Statin use and risk of tuberculosis: A systemic review of observational studies

Journal Pre-proof Statin use and risk of tuberculosis: A systemic review of observational studies Haizhen Duan, Tongying Liu, Xiaojun Zhang, Anyong Yu, Yu Cao, Li Zhang, Yuyin Zhu

PII:

S1201-9712(20)30038-2

DOI:

https://doi.org/10.1016/j.ijid.2020.01.036

Reference:

IJID 3927

To appear in:

International Journal of Infectious Diseases

Received Date:

15 November 2019

Revised Date:

11 January 2020

Accepted Date:

18 January 2020

Please cite this article as: Duan H, Liu T, Zhang X, Yu A, Cao Y, Zhang L, Zhu Y, Statin use and risk of tuberculosis: A systemic review of observational studies, International Journal of Infectious Diseases (2020), doi: https://doi.org/10.1016/j.ijid.2020.01.036

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. 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. © 2019 Published by Elsevier.

Statin use and risk of tuberculosis: A systemic review of observational studies Running title: Statins and tuberculosis Haizhen Duana,b, Tongying Liub, Xiaojun Zhangb, Anyong Yub, Yu Caoa*, Li Zhangc, Yuyin Zhuc Department of Emergency Medicine ， West China Hospital of Sichuan

ro of

a

University，Chengdu，PR China b

Department of Emergency Medicine, Affiliated hospital of Zunyi Medical

Second Department of Pulmonary Medicine, Ningbo No.2 Hospital, Ningbo,

re

c

-p

University, Zunyi, PR China

Zhejiang, PR China

lP

*Address correspondence to Yu Cao

Department of Emergency Medicine，West China Hospital of Sichuan University，

na

No.37 Guoxue Road, Chengdu , 610041，Sichuan, PR,China

ur

E-mail: [email protected], Tel:+86-28-85422288, Fax:+86-28-85422288 Abstract word count: 235

Jo

Text word count: 2428

The number of references: 36 The number of figures: 5 The number of tables:1

1

Highlights ·Six observational articles with more than 200,0000 patients were enrolled ·Statin use significantly reduced 22% risk of tuberculosis in diabetes mellitus patients (pooled RRs, 0.78; 95%CI, 0.63–0.95) with severe heterogeneity (I2=76.1%).

ro of

·Statin intake also significantly decreased 40% risk of tuberculosis in the general population (pooled RRs, 0.60; 95%CI, 0.50–0.71) with severe

-p

heterogeneity.

re

ABSTRACT

lP

Objectives: Statin intake may be linked with less risk of several infectious diseases, including tuberculosis, which is an important cause of mortality worldwide. We

na

aimed to probed into the definite impacts of statins on the risk of tuberculosis in diabetic patients and in the general population.

ur

Methods: Four databases were thoroughly searched from inception up to July 2019.

Jo

Articles in any language were enrolled if they assessed and clarified statin intake, presented risk of tuberculosis in diabetes mellitus patients or the general population, and reported odds ratios(ORs), relative risks (RRs) or hazard ratios(HRs) or contained data for relevant calculation. RRs with 95% confidence intervals (CIs) were pooled using random- effects models regardless of heterogeneity quantified by Cochran's Q and I2 statistics. 2

Results: Six observational articles with 2，073，968 patients were enrolled, including 4 cohort articles, 1 nested case-control and 1 abstract. Statin use significantly reduced 22% risk of tuberculosis in diabetes mellitus patients (pooled RRs, 0.78; 95%CI, 0.63–0.95) with severe heterogeneity (I2=76.1%). Statin intake also significantly decreased 40% risk of tuberculosis in the general population (pooled RRs, 0.60; 95%CI, 0.50–0.71) with severe heterogeneity (I2=57.7%).

ro of

Conclusions: Statin use is related to a considerably lower risk of tuberculosis both in diabetic patients and the general population. However, these conclusions should be

understood carefully given the possible remaining confounding, and call for large-size

-p

and multi-center random controlled studies in the future.

