Repeated voriconazole therapeutic drug monitoring in hospitalised patients: a real-world cohort study of out-of-range exposure and concentration response to dose adjustment

Background Longitudinal evidence on repeated voriconazole therapeutic drug monitoring (TDM) in selected hospitalised patients is limited. We examined factors associated with out-of-range exposure and described concentration changes after dose adjustment in hospitalised patients undergoing repeated TDM. Methods This retrospective, single-centre, longitudinal cohort study included hospitalised adults who received voriconazole and contributed at least three valid TDM episodes. Trough concentrations were classified as subtherapeutic (5.0 μg/mL). Bayesian mixed-effects logistic regression with weakly informative priors and patient-level random intercepts identified factors associated with each direction of out-of-range exposure. Adjacent TDM pairs with a documented dose change were analysed for return to range. Results A total of 140 patients contributed 497 episodes; 72 (14.5%) were subtherapeutic and 145 (29.2%) supratherapeutic. In the 445-episode complete-case models, subtherapeutic exposure was associated with rifampin-like inducer co-administration (OR 16.48, 95% CrI 2.38–139.44) and lower daily dose (OR 0.45 per 100 mg, 95% CrI 0.29–0.66), whereas supratherapeutic exposure was associated with the CYP2C19 poor-metaboliser phenotype (OR 5.24, 95% CrI 1.76–16.66), higher log-CRP (OR 1.69, 95% CrI 1.23–2.40), and higher total bilirubin (OR 1.09 per 10 μmol/L, 95% CrI 1.03–1.15). Concentrations returned to range after 7 of 12 (58.3%) clinically concordant dose increases, but after only 26 of 75 (34.7%) clinically concordant dose decreases. Conclusion In hospitalised patients undergoing repeated voriconazole TDM, the factors associated with subtherapeutic and supratherapeutic exposure differed by direction. After dose reduction, supratherapeutic concentrations frequently did not return to range; this descriptive pattern is compatible with several contributors, including resampling before a new steady state and voriconazole’s nonlinear metabolism. These findings may help clinicians interpret repeated TDM results in complex inpatient settings.