Estimated treatment effect of ticagrelor versus aspirin by investigator-assessed events compared with judgement by an independent event adjudication committee in the SOCRATES trial.

Abstract

Adjudication of endpoints is a standard procedure in cardiovascular clinical trials. However, several studies indicate that the benefit of adjudication in estimating treatment effect may be limited. This post hoc analysis of SOCRATES (NCT01994720) compared the treatment effects and investigated the agreement of clinical event assessment by site investigators and independent adjudicators. SOCRATES compared ticagrelor and aspirin in 13,199 patients with acute minor stroke or high-risk transient ischemic attack. The primary endpoint was stroke, myocardial infarction, or death. Stroke was the major component of the primary endpoint and a secondary endpoint. The endpoints were adjudicated by a blinded independent committee. We compared the treatment effect on the primary endpoint and stroke alone based on the investigators' and adjudicators' assessments, and investigated the agreement rate on the stroke endpoint and major hemorrhages. The hazard ratios (95% confidence interval) for ticagrelor versus aspirin therapy for the primary endpoint were 0.89 (0.78-1.01) when calculated on adjudicator-assessed events and 0.88 (0.78-1.00) for investigator-assessed events. The hazard ratios (95% confidence intervals) for stroke were 0.86 (0.75-0.99) based on the adjudicators' diagnoses and 0.85 (0.75-0.97) based on the investigators' diagnoses. The overall agreement between adjudicator- and investigator-diagnosed stroke was 91%, and for major hemorrhages was 88%. In SOCRATES, there was no clinically meaningful difference in the estimated treatment effect, on either the primary endpoint or stroke, by using investigator- or adjudicator-assessed events. Double-blind treatment outcome studies with stroke endpoints may not benefit from adjudication. ClinicalTrials.gov Identifier: NCT01994720.