BACKGROUNDThere is a need for objective measures of adherence to fully-oral bedaquiline-based regimens for rifampicin- and multidrug-resistant tuberculosis (RR/MDR-TB). Due to their prolonged half-lives, bedaquiline (BDQ) and clofazimine CFZ) are appropriate for evaluating long-term treatment adherence using a model-based approach.METHODSWe enrolled adults initiating a 9-month regimen that included BDQ and CFZ for RR/MDR-TB. Dried blood spot samples were taken at intervals and assayed for BDQ and CFZ concentrations. Demographic and clinical factors were used to predict drug concentrations using known pharmacokinetic models, assuming full adherence. Individual observed concentrations were compared to model-predicted values (1000 simulations). Observations were ranked, with those below the 5 th percentile of simulated concentrations, indicating possible poor adherence.RESULTSAdherence worsened over time, especially for BDQ (11, 21, 26, and 35% of the observations below the 5 th percentile in the first, second, fourth, and sixth month respectively). CFZ observations under the 5 th percentile were 10, 9, 7.5, 14, and 23% at months 1, 2, 4, 6, and 9.CONCLUSIONConcentrations of BDQ and CFZ below the 5 th percentile increased over time and this trend was more marked with BDQ. We hypothesise that the complexity of the thrice-weekly BDQ regimen increases the risk of suboptimal adherence.
Reséndiz‐Galván et al. (Fri,) studied this question.