Validation of the perioperative nutrition screen for prediction of postoperative outcomes.

BACKGROUND: Preoperative nutrition risk is often underrecognized and undertreated. The perioperative nutrition screen (PONS) was recently introduced as an efficient tool to rapidly screen for preoperative nutrition risk. The relationship between identification of "nutrition risk" via PONS and adverse postoperative outcomes remains undescribed. METHODS: Preoperative nutrition risk was assessed via PONS from 01/01/2019 to 09/30/2020. Key clinical outcomes were compared with individual and composite PONS components. RESULTS: A total of 3151 patients with PONS evaluations were analyzed. Multivariate regression adjusted for key covariates demonstrated positive responses for specific PONS questions was associated with adverse clinical outcomes as follows. (1) Unplanned weight loss (>10% in 6-months preoperatively) associated with a 22.4% increased length of stay (LOS) (P < 0.0001) and increased 30-day readmission rate (odds ratio [OR], 2.44, 95% CI, 1.73-3.44, P < 0.001). (2) History of <50% of previous oral intake in past week associated with a 25% increased LOS (P < 0.001). (3) Preoperative serum albumin level <3.0 g/L associated with a 29.9% increased LOS (P < 0.001) and increased 30-day readmission rate (OR, 2.66, 95% CI, 1.63-4.35, P < 0.001). (4) Low body mass index was not associated with increased LOS by adjusted analysis although was predictive by univariate analysis. CONCLUSIONS: The PONS and its individual components appear to predict risk of adverse postoperative outcomes, even independent of a validated malnutrition diagnosis. Further studies are needed to assess the impact of specific preoperative nutrition interventions on adverse outcomes predicted by PONS when delivered to patients identified via PONS screen.

Duke Scholars

Author Solomon Aronson Anesthesiology, Cardiothoracic

Author Krista Lynn Haines Trauma, Acute, and Critical Care Surgery

Author Paul Edmund Wischmeyer Anesthesiology, Critical Care Medicine