re

Keywords: Statins, Tuberculosis, Diabetes mellitus, Risk, Systemic review

lP

Introduction

Tuberculosis (TB), a chronic infectious disease induced by Mycobacterium

na

tuberculosis, is among the top 10 sources of death globally. In 2018, TB attacked more than 10,000,000 new cases and killed 1,500,000 people, including 251,000

ur

people affected by human immunodeficiency virus globally(World Health

Jo

Organization,2019). Geographically, the estimated incidence/mortality rate of TB in Korea and Taiwan is 34/2.4 and 38.9/2.1 per 100,000 people, respectively(World Health Organization,2019; Center for Disease Control, Department of Health, executive yuan, Taiwan,2018). The chronic diabetic status is a definite risk factor of TB, which is caused by various factors, as the occurrence of TB from the reactivated old foci surpasses that via fresh contact. About 5% to 30% of TB cases are 3

accompanied with diabetes mellitus (DM) (Ruslami et al., 2010). One review shows that diabetes can intensify the risk of TB by 1.5 to 7.8 times (Stevenson et al., 2007). A meta-analysis indicates diabetes raises the risk of TB among ordinary people by 3 times(Jeon and Murray, 2008), but even more in diabetic patients when diabetes is not well controlled(Baker et al., 2012). Statins are medicated by hyperlipidemic patients to cure and prevent coronary

ro of

heart diseases and strokes by blocking 3-hydroxy-3-methylglutaryl-CoA reductase, a main enzyme of cholesterol generation. Also diabetic patients take statins for various

positive outcomes. However, statins affect the immune system in multiple ways and

-p

are related to the improved outcomes of many infectious diseases(Kim et al., 2018;

re

Ma et al., 2017; Uthman et al., 2018; Yeh et al., 2018). As one infectious diseases, TB is still one major public health threat worldwide. Host cholesterol is one critical factor

lP

in the occurrence of TB in several ways in vitro (Brzostek et al., 2009; Huynh et al.,

na

2008; Miner et al., 2009), and removal of membrane cholesterol can specifically prevent TB pathogens from entering host macrophages(Gatfield and Pieters, 2000).

ur

Therefore, cholesterol reduction by statins is logically expected to have a favorable outcome during TB management. However, the anti-TB mechanism of action is not

Jo

believed to be directly due to reduction of host cholesterol, but is likely multifactorial, such as enhancing phagosomal maturation and host-protective autophagy,(Parihar et al., 2014) and,antitubercular activity of statins,(Dutta et al., 2016, 2019) and the host-directed(Guler and Brombacher, 2015). Several studies present the association between statin use and less risk of active TB in DM patients (Lai et al., 2016; Lee et 4

al., 2015; Su et al., 2017), but this finding is not supported by other studies(Kang et al., 2014; Kim et al., 2019). Recently, a population- based cohort article found the standard-course treatment did not improve the outcomes of TB patients.(Chen et al., 2019) The inconsistency among the existing findings may be attributed to the differences in participants and trial intervention. Nevertheless, there is no disagreement about the effects of statins on reducing the risk of TB in the general

ro of

population, since all extant studies proved statins can actively reduce the risk of TB in the general population (Kim et al., 2019; Lai et al., 2016; Su et al., 2017).

Regarding the above analyses, we thoroughly reviewed and meta-analyzed the

-p

existing observational studies (OSs) that explored the connection between statin use

Methodology

lP

Search strategy and selection criteria

re

and TB risk in DM patients and in the general population.

na

We searched Web of Science, Wiley Cochrane Central Register of Controlled Trials, PubMed, Scopus, and Ovid EMBASE from inception up to July 2019. The

ur

keywords included statin(s), HMG-CoA reductase inhibitor(s), atorvastatin, fluvastatin, lovastatin, pravastatin, rosuvastatin, simvastatin; tuberculosis or TB;

Jo

Mycobacterium tuberculosis. References in relevant articles were manually looked through to find out potential articles. Search was conducted independently by HD and TL and any disagreement was solved with a third reviewer (YC). The authors were contacted to ascertain the unpublished data if necessary. Inclusion and exclusion criteria 5

OSs involving participants aged ≥ 18 years; and analyzing the TB risk after statin user in DM patients or in the general population were included. Other study designs were excluded, including case report, unpublished article, comment, quasi-trial, letter and editorial. The article with the largest sample size or the latest presentation was selected if more than one articles from the same affiliation or team were qualified. Data extraction and quality assessment

ro of

Study selection and data extraction were conducted as per the Cochrane

instruction. Demographic information and Charlson comorbidity parameter were

extracted in standardized forms. Data about type of study, statin use (A standardized

