Nonlinear relationship between platelet count and 30-day in-hospital mortality in ICU stroke patients: a multicenter retrospective cohort study

Background Evidence of the relationship between platelet count and 30-day in-hospital mortality in ICU stroke patients is still scarce. Therefore, the purpose of this study was to explore the relationship between platelet count and 30-day in-hospital mortality among ICU stroke patients. Methods We conducted a multicenter retrospective cohort study using data from 8029 ICU stroke patients in the US eICU-CRD v2.0 database from 2014 to 2015. Utilizing binary logistic regression, smooth curve fitting, and subgroup analyses, we examined the link between platelet count and 30-day in-hospital mortality. Results The 30-day in-hospital mortality prevalence was 14.02%, and the mean platelet count of 223×10 9 /L. Adjusting for covariates, our findings revealed an inverse association between platelet count and 30-day in-hospital mortality (OR = 0.975, 95% CI: 0.966, 0.984). Subgroup analyses supported the robustness of these results. Moreover, a nonlinear relationship was observed between platelet count and 30-day in-hospital mortality, with the inflection point at 163×10 9 /L. On the left side of the inflection point, the effect size (OR) was 0.92 (0.89, 0.95), while on the right side, the relationship was not statistically significant. Conclusion This study establishes an independent negative association between platelet count and 30-day in-hospital mortality in ICU stroke patients. Furthermore, a nonlinear relationship with a saturation effect was identified, suggesting that maintaining the platelet count around 163×10 9 /L can reduce 30-day in-hospital mortality in these patients.