Causal association of antibiotic use and resistant tuberculosis infection: Casecontrol meta-analysis

In the world scientific literature, analytical research on tuberculosis has not been meta-analyzed. The objective of this study was to identify risk factors for tuberculosis infection with drug resistance and meta-analyze the causal relationship of prior antibiotic use.

Systematic review with meta-analysis of case-control studies, in five databases. An ex ante, exhaustive and reproducible protocol of search and selection was applied; with criteria of inclusion, exclusion and evaluation of methodological quality. Were performed a qualitative and quantitative synthesis of the articles that evaluated the previous consumption of antibiotics. The PRISMA guide applied and a meta-analysis of random effects was performed for the odds ratios, with analysis of Galbraith, Funelt Plot, Forest plot and sensitivity analysis.

We included 36 articles for the qualitative synthesis and 16 in the meta-analysis. We found a wide heterogeneity in the risk factors that include sociodemographic characteristics such as age, sex, schooling, occupation and prison; Clinics as contact with infected, absence of BCG vaccine, hospitalization, chronic comorbidities, malnutrition, HIV coinfection; And microbiological variables such as infection by Beijing genotype and therapeutic adherence. In the studies that evaluated previous consumption of antibiotic, 1880 cases and 5291 controls were studied, most with moderate or low methodological quality, with a combined measure that shows that the odds of developing resistance in patients with previous antibiotic use are 12 (95% = 6.0-23.7) times found for the non-exposed, in the meta-regression the odds were 16.6 (95% CI = 4.1-67.8) for the moderate quality studies and 5.0 (95% CI = 2.9, 8.7) for those with high methodological quality.

This meta-analysis revealed a strong causal association between the prior use of anti-tuberculosis antibiotics and drug-resistant Micobacterium tuberculosis infection.