- Email: [email protected]
Diagnostic test accuracy of antigenaemia assay for PCR-proven cytomegalovirus infection—systematic review and meta-analysis
Accepted Manuscript Diagnostic test accuracy of antigenemia assay for polymerase chain reaction proven cytomegalovirus infection: systematic review and meta-analysis Hisashi Eguchi, MD, PhD., Nobuyuki Horita, MD, PhD., Ryota Ushio, MD, PhD., Ikuma Kato, MD., Yuki Nakajima, MD, PhD, Erika Ota, PhD, Takeshi Kaneko, MD, PhD. PII:
S1198-743X(17)30269-0
DOI:
10.1016/j.cmi.2017.05.009
Reference:
CMI 948
To appear in:
Clinical Microbiology and Infection
Received Date: 2 February 2017 Revised Date:
21 April 2017
Accepted Date: 7 May 2017
Please cite this article as: Eguchi H, Horita N, Ushio R, Kato I, Nakajima Y, Ota E, Kaneko T, Diagnostic test accuracy of antigenemia assay for polymerase chain reaction proven cytomegalovirus infection: systematic review and meta-analysis, Clinical Microbiology and Infection (2017), doi: 10.1016/ j.cmi.2017.05.009. 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.
ACCEPTED MANUSCRIPT
Diagnostic test accuracy of antigenemia assay for polymerase chain reaction proven cytomegalovirus infection:
Authors Eguchi, [email protected], MD, PhD.
3,#)Nobuyuki
Ushio, [email protected], MD, PhD.
4)Ikuma
Nakajima, [email protected], MD, PhD.
6)Erika
Ota, [email protected], PhD.
3)Takeshi
Kaneko, [email protected], MD, PhD.
1)Harvard
TE D
5)Yuki
Kato, [email protected], MD.
M AN U
3)Ryota
Horita, [email protected], MD, PhD.
SC
1,2,#)Hisashi
RI PT
systematic review and meta-analysis
TH Chan School of Public Health, Boston, MA, USA.
2)Department
of Public Health, Kitasato University School of Medicine, Kanagawa,
3)Department
EP
Japan. Medicine, Yokohama, Japan.
of Pulmonology, Yokohama City University Graduate School of
AC C
Medicine, Yokohama, Japan. 4)Department
of Molecular Pathology, Yokohama City Graduate University School of
Medicine, Yokohama, Japan. 5)Department
of Stem Cell and ImmuneRegulation, Yokohama City University
Graduate School of Medicine, Yokohama, Japan. 6)Global
Health Nursing, Graduate School of Nursing Science, St. Luke's
international university, Tokyo, Japan. #)
Hisashi Eguchi and Nobuyuki Horita equally contributed as the first authors.
ACCEPTED MANUSCRIPT
Corresponding Author Nobuyuki Horita, MD, PhD.
Medicine.
Phone: +81-45-352-7962 Facsimile: +81-45-352-7963
Short title
TE D
Accuracy of Antigenemia for CMV
M AN U
E-mail: [email protected]
Word counts
EP
Abstract: 239 words. Main text: 2589 words.
Key words
AC C
Cytomegalovirus
Immunosuppression Diagnosis
Hierarchical model
Sensitivity and specificity
SC
3-9, Fukuura, Kanazawa-Ku, Yokohama, 236-0004, Japan.
RI PT
Department of Pulmonology, Yokohama City University Graduate School of
ACCEPTED MANUSCRIPT
ABSTRACT Objectives: Objectives We aimed to assess diagnostic test accuracy of antigenemia assay for PCR proven cytomegalovirus (CMV) infection.
RI PT
Methods: Methods We systematically searched studies that provide data both on sensitivity and specificity of the CMV antigenemia assay using the PCR as the reference standard. Adults, children, infants, subjects who were immunocompromised for any reason,
summary
receiver
operating
characteristics
SC
symptomatic patients, and asymptomatic individuals were all included. Hierarchical model
was
used
for
diagnostic
M AN U
meta-analysis. Study quality was assessed by Revised Tool for the Quality Assessment of Diagnostic Accuracy Studies. Protocol registration identification is CRD42016035892. Results: Results We identified 75 eligible articles including 9058 CMV PCR-positive subjects and 22232 PCR-negative subjects. The diagnostic odds ratio for positive antigenemia
TE D
was 30 (95% confidence interval (95%CI) 24-38, I2 = 28%) and the area under the hierarchical summary receiver operating characteristic curve (AUC) was 0.86 (95%CI 0.83-0.88). The summary estimates of sensitivity and specificity were 0.65 (95%CI:
EP
0.59-0.70) and 0.94 (95%CI 0.93-0.95), respectively. While a positive likelihood ratio of 10.9 (95%CI 8.5-14.0) suggested that a positive result from the antigenemia assay
AC C
greatly increased the probability of PCR-proven CMV infection, a negative likelihood ratio of 0.38 (95%CI 0.32-0.44) indicated that a negative result led to a small decrease in the probability of PCR-proven CMV infection. Sensitivity and subgroup analyses replicated these results.
Conclusions: Conclusions The antigenemia assay overlooked 35% of PCR-proven CMV infections; thus, a negative result of an antigenemia assay could not rule out a CMV infection.
ACCEPTED MANUSCRIPT
INTRODUCTION Cytomegalovirus (CMV) screening and/or close monitoring of severely immunocompromised patients are recommended(1-3). Conventional culture has been
RI PT
the gold standard for CMV diagnosis. Histopathological examination using tissue biopsy specimens is useful for the diagnosis of CMV invasive diseases such as CMV gastritis and CMV pulmonary infection, though it is often difficult to access appropriate tissue
SC
samples. Overall, quantitative real-time PCR (qrtPCR) and pp65-antigenemia assays are the most widely used tests for CMV infection diagnosis among immunocompromised
M AN U
patients. These essays can detect CMV reactivation before disease onset, permitting preemptive therapy to control symptomatic CMV disease. Among the two, qrtPCR is generally preferred for some advantages such as obtainability of standardized results, applicability for patients with decreased white blood cell count, not being labor
TE D
intensive, high throughput nature, and better transportability of specimens. The qrtPCR may have higher sensitivity for CMV infection than the antigenemia assay; however, that is still controversial. Thus, most of the recent guidelines have avoided a
EP
clear statement of qrtPCR superiority over the antigenemia assay (4-9). Many studies have compared the diagnostic accuracy of CMV antigenemia and
AC C
PCR assays. Nonetheless, there is considerable variation among the results of these studies. We carried out this systematic review and meta-analysis to evaluate the diagnostic test accuracy of the CMV antigenemia assay using PCR as the reference standard.
ACCEPTED MANUSCRIPT
METHODS
Overview
RI PT
Our study protocol, which has been registered on the International Prospective Register of Systematic Reviews as number CRD42016035892 (10), followed the methodology described in the Cochrane Handbook for Diagnostic Test Accuracy Reviews
SC
and the Preferred Reporting Items in Systematic Reviews and Meta-Analyses
M AN U
statement(11, 12).
Eligibility criteria
Type of studies
We had planned to include both two-gate case-control studies and one-gate
TE D
cohort studies as long as they could provide sufficient data concerning both the sensitivity and specificity of the CMV antigenemia assay when using the PCR as the reference standard(12, 13). However, we were unable to find case-control study, thus, we
EP
included only cohort studies. Conference abstracts and reports written in languages other than English were allowed. Studies covering only sensitivity or only specificity
AC C
were excluded. Studies for non-human subjects such as murine CMV were also excluded.
Participants
We assessed subjects who might have PCR-proven CMV infection. Pregnant women, infants, subjects who were immunocompromised for any reason, symptomatic patients, and asymptomatic individuals were all accepted. However, no included study evaluated pregnant women. Patients receiving CMV treatment could also be included.
ACCEPTED MANUSCRIPT
Index and reference test The index test was the CMV pp65-antigenemia assay using any standard
RI PT
procedure(14, 15). The antigenemia assay for cord blood was accepted. Both commercialized kits and in-house assays were allowed. The reference test was PCR including conventional/qualitative PCR, qrtPCR, nested PCR, and reverse transcription
SC
PCR(16, 17). Quantitative PCR (q-PCR) other than qrtPCR was also accepted. PCR
in addition to blood specimens.
Literature search strategy
M AN U
assays using non-blood specimens such as urine or bronchoalveolar fluid were included
In the electronic search, we systematically searched PubMed, EMBASE, the
TE D
Cochrane Library, and Web of Science on February 28th, 2016. References of previously published reviews and those of included original studies were checked by hand searching. Two investigators independently screened the candidate articles by checking
EP
the title and abstract after uploading the citation list into the software, EndNote X7 (Thomson Reuters, Philadelphia, USA). After independent screening, articles that were
AC C
still regarded as candidates by at least one investigator were then scrutinized independently through full-text reading. Final inclusion was decided after resolving discrepancies between the two investigators. Search formulas are presented
in
Supplement Text 1).
Study selection and data extraction Data were extracted by two reviewers independently and then cross-checked (HE and NH)(12, 13). Data concerning the antigenemia assay (IK) and PCR assay (RU)
ACCEPTED MANUSCRIPT
were also cross-checked again by additional reviewers. When a report presented data for multiple thresholds, we selected data using the lowest cutoff values. Studies that repeatedly evaluated the same individual were permitted for
RI PT
inclusion because repeated evaluations of CMV reactivation for the same recipients are very common. However, when an individual was evaluated twice or more, a report was marked high risk on the flow and timing domain of the quality assessment(18).
SC
If a study evaluated two or more antigenemia assays or two or more PCR assays, the most up-to-date assay for antigenemia and PCR was selected based on
M AN U
investigators discussion (HE and NH). Meta-analysis was conducted based on the assumption that each specimen was independent from another specimen. Inputting the data from one specimen twice was not allowed, because this leads to duplicate use of
TE D
same data.
Quality assessment for bias and and applicability The two investigators independently evaluated each included study by scoring
EP
seven domains of the Revised Tool for the Quality Assessment of Diagnostic Accuracy Studies (HE and NH) (18). This tool evaluated the quality of a primary diagnostic
AC C
accuracy study by checking four key domains covering patient selection, index test, reference standard, and flow of patients. Each domain was marked “high risk of bias” when the risk of bias was judged high based on signal questions (18). For sensitivity analysis, a study with at least one domain that was scored as high risk or high applicability concern was regarded as a high-risk study.
MetaMeta-analysis, statistics, and interpretation of data The main outcomes were the diagnostic test accuracy of pp65-antigenemia for
ACCEPTED MANUSCRIPT
PCR-diagnosed CMV infection evaluated by the following statistics: diagnostic odds ratio (DOR), area under the hierarchical summary receiver operating characteristic curve (AUC), and the summary estimates of the sensitivity, the specificity, the positive
RI PT
likelihood ratio (PLR) and the negative likelihood ratio (NLR). Additionally, we assessed positive and negative predictive values (PPV, NPV) based on the pre-test probability ranging from 0% to 100%.
