First-24-hour aspirin exposure and 28-day mortality in critically ill adults with cerebral embolism: A retrospective MIMIC-IV cohort study.

ObjectiveTo evaluate the association between aspirin administration during the first 24 hours of intensive care unit (ICU) admission and 28-day all-cause mortality among critically ill adults with cerebral embolism and to explore associations with early cumulative dose.MethodsThis retrospective cohort study used MIMIC-IV version 3.1 and included each patient's first eligible adult ICU admission. Aspirin exposure was defined as at least one documented administration within 24 hours. After 1:1 propensity-score matching, 125 matched pairs were obtained. Associations were examined using Kaplan-Meier methods and Cox proportional hazards models, with exploratory dose, subgroup, and sensitivity analyses. Given only 45 deaths, a parsimonious Cox model was prioritized over the approximately 40-variable saturated model.ResultsThe cohort included 568 patients (430 aspirin-exposed and 138 unexposed), with 45 deaths by day 28. Aspirin exposure was associated with lower 28-day mortality before matching (HR 0.32, 95% CI 0.18-0.58; P<0.001) and after matching (HR 0.31, 95% CI 0.13-0.73; P=0.008). Matching improved balance, but residual imbalance remained; the largest absolute post-match standardized mean difference in Table 1 was 0.177. The parsimonious adjusted model produced a similar estimate (HR 0.31, 95% CI 0.17-0.57; P<0.001). Dose analyses were exploratory and hypothesis-generating; sparse dose-specific events, nonrandom treatment selection, and time-related exposure classification prevented identification of an optimal dose. Laboratory abnormalities were examined only as exploratory surrogates; clinical bleeding was not ascertained.ConclusionsIn this single-center retrospective cohort, aspirin exposure during the first 24 hours of ICU admission was associated with lower 28-day mortality. Residual confounding by unmeasured neurologic stroke severity (because NIHSS was unavailable), treatment eligibility, and goals of care, together with the absence of adjudicated bleeding outcomes, precludes causal, safety, or dose-prescriptive conclusions. Multicenter prospective studies and adequately powered randomized controlled trials are required before definitive clinical recommendations can be made.