As an important modifiable factor that can be intervened, nutrition is closely related to the occurrence of frailty. Early identification of frailty through nutrition evaluation and reversal of its occurrence is of great significance for improving clinical outcomes. There are few available studies on the predictive value of nutrition-related parameters for frailty among older patients in the emergency department (ED) .

To evaluate the relationship between commonly used nutrition-related parameters and frailty among older adults in the ED.

Two hundred and ten people aged≥65 years were recruited from the Department of Emergency Medicine, China Rehabilitation Research Center (Beijing Bo'Ai Hospital) from January to October 2021. The demographic data were recorded. Fasting venous blood sample was collected within 24 hours after admission to measure routine indicators. The nutritional risk was assessed by Nutrition Risk Screening 2002 (NRS2002) . The basic activities of daily living were evaluated by Barthel Index (BI) . The Clinical Frailty Scale (CFS) was used to assess frailty, and individuals with CFS levels 1-4 (n=68) and those with CFS levels 5-9 (n=142) were assigned to non-frail group and frail group, respectively. Multivariable Logistic regression was used to analyze the factors associated with frailty in older patients in the ED. Hosmer-Lemeshow test and receiver operating characteristic (ROC) curve were used to evaluate the predictive validity of nutrition-related parameters for frailty in older patients in the ED. Nonparametric DeLong test was used to compare the area under the ROC curve (AUC) of each parameter.

There were statistically significant differences between frail and non-frail patients in mean age, sex ratio, mean body mass index (BMI) , prevalence of coronary heart disease, mean levels of hemoglobin (HGB) , albumin (ALB) , prealbumin (PA) , high-sensitivity C-reactive protein (hs-CRP) and 25-hydroxyvitamin D〔25 (OH) D〕, and mean score of NRS2002, as well as mean BI and length of hospital stay (P<0.05) . Multivariable Logistic regression analysis showed that higher PA〔OR=0.943, 95%CI (0.891, 0.998) , P=0.041〕 and 25 (OH) D〔OR=0.909, 95%CI (0.844, 0.979) , P=0.012〕were protective factors of frailty in older patients in the ED. The risk of frailty decreased by 44.6% for every 100 mg/L increase in PA, and decreased by 61.7% for every 10 μg/L increase in 25 (OH) D. However, higher NRS2002 score〔OR=1.701, 95%CI (1.353, 2.138) , P<0.001〕was the risk factor of frailty in older patients in the ED, and the risk of frailty increased by 70.1% for every one score increase in NRS2002 score. Hosmer-Lemeshow test showed PA (χ²=6.120, P=0.634) , 25 (OH) D (χ²=5.386, P=0.716) and NRS2002 score (χ²=4.758, P=0.446) had good goodness of fit. ROC analysis demonstrated showed that the optimal cutoff values of PA, 25 (OH) D and NRS2002 score for predicting frailty in older patients in the ED were 211.9 mg/L, 7.06 μg/L and 3 points, respectively, and the AUCs of them were 0.749, 0.670 and 0.835, respectively. Nonparametric DeLong test showed that the AUC of NRS2002 score was greater than that of PA (Z=2.241, P=0.025) and 25 (OH) D (Z=3.400, P<0.001) .

As frail patients have poor nutritional status, nutritional assessment contributes to early identification of frailty. Among the nutrition-related parameters, PA, 25 (OH) D and NRS2002 score can effectively predict frailty in older patients in the ED, and NRS2002 score may have the strongest predictive ability.