Abstract
Background
Endothelial activation and microvascular dysfunction are implicated in organ injury in critical illness, yet the prognostic value of the Endothelial Activation and Stress Index (EASIX) may vary across heterogeneous intensive care unit (ICU) populations. We assessed whether admission serum chloride, reflecting early physiologic status at ICU entry, modifies associations between EASIX and clinical outcomes.Materials and methods
We retrospectively analyzed 1,196 consecutive adult medical ICU admissions. EASIX was calculated at ICU admission and analyzed as log2(EASIX). Primary endpoints were in-hospital mortality and ICU/hospital length of stay (LOS). The secondary/exploratory endpoint was receipt of renal replacement therapy (RRT). We used multivariable regression with prespecified covariates and tested effect modification by chloride, using interaction terms and spline-based modeling.Results
After multivariable adjustment, log2(EASIX) was not independently associated with in-hospital mortality or LOS. For RRT, admission chloride significantly modified the association between log2(EASIX) and RRT. The EASIX-RRT association was most consistent within the normochloremic range and attenuated at chloride extremes; continuous spline interaction models supported non-uniform effects across the chloride spectrum.Conclusion
Among consecutively admitted adult medical ICU patients, EASIX was not independently associated with mortality or LOS after adjustment for illness severity. We observed heterogeneity in the EASIX-RRT association across admission chloride strata; however, this finding should be interpreted cautiously. These findings are hypothesis-generating and warrant external validation.