Evaluation of dried blood spot sampling for pharmacokinetic research and therapeutic drug monitoring of anti-tuberculosis drugs in children

Abstract

BackgroundDried blood spot (DBS) sampling for pharmacokinetic (PK) studies and therapeutic drug monitoring have unique advantages over venous sampling. This study aimed to evaluate a DBS method for first-line anti-tuberculosis drugs in children, and DBS sampling to assess PK parameters. MethodsParaguayan children were treated according to the revised paediatric dosing scheme of the World Health Organization. A PK curve was performed both with DBS sampling and conventional venous sampling for rifampicin, pyrazinamide and ethambutol. Passing–Bablok regression, Bland–Altman plots and predictive performance evaluation were used to assess agreement between DBS and plasma concentrations. The percentages of patients attaining population PK values for Cmax and AUC0–24h were calculated. ResultsAfter use of a conversion factor, Passing–Bablok regression showed no significant proportional or systematic bias between DBS and plasma concentrations. Bland–Altman plots showed that 95% of the ratios of the DBS predicted:observed plasma concentrations lay between 0.6 and 1.4 for rifampicin, 0.5 and 1.6 for pyrazinamide and -0.4 and 2.8 for ethambutol. DBS measurements showed acceptable predictive performance for rifampicin and pyrazinamide, but not for ethambutol. Assessment of Cmax target attainment was 62.5% for isoniazid, 25% for rifampicin, 100% for pyrazinamide and 75% for ethambutol. ConclusionFor rifampicin and pyrazinamide, the DBS method was accurate in predicting plasma concentrations, and was used successfully for PK parameter assessment. However, predicting ethambutol plasma concentrations with DBS measurement was associated with too much imprecision. Despite higher dosing, only 25% of the population reached average target adult rifampicin exposures.