SC
We used both the hierarchical summary receiver operating characteristics model and the bivariate model (19, 20). The heterogeneity, which is a degree of
M AN U
inconsistency among studies, was assessed using the I² statistic: 0% to 40% represents not important heterogeneity, 30% to 60% represents moderate heterogeneity, 50% to 90% represents substantial heterogeneity, 75% to 100% indicates considerable heterogeneity (21).To determine the overall accuracy, we calculated the DOR using the
TE D
DerSimonian-Laird random-effect model and the AUC using Holling's proportional hazard models. DOR is a measure of the effectiveness of a diagnostic test; wherein DOR = 1 means no diagnostic value, DOR > 1 means that test positive suggests disease
EP
positive, and DOR < 1 means that test negative suggests disease positive. We obtained a paired forest plot, a hierarchical summary receiver operating characteristics curve, and
AC C
summary estimates of the sensitivity and the specificity using the bivariate model (19, 20). PLR, NLR, PPV, and NPV were calculated from the summary estimate of the sensitivity and the specificity. AUC, PLR, and NLR were interpreted according to the four-grade criterion by Grimes et al. and Jones et al. (22, 23). AUC indicates how a test is accurate. AUC in the ranges of -0.75, 0.75-0.92, 0.93-0.96, and 0.97- meant "not accurate", "good", "very good", and "excellent", respectively. PLR defined by “sensitivity/ (1 - specificity)” and NLR defined by “(1 - sensitivity)/ specificity” represent how the test results change the probability of a disease. PLR value in the range of -2, 2-5, 5-10, and
ACCEPTED MANUSCRIPT
10- were recognized as showing a "not meaningful", "small", "moderate", and "large" increase of probability, respectively. NLR in the range of 0.5-, 0.2-0.5, 0.1-0.2, and -0.1 represented a "not meaningful", "small", "moderate", and "large" decrease of probability,
RI PT
respectively. PPV and NPV are proportions of patients who are correctly judged as positive or negative by the test.
The following commands of the statistics software R "mada" package were used
SC
for the meta-analysis: the “madauni” for DOR, the "phm" for AUC, and the “reitsma” for the hierarchical summary receiver operating characteristic curve and the summary
M AN U
estimates of sensitivity and specificity (19, 20). For the publication bias assessment, "funnel" and "metabias" commands of the R "meta" package were used(24).
Subgroup and sensitivity analyses
TE D
As a subgroup analysis, we focused on studies that exclusively evaluated solid organ (SOT) and hematopoietic stem cell transplantation (HSCT) recipients. Additionally, we performed subgroup analysis of studies that assessed specific
EP
commercialized kits or commercialized primary antibodies for antigenemia assays when these kits or antibodies were evaluated in more than five studies. CINAkit (Argene
AC C
Biosoft), CMV Brite Turbo Kit (IQ Products), and Clonab CMV (Biotest) satisfied this criterion. Additional analyses focusing on subgroup of studies with specific PCR procedures and of pediatric studies were done. As a sensitivity analysis, we performed subgroup analyses based on the study
quality.
ACCEPTED MANUSCRIPT
RESULTS
Study search and study characteristics
RI PT
We first identified 803 and 15 articles by electronic and hand search, respectively. Of 818 articles that met the preliminary criteria, 317, 179, and 247 were excluded through removal of duplication, title/abstract screening, and full-article
SC
scrutinizing, respectively (Figure 1, Supplement Text 2). We finally identified 75 eligible articles. Studies were reported from a variety of countries worldwide, most of which
M AN U
were European, Asian, North American, and South American nations (Table 1). Among 75 reports, three were written in Spanish and one each was written in Chinese, French, Germany, Korean, and Japanese. One was a conference abstract and one was a letter article. All other 67 were original articles in English. Patients included in these trials SOT
recipients,
HSCT
recipients,
HIV/AIDS
patients,
non-specified
TE D
were
immune-compromised population, or newborn/infants (Table 1). While 11 studies evaluated an individual only once, the remaining 64 studies
EP
evaluated individuals on multiple occasions. The median number of subjects in studies without double counting was 46 with a range of five to 1516. The median number of
AC C
subjects allowing double counting was 192 with a range of five to 7920, which totaled 31290 subjects (Table 1). Sixty-five studies had at least one domain at high risk of bias or high applicability concern (Supplement Figure 1). The risk of bias for flow and timing domain was most frequently scored "high" because of duplicate evaluation of the same subjects. Antigenemia assays were carried out by indirect immunofluorescence assay in 30 studies. Of note, CINAkit (Argene Biosoft), CMV Brite Turbo Kit (IQ Products) and Clonab CMV (Biotest) were used in 14, seven, and 15 studies, respectively. As a
ACCEPTED MANUSCRIPT
reference test, 29 studies used qrtPCR (Table 1). Seventy-one studies used only blood specimens, such as whole blood, leukocytes, plasma, and sera for PCR (Table 1). Neither the visual inspection nor the Kendall's rank correlation test (P = 0.051)
RI PT
found significant asymmetry of the funnel plot (Supplement Figure 2).
MetaMeta-analysis of all the studies
SC
Using data from all 75 studies, consisting of 9058 PCR-positive subjects and 22232 PCR-negative subjects, the DOR was 30 (95% confidence interval (95%CI) 24-38,
M AN U
I2 = 28%) and AUC was 0.86 (95%CI 0.83-0.88) (Table 2, Figure 3-A).
The summary estimates of sensitivity and specificity were 0.65 (95%CI: 0.59-0.70) and 0.94 (95%CI 0.93-0.95), respectively (Table 2). While a PLR of 10.9 (95%CI 8.5-14.0) suggested that a positive result from the antigenemia assay greatly
TE D
increased the probability of PCR-proven CMV infection, a NLR of 0.38 (95%CI 0.32-0.44) indicated that a negative result led to a small decrease in the probability of PCR-proven CMV infection (Table 2)(22).
EP
PPV and NPV depending on pre-test probability are shown in Supplement Figure 3. PPV was generally better than NPV, though values varied depending on the
AC C
pre-test probabilities (Supplement Figure 3).
Subgroup and sensitivity metameta-analyses analyses The results from the subgroup analyses based on SOT recipients and HSCT are
shown in Figure 3-B, 3-C and Table 2. The DOR based on SOT recipients and HSCT recipients were 28 (95%CI 18-42, I2 = 34%) and 25 (95%CI 15-40, I2 = 37%), respectively. The AUC estimated from data of SOT recipients and HSCT recipients were 0.86 (95%CI 0.81-0.90) and 0.84 (95%CI 0.80-0.88), respectively. We did not find significant
ACCEPTED MANUSCRIPT
differences for these results and the results from all the studies (Table 2). Studies focusing on CNAkit (DOR, 38, 95%CI 22-64) and Brite Turbo (DOR 35, 95%CI 30-41) were associated with higher DOR than all studies (Figure 3-D, Figure 3-E,
RI PT
Table 2). However, the differences were not statistically significant (Table 2).
Sensitivity analyses using data from studies without high risk was carried out. The results of this sensitivity analyses did not differ significantly from those from all
SC
the studies (Supplement Table 1). Subgroup of studies that used only artPCR revealed a DOR of 30 (95% CI 21-43, I2 = 9%) and AUC of 0.84 (95%CI 0.82-0.87). Subgroup of
M AN U
studies that used only blood derivative specimens for PCR revealed a DOR of 30 (95% CI 23-38, I2 = 26%) and AUC of 0.86 (95%CI 0.83-0.88), which was consistent with the results from the analysis that used both blood and non-blood specimens (Supplement Table 1). Based on six pediatric studies, DOR was 44 (95%CI 23-84) and the specificity
TE D
was 0.97 (95%CI 0.96-0.98) (Supplement Table 1). Among the six pediatric studies, one evaluated only neonates, one assessed only infants, one included children with a median of seven years old, one assessed children with a median of 47 months old, and the other
AC C
EP
two did not presented children’s age.
ACCEPTED MANUSCRIPT
DISCUSSION
Our systematic review detected 75 original studies that compared results from
RI PT
antigenemia assays and PCR assays to diagnose CMV infection among patients. The specificity of the antigenemia assay was consistently high and a positive test increased the probability of CMV infection. However, the sensitivity of the assay for PCR-proven
SC
CMV infection varied among the reports ranging from 0.03 to 1.00 with an interquartile range of 0.47 to 0.84. The summary estimates of sensitivity and specificity were 0.65
M AN U
(95%CI 0.59-0.70) and 0.94 (95%CI 0.93-0.95), respectively. A sensitivity of 0.65 meant that a third of PCR-proven CMV infections were not detected by pp65-antigenemia. Thus, a negative test result provided little clinical information. PPVs were generally higher than NPVs though they varied depending on the pre-test probabilities.
TE D
According to the subgroup analyses, we do not have sufficient evidence to recommend a specific commercially available antigenemia kit or primary antibody. Similarly, we do not have reason to recommend or oppose the use of the antigenemia
EP
assay for different populations.. We included studies that used PCR assays other than the standard qrtPCR. However, a sensitivity analysis focusing on studies that used
AC C
qrtPCR provided similar results to the analysis with all the studies. This meant that the PCR method did not considerably affect the results of our study. The major advantage of the antigenemia assay over the qrtPCR assay is that it
does not require expensive equipment. Even in a developed country, some national governments such as Japan do not approve the use of qrtPCR for CMV infection. Some may believe that the lowest detection limit of qrtPCR is too sensitive to distinguish CMV infection and disease and to make a judgement on preemptive therapy.
ACCEPTED MANUSCRIPT
The robustness of our analysis is supported by numerous factors such as the large numbers of included studies and subjects, minor heterogeneity among the studies suggested by I2 =28%, consistent results from many sensitivity analyses, and sound
comment on the limitations of our analysis.
RI PT
methodology following updated guidelines(12, 13). Despite these strengths, we need to We evaluated symptomatic and
asymptomatic patients collectively, Cutoff of qrtPCR and antigenemia assays for
SC
symptomatic CMV disease are still to be clarified. Our analysis also could not suggest a cutoff of qrtPCR and antigenemia assays for preemptive treatments.
M AN U
In conclusion, we carried out the first systematic review and meta-analysis to assess the diagnostic test accuracy of the CMV antigenemia assay. While this test has good PPV, the antigenemia assay overlooked 35% of PCR-proven CMV infection; thus, a negative result of an antigenemia assay could not rule out a CMV infection. Consistent
AC C
EP
supported our results.
TE D
results from sensitivity analyses, solid methodology, and narrow confidence intervals
ACCEPTED MANUSCRIPT
Add Additional information
Contribution
RI PT
H.E. contributed for study search, quality check, data extraction, and drafting.
N.H. work for conception, study search, quality check, data extraction, and analysis as a principal investigator. U.R. and I.K. revised the manuscript concerning PCR assays and
SC
antigenemia assays. Y.N. contributed for study conception and revised the manuscript
manuscript.
Competing Financial Interests
M AN U
as hematologist. E.O. helped analysis as a statistician. T.K. critically revised the
No support in the form of grants, gifts, equipment, and/or drugs was provided.