-p

data abstraction case report form was developed for statin use, duration, frequency,

re

dose，and so on), and trial properties (exclusion and inclusion criteria) were also collected. The characteristics and clinical relevance of potential articles were

na

discussing with YC.

lP

reviewed by HD and TL independently. Any conflict between them was addressed by

The quality of each OS was evaluated by HD and TL independently using the

ur

nine-star Newcastle-Ottawa scale (NOS). A score of over six stars was indicative of high quality(Gu et al., 2015). The strength of evidence (SOE) of each outcome was

Jo

graded as high, moderate, low or inadequate by HD and TL independently as per the Grading

of

Recommended

Assessment,

Development

and

Evaluation

(GRADE)(Berkman et al., 2015). Any conflict was handled by a joint reassessment of the corresponding article with YC. Statistical analysis 6

The primary and secondary end points were the risk of TB in DM patients and in the general population respectively. Since the absolute risk of TB is low in diabetic or general population, the odds ratio (OR), relative risk (RR) and hazard ratio (HR) of association can ensure similar estimates of RR(Siristatidis et al., 2013). Dichotomous data were gathered by using RR and 95% confidence interval (CI) on random-effect models. heterogeneity

was

quantified

as

severe

(I2

≥

50%),

ro of

Between-study

moderate(25%–50%) or low (<25%). Any detectable source of heterogeneity was identified via sensitivity analysis by removing the articles one by one. Subanalyses by

-p

region (Korea and Taiwan) and mean age (<60 and ≥60 years old) were also done.

re

Publication bias was judged based on visual interpretation of the asymmetry of

lP

funnel curves and on Begg's and Egger's tests. All data were analyzed on STATA 14.0 with significant level at P <0.05, except for the publication bias test (P<0.10).

na

Results

Study search and characteristics

ur

The first search returned 634 potential articles, in which 39 articles were further assessed. Four articles published in abstracts involved the already included

Jo

participants and thus were excluded. Finally, 6 OSs were enrolled, including 4 cohort articles (Kang et al., 2014; Kim et al., 2019; Lee et al., 2015; Su et al., 2017), 1 nested case-control article (Lai et al., 2016), and one abstract (SHENG-WEI PAN and WEI-JUIN SU, 2018) (Figure 1). Of the 6 OSs, the sample size varied from 13,981 to 840,899 patients, with a total of 2,073,968 patients, and the female percents ranged 7

from 31.2% to 54.0%. Three articles only included diabetic patients (Kang et al., 2014; Lee et al., 2015; SHENG-WEI PAN and WEI-JUIN SU, 2018), and three articles involved both diabetic patients and the general population (Kim et al., 2019; Lai et al., 2016; Su et al., 2017). Tables 1 list the main information of the 6 studies. Quality evaluation The 5 OSs were highly qualified, with NOS scores of ≥7 (mean, 9.0; Table S1).

ro of

Effect of statin use on the TB risk of DM patients

In the 6 OSs, statin use significantly decreased the occurrence rate of TB in DM

patients by 22% (pooled RR, 0.78; 95%CI, 0.63–0.95; P=0.015), with severe

-p

heterogeneity (I2=76.1%; P=0.001; Figure 2). The pooled RRs were widely consistent

re

across Taiwan (P<0.001) and the elder population (≥60 years old, P=0.001), but not across Korea (P=0.703) or the younger population (<60 Years old, P=0.345) (Figures

lP

3 and 4).

na

For sensitivity analysis, the removal of one study(Kang et al., 2014) gently changed the overall risk judgement (pooled RR 0.71; 95%CI,0.61–0.82), with no

ur

heterogeneity (I2=17.8%; P=0.301). No evident publication bias was revealed by funnel plot examination (Fig. S1) or

