Abstract
This retrospective cohort study examined the association between admission blood urea nitrogen (BUN) levels and mortality in critically ill acute ischemic stroke (AIS) patients from the MIMIC-IV database. Patients were stratified into quartiles based on admission BUN (Q1: ≤ 14 mg/dL; Q2:14 < BUN ≤ 19 mg/dL; Q3:19 < BUN ≤ 27 mg/dL; Q4:>27 mg/dL). Multivariable Cox regression, adjusted for demographics, comorbidities, and labs, showed elevated BUN was independently associated with higher 30-day (HR = 1.017, 95% CI: 1.010-1.025), 90-day (HR = 1.014, 1.007-1.021), and 180-day mortality (HR = 1.011, 1.004-1.018). The critically ill AIS patients in the top BUN quartile (> 27 mg/dL) had a nearly 70% higher 30-day, 45% higher 90-day, and 36% higher 180-day mortality risk than those in the bottom quartile, even after full adjustment. Restricted cubic spline analysis demonstrated a linear dose-response relationship between BUN and mortality. Subgroup analysis showed a significant interaction between BUN and CKD: elevated BUN predicted mortality in non-CKD patients but had limited prognostic value in CKD patients. Notably, BUN elevation correlated more strongly with mortality in non-CKD patients. These findings suggest BUN may be an independent predictor useful for risk stratification in critically ill AIS populations, emphasizing its prognostic utility in non-CKD individuals. Further prospective studies are warranted to validate clinical applications and explore underlying mechanisms linking BUN to adverse outcomes.