NRS2002 combined with nutritional, immune and inflammatory indicators for the nomogram to predict Sarcopenia

Abstract

Abstract Objectives Sarcopenia is a geriatric syndrome characterized by age-related loss of muscle mass and strength, with or without physical function decline. In clinical work, it is complicated to consider it as a geriatric syndrome, and the diagnostic criteria are often ignored by clinical workers. This study aims to construct a predictive model for sarcopenia using commonly used clinical indicators. Design: By collecting the basic clinical data, NRS2002 score scale, nutrition, immunity, inflammation, and other blood indicators of the subjects, the diagnosis and prediction model of sarcopenia was established. The LASSO regression method was used to screen the variables and select predictors. logistic regression analysis was used to construct the modal map, and the discriminant ability of the model was determined by calculating the area under the curve (AUC). Finally, the training set and validation set were randomly split for internal verification, and the AUC was used to judge the verification effect. Participants: The study was conducted from June 2023 to September 2022 in the First Affiliated Hospital of Kunming Medical University; Elderly inpatients over 60 years old were included, and sarcopenia was diagnosed using the Asian Working Group for Sarcopenia (AWGS2019) diagnostic criteria. NRS2002 score, nutrition, immunity, and inflammation indexes were collected to construct the model. Results Four variables were selected and screened by the LASSO regression method, and a diagnostic and prediction model was established based on these variables. The AUC of the prediction model was 0.80. In the internal validation, the total number of samples was randomly divided into training set and validation set according to a 0.85 split ratio, and the ROC curve was used to verify the results, and the AUC was 0.8047 and 0.9065 respectively. Finally, the model was used to correct the curve, and the curve fit was good, the mean absolute error (MAE) was 0.014, and the prediction effect was good. The model can be used to diagnose and predict sarcopenia in clinical patients. Conclusion In this study, NRS2002 combined with BMI, lymphocyte count, and BNP were used to construct a diagnosis and prediction model for sarcopenia, which has important value for the prediction of sarcopenia.