Development and validation of an LC-MS/MS multiplex assay for the quantification of bedaquiline, n-desmethyl bedaquiline, linezolid, levofloxacin, and clofazimine in dried blood spots

•LC-MS/MS method to simultaneously quantify five second-line TB drugs in DBS.•Single procedure (PPT and SPE) extracting polar and hydrophobic analytes together.•Drug concentrations measured accurately from a single drop of blood (∼40 µL).•Strong correlation and agreement between plasma and estimated plasma drug levels.•DBS method a good alternative to traditional methods for remote, resource-poor areas. Dried blood spot (DBS) assays to quantify novel and repurposed drugs for the treatment of rifampicin-resistant tuberculosis (RR-TB) would facilitate pharmacokinetic studies and therapeutic drug monitoring in low-middle income settings, considering their ease of application and simple sample storage requirements. We describe a DBS method for the simultaneous quantification of bedaquiline and metabolite N-desmethyl bedaquiline, linezolid, levofloxacin, and clofazimine. The analytes were extracted from the matrix and isolated by solid-phase extraction. Two LC-MS/MS systems were used, optimized for the separate analysis of the more polar compounds (linezolid and levofloxacin), and less polar compounds (bedaquiline, N-desmethyl bedaquiline, and clofazimine), employing gradient elution. Electrospray ionization and multiple reaction monitoring were used to quantify the analytes on a Sciex API3200 and an API5500 triple quadrupole mass spectrometer, for the more polar and less polar analytes, respectively. Isotopically labelled internal standards were used to compensate for variability in the quantification of each analyte. The method was validated according to international guidelines and applied to samples from a clinical trial. We performed correlation and agreement analysis of the DBS assay and in-house plasma methods using Deming regressions and Bland–Altman plots. Coefficients of correlation between measured plasma and DBS concentrations ranged from 0.866 (95% CI: 0.817–0.902) to 0.989 (95% CI: 0.985–0.992). More than 67% of the samples showed a difference between the observed and estimated plasma concentrations within 20% of their means, meeting EMA requirements for method reproducibility and demonstrating the interchangeability of our DBS and plasma LC-MS/MS methods.