Blood Urea Nitrogen‐to‐Albumin Ratio and 90‐Day Functional Outcomes After Acute Ischemic Stroke: An Ambispective Multicenter Cohort Study With Exposure–Response Analysis
Mingquan Li, Meng Cao, Ning Gu, Xingyu Wu, Yanli Zhang, Yanhang Yin, Wen Xie, Fuling Yan, Ling Liu, Liqin Luan, Pingmin WeiBackground
The blood urea nitrogen to serum albumin ratio (BAR) is an emerging prognostic biomarker, but its value for predicting functional outcomes after acute ischemic stroke (AIS) is unestablished. We investigated whether admission BAR is associated with poor 90‐day functional outcomes (modified Rankin Scale score 3–6) in middle‐aged and older AIS patients.
Methods
This multicenter ambispective cohort study (2017–2024) enrolled AIS patients from three tertiary hospitals in Nanjing. After exclusions, 4390 patients were analyzed. Continuous variables are presented as mean ± standard deviation or median (interquartile range). Propensity score matching (1:1) substantially improved covariate balance, as residual imbalance remained for NIHSS, CKD and BMI (Table 3). The primary outcome was 90‐day functional status. Modified Poisson regression models with robust error variances were used in both the overall cohort and the propensity score–matched analysis.
Results
Among 4390 patients (mean age 63.8 ± 9.9 years; 70.1% male), 1579 (35.97%) had poor outcomes. The mean BAR was 4.32 ± 2.45. High BAR (≥ 5.12) was linked to a higher incidence of poor outcomes vs. low BAR (52.0% vs. 32.0%). In the matched cohort ( n = 1702), each unit increase in BAR was not significantly associated with poor outcome in the fully adjusted model (adjusted RR = 1.01, 95% CI: 1.00–1.03, p = 0.125), whereas high BAR (≥ 5.12) was associated with a 23% increased risk of poor outcome in the fully adjusted model (adjusted RR = 1.23, 95% CI: 1.11–1.36, p < 0.001). The absolute risk difference between high and low BAR groups was 8.9% (95% CI: 4.7%–13.2%), corresponding to a number needed to evaluate (NNE; i.e., the number of high‐BAR patients that would need to be identified to observe one additional poor outcome) of approximately 11. A nonlinear dose–response relationship was observed in the non‐CKD subgroup ( p < 0.001), whereas the association appeared approximately linear in the overall cohort ( p = 0.12). Mediation analysis showed a nonsignificant total effect with nonsignificant indirect pathways.
Conclusion
The overall cohort analysis was nonsignificant, whereas a significant association was observed in the matched subsample. This discrepancy suggests that BAR’s prognostic value may be subgroup‐specific and hypothesis‐generating. External validation in independent cohorts is needed before BAR can be considered for clinical risk stratification in AIS.