Jo

Egger's test (P=0.123). The low GRADE implied that statins resisted TB. Effect of statin use on TB risk in the general population The statin-use versus non-statin use significantly lowered the risk of TB in general patients by 40% (pooled RR, 0.60; 95%CI, 0.50–0.71; P<0.001), with severe heterogeneity (I2=57.7%, P=0.094, Figure 5). The low GRADE proved that statins 8

protected the general population from TB. Discussion This review and meta-analysis involving 2,073,968 patients suggests statin use is related with a considerable 22% and 40% decrease in the TB risk among diabetic patients and the general population respectively. Whether risk of TB differed among regions and among ages was explored via subgroup analyses. The a-priori subgroup

ro of

analysis by regions implied statin use versus non-statin use largely relieved the TB risk in the studies conducted in Taiwan, but not in Korea. Of the two trails(Kang et al.,

2014; Kim et al., 2019) from Korea, one study only involved new cases of type 2 DM

-p

and presented no data on death(Kang et al., 2014), which can be a bias from the

re

competing risks between death and TB that may lower the TB occurrence in statin users. The other study found statin intake relieved the risk of active TB in general

lP

people, but was weakened by diabetes (Kim et al., 2019). They thought the studies

na

from Taiwan may have a major limitation of inadequate control of confounding factors, which may account for the disagreement regarding the TB risk depending on

ur

whether diabetes occurred. In terms of age, the risk of TB after statin use versus non-statin intake significantly decreased only in elder population, which may be

Jo

because statins were often medicated for several positive outcomes in the elderly. Other drugs also have the anti-TB effect for elder population(Lee et al., 2015). Besides regions and ages，some other key and potential confounders, including prior TB status, body mass index(BMI) or obesity, cholesterol levels, smoking status, validity of TB diagnosis, presence of cardiovascular disease (CVD), and mortality 9

which may play important roles in the association between statin use and TB, were not available because of limited data in the review. The lack of these confounding factors might deviate us from more accurate estimate for the association. Latent tuberculous infection was not evaluated by the most selected studies. Recently, Lin et al(Lin et al., 2018) reported that the protective effect of obesity on TB risk was weakened by a mediation effect of diabetes. The validity of TB diagnosis based on the

ro of

ICD codes and the prescription criteria in most selected studies might be challenged

since it might include TB cases that were not bacteriologically confirmed and cause misclassification for active TB. Meanwhile, comorbidities such as CVD, at least in

-p

part smoking-related, could be high risk factors for TB development among statin

re

users, which might lead to an overestimation of the risk of TB among statin users. As for the duration and dose response relationship between statin exposure and

lP

the incident TB disease risk, only two studies(Lai et al., 2016; Su et al., 2017)

na

reported them respectively. Lai et al(Lai et al., 2016) suggested the length of statin therapy affected the TB protection and a trend toward having temporal association

ur

between the decreased TB risk and the time period after statin discontinuation by conducting a sensitivity analysis using different censoring cutoffs. Su et al(Su et al.,

Jo

2017) revealed a dose-dependent benefit of statins on the TB risk and found that there was no protective effect on TB risk by statin use of < 180 cDDDs. There are some probable mechanisms underlying statin use may lower the risk of TB. Evidence implies that cholesterol metabolization by M. tuberculosis occurs throughout the tuberculosis infection in the human host. Such sterol metabolism is 10

crucial in the disease progression (Miner et al., 2009). Statins reduce macrophage cholesterol by multiple mechanisms, including the decreased cholesterol formation, activated efflux and blocked ester accumulation(Argmann et al., 2005; Qiu and Hill, 2008). Statin intake prevented phagocytosis in macrophages in vitro due to the cholesterol-lowering effect (Loike et al., 2004). Other than cholesterol reduction, statins may play beneficial roles, such as anti-inflammation and immune-modulatory

ro of

for the host during tuberculosis. These potentially conflicting immune-modulating

activities of statins include promotion of autophagy and phagosome maturation in macrophages(Dutta et al., 2019; Parihar et al., 2014), as well as induction of NKT

-p

cells and promotion of the secretion of pro-inflammatory cytokines, such as IL-1β

re

and IL-12p70(Guerra-De-Blas et al., 2019). On the other hand, recent studies have suggested that statins may prevent induction of trained immunity(Bekkering et al.,

lP

2018), which is important in resistance to Mtb infection(Joosten et al., 2018).

na

Like other reviews and meta-analyses, the heterogeneity of OSs was unavoidable, given the discrepancy in study design, patient information and statistical method.

