Analysis of adverse events in a dose titration study

The authors have examined the analysis of adverse event data from an efficacy dose escalation trial. Unlike the analysis of efficacy data, the assumption that when a patient experiences an adverse event at a given dose, he or she will experience the same at a greater dosage level was not applicable in the analysis of adverse event data. Because the time effect is confounded with the dose effect in a dose escalation design, any assessment of a dose-effect relationship from such a scheme is found to be preliminary and suspect. For drugs that need to be dosed with a titration schedule, a time-dose-specific incidence of an adverse event provides more useful information than a dose-specific incidence. The pace of dose titration, which was found to be important in the manifestation of an adverse event, also needs to be specified. These aspects are illustrated with data from a specially designed trial. The entire study contained a placebo arm and three arms of an active drug randomized in a parallel comparative fashion. Within each of the three active drug arms, a forced titration scheme was used to raise the dose to different levels, which distinguished the three arms. With an efficacy dose titration design, the dose-response relationship for adverse events cannot be determined without incorporating a placebo arm and other arms with different maximum allowable doses. For drugs that need to be administered with a titration scheme, incidence of adverse events needs to be presented with the dosage, the time, and the pace of titration.