Statin Use May Be Associated With Reduced Active Tuberculosis Infection: A Meta-Analysis of Observational Studies

• Main Authors: Xiaofei Li, Lina Sheng, Lanqing Lou
• Format: Article
• Language: English
• Published: Frontiers Media S.A. 2020-04-01
• Series: Frontiers in Medicine
• Subjects: statin; tuberculosis; infection; diabetes; meta-analysis
• Online Access: https://www.frontiersin.org/article/10.3389/fmed.2020.00121/full

Background: Tuberculosis remains one of the leading causes of mortality among the infectious diseases, while statins were suggested to confer anti-infective efficacy in experimental studies. We aimed to evaluate the association between statin use and tuberculosis infection in a meta-analysis.Method: Relevant studies were obtained via systematically search of PubMed and Embase databases. A random or a fixed effect model was applied to pool the results according to the heterogeneity among the included studies. Subgroup analyses according to the gender and diabetic status of the participants were performed. We assessed the quality of evidence with the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.Results: Nine observational studies were included. Significant heterogeneity was detected among the studies (p for Cochrane's Q test <0.001, I2 = 93%). The GRADE approach showed generally low quality of evidence. Pooled results showed that statin use was associated with reduced active tuberculosis infection (risk ratio [RR]: 0.60, 95% confidence interval [CI]: 0.45 to 0.75, p < 0.001). Subgroup analyses showed that the negative association between statin use and active tuberculosis infection was consistent in men (RR: 0.63, p = 0.01) and women (RR: 0.58, p < 0.001), in participants with (RR: 0.63, p = 0.02) and without diabetes (RR: 0.50, p < 0.001), in retrospective cohort studies (RR: 0.56, p = 0.02), prospective cohort studies (RR: 0.76, p = 0.03), nested case-controls studies (RR: 0.57, p < 0.001), and case-control studies (RR: 0.60, p < 0.001), and in studies with statin used defined as any use within 1 year (RR: 0.59, p < 0.001) or during follow-up (RR: 0.61, p < 0.001). Significant publication bias was detected (p for Egger's regression test = 0.046). Subsequent “trim and fill” analyses retrieved an unpublished study to generate symmetrical funnel plots, and meta-analysis incorporating this study did not significantly affect the results (RR: 0.72, 95% CI: 0.68 to 0.76, p < 0.001).Conclusions: Statin use may be associated with reduced active tuberculosis infection. Randomized controlled trials (RCTs) are needed to confirm the potential preventative role of statin use on tuberculosis infection.