ur

With our strict criteria, the enrolled studies stand for a thorough try to pool relevant published and upcoming studies. Thus, we adopted the summary-level judgments of

Jo

single study effects. Also the traditional statistical methods used in the OSs did not effectively handle the influence of undetected confounding factors on the overall effect judgement. We explored the origins of heterogeneousness through sensitivity analyses. The incorporated RR for TB of the six OSs in diabetic patients was smaller than that reported by Kang et al.(Yang et al., 2015)(0.78 vs. 0.98). The possible reason 11

was discussed above. This review has some advances, such as the extensive search strategy, adoption of Cochrane instruction, and large sample size. However, it also has some disadvantages. Firstly, all the included studies were from Asia and did not involve RCT or unpublished article, which may bias our findings. Secondly, little data were found about the dose and duration effect of statins, case numbers, or person years for

ro of

dose-reaction meta-analyses. As maybe expected, the risks of TB observed in diabetic patients were highly heterogeneous among the OSs. Furthermore, it was not blinded

during quality assessment and study selection in the meta-analysis. In conclusion,

-p

statin use is linked with a less risk of TB both in diabetic patients and the general

re

population. However, these conclusions should be understood carefully regarding the potential remaining confounding. Thus, large-size and multi- center RCTs are needed.

lP

Contributions

na

XZ, AY, LZ and YZ done the statistical analysis, HD and TL performed the database research, YC reviewed the conflicts in the databases obtained by HD and TL. HD

ur

wrote the first draft of the manuscript; TL and YC reviewed the manuscript critically for contents. All authors are in agreement with the content of the manuscript and

Jo

approved the final manuscript. Funding

The authors did not receive any funding, be it from a public, private or not for-profit institution. Ethical approval 12

Not required. Conflict of interest The authors declare no competing interests. Acknowledgements

ro of

None.

References World

Health

Organization.

Global

Tuberculosis

Report.

https://www.who.int/tb/publications/global_report/en/.

2019.

Available

from:

-p

Center for Disease Control, Department of Health, executive yuan, Taiwan. Tuberculosis annual report. 2018. Available from: https://www.cdc.gov.tw/File/Get/lnTbCo2CdtjzYqWZWFEL8Q.

Argmann CA, Edwards JY, Sawyez CG, O'Neil CH, Hegele RA, Pickering JG, et al. Regulation of macrophage

ABCA1-mediated

re

cholesterol efflux through hydroxymethylglutaryl-CoA reductase inhibition: a role for RhoA in cholesterol

efflux.

J

Biol

Chem

2005;280:22212-21,

doi:http://dx.doi.org/10.1074/jbc.M502761200.

mellitus:

a

lP

Baker MA, Lin HH, Chang HY, Murray MB. The risk of tuberculosis disease among persons with diabetes prospective

cohort

study.

Clin

Infect

Dis

2012;54:818-25,

doi:http://dx.doi.org/10.1093/cid/cir939.

Bekkering S, Arts R, Novakovic B, Kourtzelis I, van der Heijden C, Li Y, et al. Metabolic Induction of Trained through

the

na

Immunity

Mevalonate

Pathway.

Cell

2018;172:135-46.e9,

doi:http://dx.doi.org/10.1016/j.cell.2017.11.025. Berkman ND, Lohr KN, Ansari MT, Balk EM, Kane R, McDonagh M, et al. Grading the strength of a body of

ur

evidence when assessing health care interventions: an EPC update. J Clin Epidemiol 2015;68:1312-24, doi:http://dx.doi.org/10.1016/j.jclinepi.2014.11.023. Brzostek A, Pawelczyk J, Rumijowska-Galewicz A, Dziadek B, Dziadek J. Mycobacterium tuberculosis is able to and

utilize

cholesterol.

J

Bacteriol

2009;191:6584-91,

Jo

accumulate

doi:http://dx.doi.org/10.1128/JB.00488-09.

Chen YT, Kuo SC, Chao PW, Chang YY. Use of lipid-lowering agents is not associated with improved outcomes for tuberculosis patients on standard-course therapy: A population-based cohort study. PLoS One 2019;14:e0210479, doi:http://dx.doi.org/10.1371/journal.pone.0210479.

