Simultaneous quantification of first-line anti-tuberculosis drugs and metabolites in human plasma

Abstract

Tuberculosis (TB) currently kills more people than any other infectious disease worldwide, the highest burden being in Africa and Asia (1). Therapy recommended for drug sensitive TB consists of a cocktail of isoniazid (INH), rifampicin (RIF), pyrazinamide (PZA) and ethambutol (EMB), all given in a 2-month intensive phase, followed by only INH and RIF in a 4-month continuation phase. Clinical studies seeking to optimize dosing, gain more knowledge on the pharmacokinetics and pharmacodynamics of the drugs and compare current therapy to alternative regimens are required (2, 3). Therapeutic drug monitoring (TDM) is frequently carried out in cases responding poorly to therapy (3, 4). Both clinical studies and TDM require bioanalytical methods for quantifying drug concentrations in biological fluids. Several methods have been developed, mostly analysing individual drugs but a few analyse combinations. Ideally, quantification of all four drugs in one method is desirable as it is economical and allows high throughput. A method was developed and validated for the quantification of first line anti-tuberculosis drugs EMB, INH, PZA and RIF and the metabolites N-acetyl isoniazid (AcINH) and 25- desacetyl rifampicin (desRIF). Sample preparation consisted of protein precipitation, followed by high performance liquid chromatography (HPLC) with tandem mass spectrometry (MS/MS) detection. Deuterated internal standards for each analyte (AcINH-d4, desRIF-d3, EMB-d4, INH-d4, PZA-15N,d3 and RIF-d3) were used. Mean recoveries of the analytes from plasma were as follows: AcINH 106.5%, DesRIF 123.2%, EMB 105.3%, INH 110.1%, PZA 132.0% and RIF 127.7%. Sample preparation was followed by reverse phase liquid chromatography on an Agilent 1200 series HPLC system using an Agilent Poroshell 120 EC- C18 2.7µm, 4.6 X 50mm analytical column. Separation of all analytes was achieved using a mobile phase gradient consisting of an aqueous mobile phase A (0.05% formic acid in water) and an organic mobile phase B (0.05% formic acid in a mixture of methanol and acetonitrile, 1:1). A T-junction splitter was used to reduce the mobile phase flow to the ion source by about 30%. Retention times for AcINH, desRIF, EMB, INH, PZA and RIF were 2.45, 5.40, 1.75, 2.22, 4.30 and 5.68 minutes respectively. An AB Sciex API 4000 triple quadrupole mass spectrometer at unit mass resolution in the multiple reaction monitoring (MRM) mode was used for detection, monitoring the following transitions for the six analytes: AcINH 180 → 121, desRIF 784 → 752, EMB 205 → 116, INH 138 → 79, PZA 124 → 81 and RIF 823 → 792. An electrospray ionisation (ESI) source in the positive ion mode was used to couple the mass spectrometer to the LC system. Accuracy and precision were assessed over three consecutive and independent runs. The calibration curves fit quadratic regressions for all analytes, with weighting of 1/x (where x=concentration) for all except PZA which had a weighting of 1/x2 over the calibration range. Calibration ranges in µg/ml were as follows: AcINH 0.050 – 12.5, desRIF 0.040 – 10.0, EMB 0.020 – 5.00, INH 0.100 – 25.0, PZA 0.32. – 80.0 and RIF 0.120 – 30.0, based on peak area ratios. A 1:4 dilution of the QC Dilute sample showed that concentrations of up to 20.0 µg/ml for AcINH, 16.0 µg/ml for desRIF, 8.00 µg/ml for EMB, 40.0 µg/ml for INH, 128 µg/ml for PZA and 48.0 µg/ml for RIF in plasma could be analysed reliably when diluted into the calibration range. No significant carry-over was observed for all analytes. The method was shown to be reproducible when human plasma samples from six different sources were analysed and endogenous matrix components had no significant effect on the assay. All analytes were stable in plasma for at least four hours on ice, and when subjected to three freeze-thaw cycles. Reinjection reproducibility experiments showed that all analytes except PZA could be reliably analysed by re-injecting an entire batch after about 48 hours. Quantification of AcINH, INH and RIF was not significantly affected by 2% hemolysis of sample while desRIF, EMB and PZA were significantly affected. Data was analysed using Analyst ® version 1.6.2 software. With wide calibration ranges, the assay is suitable for both routine TDM and PK studies. Concurrent analysis of metabolites allows inferences to be made on the PK of the two main TB drugs. The total run time of 6.5 minutes per sample combined with the simple sample preparation procedure, the method is more economical on both time and resources than single analyte assays.