TE D
Conflict of interest disclosure
AC C
EP
The authors report no conflicts of interest.
ACCEPTED MANUSCRIPT
References 1. Kalil AC, Florescu DF. Prevalence and mortality associated with cytomegalovirus infection in nonimmunosuppressed patients in the intensive care unit. Crit Care
RI PT
Med 2009; 37: 2350-2358.
2. Cantisán S, Torre-Cisneros J, Lara R, Zarraga S, Montejo M, Solana R. Impact of cytomegalovirus on early immunosenescence of CD8+ T lymphocytes after solid
SC
organ transplantation. J Gerontol A Biol Sci Med Sci 2013; 68: 1-5.
3. Ellington SR, Clarke KE, Kourtis AP. Cytomegalovirus Infection in Human
M AN U
Immunodeficiency Virus (HIV)-Exposed and HIV-Infected Infants: A Systematic Review. J Infect Dis 2016; 213: 891-900.
4. Kotton CN, Kumar D, Caliendo AM, Asberg A, Chou S, Danziger-Isakov L, Humar A, Group TSICC. Updated international consensus guidelines on the management of
TE D
cytomegalovirus in solid-organ transplantation. Transplantation 2013; 96: 333-360. 5. Pilmore H, Pussell B, Goodman D. KHA-CARI guideline: cytomegalovirus disease and kidney transplantation. Nephrology (Carlton) 2011; 16: 683-687.
EP
6. O'Brien SM, Keating MJ, Mocarski ES. Updated guidelines on the management of cytomegalovirus reactivation in patients with chronic lymphocytic leukemia treated
AC C
with alemtuzumab. Clin Lymphoma Myeloma 2006; 7: 125-130. 7. Preiksaitis JK, Brennan DC, Fishman J, Allen U. Canadian society of transplantation consensus workshop on cytomegalovirus management in solid organ transplantation final report. Am J Transplant 2005; 5: 218-227.
8. Caliendo AM. Approach to the diagnosis of cytomegalovirus infection. UpToDate. 2016. Topic
8305
Version
15.0.
Available
from:
https://www.uptodate.com/contents/approach-to-the-diagnosis-of-cytomegalovirus-in fection?source=search_result&search=CMV%20diagnosis&selectedTitle=2~150.
ACCEPTED MANUSCRIPT
Accessed on 31st October, 2016. 9. Caliendo AM. Overview of diagnostic tests for cytomegalovirus infection. UpToDate. 2016. Topic
8291
Version
14.0.
Available
from:
RI PT
https://www.uptodate.com/contents/overview-of-diagnostic-tests-for-cytomegalovirus -infection?source=search_result&search=CMV%20diagnosis&selectedTitle=1~150. Accessed on 31st October, 2016.
SC
10. Booth A, Clarke M, Dooley G, Ghersi D, Moher D, Petticrew M, Stewart L. The nuts and
Syst Rev 2012; 1: 2.
M AN U
bolts of PROSPERO: an international prospective register of systematic reviews.
11. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009; 339: b2535.
TE D
12. Leeflang MM, Deeks JJ, Takwoingi Y, Macaskill P. Cochrane diagnostic test accuracy reviews. Syst Rev 2013; 2: 82.
13. Leeflang MM, Deeks JJ, Gatsonis C, Bossuyt PM, Group CDTAW. Systematic reviews of
EP
diagnostic test accuracy. Annals of internal medicine 2008; 149: 889-897. 14. Boeckh M, Gooley TA, Myerson D, Cunningham T, Schoch G, Bowden RA.
AC C
Cytomegalovirus pp65 antigenemia-guided early treatment with ganciclovir versus ganciclovir at engraftment after allogeneic marrow transplantation: a randomized double-blind study. Blood 1996; 88: 4063-4071.
15. Landry ML, Ferguson D. Comparison of quantitative cytomegalovirus antigenemia assay with culture methods and correlation with clinical disease. J Clin Microbiol 1993; 31: 2851-2856. 16. Machida U, Kami M, Fukui T, Kazuyama Y, Kinoshita M, Tanaka Y, Kanda Y, Ogawa S, Honda H, Chiba S, Mitani K, Muto Y, Osumi K, Kimura S, Hirai H. Real-time
ACCEPTED MANUSCRIPT
automated PCR for early diagnosis and monitoring of cytomegalovirus infection after bone marrow transplantation. J Clin Microbiol 2000; 38: 2536-2542. 17. Limaye AP, Huang ML, Leisenring W, Stensland L, Corey L, Boeckh M. Cytomegalovirus
RI PT
(CMV) DNA load in plasma for the diagnosis of CMV disease before engraftment in hematopoietic stem-cell transplant recipients. J Infect Dis 2001; 183: 377-382.
18. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, Leeflang MM,
SC
Sterne JA, Bossuyt PM, Group Q-. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Annals of internal medicine 2011; 155:
19.
Doebler
M AN U
529-536. P.
Package
‘mada’.
Available
from:
https://cran.r-project.org/web/packages/mada/mada.pdf#search='mada+r'. Accessed on November 5th, 2015.
TE D
20. Doebler P., & Holling H. Meta-analysis of diagnostic accuracy with mada. Available from: https://cran.r-project.org/web/packages/mada/vignettes/mada.pdf#search='mada+r+ systematic+review+diagnostic+accuracy'. Accessed on November 5th, 2015.
EP
21. Higgins PJ, Green S. Cochrane Handbook for Systematic Reviews of Interventions (Version 5.1.0). (2011). Available from: http://handbook.cochrane.org/front_page.htm.
AC C
Accessed on December 4th, 2015). 22. Grimes DA, Schulz KF. Refining clinical diagnosis with likelihood ratios. Lancet 2005; 365: 1500-1505.
23. Jones CM, Athanasiou T. Summary receiver operating characteristic curve analysis techniques in the evaluation of diagnostic tests. Ann Thorac Surg 2005; 79: 16-20. 24.
Schwarzer
G.
Package
'meta'.
2016,
Available
ftp://cran.r-project.org/pub/R/web/packages/meta/meta.pdf#search='r+meta'. Accessed on October 30th, 2016.
from:
ACCEPTED MANUSCRIPT
Table 1. Characteristics of included studies.
Study
Country
Subject background
Antigenemia assay (primary antibody or
Author (year)
(publication type)
Allice(2006)
Italy
SOT
Ind EIA (1C3+AYM-,1Argene Biosoft)
Ashokkumar(2015)
USA (conf abst)
liver/intestine T, child
NS
Beckmann(2011)
Switzerland
OT
Ind IF (CINAkit, Argene Biosoft)
Bergallo(2008)
Italy
renal T
NS
Boaretti(2013)
Italy
SOT
Ind IF (CINAkit, Argene Biosoft)
Boeckh(1997)
USA
BMT
Boeckh(2004)
USA
Boivin(1998)
PCR assay, target
RI PT
commercialized kit))
PCR
Subject
Subject
High
specimen
w/o rep
w rep
risk
Bld
48
375
Y
PCR
W-Bld
39
39
N
qrtPCR, UL111a
W-Bld
81
720
Y
rev tras PCR, IE1(UL123)
Leu
14
29
Y
qrtPCR(AmpliP), UL54
Pla
45
266
Y
IF (Clonab CMV, Biotest)
cPCR, IE1
Bld
29
328
Y
Immunosuppressed
IF (Brite kit, Biotest)
qrtPCR, UL123/125/126, UL55(gB)
Pla
NS
1902
Y
Canada
HIV
(1C3 clone, Argene)
cPCR(AmpliT), UL54
Bld
168
172
Y
Boland(1992)
Netherlands
heart/lung T
Ind PO (Mab C10/C11, in-house)
cPCR, IEA
Leu
21
201
Y
Bonon(2006)
Brazil
HSCT
Ind PO (Clonab CMV, Biotest)
nPCR, mIE
Bld
46
46
N
Breda(2013)
Brazil
HSCT
Ind IF (Brite Turbo Kit)
qrtPCR, UL123
W-Bld, Pla
21
196
Y
Camargo(2001)
Brazil
heart T
Ind PO (Clonab CMV, Biotest)
qPCR, LA gp64
Bld
51
162
Y
Cardenoso(2013)
Switzerland
HSCT
Ind IF (Brite Turbo Kit)
qrtPCR(AmpliP), UL54
Pla
135
377
Y
Cathomas(1994)
Switzerland
renal T
Ind PO (Clonab CMV, Biotest)
cPCR, mIE
Leu
18
185
Y
APAAP Ind (Clonab CMV, Biotest)
qrtPCR(Artus),
W-Bld
131
555
Y
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Bld
NS
164
Y
308
192
Y
M AN U
TE D
EP AC C
(Germany)
SC
qrtPCR, UL123
Choi(2009)
Korea
HSCT
Dávila(2005)
Spain
Liver T
Deback(2007)
France
Immunosuppressed
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, HXFL4
W-Bld
Eckart(1996)
France
renal T
IF
cPCR, IE1
Pla
5
5
N
Flexman(2001)
Australia
BMT, SOT
IF (Monofluor Kit CMV, Sanofi)
qPCR(AmpliM),
Pla
NS
469
Y
ACCEPTED MANUSCRIPT France
SOT
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, UL83
W-Bld
61
165
Y
Guiver(2001)
UK
lung/heart T
IF (Mab, Biosoft)
qrtPCR, gpB
Bld
25
362
Y
Hadaya(2003)
Switzerland
SOT
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Pla
23
321
Y
Halfon(2011)
France
HSCT
Ind IF (CINAkit, Argene Biosoft)
qrtPCR(Artus),
Bld
93
244
Y
Hebart(1996)
Germany
BMT
APAAP (Clonab CMV, Biotest)
cPCR, IEA
Pla
20
109
Y
Hoe(2007)
Korea (Korean)
OT
Ind APAAP (Clonab CMV, Biotest)
qrtPCR, gB
W-Bld
84
343
Y
Juillet(2010)
France (French)
Immunosuppressed
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, UL83
W-Bld
5
32
Y
Khansarinejad(2015)
Iran
HSCT, child
Ind IF (Brite Turbo Kit)
qrtPCR, UL83
Pla
82
1179
Y
Kidd(1993)
UK
Immunosuppressed
Ind IF (Mab; FITC, Tago)
