Comparison of Two Malnutrition Assessment Scales in Predicting Postoperative Complications in Elderly Patients Undergoing Noncardiac Surgery

Abstract

Background: The present study was designed to investigate the relationship between two malnutrition assessment scales, perioperative nutrition screen (PONS) and Nutritional Risk Screening 2002 (NRS2002), with postoperative complications in elderly patients after noncardiac surgery. Methods: This was a secondary analysis of a prospective cohort study. Elderly patients (65–90 years) undergoing noncardiac surgery were enrolled in Peking University First Hospital. Malnutrition was screened by PONS and NRS2002 at the day before surgery. Multivariable analysis was employed to analyze the relationship between PONS and NRS2002 and postoperative 30-day complications. Receiver operating characteristic (ROC) curve was generated to evaluate the predictive value of PONS and NRS2002 in predicting postoperative complications. Results: A total of 915 patients with mean age of 71.6 ± 5.2 years were consecutively enrolled from September 21, 2017, to April 10, 2019. The incidence of malnutrition was 27.3% (250/915) by PONS ≥ 1 and 53.6% (490/915) by NRS2002 ≥ 3. The overall incidence of complications within postoperative 30 days was 45.8% (419/915). After confounders were adjusted, malnutrition by PONS ≥ 1 (OR 2.308, 95% CI 1.676–3.178, P < 0.001), but not NRS2002 ≥ 3 (OR 1.313, 95% CI 0.973–1.771, P = 0.075), was related with an increased risk of postoperative complications. ROC curve analysis showed that the performances of PONS [area under the ROC curve (AUC) 0.595, 95% CI 0.558–0.633] showed very weak improvement in predicting postoperative complications than NRS2002 score (AUC 0.577, 95% CI 0.540–0.614). Conclusion: The present study found that malnutrition diagnosed by PONS was related with an increased risk of postoperative complications. The performances of PONS and NRS2002 were poor in predicting overall postoperative complications. Clinical Trial Registration: www.chictr.org.cn, identifier: ChiCTR-OOC-17012734.