Stopping the Randomized Aldactone Evaluation Study Early for Efficacy

The Randomized Aldactone Evaluation Study (RALES) was a randomized double-blind placebo-controlled trial designed to test the hypothesis that addition of daily spironolactone to standard therapy would reduce the risk of all-cause mortality in patients with severe heart failure as a result of systolic left ventricular dysfunction. The Data Safety Monitoring Board (DSMB) for RALES reviewed data on safety and efficacy throughout the trial using pre-specified statistical stopping boundaries for efficacy. To ensure that the data were complete, the DSMB requested successive “mortality sweeps.” At the time of these sweeps, all RALES investigators determined the vital status of participants at their clinics. Therefore, the data that the DSMB saw included a much higher percentage of the deaths than would have been observed without these sweeps. At the DSMB’s fifth meeting, the data showed 351 deaths in the placebo group and 269 in the spironolactone group for an estimated hazard ratio of 0.78 (p = 0.00018). The board recommended early termination of the trial because the observed Z-value of 3.75 exceeded the pre-specified critical value of 2.79 and the data on mortality showed consistency among subgroups and across time. The sweeps had identified 31 deaths that likely would not have been reported by the time of the meeting. Subsequent data collection identified an additional 46 deaths that had occurred by the time the study ended. Even when the endpoint of a randomized clinical trial is mortality, routine methods of data collection and reporting are unlikely to identify all events in a timely manner. The experience from RALES provides an example of the importance of active follow-up of patients to ensure that a DSMB is observing a high proportion of the events that have actually occurred.