The impact of geriatric nutritional risk index on one-year outcomes in hospitalized elderly patients with heart failure

BACKGROUND: Strategies that accurately predict outcomes in elderly heart failure (HF) patients have not been sufficiently established. In previous reports, nutritional status, ability to perform activities of daily living (ADL), and lower limb muscle strength are known prognostic factors associated with cardiac rehabilitation (CR). In the present study, we investigated which CR factors can accurately predict one-year outcomes in elderly patients with HF among the above factors. METHODS: Hospitalized patients with HF over 65 years of age from January 2016 to January 2022 were retrospectively enrolled in the Yamaguchi Prefectural Grand Medical (YPGM) Center. They were consequently recruited to this single-center retrospective cohort study. Nutritional status, ADL, and lower limb muscle strength were assessed by geriatric nutritional risk index (GNRI), Barthel index (BI), and short physical performance battery (SPPB) at discharge, respectively. One year after discharge, the primary and secondary outcomes were evaluated by all-cause death or HF readmission and major adverse cardiac and cerebrovascular events (MACCE), respectively. RESULTS: Overall, 1,078 HF patients were admitted to YPGM Center. Of those, 839 (median age 84.0, 52% female) met the study criteria. During the follow-up of 228.0 days, 72 patients reached all-cause death (8%), 215 experienced HF readmission (23%), and 267 reached MACCE (30%: 25 HF death, six cardiac death, and 13 strokes). A multivariate Cox proportional hazard regression analysis revealed that the GNRI predicted the primary outcome (Hazard ratio [HR]: 0.957; 95% confidence interval [CI]: 0.934–0.980; p < 0.001) and the secondary outcome (HR: 0.963; 95%CI: 0.940–0.986; p = 0.002). Furthermore, a multiple logistic regression model using the GNRI most accurately predicted the primary and secondary outcomes compared to those with the SPPB or BI models. CONCLUSION: A nutrition status model using GNRI provided a better predictive value than ADL ability or lower limb muscle strength. It should be recognized that HF patients with a low GNRI at discharge may have a poor prognosis at one year.