Age and the association between admission platelet count and in-hospital mortality in critically ill children

To investigate whether the prognostic weight of admission coagulation markers for in-hospital mortality varies with age in critically ill children. We retrospectively analyzed children admitted to the intensive care unit (ICU) in the Pediatric Intensive Care database, with in-hospital mortality as the primary outcome. Inclusion required at least one valid coagulation measurement from 6 h before to 24 h after ICU admission. Logistic regression examined interactions between age and six coagulation markers—platelet count, prothrombin time (PT), activated partial thromboplastin time (APTT), international normalized ratio (INR), fibrinogen, and D-dimer—adjusting for sex, ICU, admission lactate, vasoactive prescriptions, and diagnostic category. Sensitivity analyses assessed nonlinear relationships, age grouping, additional laboratory adjustment, exclusions, missing data, and selection bias. A total of 11,508 children were included, with 693 in-hospital deaths (6.0%). The primary platelet model included 11,330 children and 657 deaths. Only platelet count met the Bonferroni threshold for an adjusted age interaction (odds ratio [OR] 0.957, 95% confidence interval [CI] 0.928–0.986 per 1-SD higher platelet count per additional year; P = 0.004). With a nonlinear platelet main effect and a linear platelet×age interaction, the interaction attenuated to OR 0.969 (95% CI 0.942–0.997; nominal P = 0.032). The unadjusted INR interaction was significant (OR 0.977, 95% CI 0.964–0.990; P < 0.001), but the adjusted interaction was not (OR 0.999, 95% CI 0.987–1.011; P = 0.877). Age-stratified platelet associations were generally stronger in older groups but were not monotonic. Overall interaction tests were not significant with spline models for platelet count, age, or both ( P = 0.174, 0.065, and 0.629), or with three broad age groups ( P = 0.088). Evidence for age-related variation in the platelet–mortality association was sensitive to model specification. The finding remains exploratory and requires external validation before considering age-specific platelet weighting in clinical scores.