Clinical and Dynamic Evaluation of the C-Reactive Protein–to–Albumin Ratio in Acute Ischemic Stroke: Interaction with Neurological Severity and Limited Prognostic Utility

Background The C-reactive protein–to–albumin ratio has been proposed as a composite inflammatory biomarker associated with adverse outcomes in critically ill patients. However, whether the C-reactive protein–to–albumin ratio provides independent or incremental prognostic information beyond established clinical severity markers in acute ischemic stroke remains uncertain. Methods This retrospective observational cohort study included 146 consecutive adult patients admitted with acute ischemic stroke and managed in an intensive care setting. The C-reactive protein–to–albumin ratio was calculated at emergency department admission and during intensive care follow-up. Multivariable logistic regression models were constructed to evaluate associations between the C-reactive protein–to–albumin ratio, sepsis at admission, and ninety-day mortality, adjusting for age, National Institutes of Health Stroke Scale score, sepsis status, and admission glucose. Incremental prognostic value was assessed using calibration analysis, net reclassification improvement, integrated discrimination improvement, interaction modelling, National Institutes of Health Stroke Scale–stratified analyses, and penalized regression. Results Sepsis was present in 25.3% of patients, and ninety-day mortality occurred in 21.9%. Admission C-reactive protein–to–albumin ratio was not independently associated with sepsis after adjustment (odds ratio 1.023, 95% confidence interval 0.994–1.053; p = 0.117), whereas age and National Institutes of Health Stroke Scale score remained significant predictors. Similarly, the C-reactive protein–to–albumin ratio measured during intensive care was associated with mortality in univariate analysis but lost significance in multivariable models (odds ratio 1.026, 95% confidence interval 0.982–1.070; p = 0.246). Sepsis status and age consistently emerged as the strongest independent predictors of mortality. No significant interaction was observed between the C-reactive protein–to–albumin ratio and National Institutes of Health Stroke Scale score (p = 0.802), and the C-reactive protein–to–albumin ratio was not associated with mortality within National Institutes of Health Stroke Scale–stratified subgroups. Penalized regression did not retain the C-reactive protein–to–albumin ratio as a stable predictor. Addition of the C-reactive protein–to–albumin ratio did not improve model calibration or risk reclassification (net reclassification improvement − 0.057; integrated discrimination improvement 0.011). Conclusions In acute ischemic stroke, the C-reactive protein–to–albumin ratio reflects systemic inflammatory burden but does not provide independent or incremental prognostic value beyond established clinical severity markers. Dynamic assessment of the C-reactive protein–to–albumin ratio does not enhance mortality prediction. The C-reactive protein–to–albumin ratio may therefore be best interpreted as a complementary, rule-in biomarker rather than a primary risk stratification tool.