Geriatric Nutrition Risk Index as a predictor of cardiovascular and all-cause mortality in older Americans with diabetes

Abstract

Abstract Background and aims Few studies have examined the relationship between malnutrition, as defined by the Geriatric Nutrition Risk Index (GNRI), and all-cause mortality and cardiovascular mortality events, particularly in persons with diabetes. The study aimed at the association between GNRI and all-cause mortality and cardiovascular mortality in older Americans with diabetes. Methods Data from this retrospective study were obtained from the National Health and Nutrition Examination (NHANES) 1999–2016. Using data from The NHANES Public-Use Linked Mortality Files to assess all-cause mortality (ACM) and cardiovascular mortality (CVM). After excluding participants younger than 60 years and without diabetes, and with missing follow-up data, 4400 cases were left in this study. Persons with diabetes were divided by GNRI into 3 groups: GNRI ≥ 98; 92 ≤ GNRI < 98; and GNRI < 92; (No; Low; Moderate/Severe (M/S) group). We used Cox proportional hazard regression model to explore the predictive role of GNRI on ACM and CVM in elderly persons with diabetes. Restricted cubic splines to investigate the existence of a dose–response linear relationship between them. Result During a median follow-up period of 89 months, a total of 538 (12.23%) cardiovascular deaths occurred and 1890 (42.95%) all-cause deaths occurred. Multifactorial COX regression analysis showed all-cause mortality (hazard ratio [HR]: 2.58, 95% CI: 1.672–3.994, p < 0.001) and cardiovascular mortality (HR: 2.29, 95% CI: 1.063–4.936, p = 0.034) associated with M/S group risk of malnutrition in GNRI compared to no group. A negative association between GNRI and all-cause mortality was observed across gender and ethnicity. However, the same negative association between GNRI and cardiovascular mortality was observed only for males (HR:0.94, 95% CI:0.905–0.974, p < 0.001) and other races (HR:0.92, 95% CI:0.861–0.976, p = 0.007). And there was no significant correlation between low malnutrition and cardiovascular mortality (p = 0.076). Restricted cubic splines showed a nonlinear relationship between GNRI and all-cause mortality and cardiovascular mortality (non-linear p < 0.001, non-linear p = 0.019). Conclusions Lower GNRI levels are associated with mortality in older patients with diabetes. GNRI may be a predictor of all-cause mortality and cardiovascular mortality risk in older patients with diabetes.