Integrating skeletal muscle index and body roundness index for predicting functional outcomes in acute stroke patients: a prospective observational study.

BACKGROUND

After stroke, many patients experience dysphagia, anorexia, and metabolic stress, which may lead to malnutrition and accelerated loss of skeletal muscle mass. Sarcopenia and body fat distribution abnormalities significantly impact functional outcomes in acute stroke patients. While the skeletal muscle index (SMI) and body roundness index (BRI) have been studied individually, their combined predictive value for poor prognosis remains unclear. This study evaluates the integration of SMI and BRI for predicting unfavorable functional outcomes in acute stroke patients.

METHODS

A single-center, prospective cohort study was conducted on 123 acute ischemic stroke patients admitted within 3 days of onset. In this acute cohort, standardized strength/performance testing at admission was not feasible due to hemiparesis. Therefore, low skeletal muscle mass (L3-SMI) was used as the primary exposure. SMI was measured at the L3 vertebra using MRI, with sex-specific thresholds informed by EWGSOP2/AWGS muscle-quantity criteria. BRI was calculated based on waist circumference and height. Functional outcomes were assessed at 90 days post-discharge using the modified Rankin scale (mRS). Multivariate logistic regression and receiver operating characteristic (ROC) analyses were used to evaluate the independent and combined predictive abilities of SMI and BRI.

RESULTS

Patients with sarcopenia had significantly lower SMI (33.722 ± 3.307 cm2/m2) compared to non-sarcopenia patients (47.484 ± 5.934 cm2/m2, p < 0.001). Univariate analysis showed that lower SMI (OR = 0.90, 95% CI: 0.84-0.95, p < 0.001) and higher BRI (OR = 1.86, 95% CI: 1.21-2.85, p = 0.005) were associated with poor outcomes. Multivariate regression confirmed that sarcopenia (OR = 33.470, 95% CI: 7.118-157.394, p < 0.001) and BRI (OR = 2.200, 95% CI: 1.212-3.992, p = 0.010) independently predicted unfavorable outcomes. Combining SMI and BRI achieved an AUC of 0.933, demonstrating superior predictive performance compared to individual metrics. Decision curve analysis further highlighted the clinical utility of the combined model.

CONCLUSION

The integration of SMI and BRI suggests a promising, hypothesis-generating model for identifying patients at risk of unfavorable outcomes after acute stroke, which may support early recognition and individualized management. These exploratory findings require further validation in larger, multicenter studies to confirm their robustness and generalizability.