A pilot study of the end point verification system in the asymptomatic carotid atherosclerosis study

Abstract

Ipsilateral transient ischemic attack (TIA) and cerebral infarction (CI) are primary end points of the Asymptomatic Carotid Atherosclerosis Study (ACAS), an ongoing multicenter, prospective, randomized trial designed to assess the effectiveness of carotid endarterectomy for patients with asymptomatic, hemodynamically significant stenosis of the internal or common carotid artery. The evanescent nature and absence of residual physical findings in TIAs pose special difficulties that contribute to interobserver disagreement. A standardized definition was established to be applied accurately and uniformly at the participating centers, including a questionnaire on six cardinal neurological symptoms, a computerized algorithm for characterization of events, and a verification system involving the independent evaluations of three reviewers. Modifications to the end point verification protocol after pilot testing in 115 patients with one or more symptoms enabled the algorithm to distinguish vascular from nonvascular events with a sensitivity of 83%, a specificity of 69%, positive predictive value of 71%, and negative predictive value of 82%. ACAS has developed a system for detecting and diagnosing end points. Using the algorithm as a guideline, consensus between two institutional physicians and a blinded independent reviewer must be made for an end point to be declared. If no consensus can be reached, a panel of two blinded external reviewers confers to arrive at an adjudication of the case. All potential end points are reviewed retrospectively by an end points committee for final adjudication.