Dutta NK, Bruiners N, Pinn ML, Zimmerman MD, Prideaux B, Dartois V, et al. Statin adjunctive therapy shortens the

duration

of

TB

treatment

in

mice.

J

Antimicrob

Chemother

2016;71:1570-7,

doi:http://dx.doi.org/10.1093/jac/dkw014. Dutta NK, Bruiners N, Zimmerman MD, Tan S, Dartois V, Gennaro ML, et al. Adjunctive host-directed therapy with

statins

improves

tuberculosis-related 13

outcomes

in

mice.

J

Infect

Dis

2019,

doi:http://dx.doi.org/10.1093/infdis/jiz517. Gatfield J, Pieters J. Essential role for cholesterol in entry of mycobacteria into macrophages. Science 2000;288:1647-50, doi:http://dx.doi.org/10.1126/science.288.5471.1647. Gu WJ, Wang F, Tang L, Liu JC. Single-dose etomidate does not increase mortality in patients with sepsis: a systematic review and meta-analysis of randomized controlled trials and observational studies. Chest 2015;147:335-46, doi:http://dx.doi.org/10.1378/chest.14-1012. Guerra-De-Blas P, Bobadilla-Del-Valle M, Sada-Ovalle I, Estrada-García I, Torres-González P, López-Saavedra A, et al. Simvastatin Enhances the Immune Response Against Mycobacterium tuberculosis. Front Microbiol 2019;10:2097, doi:http://dx.doi.org/10.3389/fmicb.2019.02097. Guler R, Brombacher F. Host-directed drug therapy for tuberculosis. Nat Chem Biol 2015;11:748-51, doi:http://dx.doi.org/10.1038/nchembio.1917. Huynh KK, Gershenzon E, Grinstein S. Cholesterol accumulation by macrophages impairs phagosome maturation. J Biol Chem 2008;283:35745-55, doi:http://dx.doi.org/10.1074/jbc.M806232200.

ro of

Jeon CY, Murray MB. Diabetes mellitus increases the risk of active tuberculosis: a systematic review of 13 observational studies. PLoS Med 2008;5:e152, doi:http://dx.doi.org/10.1371/journal.pmed.0050152.

Joosten SA, van Meijgaarden KE, Arend SM, Prins C, Oftung F, Korsvold GE, et al. Mycobacterial growth inhibition

is

associated

with

trained

innate

immunity.

J

doi:http://dx.doi.org/10.1172/JCI97508.

Clin

Invest

2018;128:1837-51,

-p

Kang YA, Choi NK, Seong JM, Heo EY, Koo BK, Hwang SS, et al. The effects of statin use on the development of

tuberculosis among patients with diabetes mellitus. Int J Tuberc Lung Dis 2014;18:717-24, doi:http://dx.doi.org/10.5588/ijtld.13.0854.

propensity

score-matched

re

Kim MC, Yun SC, Lee SO, Choi SH, Kim YS, Woo JH, et al. Statins increase the risk of herpes zoster: A analysis.

PLoS

One

2018;13:e0198263,

doi:http://dx.doi.org/10.1371/journal.pone.0198263.

Diabetes:

lP

Kim MC, Yun SC, Lee SO, Choi SH, Kim YS, Woo JH, et al. Association between Tuberculosis, Statin Use, and A Propensity

Score-Matched

Analysis.

Am

J

Trop

Med

Hyg

2019;101:350-6,

doi:http://dx.doi.org/10.4269/ajtmh.18-0983.

Lai CC, Lee MT, Lee SH, Hsu WT, Chang SS, Chen SC, et al. Statin treatment is associated with a decreased risk

na

of active tuberculosis: an analysis of a nationally representative cohort. Thorax 2016;71:646-51, doi:http://dx.doi.org/10.1136/thoraxjnl-2015-207052. Lee MY, Lin KD, Hsu WH, Chang HL, Yang YH, Hsiao PJ, et al. Statin, calcium channel blocker and Beta blocker

ur

therapy may decrease the incidence of tuberculosis infection in elderly Taiwanese patients with type 2 diabetes. Int J Mol Sci 2015;16:11369-84, doi:http://dx.doi.org/10.3390/ijms160511369. Lin HH, Wu CY, Wang CH, Fu H, Lönnroth K, Chang YC, et al. Association of Obesity, Diabetes, and Risk of Two

Population-Based

Cohorts.

