Platelet trajectory patterns and 28-day mortality in ICU stroke patients: a retrospective cohort study based on the MIMIC-IV database

Le résumé fourni par la source

Abstract Background Stroke remains a leading cause of death and disability worldwide. Platelet abnormalities are common in critically ill patients, but dynamic trajectory patterns and their prognostic value in intensive care unit (ICU) stroke patients remain unclear. Methods This retrospective cohort study included 1,562 ICU stroke patients from the Medical Information Mart for Intensive Care (MIMIC-IV) database (version 3.1). Latent class trajectory models (LCTM) were used to identify platelet trajectory patterns during the first 5 days of ICU admission. Cox proportional hazards regression was performed to assess the association between trajectory groups and 28-day mortality. Exploratory effect decomposition analysis was conducted to explore the relative contribution of baseline platelet levels and subsequent platelet decline to prognosis. Results Four distinct platelet trajectory groups were identified: High-Declining (14.7%), Moderate-Low Stable (52.4%), Low-Rising (4.5%), and Moderate-High Stable (28.4%). The High-Declining group had the highest 28-day mortality (34.1% vs. 20.4% in the Moderate-Low Stable group, P < 0.001). After multivariable adjustment, the High-Declining group was associated with a 56% increase in mortality risk compared with the Moderate-Low Stable group (hazard ratio [HR] = 1.56, 95% confidence interval [CI]: 1.17–2.07, P = 0.002). Exploratory effect decomposition analysis revealed a suppression pattern: the direct protective effect of higher Day 2 platelet was offset by an indirect harmful effect mediated through greater subsequent platelet decline (Average Component Mediation Effect [ACME] = 0.017 on the probability-difference scale, 95% CI: 0.010–0.026, P < 0.001), suggesting that rapid platelet decline may reflect a high-risk phenotypic pattern associated with this trajectory group. Conclusions Platelet trajectory patterns are associated with 28-day mortality and may serve as prognostic indicators in ICU stroke patients. A high baseline platelet level with rapid decline represents the highest-risk phenotype and, if confirmed in future prospective studies, may help identify patients requiring closer clinical evaluation to screen for underlying deteriorating conditions.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.