CLINICAL TRIALS Considering Recurrent Events in Clinical Trials Statistical Analysis

摘要

Many chronic diseases are characterized by non-fatal recurrent events. Examples of such include asthma attacks in asthma, epileptic seizures in epilepsy, and hospitalizations for worsening condition in heart failure. Typically, in clinical trials for heart failure, composite outcomes (eg, heart failure hospitalization or cardiovascular death) are adopted as the primary endpoint and then analyzed as a time-to-first-event analysis using the Cox proportional-hazards model. These composite outcomes combine fatal and non-fatal events, thereby providing more comprehensive information about the impact of the treatments compared. Combining multiple endpoints into one composite outcome also increases the event rate and avoids the multiplicity issues surrounding the analysis of multiple endpoints. However, composite outcomes that only consider the first event are suboptimal for a chronic disease such as heart failure, which is characterized by recurrent heart failure hospitalizations, since repeat events within individuals are ignored in analyzes. Recurrent hospitalizations are an indication of worsening condition, so analyzing all these repeat events within individuals is more representative of disease progression. Furthermore, recurrent hospitalizations are distressing for patients and thus outcomes that consider all these events more accurately estimate the effect of treatment on the true burden of disease. If we consider the CHARM-Preserved trial, we can examine the impact of analyzing only the time-to-first event, ignoring repeat hospitalizations.