Clin

Infect

Dis

2018;66:699-705,

Jo

Tuberculosis:

doi:http://dx.doi.org/10.1093/cid/cix852.

Loike JD, Shabtai DY, Neuhut R, Malitzky S, Lu E, Husemann J, et al. Statin inhibition of Fc receptor-mediated phagocytosis by macrophages is modulated by cell activation and cholesterol. Arterioscler Thromb Vasc Biol 2004;24:2051-6, doi:http://dx.doi.org/10.1161/01.ATV.0000143858.15909.29.

Ma X, Sun D, Li C, Ying J, Yan Y. Statin use and virus-related cirrhosis: A systemic review and meta-analysis. Clin Res Hepatol Gastroenterol 2017;41:533-42, doi:http://dx.doi.org/10.1016/j.clinre.2017.07.004. Miner MD, Chang JC, Pandey AK, Sassetti CM, Sherman DR. Role of cholesterol in Mycobacterium tuberculosis infection. Indian J Exp Biol 2009;47:407-11. Parihar SP, Guler R, Khutlang R, Lang DM, Hurdayal R, Mhlanga MM, et al. Statin therapy reduces the 14

mycobacterium tuberculosis burden in human macrophages and in mice by enhancing autophagy and phagosome maturation. J Infect Dis 2014;209:754-63, doi:http://dx.doi.org/10.1093/infdis/jit550. Qiu G, Hill JS. Atorvastatin inhibits ABCA1 expression and cholesterol efflux in THP-1 macrophages by an LXR-dependent

pathway.

J

Cardiovasc

Pharmacol

2008;51:388-95,

doi:http://dx.doi.org/10.1097/FJC.0b013e318167141f. Ruslami R, Aarnoutse RE, Alisjahbana B, van der Ven AJ, van Crevel R. Implications of the global increase of diabetes for tuberculosis control and patient care. Trop Med Int Health 2010;15:1289-99, doi:http://dx.doi.org/10.1111/j.1365-3156.2010.02625.x. SHENG-WEI PAN JF, WEI-JUIN SU YC. Reducing tuberculosis but contracting shingles in patients with type 2 diabetes receiving statin vs nonstatin lipid lowering agents. Respirology 2018;23:90–334. Siristatidis C, Sergentanis TN, Kanavidis P, Trivella M, Sotiraki M, Mavromatis I, et al. Controlled ovarian hyperstimulation for IVF: impact on ovarian, endometrial and cervical cancer--a systematic review and meta-analysis. Hum Reprod Update 2013;19:105-23, doi:http://dx.doi.org/10.1093/humupd/dms051.

tuberculosis:

a

neglected

threat

to

public

health.

doi:http://dx.doi.org/10.1177/1742395307081502.

ro of

Stevenson CR, Critchley JA, Forouhi NG, Roglic G, Williams BG, Dye C, et al. Diabetes and the risk of Chronic

Illn

2007;3:228-45,

Su VY, Su WJ, Yen YF, Pan SW, Chuang PH, Feng JY, et al. Statin Use Is Associated With a Lower Risk of TB. Chest 2017;152:598-606, doi:http://dx.doi.org/10.1016/j.chest.2017.04.170.

-p

Uthman OA, Nduka C, Watson SI, Mills EJ, Kengne AP, Jaffar SS, et al. Statin use and all-cause mortality in people living with HIV: a systematic review and meta-analysis. BMC Infect Dis 2018;18:258, doi:http://dx.doi.org/10.1186/s12879-018-3162-1.

in

patients

with

hepatitis

C

re

Yang YH, Chen WC, Tsan YT, Chen MJ, Shih WT, Tsai YH, et al. Statin use and the risk of cirrhosis development virus

infection.

J

Hepatol

2015;63:1111-7,

doi:http://dx.doi.org/10.1016/j.jhep.2015.07.006.

lP

Yeh JJ, Lin CL, Hsu CY, Shae Z, Kao CH. Statin for Tuberculosis and Pneumonia in Patients with Asthma⁻ Chronic Pulmonary Disease Overlap Syndrome: A Time-Dependent Population-Based Cohort

Figure legend

Jo

ur

na

Study. J Clin Med 2018;7, doi:http://dx.doi.org/10.3390/jcm7110381.