qPCR, gB
Leu
NS
295
Y
Ksouri(2006)
Tunisia
BMT
Ind IF (CINAkit, Argene Biosoft)
cPCR,
Pla
18
126
Y
Landini(1995)
Italy
SOT
Ind IF (Mab, Biotest)
cPCR, IE1
Leu
23
104
Y
Lazzarotto(1996)
Italy
OT, AIDS
Ind IF (Mab, Argene)
cPCR, mIEA
Leu
NS
48
Y
Li(2003)
USA
BMT, SOT
Inirect IF (CMV Brite, Biotest)
qrtPCR, IEA
Bld
NS
298
Y
Lo(1997)
Hong Kong
renal T
APAAP Ind IF (Clonab CMV, Biotest)
nPCR, mtr II
Leu
22
108
Y
Lomas(1992)
Spain (Spanish)
HIV, Kaposi, sarcoma
IF (Mab)
cPCR, mIE
Bld
25
25
N
Luchsinger(2015)
Chile (Spanish)
HIV
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, gB
Bld
179
179
N
Madhavan(2007)
India
renal T, Immunosuppressed
Ind IF (Clonab CMV, Biotest)
qrtPCR, mtrII, UL83, gO
Leu
NS
70
Y
Mansy(1999)
Belgium
in viral Dept
Ind IF, PO, ABC, APAAP
nPCR, pp150 (UL32)
Bld
NS
40
Y
Marenzi(1996)
Italy
AIDS
IF (Clonatec, Biosoft)
nPCR, IEA
Ser
21
36
Y
Martiny(2011)
Brazil
Immunosuppressed
Ind IF (Brite Turbo Kit)
nPCR, gB
Bld
NS
216
Y
Masaoka(2001)
Japan
Immunosuppressed
Dir IF (Mitsubishi kit, Yuka Medias)
qPCR(AmpliM), UL54
Ser
115
106
Y
Matsunaga(1999)
Japan
BMT
PO (Mab)
nPCR, IE
Ser
26
26
N
Mauricio(2011)
Chile (Spanish)
SOT, BMT, child
Ind IF (Mab, Argene Biosoft)
qrtPCR,
Bld
68
219
Y
Milan(2011)
Brazil
liver T
Ind PO (C10/C11, Biotest AG)
nPCR,
Bld
94
465
Y
Miller(2010)
USA
HSCT, SOT
Ind IF (Brite Turbo Kit)
qrtPCR, UL54
Pla
NS
61
Y
AC C
EP
TE D
M AN U
SC
RI PT
Garrigue(2006)
ACCEPTED MANUSCRIPT Japan (Japanese)
SCT
Ind ALP stain (Mab C10/11)
qrtPCR,
Pla
13
120
Y
Muller(2007)
Germany
HSCT
Ind IF (Mab, Virion)
qPCR(AmpliM), UL54
Pla
40
435
Y
Nitsche(2003)
Germany (letter)
SCT
NS
qrtPCR, mIE
Pla
77
1122
Y
Pang(2003)
Canada
SOT
IF (Clonab CMV, Biotest AG)
qrtPCR, gB
Pla
66
404
Y
Pellegrin(2000)
France
renal T
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Pla
30
283
Y
Peres(2010)
Brazil
HSCT
Ind PO (Mab, IQP; HRP, Biotest)
qrtPCR, US17
Bld
30
30
N
Piiparinen(2001)
Finland
liver/renal T
PO (Mab, Biotest Pharma)
qPCR(AmpliM), UL54
Pla
NS
253
Y
Rautenberg(1999)
Germany
SOT
Ind APAAP (Clonab CMV, Biotest)
nPCR, UL122
Bld
26
69
Y
Rayes(2005)
Germany
renal T
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Pla
72
389
Y
Rhee(2011)
Korea
renal T
Ind APAAP (Clonab CMV, Biotest)
qrtPCR(Artus),
Pla
111
899
Y
Sakamaki(1997)
Japan
BMT
Dir PO (HRP-C7Mab, Teijin)
cPCR, mIE
BALF
30
43
Y
Sanghavi(2008)
USA
OT
Ind EIA (Clonatec 1C3, Biosoft)
qrtPCR, US17, UL54
Bld
12
3422
Y
Schirm(1999)
Netherlands
Immunosuppressed
Ind PO (Mab C10/C11)
cPCR, IE
Bld
55
280
Y
Schmidt(1994)
Germany
BMT
Ind PO (Clonab CMV, Biotest)
cPCR, IE
Buf C
50
96
Y
Schmidt(1995)
Germany
Liver T
Ind APAAP (Clonab CMV, Biotest)
cPCR, β2mg
Buf C
30
264
Y
Schvoerer(2005)
France
OT
IF (Biotest)
qrtPCR, LA(HXFL4)
Pla
16
145
Y
Seehofer(2004)
Germany
Liver T
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Bld
135
719
Y
Solano(2001)
Spain
ASCT
Ind IF (Chemicon Int'l)
cPCR(AmpliT),
Pla
43
548
Y
Sowmya(2006)
India
Immunosuppressed
Ind IF (DAKO)
nPCR, mtr II, gO, UL83
Buf C, Urn
74
92
Y
Sun(2009)
China
infant
Ind EIA (Chemicon, BD Biosciences)
qrtPCR, UL123
Urn
126
126
N
Tantivanich(2002)
Thailand
HIV s/o, child
Ind PO (AAC10, DAKO)
nPCR,
Leu
100
100
N
Thorne(2007)
USA
Immunosuppressed
Ind IF (Brite Turbo Kit)
qrtPCR, IEA
W-Bld
49
106
Y
Toulemonde(2000)
France
SOT, SCT
Ind IF (CINAkit, Argene Biosoft)
qc-PCR, (Us Region(cPCR))
Leu
110
443
Y
Velzing(1994)
Netherlands
Immunosuppressed
Ind PO (Clonab CMV, Biotest)
cPCR, mIEA
Bld
36
37
Y
Vlieger(1992)
Netherlands
BMT
PO (Mab, in-house)
nPCR, IE
Leu
22
128
Y
AC C
EP
TE D
M AN U
SC
RI PT
Mori(2011)
ACCEPTED MANUSCRIPT Wang(2013)
China (Chinese)
neonate
Dir IF (Mab FITC, Millipore)
nPCR, gB
Throat,
51
51
N
Cord Bld Germany
OT
Ind APAAP (C10/11, Biotest)
nPCR, IE
Leu
22
22
Y
Wie(1996)
Korea
BMT
PO (CMV-Vue, Incstar Corp)
nPCR, IE
Bld
28
197
Y
Wirgart(1996)
Sweden
renal T
Ind PO (Clonab CMV, Biotest)
nPCR, IE
Pla
85
471
Y
Xue(2009)
China
Renal T
Ind IF (Brite Turbo Kit)
cPCR,
Pla
1516
7920
Y
Yakushiji(2002)
Japan
ASCT
Dir PO (HRP-C7Mab)
qrtPCR, US17
Pla
51
516
Y
SC
RI PT
Weber(1994)
[Publication type] conf abst: conference abstract. If not specified, English full article.
M AN U
[Background] HSCT: hematopoietic stem cell transplantation. ASCT: Allograft stem cell transplantation. PBSCT: peripheral blood stem cell transplantation. BMT: bone marrow transplantation. SCT: stem cell transplantation. T: transplantation. SOT: solid organ transplantation. ImnS: immunosuppressed. HIV: human immunodeficiency virus. AIDS: acquired immune deficiency syndrome.
[Antigenemia assay] ( ): primary antibody or assay kit are presented in brackets. Ind: indirect. Dir: direct. IF: immunofluorescence. APAAP: alkaline phosphatase anti-alkaline phosphatase. PO: peroxidase. ABC: Avidin Biotin Complex. HRP: Horse radish peroxidase. Mab: monoclonal antibody. FITC:
TE D
fluorescein isothiocyanate. Chemicon Int'l: Chemicon International. IQP: IQ Products [PCR assay] cPCR: conventional/qualitative PCR. nPCR: nested PCR. qrtPCR: quantitative real-time PCR. qPCR: quantitative PCR OTHER THAN real-time
II.
EP
PCR. rev tras PCR: reverse transcription PCR. IE immediate early. mIE major immediate early. gB: glycoprotein B. mtr II: morphological transforming region
AC C
[PCR specimen] Bld: blood. W-BLD: whole blood. Leu: leukocyte. Pla: plasma. Buf C: buffy coat. BALF: broncho-alveolar lavage fluid. [Subject w/o rep] a number of subjects without repeated evaluation of same subjects. [Subject w rep] a number of subjects allowing repeated evaluation of same subjects. [High risk] Y: a study had at least one domains of high risk or high applicability concern in the Revised Tool for the Quality Assessment of Diagnostic Accuracy scoring sheet. N: a study does not.
ACCEPTED MANUSCRIPT
Table 2. Summary of diagnostic test accuracy.
SOT
HSCT
CNAkit
Brite Turbo
Clonab
Studies
75
26
18
14
7
15
PCR positive
9058
5123
1580
1218
PCR negative
22232
10219
3635
3025
DOR
30 (24-38)
28 (18-42)
25 (15-40)
38 (22-64)
I2 = 28%
I2 = 34%
I2 = 37%
AUC
0.86 (0.83-0.88)
0.86 (0.81-0.90)
0.84 (0.80-0.88)
Sensitivity
0.65 (0.59-0.70)
0.65 (0.56-0.73)
0.60 (0.49-0.70)
Specificity
0.94 (0.93-0.95)
0.93 (0.91-0.95)
PLR
10.9 (8.5-14.0)
NLR
0.38 (0.32-0.44)
RI PT
All
1082
6720
2660
35 (30-41)
24 (13-44)
I2 = 22%
I2 = 0%
I2 = 32%
0.88 (0.99-0.92)
0.83 (0.78-0.88)
0.84 (0.78-0.90)
0.65 (0.55-0.75)
0.54 (0.34-0.73)
0.64 (0.52-0.75)
0.94 (0.91-0.97)
0.95 (0.91-0.97)
0.97 (0.94-0.98)
0.94 (0.89-0.96)
9.7 (6.7-14.0)
10.3 (5.3-20.1)
11.9 (6.9-20.6)
16.4 (7.4-33.8)
9.9 (5.6-17.3)
0.38 (0.29-0.47)
0.43 (0.32-0.55)
0.37 (0.27-0.48)
0.47 (0.28-0.68)
0.38 (0.27-0.52)
M AN U
SC
3335
TE D
PCR positive/negative: a number of subjects allowing repeated evaluation of same subjects. DOR: diagnostic odds ratio.
EP
AUC: area under hierarchical summary receiver operating characteristics curve. PLR: Positive likelihood ratio.
SOT: solid organ transplantation recipient.
AC C
NLR: Negative likelihood ratio.
HSCT: hematopoietic stem cell transplantation recipient. CNAkit: CINAkit (Argene Biosoft).
Brite Turbo: CMV Brite Turbo Kit (IQ Products). Clonab: Clonab CMV (Biotest).
ACCEPTED MANUSCRIPT
Figure 1. The Preferred Reporting Items in Systematic Reviews and Meta-Analyses
AC C
EP
TE D
M AN U
SC
RI PT
flow chart for study search.
ACCEPTED MANUSCRIPT
Figure 2. A paired forest plot antigenemia assay for polymerase chain reaction proven cytomegalovirus infection.
AC C
EP
TE D
M AN U
SC
RI PT
TP: true positive. FP: false positive. FN: false negative. TN: true negative.
ACCEPTED MANUSCRIPT
Figure 3. Hierarchical summary receiver operating characteristic curves.
Confidence regions are filled with gray.
RI PT
HSCT: hematopoietic stem cell transplantation
of each study not a confidence region. CNAkit: CINAkit (Argene Biosoft).
AC C
EP
TE D
Clonab: Clonab CMV (Biotest).
