Association between baseline platelet count and 30-day in-hospital mortality in cardiac surgery patients admitted to intensive care units: a retrospective cohort study

To examine the association between baseline platelet count and 30-day in-hospital mortality in cardiac surgery patients requiring ICU care. We conducted a retrospective multicenter cohort study using the eICU Collaborative Research Database (2014–2015). Adult cardiac surgery patients admitted to the ICU were included. Baseline platelet count was defined as the first platelet measurement obtained within 24 h of ICU admission. The primary outcome was 30-day in-hospital mortality. Associations were assessed using multivariable logistic regression, generalized additive models, and piecewise logistic regression. A total of 8203 patients were included, of whom 304 (3.7%) died during hospitalization within 30 days of ICU admission. In the fully adjusted model, baseline platelet count was not linearly associated with 30-day in-hospital mortality (OR per 10 × 109/L increase, 0.995; 95% CI 0.978–1.012; P = 0.584). However, nonlinear modeling suggested an inflection point at approximately 84 × 109/L. Below this value, higher platelet count was associated with lower mortality (OR, 0.595; 95% CI 0.515–0.689; P < 0.001); above this value, no significant association was observed. Similar nonlinear patterns were observed across selected subgroups. In this multicenter retrospective cohort of cardiac surgery patients requiring ICU care, baseline platelet count showed a nonlinear association with 30-day in-hospital mortality. Lower platelet counts, particularly in the range of severe thrombocytopenia, were associated with higher mortality. These findings may help identify patients at higher short-term in-hospital risk; however, the estimated threshold should be interpreted cautiously, and no causal or therapeutic inference can be made.