Simulating therapeutic drug monitoring results for dose individualisation to maintain investigator blinding in a randomised controlled trial

BACKGROUND: Therapeutic drug monitoring (TDM) is essential practice when dosing drugs with a narrow therapeutic index in order to achieve a plasma drug concentration within a narrow target range above the efficacy concentration but below the toxicity concentration. However, TDM with dose individualisation is challenging during a double-blind clinical trial with laboratory staff and investigators blinded to treatment arm allocation. METHODS: Drug concentrations were simulated for participants in the placebo arm by an unblinded independent statistician, utilising the measured values from the treatment arm participants. Simulated and actual concentrations were re-blinded and passed on to a dose-adjusting investigator, who made dose adjustment recommendations but was not directly responsible for clinical care of participants. RESULTS: A total of 257 sham lithium plasma concentrations were simulated utilising 242 true lithium plasma concentrations in real time as the trial progressed. The simulated values had a median (interquartile range) of 0.59 (0.46, 0.72) compared to 0.53 (0.39, 0.72) in the treatment arm. Blinding of the laboratory staff and dose-adjusting investigator was maintained successfully. CONCLUSIONS: We succeeded in simulating sham lithium plasma concentrations while maintaining blinding. Our simulated values have a smaller range than the observed data, which can be explained by the challenges with respect to drug adherence and dose timing that were experienced. TRIAL REGISTRATION: Pan African Clinical Trials Registry, PACTR201310000635418. Registered on 30 August 2013.