M AN U
Brite Turbo: CMV Brite Turbo Kit (IQ Products).
SC
An open circles indicates data from an original study. The circle size represent a weight
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
S1198-743X(17)30269-0
DOI:
10.1016/j.cmi.2017.05.009
Reference:
CMI 948
To appear in:
Clinical Microbiology and Infection
Received Date: 2 February 2017 Revised Date:
21 April 2017
Accepted Date: 7 May 2017
Please cite this article as: Eguchi H, Horita N, Ushio R, Kato I, Nakajima Y, Ota E, Kaneko T, Diagnostic test accuracy of antigenemia assay for polymerase chain reaction proven cytomegalovirus infection: systematic review and meta-analysis, Clinical Microbiology and Infection (2017), doi: 10.1016/ j.cmi.2017.05.009. 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.
ACCEPTED MANUSCRIPT
Diagnostic test accuracy of antigenemia assay for polymerase chain reaction proven cytomegalovirus infection:
Authors Eguchi, [email protected], MD, PhD.
3,#)Nobuyuki
Ushio, [email protected], MD, PhD.
4)Ikuma
Nakajima, [email protected], MD, PhD.
6)Erika
Ota, [email protected], PhD.
3)Takeshi
Kaneko, [email protected], MD, PhD.
1)Harvard
TE D
5)Yuki
Kato, [email protected], MD.
M AN U
3)Ryota
Horita, [email protected], MD, PhD.
SC
1,2,#)Hisashi
RI PT
systematic review and meta-analysis
TH Chan School of Public Health, Boston, MA, USA.
2)Department
of Public Health, Kitasato University School of Medicine, Kanagawa,
3)Department
EP
Japan. Medicine, Yokohama, Japan.
of Pulmonology, Yokohama City University Graduate School of
AC C
Medicine, Yokohama, Japan. 4)Department
of Molecular Pathology, Yokohama City Graduate University School of
Medicine, Yokohama, Japan. 5)Department
of Stem Cell and ImmuneRegulation, Yokohama City University
Graduate School of Medicine, Yokohama, Japan. 6)Global
Health Nursing, Graduate School of Nursing Science, St. Luke's
international university, Tokyo, Japan. #)
Hisashi Eguchi and Nobuyuki Horita equally contributed as the first authors.
ACCEPTED MANUSCRIPT
Corresponding Author Nobuyuki Horita, MD, PhD.
Medicine.
Phone: +81-45-352-7962 Facsimile: +81-45-352-7963
Short title
TE D
Accuracy of Antigenemia for CMV
M AN U
E-mail: [email protected]
Word counts
EP
Abstract: 239 words. Main text: 2589 words.
Key words
AC C
Cytomegalovirus
Immunosuppression Diagnosis
Hierarchical model
Sensitivity and specificity
SC
3-9, Fukuura, Kanazawa-Ku, Yokohama, 236-0004, Japan.
RI PT
Department of Pulmonology, Yokohama City University Graduate School of
ACCEPTED MANUSCRIPT
ABSTRACT Objectives: Objectives We aimed to assess diagnostic test accuracy of antigenemia assay for PCR proven cytomegalovirus (CMV) infection.
RI PT
Methods: Methods We systematically searched studies that provide data both on sensitivity and specificity of the CMV antigenemia assay using the PCR as the reference standard. Adults, children, infants, subjects who were immunocompromised for any reason,
summary
receiver
operating
characteristics
SC
symptomatic patients, and asymptomatic individuals were all included. Hierarchical model
was
used
for
diagnostic
M AN U
meta-analysis. Study quality was assessed by Revised Tool for the Quality Assessment of Diagnostic Accuracy Studies. Protocol registration identification is CRD42016035892. Results: Results We identified 75 eligible articles including 9058 CMV PCR-positive subjects and 22232 PCR-negative subjects. The diagnostic odds ratio for positive antigenemia
TE D
was 30 (95% confidence interval (95%CI) 24-38, I2 = 28%) and the area under the hierarchical summary receiver operating characteristic curve (AUC) was 0.86 (95%CI 0.83-0.88). The summary estimates of sensitivity and specificity were 0.65 (95%CI:
EP
0.59-0.70) and 0.94 (95%CI 0.93-0.95), respectively. While a positive likelihood ratio of 10.9 (95%CI 8.5-14.0) suggested that a positive result from the antigenemia assay
AC C
greatly increased the probability of PCR-proven CMV infection, a negative likelihood ratio of 0.38 (95%CI 0.32-0.44) indicated that a negative result led to a small decrease in the probability of PCR-proven CMV infection. Sensitivity and subgroup analyses replicated these results.
Conclusions: Conclusions The antigenemia assay overlooked 35% of PCR-proven CMV infections; thus, a negative result of an antigenemia assay could not rule out a CMV infection.
ACCEPTED MANUSCRIPT
INTRODUCTION Cytomegalovirus (CMV) screening and/or close monitoring of severely immunocompromised patients are recommended(1-3). Conventional culture has been
RI PT
the gold standard for CMV diagnosis. Histopathological examination using tissue biopsy specimens is useful for the diagnosis of CMV invasive diseases such as CMV gastritis and CMV pulmonary infection, though it is often difficult to access appropriate tissue
SC
samples. Overall, quantitative real-time PCR (qrtPCR) and pp65-antigenemia assays are the most widely used tests for CMV infection diagnosis among immunocompromised
M AN U
patients. These essays can detect CMV reactivation before disease onset, permitting preemptive therapy to control symptomatic CMV disease. Among the two, qrtPCR is generally preferred for some advantages such as obtainability of standardized results, applicability for patients with decreased white blood cell count, not being labor
TE D
intensive, high throughput nature, and better transportability of specimens. The qrtPCR may have higher sensitivity for CMV infection than the antigenemia assay; however, that is still controversial. Thus, most of the recent guidelines have avoided a
EP
clear statement of qrtPCR superiority over the antigenemia assay (4-9). Many studies have compared the diagnostic accuracy of CMV antigenemia and
AC C
PCR assays. Nonetheless, there is considerable variation among the results of these studies. We carried out this systematic review and meta-analysis to evaluate the diagnostic test accuracy of the CMV antigenemia assay using PCR as the reference standard.
ACCEPTED MANUSCRIPT
METHODS
Overview
RI PT
Our study protocol, which has been registered on the International Prospective Register of Systematic Reviews as number CRD42016035892 (10), followed the methodology described in the Cochrane Handbook for Diagnostic Test Accuracy Reviews
SC
and the Preferred Reporting Items in Systematic Reviews and Meta-Analyses
M AN U
statement(11, 12).
Eligibility criteria
Type of studies
We had planned to include both two-gate case-control studies and one-gate
TE D
cohort studies as long as they could provide sufficient data concerning both the sensitivity and specificity of the CMV antigenemia assay when using the PCR as the reference standard(12, 13). However, we were unable to find case-control study, thus, we
EP
included only cohort studies. Conference abstracts and reports written in languages other than English were allowed. Studies covering only sensitivity or only specificity
AC C
were excluded. Studies for non-human subjects such as murine CMV were also excluded.
Participants
We assessed subjects who might have PCR-proven CMV infection. Pregnant women, infants, subjects who were immunocompromised for any reason, symptomatic patients, and asymptomatic individuals were all accepted. However, no included study evaluated pregnant women. Patients receiving CMV treatment could also be included.
ACCEPTED MANUSCRIPT
Index and reference test The index test was the CMV pp65-antigenemia assay using any standard
RI PT
procedure(14, 15). The antigenemia assay for cord blood was accepted. Both commercialized kits and in-house assays were allowed. The reference test was PCR including conventional/qualitative PCR, qrtPCR, nested PCR, and reverse transcription
SC
PCR(16, 17). Quantitative PCR (q-PCR) other than qrtPCR was also accepted. PCR
in addition to blood specimens.
Literature search strategy
M AN U
assays using non-blood specimens such as urine or bronchoalveolar fluid were included
In the electronic search, we systematically searched PubMed, EMBASE, the
TE D
Cochrane Library, and Web of Science on February 28th, 2016. References of previously published reviews and those of included original studies were checked by hand searching. Two investigators independently screened the candidate articles by checking
EP
the title and abstract after uploading the citation list into the software, EndNote X7 (Thomson Reuters, Philadelphia, USA). After independent screening, articles that were
AC C
still regarded as candidates by at least one investigator were then scrutinized independently through full-text reading. Final inclusion was decided after resolving discrepancies between the two investigators. Search formulas are presented
in
Supplement Text 1).
Study selection and data extraction Data were extracted by two reviewers independently and then cross-checked (HE and NH)(12, 13). Data concerning the antigenemia assay (IK) and PCR assay (RU)
ACCEPTED MANUSCRIPT
were also cross-checked again by additional reviewers. When a report presented data for multiple thresholds, we selected data using the lowest cutoff values. Studies that repeatedly evaluated the same individual were permitted for
RI PT
inclusion because repeated evaluations of CMV reactivation for the same recipients are very common. However, when an individual was evaluated twice or more, a report was marked high risk on the flow and timing domain of the quality assessment(18).
SC
If a study evaluated two or more antigenemia assays or two or more PCR assays, the most up-to-date assay for antigenemia and PCR was selected based on
M AN U
investigators discussion (HE and NH). Meta-analysis was conducted based on the assumption that each specimen was independent from another specimen. Inputting the data from one specimen twice was not allowed, because this leads to duplicate use of
TE D
same data.
Quality assessment for bias and and applicability The two investigators independently evaluated each included study by scoring
EP
seven domains of the Revised Tool for the Quality Assessment of Diagnostic Accuracy Studies (HE and NH) (18). This tool evaluated the quality of a primary diagnostic
AC C
accuracy study by checking four key domains covering patient selection, index test, reference standard, and flow of patients. Each domain was marked “high risk of bias” when the risk of bias was judged high based on signal questions (18). For sensitivity analysis, a study with at least one domain that was scored as high risk or high applicability concern was regarded as a high-risk study.
MetaMeta-analysis, statistics, and interpretation of data The main outcomes were the diagnostic test accuracy of pp65-antigenemia for
ACCEPTED MANUSCRIPT
PCR-diagnosed CMV infection evaluated by the following statistics: diagnostic odds ratio (DOR), area under the hierarchical summary receiver operating characteristic curve (AUC), and the summary estimates of the sensitivity, the specificity, the positive
RI PT
likelihood ratio (PLR) and the negative likelihood ratio (NLR). Additionally, we assessed positive and negative predictive values (PPV, NPV) based on the pre-test probability ranging from 0% to 100%.