15

ro of -p

Jo

ur

na

lP

re

Figure 1. Flow diagram of study selection.

Figure 2. Forest plot for the association between statin use and the risk of tuberculosis (X-axis: log scale; solid square: relative risk; horizontal lines: 95% CIs. The same in other figures). 16

ro of

-p

Figure 3. Forest plot for the association between statin use and the risk of tuberculosis

Jo

ur

na

lP

re

according to regions

Figure 4. Forest plot for the association between statin use and the risk of tuberculosis according to ages

17

ro of

Figure 5. Forest plot for the association between statin use and the risk of tuberculosis

Jo

ur

na

lP

re

-p

in general patients

18

Study

Baseline years

Design

Kang, 2014

Korea

Cohort

2007-2010

Sample

Mean age

Sex

size

(years)

(%Female)

840899

56.3

f

40.9

Taiwan

Cohort

1998-2009

13981

>65 years

Lai, 2016

Taiwan

Nested

1999-2011

817898

54.0 31.2

Jo ur

Su, 2017

Taiwan

Cohort

2000-2013

305142

58.02

Population

Effect estimates

At

Adjusting confounders (Yes or No)

least one prescription

DM

of any type of statin during

aHR(95%CI):

Yes

0.98(0.89-1.07)

the study period * 2

NR

DM

* 2

Having statins prescription

General and

aRR(95%CI)[General]:

record ≥7 days(Current user

DM

0.66(0.56-0.78);

aRR(95%CI):

Yes

0.76(0.60-0.97)

status

na l

60.2

Definition of statin user

definition

* 1

Pr

Lee, 2015

case-control

TB

pr

Region

e-

Study

oo

Table 1 Detailed demographic characteristics of studies included in the meta-analysis

α

Yes

aRR (95%CI) [DM]: vs

Recent user

0.70(0.54-0.91)

β γ status vs Past user status δ vs Chronic use status ) 49.3

* 2

A cumulative prescription

General and

aHR (95%CI)[General]:

record for ≥ 30 days of one

DM

0.53(0.47-0.61);

or more of the following

aHR (95%CI)[DM]:

compounds:

simvastatin,

0.60(0.47-0.78)

lovastatin,

pravastatin,

fluvastatin,

atorvastatin,

cerivastatin,

rosuvastatin,

and pitavastatin

19

Yes

Korea

Cohort

2001-2013

2003-2013

40012

56036

NR

52.5

NR

NR

f

Cohort

51.1

* 3

NR

DM

aHR(95%CI):0.66(0.44-0.

Received prescriptions

General and

aHR (95%CI)[General]:

for statins continuously for

DM

0.67(0.46-0.98);

at least 7 days

aHR (95%CI)[DM]: 1.05(0.66-1.67)

Pr

TB; tuberculosis；DM: diabetes mellitus; aHR, adjusted hazard ratio; aRR, adjusted risk ratios; CI, confidence interval; NR, not reported. 1*Both the use of ICD-10 codes for TB (A15–A19, U88.0–U88.1) to identify the disease and the satisfaction of prescription criteria, based on a prescription of at

na l

least one of the following: 1) isoniazid and rifamycin, 2) ethambutol, 3)pyrazinamide, 4) prothionamide, 5) cycloserine, or 6) paraaminosalicylic acid. 2 * ICD-9-CM codes of TB plus the prescription of more than two anti-tuberculosis medications for more than 28 days.

Jo ur

3* ICD-10codes, along with two or more anti-TB medications of 1) isoniazid and rifamycin, 2) ethambutol, 3)pyrazinamide, 4) prothionamide, 5) cycloserine, or 6) paraaminosalicylic acid that were prescribed within 3 months of the diagnosis of TB. α

referred to patients with a statin prescription that was filled within 30 days of the index date；β Referred to patients with a statin prescription filled between 31

and 90 days prior to the index date；γ Referred to patients with a statin prescription filled between 91 days and 1 year prior to the index date；δ Referred to patients with a cumulative prescription length of more than 90 days in the year in which the TB diagnosis was made.

20

Yes

99)

e-

pr

Kim, 2019

Taiwan

oo

Pan, 2018( abstract)

Yes