Background and Purpose—Adverse events (AE) in trial populations present a major burden to researchers and patients, yet most events are unrelated to investigational treatment. We aimed to develop a coherent list of expected AEs, whose incidence can be predicted by patient characteristics that will inform future trials and perhaps general poststroke care.Methods—We analyzed raw AE data from patients participating in acute ischemic stroke trials. We identified events that occurred with a lower 99% confidence bound greater than nil. Among these, we applied receiver operating characteristic principles to select the fewest types of events that together represented the greatest number of reports. Using ordinal logistic regression, we modeled the incidence of these events as a function of patient age, sex, baseline National Institutes of Health Stroke Scale, and multimorbidity status, defining P