Association between nutritional status and mortality/neurological outcomes in stroke patients: A systematic review...

To systematically evaluate the association between Nutritional Risk Screening 2002 (NRS-2002), Controlling Nutritional Status (CONUT) score, Geriatric Nutritional Risk Index (GNRI), and 3-month mortality and poor neurological outcomes (modified Rankin Scale [mRS] ≥3) in stroke patients.

A systematic search of PubMed, Embase, and other databases up to April 2025 identified 11 observational studies (6 prospective cohorts; n = 7696). Study quality was assessed using the Newcastle-Ottawa Scale. Random-effects models were used to calculate pooled odds ratios (ORs) with 95 % confidence intervals (CIs). Heterogeneity, sensitivity, and publication bias were assessed. Meta-regression explored sources of heterogeneity. A gradient boosting classifier and Bayesian MCMC simulations were used for supplementary modeling.

NRS-2002 ≥ 3 (OR=3.42, 95 % CI: 2.59-4.51), CONUT ≥5 (OR=3.66, 95 % CI: 2.47-5.43), and GNRI <98 (OR=2.68, 95 % CI: 1.86-3.84) were significantly associated with poor functional outcomes. These indices also predicted higher 3-month mortality: NRS-2002 ≥ 3 (OR=4.13), CONUT ≥5 (OR=3.57), GNRI <98 (OR=2.93). Heterogeneity ranged from moderate to high (I²=42.1-68.9 %). Meta-regression implicated regional and clinical factors as sources of variability. Predictive modeling (AUROC = 0.81) identified GNRI <92, age ≥75, and NIHSS as key mortality predictors, consistent with SHAP and Bayesian analyses.

Malnutrition-particularly as defined by NRS-2002 ≥ 3, CONUT ≥5, and GNRI <98-is strongly linked to early mortality and poor recovery after stroke. GNRI showed high predictive value in older patients. Integrating nutritional screening into acute stroke care may enable early, cost-effective interventions to improve outcomes.