic prophylaxis for bacterial infections in afebrile neutropenic patients following chemotherapy Antifungal interventions for the primary prevention of cryptococcal disease in adults with HIV Antibiotic prophylaxis for spontaneous bacterial peritonitis in cirrhotic patients with ascites without gastro-intestinal bleeding Stem cell treatment for acute myocardial infarction HMG CoA reductase inhibitors (statins) for people with chronic kidney disease not requiring dialysis Cilostazol versus aspirin for secondary prevention of vascular after stroke of arterial origin Vaccines for preventing rotavirus diarrhoea: vaccines in use Vaccines for preventing rotavirus diarrhoea: vaccines in use
TE D
CD004183
AC C
EP
Footnote: N= number; RCTs: randomized clinical trials; ROR: relative odds ratio (Odds Ratio from Trials from More developed countries vs Trials from less developed countries); RV5: pentavalent rotavirus vaccine; RV1: monovalent rotavirus vaccine; TB: tuberculosis; TIA: transient ischemic attack; TMP-SMX: trimethoprim-sulfamethoxazole; vs: versus
ACCEPTED MANUSCRIPT
Table 2: Individual meta-analyses with nominally significant relative differences between more developed and less developed
ROR (95% CIs)
sOR_MD (95% sOR_LD (95% CIs) CIs)
0.38 (0.18-0.83)
0.75 (0.55-1.03)
1.96 (0.96-3.98)
Stem cells compared to no stem cells,
Adverse effects
0.06 (0.02-0.20)
0.84 (0.48-1.45)
13.01 (4.90-34.57)
Any zidovudine vs placebo/no treatment
Neonatal haematological toxicity
2.77 (1.23-6.22)
2.09 (1.22-3.59)
0.76 (0.41-1.38)
Gastrointestinal complications
0.003 (0.001-0.06)
1.15 (0.34-3.81)
441.20 (24.93-7808)
Third generation cephalosporins versus conventional therapy
Diarrhoea
3.92 (1.35-11.34)
2.50 (1.54-4.05)
0.64 (0.25-1.64)
Rotarix vs placebo: after dose 2
Vomitingreactogenicity
0.66 (0.44-0.99)
0.68 (0.47-0.98)
1.03 (0.86-1.24)
M AN U
PDE4 inhibitors vs placebo
TE D
Phosphodiesterase 4 inhibitors for chronic obstructive pulmonary disease CD006536 Stem cell treatment for acute myocardial infarction CD000102 Interventions for reducing the risk of mother-to-child transmission of HIV infection CD001556c Preoperative chemotherapy for resectable thoracic esophageal cancer CD001832 Third generation cephalosporins versus conventional antibiotics for treating acute bacterial meningitis CD008521b Vaccines for preventing rotavirus diarrhoea: vaccines in use
Description of Harm-related endpoint Non fatal serious adverse events
SC
Comparison
Preoperative chemotherapy vs surgery alone
AC C
CD002309
Title
EP
Review
RI PT
countries
ACCEPTED MANUSCRIPT
Abbreviations: ROR: relative odds ratio; sOR_MD: summary odds ratio in more developed countries; sROR_LD: summary odds
AC C
EP
TE D
M AN U
SC
RI PT
ratio in less developed countries; CIs: confidence intervals; Footnote: ORs and RORs were calculated by random effects models
ACCEPTED MANUSCRIPT
Figure 1: Flow Chart
RI PT
Screened 131 Systematic Reviews
SC
Excluded 89 Systematic Reviews No adverse events (n=40) No pharmacologic interventions (n=30) No More developed and Less developed countries (n=18) No country information (n=1)
Discontinuations due to Adverse Events: 12 ma; 108 RCTs (70 from MD and 38 from LD countries)
TE D
Severe Adverse Events: 12 ma; 143 RCTs (99 from MD and 44 from LD countries)
M AN U
Included 42 eligible Systematic Reviews; 128 meta-analyses on Harm-related endpoints with 1186 study-level analyses from 474 independent RCTs (337 from MD and 137 from LD countries)a b
AC C
EP
Organ System Specific Adverse Events: 5 ma; 20 RCTs (11 from MD and 9 from LD countries)
Individual adverse events (n=60) 14 adverse events with ≥2 ma (43 ma) 46 adverse events with 1 ma (46 ma)
Any Adverse Event: 9 ma; 81 RCTs (71 from MD and 10 from LD countries)
All Discontinuations: 1 ma; 7 RCTs (6 from MD and 1 from LD countries)
89 ma, 767 RCTS corresponding to 311 independent RCTs (222 from MD and 89 from LD countries)
Footnotes: a
Each RCT could have contributed data to more than one study-level analyses, for the different harm-related endpoints.
b
We excluded from these 128 meta-analyses data from 47 RCTs that were from mixed-country settings (n=38); Eastern European countries (n=4) or did not report country (n=5).