Validation of heart failure events in ALLHAT participants assigned to doxazosin

Abstract

The Antihypertensive and Lipid Lowering Treatment to Prevent Heart Attack Trial (ALLHAT) is a randomized, double-blind, active-controlled trial designed to compare the incidence of coronary heart disease (CHD) between high risk hypertensive participants randomized to a diuretic and those randomized to one of three alternative antihypertensive drugs: alpha blocker, ACE inhibitor, and calcium channel blocker. Following independent data reviews, the Director of the National Heart, Lung and Blood Institute (NHLBI) accepted a recommendation to discontinue the alpha blocker (doxazosin) arm of the antihypertensive trial. This recommendation was based on the very low probability of doxazosin showing benefit over chlorthalidone in the primary endpoint at the end of the trial. In addition, the doxazosin arm showed significantly increased reports of the secondary endpoint, major cardiovascular disease, with a doubling of risk of congestive heart failure (CHF) (RR=2.04, P<.001). Because the diagnosis of heart failure is based on the presence of subjective, non-specific signs and symptoms, the Data Safety and Monitoring Board proposed validation of the reported cases of hospitalized and fatal CHF. Accordingly, a more complete evaluation of event reports and documentation was undertaken. The purpose of this study is to describe the methods used to validate the CHF event reports in the doxazosin and chlorthalidone arms. The Endpoints Subcommittee reviewed a random sample of 50 CHF cases to determine the proportion meeting ALLHAT criteria. Of the 39 cases with sufficient data, 33 (85%) met the standard. The coordinating center staff then reviewed half of all reported hospitalized or fatal CHF cases for comparison of ejection fraction data. Of those with ejection fractions given, two thirds had ejection fractions of 40% or lower. The CHF cases in the doxazosin and chlorthalidone treatment groups were compared for baseline data, pre and post-event pharmacologic treatment, and case fatality rates. The extensive verification process of the reported CHF cases in the doxazosin and chlorthalidone arms supported the finding of a significantly higher rate of CHF in the doxazosin arm compared to the chlorthalidone arm.