Browsing by Subject "bioanalytics"

Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Open Access
    Development of novel bioanalytical assays and an objective adherence measure of key second-line tuberculosis drugs
    (2026) Kenfack Teponnou, Gerard Aime; Wiesner, Joachim; Maartens, Gary; Court, Richard
    Drug resistance prevails as a challenge in eradicating tuberculosis. Poor adherence and limited understanding of drug pharmacology impede treatment strategies. Therapeutic drug monitoring, relying on the measurement of drug concentrations, aims to optimise drug exposure and medication adherence. Traditional methods analyse drugs in plasma/serum samples, processed and stored by cumbersome and expensive procedures, which are unsuitable in most high burden settings. Dried blood spot assays, conversely, use smaller volumes, less invasive sampling techniques, and easier processing and storage procedures. We developed a dried blood spot assay for quantifying bedaquiline and N-Desmethyl bedaquiline, linezolid, levofloxacin, and clofazimine, and applied it to clinical study samples. Analytes were extracted from dried blood spots through protein precipitation, purified by solidphase extraction, and quantified using LC-MS/MS. Analyte separation was on two chromatographic systems for the polar (levofloxacin and linezolid) and less polar (bedaquiline, N-Desmethyl bedaquiline, and clofazimine) compounds. The method was validated against international guidelines, over 0.0181-4.94 µg/mL (bedaquiline), 0.00905-2.47 µg/mL (N-Desmethyl bedaquiline), 0.00814-2.22 µg/mL (clofazimine), 0.0741-20.2 µg/mL (levofloxacin), and 0.113-30.9 µg/mL (linezolid). Paired dried blood spot and plasma samples (141) from the BEAT-TB clinical trial were analysed with the dried blood spot and existing plasma methods. Deming regressions were used to correlate the dried blood spot and plasma concentrations. Estimated plasma concentrations were derived from Deming regression equations. Bland-Altman plots were used to determine the magnitude of bias between estimated and measured plasma concentrations. The median AUC0-24 ratio compared the exposure calculated from the estimated and measured plasma concentrations. The non-parametric Wilcoxon signed-rank test was used to assess the statistical significance of the agreement. Dried blood spot and plasma concentrations correlated well (r > 0.86); percentage differences between mean estimated and measured plasma concentrations ranged from −1.1 to 2.1%; and > 67% of paired samples were within 20% of their mean. Median AUC0-24 ratios ([plasma]/ [estimated plasma]) were 1.1 (bedaquiline), 1.0 (N-Desmethyl bedaquiline), 1.2 (clofazimine), 1.0 (levofloxacin), and 1.0 (linezolid). The test statistic supports agreement between plasma and estimated AUCs for all analytes except Clofazime, likely due to interpatient variability in AUC, other biological factors, and limitations associated with the relatively small sample size. The strong correlation between dried blood spot and plasma concentrations and the negligible bias between estimated and measured plasma concentrations were consistent with the agreement observed between estimated and measured plasma AUCs, demonstrating the method's potential for use in pharmacokinetic studies and adherence monitoring. Future work will explore heel and finger-prick sampling to extend the use of our method into resource-constrained settings