Acquired rifampicin resistance during first TB treatment: magnitude, relative importance, risk factors and keys to control in low-income settings.

Background: The incidence of acquired rifampicin resistance (RIF-ADR; RR) during first-line treatment varies. Objectives: Compare clinically significant RIF-ADR versus primary and reinfection RR, between regimens (daily versus no rifampicin in the continuation phase; daily versus intermittent rifampicin in the continuation phase) and between rural Bangladesh and Kinshasa, Democratic Republic of Congo. Methods: From patients with treatment failure, relapse, or lost to follow-up, both the outcome and baseline sputum sample were prospectively collected for rpoB sequencing to determine whether RR was present in both samples (primary RR) or only at outcome (RIF-ADR or reinfection RR). Results: The most frequent cause of RR at outcome was primary RR (62.9%; 190/302). RIF-ADR was more frequent with the use of rifampicin throughout versus only in the intensive phase (difference: 3.1%; 95% CI: 0.2-6.0). The RIF-ADR rate was higher with intermittent versus daily rifampicin in the continuation phase (difference: 3.9%; 95% CI: 0.4-7.5). RIF-ADR after rifampicin-throughout treatment was higher when resistance to isoniazid was also found compared with isoniazid-susceptible TB. The estimated RIF-ADR rate was 0.5 per 1000 with daily rifampicin during the entire treatment. Reinfection RR was more frequent in Kinshasa than in Bangladesh (difference: 51.0%; 95% CI: 34.9-67.2). Conclusions: RR is less frequently created when rifampicin is used only during the intensive phase. Under control programme conditions, the RIF-ADR rate for the WHO 6 month rifampicin daily regimen was as low as in affluent settings. For RR-TB control, first-line regimens should be sturdy with optimal rifampicin protection. RIF-ADR prevention is most needed where isoniazid-polyresistance is high, (re)infection control where crowding is extreme.