Background
Osteoporosis is a very common bone disease in older people. Malnutrition can lead to osteoporosis. The Geriatric Nutritional Risk Index (GNRI) is a tool used to assess nutritional status and is an important predictor of many diseases. Therefore, this study investigated the association between the GNRI and the presence of osteoporosis and assessed the value of this index for predicting osteoporosis in patients with fragility fractures.
Methods
A total of 1172 inpatients with fragility fractures from January 2017 to March 2023 were retrospectively analyzed. This included 806 cases in the osteoporosis group and 366 cases in the non-osteoporosis group. The patients' general and laboratory data were collected, along with their bone mineral density (BMD) measurements. GNRI was calculated based on ideal body weight and serum albumin levels. Correlation analysis was performed to determine the relationship between GNRI and BMD and bone metabolism markers. Descriptive analysis and logistic regression analysis were performed for osteoporosis characteristics and its risk factors. Receiver operating characteristic (ROC) curve was developed to predict the cut-off value.
Results
Univariate analysis showed significant differences between the osteoporosis group and the non-osteoporosis group in sex, age, height, weight, BMI, history of diabetes and gastrointestinal diseases, hemoglobin concentration, albumin concentration, prealbumin concentration, GNRI, blood glucose concentration, osteocalcin, β-isomerized C-terminal telopeptides (β-CTX), procollagen of type I N-propeptide (PINP), BMD and T-score. Spearman's correlation analysis showed that GNRI was positively associated with BMD and T-score at all bone sites (r = 0.272–0.397, P < 0.05). GNRI was negatively associated with procollagen of type I N-propeptide (r=-0.14, P = 0.025). Further logistic regression showed that sex, age, BMI, GNRI, albumin and diabetes were independent risk factors for osteoporosis. According to the results of the receiver operating characteristic curve, the predictive accuracy of osteoporosis was high, with an area under the curve (AUC) of 0.644, sensitivity of 52.4%, specificity of 71.3% and threshold value of 97.31.
Conclusion
Gender, age, BMI, GNRI, albumin and diabetes were independent risk factors. GNRI was positively correlated with BMD and inversely correlated with osteoporosis in patients with fragility fractures. In addition, the incidence of osteoporosis increased when GNRI was less than 97.31.