Methodologic issues in terminating enrollment of a subgroup of patients in a multicenter randomized trial

Abstract

The National Emphysema Treatment Trial (NETT) was a multicenter randomized controlled trial comparing medical treatment plus lung-volume-reduction surgery (LVRS) to medical treatment alone for the treatment of severe emphysema. The primary outcomes specified for the trial were mortality from all causes and change in functional status as indicated by the change in maximum exercise capacity measured two years after randomization. A secondary objective of the trial was to define criteria to identify subgroups of patients at risk of harm or benefit from LVRS. Stopping guidelines for safety and efficacy based on 30-day mortality and a combination of overall mortality and functional status at two years were specified at the inception of the trial. Although specific subgroups of patients likely to benefit were not identified in advance, several clinical factors were specified as likely to be important in defining subgroups with differential outcome. In May 2001, with 40% of expected deaths accrued, the Data and Safety Monitoring Board determined that a subgroup of patients was at significantly higher risk of 30-day mortality from LVRS without counterbalancing evidence of functional benefit, and recommended that the protocol be modified to exclude further randomization of such patients. The trial's sponsor, the National Heart, Lung and Blood Institute, accepted the recommendation, which was rapidly communicated to participating clinics. This paper describes the operational aspects of identification of the subgroup and implementation of the recommendation to continue the trial, but to terminate enrollment of new patients in the subgroup. These aspects include notification of the investigators, the institutional review boards, the Research Group, the patients and the medical community. We also describe the repercussions of the publication and the misinterpretations of the results based on media coverage.