Integrase inhibitors versus efavirenz combination antiretroviral therapies for TB/HIV coinfection: a meta-analysis of randomized controlled trials

Background Integrase inhibitors (INIs)-based antiretroviral therapies (ART) are more recommended than efavirenz (EFV)-based ART for people living with HIV/AIDS (PLWHA). Yet, the advantage of integrase inhibitors in treating TB/HIV coinfection is uncertain. Therefore, the objective of this systematic review is to evaluate the effects and safety of INIs- versus EFV-based ART in TB/HIV coinfection, and demonstrate the feasibility of the regimens. Methods Four electronic databases were systematically searched through September 2020. Fixed-effects models were used to calculate pooled effect size for all outcomes. The primary outcomes were virologic suppression and bacteriology suppression for INIs- versus EFV-based ART. Secondary outcomes included CD4+ cell counts change from baseline, adherence and safety. Results Three trials (including 672 TB/HIV patients) were eligible. ART combining INIs and EFV had similar effects for all outcomes, with none of the point estimates argued against the INIs-based ART on TB/HIV patients. Compared to EFV-based ART as the reference group, the RR was 0.94 (95% CI 0.85 to 1.05) for virologic suppression, 1.00 (95% CI 0.95 to 1.05) for bacteriology suppression, 0.98 (95% CI 0.95 to 1.01) for adherence. The mean difference in CD4+ cell counts increase between the two groups was 14.23 cells/μl (95% CI 0− 6.40 to 34.86). With regard to safety (adverse events, drug-related adverse events, discontinuation for drugs, grade 3–4 adverse events, IRIS (grade 3–4), and death), INIs-based regimen was broadly similar to EFV-based regimens. The analytical results in all sub-analyses of raltegravir- (RAL) and dolutegravir (DTG) -based ART were valid. Conclusion This meta-analysis demonstrates similar efficacy and safety of INIs-based ART compared with EFV-based ART. This finding supports INIs-based ART as a first-line treatment in TB/HIV patients. The conclusions presented here still await further validation owing to insufficient data.