Chronic heart failure is a syndrome occurring at the end-stage of multiple cardiovascular diseases. In the condition, nutritional and metabolic problems such as loss of appetite, diarrhea, abdominal distension, and constipation are highly prevalent, which in turn affect the prognosis of heart failure. The relationship of nutritional assessment results with prognosis in chronic heart failure has been studied extensively, while nutritional assessment for older adults with chronic heart failure has been rarely studied, and there is no clinically recognized assessment method.

To perform a comparative analysis of four nutritional assessment methods in terms of clinical prognosis prediction in elderly patients with chronic heart failure.

Eligible older inpatients with chronic heart failure (n=199) were recruited from Department of Cardiology, ICU, and Department of Geriatrics, Linyi People's Hospital from June 2018 to June 2020. Data were collected via reviewing the medical records and telephone-based follow-ups, including sex, age, height, weight, serum albumin (ALB) level, BMI, Geriatric Nutritional Risk Index (GNRI) , and result of Nutrition Risk Screening 2002 (NRS2002) , as well as prognosis〔containing three classifications: in-hospital deaths (n=43) and in-hospital survivors (n=156) ; one-year deaths (n=51) and one-year survivors (n=148) ; readmission within half a year (n=69) and readmission after half a year (n=130) 〕. Multivariate Logistic regression analysis was used to explore the prognostic factors of chronic heart failure. The analysis of ROC curve with AUC value was carried out to comparatively estimate prognosis predictive values of the nutritional assessment methods.

There were significant differences in mean age, serum ALB, GNRI and NRS2002 score between in-hospital deaths and survivors (P<0.05) . The mean age, height, serum ALB, GNRI and NRS2002 score were also significantly different between one-year deaths and survivors (P<0.05) . Those with readmission within half a year had significantly different mean BMI, serum ALB, GNRI and NRS2002 score compared with those with readmission after half a year (P<0.05) . For predicting in-hospital death, the AUC of serum ALB was 0.76〔95%CI (0.68, 0.84) , P<0.001〕, and that of NRS2002 score was 0.80〔95%CI (0.73, 0.86) , P<0.001〕. In predicting one-year death, the AUC of serum ALB was 0.75〔95%CI (0.67, 0.82) , P<0.001〕, and that of NRS2002 score was 0.82〔95%CI (0.76, 0.88) , P<0.001〕. The AUC of NRS2002 score in predicting readmission within half a year was 0.73〔95%CI (0.65, 0.80) , P<0.001〕.

On the whole, NRS2002 score could be the first choice for prognostic assessment in elderly patients with chronic heart failure, for it was more effective in predicting the risks of in-hospital death, one-year death and readmission within half a year than serum ALB level, GNRI and BMI.