SC
We used both the hierarchical summary receiver operating characteristics model and the bivariate model (19, 20). The heterogeneity, which is a degree of
M AN U
inconsistency among studies, was assessed using the I² statistic: 0% to 40% represents not important heterogeneity, 30% to 60% represents moderate heterogeneity, 50% to 90% represents substantial heterogeneity, 75% to 100% indicates considerable heterogeneity (21).To determine the overall accuracy, we calculated the DOR using the
TE D
DerSimonian-Laird random-effect model and the AUC using Holling's proportional hazard models. DOR is a measure of the effectiveness of a diagnostic test; wherein DOR = 1 means no diagnostic value, DOR > 1 means that test positive suggests disease
EP
positive, and DOR < 1 means that test negative suggests disease positive. We obtained a paired forest plot, a hierarchical summary receiver operating characteristics curve, and
AC C
summary estimates of the sensitivity and the specificity using the bivariate model (19, 20). PLR, NLR, PPV, and NPV were calculated from the summary estimate of the sensitivity and the specificity. AUC, PLR, and NLR were interpreted according to the four-grade criterion by Grimes et al. and Jones et al. (22, 23). AUC indicates how a test is accurate. AUC in the ranges of -0.75, 0.75-0.92, 0.93-0.96, and 0.97- meant "not accurate", "good", "very good", and "excellent", respectively. PLR defined by “sensitivity/ (1 - specificity)” and NLR defined by “(1 - sensitivity)/ specificity” represent how the test results change the probability of a disease. PLR value in the range of -2, 2-5, 5-10, and
ACCEPTED MANUSCRIPT
10- were recognized as showing a "not meaningful", "small", "moderate", and "large" increase of probability, respectively. NLR in the range of 0.5-, 0.2-0.5, 0.1-0.2, and -0.1 represented a "not meaningful", "small", "moderate", and "large" decrease of probability,
RI PT
respectively. PPV and NPV are proportions of patients who are correctly judged as positive or negative by the test.
The following commands of the statistics software R "mada" package were used
SC
for the meta-analysis: the “madauni” for DOR, the "phm" for AUC, and the “reitsma” for the hierarchical summary receiver operating characteristic curve and the summary
M AN U
estimates of sensitivity and specificity (19, 20). For the publication bias assessment, "funnel" and "metabias" commands of the R "meta" package were used(24).
Subgroup and sensitivity analyses
TE D
As a subgroup analysis, we focused on studies that exclusively evaluated solid organ (SOT) and hematopoietic stem cell transplantation (HSCT) recipients. Additionally, we performed subgroup analysis of studies that assessed specific
EP
commercialized kits or commercialized primary antibodies for antigenemia assays when these kits or antibodies were evaluated in more than five studies. CINAkit (Argene
AC C
Biosoft), CMV Brite Turbo Kit (IQ Products), and Clonab CMV (Biotest) satisfied this criterion. Additional analyses focusing on subgroup of studies with specific PCR procedures and of pediatric studies were done. As a sensitivity analysis, we performed subgroup analyses based on the study
quality.
ACCEPTED MANUSCRIPT
RESULTS
Study search and study characteristics
RI PT
We first identified 803 and 15 articles by electronic and hand search, respectively. Of 818 articles that met the preliminary criteria, 317, 179, and 247 were excluded through removal of duplication, title/abstract screening, and full-article
SC
scrutinizing, respectively (Figure 1, Supplement Text 2). We finally identified 75 eligible articles. Studies were reported from a variety of countries worldwide, most of which
M AN U
were European, Asian, North American, and South American nations (Table 1). Among 75 reports, three were written in Spanish and one each was written in Chinese, French, Germany, Korean, and Japanese. One was a conference abstract and one was a letter article. All other 67 were original articles in English. Patients included in these trials SOT
recipients,
HSCT
recipients,
HIV/AIDS
patients,
non-specified
TE D
were
immune-compromised population, or newborn/infants (Table 1). While 11 studies evaluated an individual only once, the remaining 64 studies
EP
evaluated individuals on multiple occasions. The median number of subjects in studies without double counting was 46 with a range of five to 1516. The median number of
AC C
subjects allowing double counting was 192 with a range of five to 7920, which totaled 31290 subjects (Table 1). Sixty-five studies had at least one domain at high risk of bias or high applicability concern (Supplement Figure 1). The risk of bias for flow and timing domain was most frequently scored "high" because of duplicate evaluation of the same subjects. Antigenemia assays were carried out by indirect immunofluorescence assay in 30 studies. Of note, CINAkit (Argene Biosoft), CMV Brite Turbo Kit (IQ Products) and Clonab CMV (Biotest) were used in 14, seven, and 15 studies, respectively. As a
ACCEPTED MANUSCRIPT
reference test, 29 studies used qrtPCR (Table 1). Seventy-one studies used only blood specimens, such as whole blood, leukocytes, plasma, and sera for PCR (Table 1). Neither the visual inspection nor the Kendall's rank correlation test (P = 0.051)
RI PT
found significant asymmetry of the funnel plot (Supplement Figure 2).
MetaMeta-analysis of all the studies
SC
Using data from all 75 studies, consisting of 9058 PCR-positive subjects and 22232 PCR-negative subjects, the DOR was 30 (95% confidence interval (95%CI) 24-38,
M AN U
I2 = 28%) and AUC was 0.86 (95%CI 0.83-0.88) (Table 2, Figure 3-A).
The summary estimates of sensitivity and specificity were 0.65 (95%CI: 0.59-0.70) and 0.94 (95%CI 0.93-0.95), respectively (Table 2). While a PLR of 10.9 (95%CI 8.5-14.0) suggested that a positive result from the antigenemia assay greatly
TE D
increased the probability of PCR-proven CMV infection, a NLR of 0.38 (95%CI 0.32-0.44) indicated that a negative result led to a small decrease in the probability of PCR-proven CMV infection (Table 2)(22).
EP
PPV and NPV depending on pre-test probability are shown in Supplement Figure 3. PPV was generally better than NPV, though values varied depending on the
AC C
pre-test probabilities (Supplement Figure 3).
Subgroup and sensitivity metameta-analyses analyses The results from the subgroup analyses based on SOT recipients and HSCT are
shown in Figure 3-B, 3-C and Table 2. The DOR based on SOT recipients and HSCT recipients were 28 (95%CI 18-42, I2 = 34%) and 25 (95%CI 15-40, I2 = 37%), respectively. The AUC estimated from data of SOT recipients and HSCT recipients were 0.86 (95%CI 0.81-0.90) and 0.84 (95%CI 0.80-0.88), respectively. We did not find significant
ACCEPTED MANUSCRIPT
differences for these results and the results from all the studies (Table 2). Studies focusing on CNAkit (DOR, 38, 95%CI 22-64) and Brite Turbo (DOR 35, 95%CI 30-41) were associated with higher DOR than all studies (Figure 3-D, Figure 3-E,
RI PT
Table 2). However, the differences were not statistically significant (Table 2).
Sensitivity analyses using data from studies without high risk was carried out. The results of this sensitivity analyses did not differ significantly from those from all
SC
the studies (Supplement Table 1). Subgroup of studies that used only artPCR revealed a DOR of 30 (95% CI 21-43, I2 = 9%) and AUC of 0.84 (95%CI 0.82-0.87). Subgroup of
M AN U
studies that used only blood derivative specimens for PCR revealed a DOR of 30 (95% CI 23-38, I2 = 26%) and AUC of 0.86 (95%CI 0.83-0.88), which was consistent with the results from the analysis that used both blood and non-blood specimens (Supplement Table 1). Based on six pediatric studies, DOR was 44 (95%CI 23-84) and the specificity
TE D
was 0.97 (95%CI 0.96-0.98) (Supplement Table 1). Among the six pediatric studies, one evaluated only neonates, one assessed only infants, one included children with a median of seven years old, one assessed children with a median of 47 months old, and the other
AC C
EP
two did not presented children’s age.
ACCEPTED MANUSCRIPT
DISCUSSION
Our systematic review detected 75 original studies that compared results from
RI PT
antigenemia assays and PCR assays to diagnose CMV infection among patients. The specificity of the antigenemia assay was consistently high and a positive test increased the probability of CMV infection. However, the sensitivity of the assay for PCR-proven
SC
CMV infection varied among the reports ranging from 0.03 to 1.00 with an interquartile range of 0.47 to 0.84. The summary estimates of sensitivity and specificity were 0.65
M AN U
(95%CI 0.59-0.70) and 0.94 (95%CI 0.93-0.95), respectively. A sensitivity of 0.65 meant that a third of PCR-proven CMV infections were not detected by pp65-antigenemia. Thus, a negative test result provided little clinical information. PPVs were generally higher than NPVs though they varied depending on the pre-test probabilities.
TE D
According to the subgroup analyses, we do not have sufficient evidence to recommend a specific commercially available antigenemia kit or primary antibody. Similarly, we do not have reason to recommend or oppose the use of the antigenemia
EP
assay for different populations.. We included studies that used PCR assays other than the standard qrtPCR. However, a sensitivity analysis focusing on studies that used
AC C
qrtPCR provided similar results to the analysis with all the studies. This meant that the PCR method did not considerably affect the results of our study. The major advantage of the antigenemia assay over the qrtPCR assay is that it
does not require expensive equipment. Even in a developed country, some national governments such as Japan do not approve the use of qrtPCR for CMV infection. Some may believe that the lowest detection limit of qrtPCR is too sensitive to distinguish CMV infection and disease and to make a judgement on preemptive therapy.
ACCEPTED MANUSCRIPT
The robustness of our analysis is supported by numerous factors such as the large numbers of included studies and subjects, minor heterogeneity among the studies suggested by I2 =28%, consistent results from many sensitivity analyses, and sound
comment on the limitations of our analysis.
RI PT
methodology following updated guidelines(12, 13). Despite these strengths, we need to We evaluated symptomatic and
asymptomatic patients collectively, Cutoff of qrtPCR and antigenemia assays for
SC
symptomatic CMV disease are still to be clarified. Our analysis also could not suggest a cutoff of qrtPCR and antigenemia assays for preemptive treatments.
M AN U
In conclusion, we carried out the first systematic review and meta-analysis to assess the diagnostic test accuracy of the CMV antigenemia assay. While this test has good PPV, the antigenemia assay overlooked 35% of PCR-proven CMV infection; thus, a negative result of an antigenemia assay could not rule out a CMV infection. Consistent
AC C
EP
supported our results.
TE D
results from sensitivity analyses, solid methodology, and narrow confidence intervals
ACCEPTED MANUSCRIPT
Add Additional information
Contribution
RI PT
H.E. contributed for study search, quality check, data extraction, and drafting.
N.H. work for conception, study search, quality check, data extraction, and analysis as a principal investigator. U.R. and I.K. revised the manuscript concerning PCR assays and
SC
antigenemia assays. Y.N. contributed for study conception and revised the manuscript
manuscript.
Competing Financial Interests
M AN U
as hematologist. E.O. helped analysis as a statistician. T.K. critically revised the
No support in the form of grants, gifts, equipment, and/or drugs was provided.
TE D
Conflict of interest disclosure
AC C
EP
The authors report no conflicts of interest.
ACCEPTED MANUSCRIPT
References 1. Kalil AC, Florescu DF. Prevalence and mortality associated with cytomegalovirus infection in nonimmunosuppressed patients in the intensive care unit. Crit Care
RI PT
Med 2009; 37: 2350-2358.
2. Cantisán S, Torre-Cisneros J, Lara R, Zarraga S, Montejo M, Solana R. Impact of cytomegalovirus on early immunosenescence of CD8+ T lymphocytes after solid
SC
organ transplantation. J Gerontol A Biol Sci Med Sci 2013; 68: 1-5.
