OBJECTIVES

We conducted an updated network meta-analysis to elucidate the best regimen for latent tuberculosis infection (LTBI).

METHODS

We searched the PubMed, Embase, and Cochrane Library databases on 16 August 2021 to perform an updated network meta-analysis. Only randomised controlled trials on populations with LTBI that reported the efficacy for preventing incident tuberculosis or the completion rates of treatment regimens were included. The Cochrane Collaboration tool was used to assess the risk of bias. We tested for possible global inconsistency with a χ test and local inconsistency by calculating inconsistency factors for each comparison in closed loops. The probability of each regimen being at each possible rank was estimated. Comparison-adjusted funnel plots were obtained to assess publication bias, and sensitivity analysis was performed. The major outcomes were the efficacy for preventing incident tuberculosis and the completion rates of treatment regimens.

RESULTS

We identified 27 studies that matched our inclusion criteria; the risk of bias was mostly low. Rifampicin for four months (RFMP-4) was the most likely to be effective (probability: 56.3%) and the second most likely treatment to be completed (probability: 22.4%). By applying a multidimensional scaling approach for ranking based on a scatterplot with the surface under the cumulative ranking values for efficacy and completion rates, RFMP-4 was deemed the best choice for treating LTBI. Similar results were demonstrated after sensitivity analysis.

CONCLUSION

This updated network meta-analysis revealed RFMP-4 to be the best choice for treating LTBI, per simultaneous consideration of efficacy and completion rates.