Development and validation of a simple LC-MS/MS method for simultaneous determination of moxifloxacin, levofloxacin, prothionamide, pyrazinamide and ethambutol in human plasma.

Treatment of multidrug-resistant tuberculosis (MDR-TB) is challenging due to high treatment failure rate and adverse drug events. This study aimed to develop and validate a simple LC-MS/MS method for simultaneous measurement of five TB drugs in human plasma and to facilitate therapeutic drug monitoring (TDM) in MDR-TB treatment to increase efficacy and reduce toxicity. Moxifloxacin, levofloxacin, prothionamide, pyrazinamide and ethambutol were prepared in blank plasma from healthy volunteers and extracted using protein precipitation reagent containing trichloroacetic acid. Separation was achieved on an Atlantis T3 column with gradient of 0.1% formic acid in water and acetonitrile. Drug concentrations were determined by dynamic multiple reaction monitoring in positive ion mode on a LC-MS/MS system. The method was validated according to the United States' Food and Drug Administration (FDA) guideline for bioanalytical method validation. The calibration curves for moxifloxacin, levofloxacin, prothionamide, pyrazinamide and ethambutol were linear, with the correlation coefficient values above 0.993, over a range of 0.1-5, 0.4-40, 0.2-10, 2-100 and 0.2-10 mg/L, respectively. Validation showed the method to be accurate and precise with bias from 6.5% to 18.3% for lower limit of quantification and -5.8% to 14.6% for LOW, medium (MED) and HIGH drug levels, and with coefficient of variations within 11.4% for all levels. Regarding dilution integrity, the bias was within 7.2% and the coefficient of variation was within 14.9%. Matrix effect (95.7%-112.5%) and recovery (91.4%-109.7%) for all drugs could be well compensated by their isotope-labelled internal standards. A benchtop stability test showed that the degradation of prothionamide was over 15% after placement at room temperature for 72 h. Clinical samples (n = 224) from a cohort study were analyzed and all concentrations were within the analytical range. The signal of prothionamide was suppressed in samples with hemolysis which was solved by sample dilution. As the method is robust and sample preparation is simple, it can easily be implemented to facilitate TDM in programmatic MDR-TB treatment.