3. Ellington SR, Clarke KE, Kourtis AP. Cytomegalovirus Infection in Human
M AN U
Immunodeficiency Virus (HIV)-Exposed and HIV-Infected Infants: A Systematic Review. J Infect Dis 2016; 213: 891-900.
4. Kotton CN, Kumar D, Caliendo AM, Asberg A, Chou S, Danziger-Isakov L, Humar A, Group TSICC. Updated international consensus guidelines on the management of
TE D
cytomegalovirus in solid-organ transplantation. Transplantation 2013; 96: 333-360. 5. Pilmore H, Pussell B, Goodman D. KHA-CARI guideline: cytomegalovirus disease and kidney transplantation. Nephrology (Carlton) 2011; 16: 683-687.
EP
6. O'Brien SM, Keating MJ, Mocarski ES. Updated guidelines on the management of cytomegalovirus reactivation in patients with chronic lymphocytic leukemia treated
AC C
with alemtuzumab. Clin Lymphoma Myeloma 2006; 7: 125-130. 7. Preiksaitis JK, Brennan DC, Fishman J, Allen U. Canadian society of transplantation consensus workshop on cytomegalovirus management in solid organ transplantation final report. Am J Transplant 2005; 5: 218-227.
8. Caliendo AM. Approach to the diagnosis of cytomegalovirus infection. UpToDate. 2016. Topic
8305
Version
15.0.
Available
from:
https://www.uptodate.com/contents/approach-to-the-diagnosis-of-cytomegalovirus-in fection?source=search_result&search=CMV%20diagnosis&selectedTitle=2~150.
ACCEPTED MANUSCRIPT
Accessed on 31st October, 2016. 9. Caliendo AM. Overview of diagnostic tests for cytomegalovirus infection. UpToDate. 2016. Topic
8291
Version
14.0.
Available
from:
RI PT
https://www.uptodate.com/contents/overview-of-diagnostic-tests-for-cytomegalovirus -infection?source=search_result&search=CMV%20diagnosis&selectedTitle=1~150. Accessed on 31st October, 2016.
SC
10. Booth A, Clarke M, Dooley G, Ghersi D, Moher D, Petticrew M, Stewart L. The nuts and
Syst Rev 2012; 1: 2.
M AN U
bolts of PROSPERO: an international prospective register of systematic reviews.
11. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. BMJ 2009; 339: b2535.
TE D
12. Leeflang MM, Deeks JJ, Takwoingi Y, Macaskill P. Cochrane diagnostic test accuracy reviews. Syst Rev 2013; 2: 82.
13. Leeflang MM, Deeks JJ, Gatsonis C, Bossuyt PM, Group CDTAW. Systematic reviews of
EP
diagnostic test accuracy. Annals of internal medicine 2008; 149: 889-897. 14. Boeckh M, Gooley TA, Myerson D, Cunningham T, Schoch G, Bowden RA.
AC C
Cytomegalovirus pp65 antigenemia-guided early treatment with ganciclovir versus ganciclovir at engraftment after allogeneic marrow transplantation: a randomized double-blind study. Blood 1996; 88: 4063-4071.
15. Landry ML, Ferguson D. Comparison of quantitative cytomegalovirus antigenemia assay with culture methods and correlation with clinical disease. J Clin Microbiol 1993; 31: 2851-2856. 16. Machida U, Kami M, Fukui T, Kazuyama Y, Kinoshita M, Tanaka Y, Kanda Y, Ogawa S, Honda H, Chiba S, Mitani K, Muto Y, Osumi K, Kimura S, Hirai H. Real-time
ACCEPTED MANUSCRIPT
automated PCR for early diagnosis and monitoring of cytomegalovirus infection after bone marrow transplantation. J Clin Microbiol 2000; 38: 2536-2542. 17. Limaye AP, Huang ML, Leisenring W, Stensland L, Corey L, Boeckh M. Cytomegalovirus
RI PT
(CMV) DNA load in plasma for the diagnosis of CMV disease before engraftment in hematopoietic stem-cell transplant recipients. J Infect Dis 2001; 183: 377-382.
18. Whiting PF, Rutjes AW, Westwood ME, Mallett S, Deeks JJ, Reitsma JB, Leeflang MM,
SC
Sterne JA, Bossuyt PM, Group Q-. QUADAS-2: a revised tool for the quality assessment of diagnostic accuracy studies. Annals of internal medicine 2011; 155:
19.
Doebler
M AN U
529-536. P.
Package
‘mada’.
Available
from:
https://cran.r-project.org/web/packages/mada/mada.pdf#search='mada+r'. Accessed on November 5th, 2015.
TE D
20. Doebler P., & Holling H. Meta-analysis of diagnostic accuracy with mada. Available from: https://cran.r-project.org/web/packages/mada/vignettes/mada.pdf#search='mada+r+ systematic+review+diagnostic+accuracy'. Accessed on November 5th, 2015.
EP
21. Higgins PJ, Green S. Cochrane Handbook for Systematic Reviews of Interventions (Version 5.1.0). (2011). Available from: http://handbook.cochrane.org/front_page.htm.
AC C
Accessed on December 4th, 2015). 22. Grimes DA, Schulz KF. Refining clinical diagnosis with likelihood ratios. Lancet 2005; 365: 1500-1505.
23. Jones CM, Athanasiou T. Summary receiver operating characteristic curve analysis techniques in the evaluation of diagnostic tests. Ann Thorac Surg 2005; 79: 16-20. 24.
Schwarzer
G.
Package
'meta'.
2016,
Available
ftp://cran.r-project.org/pub/R/web/packages/meta/meta.pdf#search='r+meta'. Accessed on October 30th, 2016.
from:
ACCEPTED MANUSCRIPT
Table 1. Characteristics of included studies.
Study
Country
Subject background
Antigenemia assay (primary antibody or
Author (year)
(publication type)
Allice(2006)
Italy
SOT
Ind EIA (1C3+AYM-,1Argene Biosoft)
Ashokkumar(2015)
USA (conf abst)
liver/intestine T, child
NS
Beckmann(2011)
Switzerland
OT
Ind IF (CINAkit, Argene Biosoft)
Bergallo(2008)
Italy
renal T
NS
Boaretti(2013)
Italy
SOT
Ind IF (CINAkit, Argene Biosoft)
Boeckh(1997)
USA
BMT
Boeckh(2004)
USA
Boivin(1998)
PCR assay, target
RI PT
commercialized kit))
PCR
Subject
Subject
High
specimen
w/o rep
w rep
risk
Bld
48
375
Y
PCR
W-Bld
39
39
N
qrtPCR, UL111a
W-Bld
81
720
Y
rev tras PCR, IE1(UL123)
Leu
14
29
Y
qrtPCR(AmpliP), UL54
Pla
45
266
Y
IF (Clonab CMV, Biotest)
cPCR, IE1
Bld
29
328
Y
Immunosuppressed
IF (Brite kit, Biotest)
qrtPCR, UL123/125/126, UL55(gB)
Pla
NS
1902
Y
Canada
HIV
(1C3 clone, Argene)
cPCR(AmpliT), UL54
Bld
168
172
Y
Boland(1992)
Netherlands
heart/lung T
Ind PO (Mab C10/C11, in-house)
cPCR, IEA
Leu
21
201
Y
Bonon(2006)
Brazil
HSCT
Ind PO (Clonab CMV, Biotest)
nPCR, mIE
Bld
46
46
N
Breda(2013)
Brazil
HSCT
Ind IF (Brite Turbo Kit)
qrtPCR, UL123
W-Bld, Pla
21
196
Y
Camargo(2001)
Brazil
heart T
Ind PO (Clonab CMV, Biotest)
qPCR, LA gp64
Bld
51
162
Y
Cardenoso(2013)
Switzerland
HSCT
Ind IF (Brite Turbo Kit)
qrtPCR(AmpliP), UL54
Pla
135
377
Y
Cathomas(1994)
Switzerland
renal T
Ind PO (Clonab CMV, Biotest)
cPCR, mIE
Leu
18
185
Y
APAAP Ind (Clonab CMV, Biotest)
qrtPCR(Artus),
W-Bld
131
555
Y
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Bld
NS
164
Y
308
192
Y
M AN U
TE D
EP AC C
(Germany)
SC
qrtPCR, UL123
Choi(2009)
Korea
HSCT
Dávila(2005)
Spain
Liver T
Deback(2007)
France
Immunosuppressed
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, HXFL4
W-Bld
Eckart(1996)
France
renal T
IF
cPCR, IE1
Pla
5
5
N
Flexman(2001)
Australia
BMT, SOT
IF (Monofluor Kit CMV, Sanofi)
qPCR(AmpliM),
Pla
NS
469
Y
ACCEPTED MANUSCRIPT France
SOT
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, UL83
W-Bld
61
165
Y
Guiver(2001)
UK
lung/heart T
IF (Mab, Biosoft)
qrtPCR, gpB
Bld
25
362
Y
Hadaya(2003)
Switzerland
SOT
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Pla
23
321
Y
Halfon(2011)
France
HSCT
Ind IF (CINAkit, Argene Biosoft)
qrtPCR(Artus),
Bld
93
244
Y
Hebart(1996)
Germany
BMT
APAAP (Clonab CMV, Biotest)
cPCR, IEA
Pla
20
109
Y
Hoe(2007)
Korea (Korean)
OT
Ind APAAP (Clonab CMV, Biotest)
qrtPCR, gB
W-Bld
84
343
Y
Juillet(2010)
France (French)
Immunosuppressed
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, UL83
W-Bld
5
32
Y
Khansarinejad(2015)
Iran
HSCT, child
Ind IF (Brite Turbo Kit)
qrtPCR, UL83
Pla
82
1179
Y
Kidd(1993)
UK
Immunosuppressed
Ind IF (Mab; FITC, Tago)
qPCR, gB
Leu
NS
295
Y
Ksouri(2006)
Tunisia
BMT
Ind IF (CINAkit, Argene Biosoft)
cPCR,
Pla
18
126
Y
Landini(1995)
Italy
SOT
Ind IF (Mab, Biotest)
cPCR, IE1
Leu
23
104
Y
Lazzarotto(1996)
Italy
OT, AIDS
Ind IF (Mab, Argene)
cPCR, mIEA
Leu
NS
48
Y
Li(2003)
USA
BMT, SOT
Inirect IF (CMV Brite, Biotest)
qrtPCR, IEA
Bld
NS
298
Y
Lo(1997)
