Pooled analysis of adverse event collection from 4 acute coronary syndrome trials

BackgroundAdverse event collection in randomized clinical trials establishes drug safety. Although costly and regulated, it is rarely studied.MethodsAdverse event data from 4 clinical trials (APPRAISE-2, PLATO, TRACER, TRILOGY ACS) comprising 48,118 participants with acute coronary syndromes were pooled to compare patterns and determinants of reporting. Events were classified as serious (SAE) or nonserious (AE) from hospital discharge to 1 year; study end points were excluded.ResultsIn total, 84,901 events were reported. Of those, 12,266 (14.4%) were SAEs and 72,635 (85.6%) were AEs. Of all participants, 7,823 (16.3%) had SAEs, 18,124 (37.7%) had only AEs, and 22,171 (46.1%) had neither. Nonserious adverse events were distributed across system organ classes: general disorders (11%), infection (10%), gastrointestinal (10%), respiratory (9%), cardiovascular (8.4%), and other (35%). Serious adverse events had a higher proportion of cardiovascular causes (14.0%). Event reporting was highest after hospital discharge, decreasing rapidly during the following 3 months. In a Cox proportional hazards model, chronic obstructive pulmonary disease (hazard ratio 1.58, 95% CI 1.44-1.74), heart failure (1.55, 1.40-1.70), older age, and female sex were independent predictors of more SAEs, whereas enrollment in Eastern Europe (0.63, 0.58-0.69) or Asia (0.84, 0.75-0.94) were independent predictors of fewer SAEs.ConclusionsHalf of all participants reported adverse events in the year after acute coronary syndrome; most were AEs and occurred within 3 months. The high volume of events, as well as the variation in SAE reporting by characteristics and enrollment region, indicates that efforts to refine event collection in large trials are warranted.