Background
Voriconazole (VOR) levels are highly variable, with potential implications to both efficacy and safety. We hypothesized that VOR therapeutic drug monitoring (TDM) will decrease the incidence of treatment failures and adverse events (AEs).
Methods
We initiated a prospective, randomized, non-blinded multicenter study to compare clinical outcomes in adult patients randomized to standard dosing (clinician-driven) vs. TDM (doses adjusted based on levels). VOR trough levels were obtained on day 5, 14, 28, and 42 (or at completion of drug; ± 3 days). Real-time dose adjustments were made to maintain a range between 1–5 μg/mL on the TDM-arm, while levels were assessed retrospectively in the standard arm. Patient questionnaires were administered to assess subjective AEs.
Results
The study was discontinued prematurely, after 29 patients were enrolled. Seventeen (58.6%) patients experienced 38 AEs: visual changes (22/38, 57.9%), neurological symptoms (13/38, 34.2%), and liver abnormalities (3/38, 7.9%). VOR was discontinued in 7 (25%) patients because of an AE (4 standard-arm, 3 TDM-arm). VOR levels were frequently out of range in the standard-arm (8 tests >5 μg/mL; 9 tests < 1 μg/mL). Three dose changes occurred in the TDM-arm for VOR levels <1 μg/mL. Levels decreased over time in the standard-arm, with mean VOR levels lower at end of therapy compared to TDM (1.3 vs. 4.6 μg/mL, P = 0.008).
Conclusions
VOR TDM has become widespread clinical practice, based on known variability in drug levels, which impaired accrual in this study. Although comparative conclusions are limited, observations of variability and waning levels over time support TDM.