Hong Kong
renal T
APAAP Ind IF (Clonab CMV, Biotest)
nPCR, mtr II
Leu
22
108
Y
Lomas(1992)
Spain (Spanish)
HIV, Kaposi, sarcoma
IF (Mab)
cPCR, mIE
Bld
25
25
N
Luchsinger(2015)
Chile (Spanish)
HIV
Ind IF (CINAkit, Argene Biosoft)
qrtPCR, gB
Bld
179
179
N
Madhavan(2007)
India
renal T, Immunosuppressed
Ind IF (Clonab CMV, Biotest)
qrtPCR, mtrII, UL83, gO
Leu
NS
70
Y
Mansy(1999)
Belgium
in viral Dept
Ind IF, PO, ABC, APAAP
nPCR, pp150 (UL32)
Bld
NS
40
Y
Marenzi(1996)
Italy
AIDS
IF (Clonatec, Biosoft)
nPCR, IEA
Ser
21
36
Y
Martiny(2011)
Brazil
Immunosuppressed
Ind IF (Brite Turbo Kit)
nPCR, gB
Bld
NS
216
Y
Masaoka(2001)
Japan
Immunosuppressed
Dir IF (Mitsubishi kit, Yuka Medias)
qPCR(AmpliM), UL54
Ser
115
106
Y
Matsunaga(1999)
Japan
BMT
PO (Mab)
nPCR, IE
Ser
26
26
N
Mauricio(2011)
Chile (Spanish)
SOT, BMT, child
Ind IF (Mab, Argene Biosoft)
qrtPCR,
Bld
68
219
Y
Milan(2011)
Brazil
liver T
Ind PO (C10/C11, Biotest AG)
nPCR,
Bld
94
465
Y
Miller(2010)
USA
HSCT, SOT
Ind IF (Brite Turbo Kit)
qrtPCR, UL54
Pla
NS
61
Y
AC C
EP
TE D
M AN U
SC
RI PT
Garrigue(2006)
ACCEPTED MANUSCRIPT Japan (Japanese)
SCT
Ind ALP stain (Mab C10/11)
qrtPCR,
Pla
13
120
Y
Muller(2007)
Germany
HSCT
Ind IF (Mab, Virion)
qPCR(AmpliM), UL54
Pla
40
435
Y
Nitsche(2003)
Germany (letter)
SCT
NS
qrtPCR, mIE
Pla
77
1122
Y
Pang(2003)
Canada
SOT
IF (Clonab CMV, Biotest AG)
qrtPCR, gB
Pla
66
404
Y
Pellegrin(2000)
France
renal T
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Pla
30
283
Y
Peres(2010)
Brazil
HSCT
Ind PO (Mab, IQP; HRP, Biotest)
qrtPCR, US17
Bld
30
30
N
Piiparinen(2001)
Finland
liver/renal T
PO (Mab, Biotest Pharma)
qPCR(AmpliM), UL54
Pla
NS
253
Y
Rautenberg(1999)
Germany
SOT
Ind APAAP (Clonab CMV, Biotest)
nPCR, UL122
Bld
26
69
Y
Rayes(2005)
Germany
renal T
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Pla
72
389
Y
Rhee(2011)
Korea
renal T
Ind APAAP (Clonab CMV, Biotest)
qrtPCR(Artus),
Pla
111
899
Y
Sakamaki(1997)
Japan
BMT
Dir PO (HRP-C7Mab, Teijin)
cPCR, mIE
BALF
30
43
Y
Sanghavi(2008)
USA
OT
Ind EIA (Clonatec 1C3, Biosoft)
qrtPCR, US17, UL54
Bld
12
3422
Y
Schirm(1999)
Netherlands
Immunosuppressed
Ind PO (Mab C10/C11)
cPCR, IE
Bld
55
280
Y
Schmidt(1994)
Germany
BMT
Ind PO (Clonab CMV, Biotest)
cPCR, IE
Buf C
50
96
Y
Schmidt(1995)
Germany
Liver T
Ind APAAP (Clonab CMV, Biotest)
cPCR, β2mg
Buf C
30
264
Y
Schvoerer(2005)
France
OT
IF (Biotest)
qrtPCR, LA(HXFL4)
Pla
16
145
Y
Seehofer(2004)
Germany
Liver T
Ind IF (CINAkit, Argene Biosoft)
qPCR(AmpliM), UL54
Bld
135
719
Y
Solano(2001)
Spain
ASCT
Ind IF (Chemicon Int'l)
cPCR(AmpliT),
Pla
43
548
Y
Sowmya(2006)
India
Immunosuppressed
Ind IF (DAKO)
nPCR, mtr II, gO, UL83
Buf C, Urn
74
92
Y
Sun(2009)
China
infant
Ind EIA (Chemicon, BD Biosciences)
qrtPCR, UL123
Urn
126
126
N
Tantivanich(2002)
Thailand
HIV s/o, child
Ind PO (AAC10, DAKO)
nPCR,
Leu
100
100
N
Thorne(2007)
USA
Immunosuppressed
Ind IF (Brite Turbo Kit)
qrtPCR, IEA
W-Bld
49
106
Y
Toulemonde(2000)
France
SOT, SCT
Ind IF (CINAkit, Argene Biosoft)
qc-PCR, (Us Region(cPCR))
Leu
110
443
Y
Velzing(1994)
Netherlands
Immunosuppressed
Ind PO (Clonab CMV, Biotest)
cPCR, mIEA
Bld
36
37
Y
Vlieger(1992)
Netherlands
BMT
PO (Mab, in-house)
nPCR, IE
Leu
22
128
Y
AC C
EP
TE D
M AN U
SC
RI PT
Mori(2011)
ACCEPTED MANUSCRIPT Wang(2013)
China (Chinese)
neonate
Dir IF (Mab FITC, Millipore)
nPCR, gB
Throat,
51
51
N
Cord Bld Germany
OT
Ind APAAP (C10/11, Biotest)
nPCR, IE
Leu
22
22
Y
Wie(1996)
Korea
BMT
PO (CMV-Vue, Incstar Corp)
nPCR, IE
Bld
28
197
Y
Wirgart(1996)
Sweden
renal T
Ind PO (Clonab CMV, Biotest)
nPCR, IE
Pla
85
471
Y
Xue(2009)
China
Renal T
Ind IF (Brite Turbo Kit)
cPCR,
Pla
1516
7920
Y
Yakushiji(2002)
Japan
ASCT
Dir PO (HRP-C7Mab)
qrtPCR, US17
Pla
51
516
Y
SC
RI PT
Weber(1994)
[Publication type] conf abst: conference abstract. If not specified, English full article.
M AN U
[Background] HSCT: hematopoietic stem cell transplantation. ASCT: Allograft stem cell transplantation. PBSCT: peripheral blood stem cell transplantation. BMT: bone marrow transplantation. SCT: stem cell transplantation. T: transplantation. SOT: solid organ transplantation. ImnS: immunosuppressed. HIV: human immunodeficiency virus. AIDS: acquired immune deficiency syndrome.
[Antigenemia assay] ( ): primary antibody or assay kit are presented in brackets. Ind: indirect. Dir: direct. IF: immunofluorescence. APAAP: alkaline phosphatase anti-alkaline phosphatase. PO: peroxidase. ABC: Avidin Biotin Complex. HRP: Horse radish peroxidase. Mab: monoclonal antibody. FITC:
TE D
fluorescein isothiocyanate. Chemicon Int'l: Chemicon International. IQP: IQ Products [PCR assay] cPCR: conventional/qualitative PCR. nPCR: nested PCR. qrtPCR: quantitative real-time PCR. qPCR: quantitative PCR OTHER THAN real-time
II.
EP
PCR. rev tras PCR: reverse transcription PCR. IE immediate early. mIE major immediate early. gB: glycoprotein B. mtr II: morphological transforming region
AC C
[PCR specimen] Bld: blood. W-BLD: whole blood. Leu: leukocyte. Pla: plasma. Buf C: buffy coat. BALF: broncho-alveolar lavage fluid. [Subject w/o rep] a number of subjects without repeated evaluation of same subjects. [Subject w rep] a number of subjects allowing repeated evaluation of same subjects. [High risk] Y: a study had at least one domains of high risk or high applicability concern in the Revised Tool for the Quality Assessment of Diagnostic Accuracy scoring sheet. N: a study does not.
ACCEPTED MANUSCRIPT
Table 2. Summary of diagnostic test accuracy.
SOT
HSCT
CNAkit
Brite Turbo
Clonab
Studies
75
26
18
14
7
15
PCR positive
9058
5123
1580
1218
PCR negative
22232
10219
3635
3025
DOR
30 (24-38)
28 (18-42)
25 (15-40)
38 (22-64)
I2 = 28%
I2 = 34%
I2 = 37%
AUC
0.86 (0.83-0.88)
0.86 (0.81-0.90)
0.84 (0.80-0.88)
Sensitivity
0.65 (0.59-0.70)
0.65 (0.56-0.73)
0.60 (0.49-0.70)
Specificity
0.94 (0.93-0.95)
0.93 (0.91-0.95)
PLR
10.9 (8.5-14.0)
NLR
0.38 (0.32-0.44)
RI PT
All
1082
6720
2660
35 (30-41)
24 (13-44)
I2 = 22%
I2 = 0%
I2 = 32%
0.88 (0.99-0.92)
0.83 (0.78-0.88)
0.84 (0.78-0.90)
0.65 (0.55-0.75)
0.54 (0.34-0.73)
0.64 (0.52-0.75)
0.94 (0.91-0.97)
0.95 (0.91-0.97)
0.97 (0.94-0.98)
0.94 (0.89-0.96)
9.7 (6.7-14.0)
10.3 (5.3-20.1)
11.9 (6.9-20.6)
16.4 (7.4-33.8)
9.9 (5.6-17.3)
0.38 (0.29-0.47)
0.43 (0.32-0.55)
0.37 (0.27-0.48)
0.47 (0.28-0.68)
0.38 (0.27-0.52)
M AN U
SC
3335
TE D
PCR positive/negative: a number of subjects allowing repeated evaluation of same subjects. DOR: diagnostic odds ratio.
EP
AUC: area under hierarchical summary receiver operating characteristics curve. PLR: Positive likelihood ratio.
SOT: solid organ transplantation recipient.
AC C
NLR: Negative likelihood ratio.
HSCT: hematopoietic stem cell transplantation recipient. CNAkit: CINAkit (Argene Biosoft).
Brite Turbo: CMV Brite Turbo Kit (IQ Products). Clonab: Clonab CMV (Biotest).
ACCEPTED MANUSCRIPT
Figure 1. The Preferred Reporting Items in Systematic Reviews and Meta-Analyses
AC C
EP
TE D
M AN U
SC
RI PT
flow chart for study search.
ACCEPTED MANUSCRIPT
Figure 2. A paired forest plot antigenemia assay for polymerase chain reaction proven cytomegalovirus infection.
AC C
EP
TE D
M AN U
SC
RI PT
TP: true positive. FP: false positive. FN: false negative. TN: true negative.
ACCEPTED MANUSCRIPT
Figure 3. Hierarchical summary receiver operating characteristic curves.
Confidence regions are filled with gray.
RI PT
HSCT: hematopoietic stem cell transplantation
of each study not a confidence region. CNAkit: CINAkit (Argene Biosoft).
AC C
EP
TE D
Clonab: Clonab CMV (Biotest).
M AN U
Brite Turbo: CMV Brite Turbo Kit (IQ Products).
SC
An open circles indicates data from an original study. The circle size represent a weight
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT