Maria Cristina Mele

Background: Hospital malnutrition affects nearly 30% of patients in medical wards and correlates with worse outcomes. An early assessment is necessary to stratify the risk of short-term outcomes and mortality. The predictive role of COntrolling NUTritional status (CONUT) score in this context has not yet been elucidated in Western countries. We aimed to test CONUT at admission as a predictive score of hospital outcomes, in an Internal Medicine and Gastroenterology Department of an Italian Tertiary Care University hospital. Methods: We prospectively enrolled patients admitted to our center, stratifying them into the four CONUT classes (normal = 0–1; mild = 2–4; moderate = 5–8; severe = 9–12 points) according to serum albumin (g/dL), total lymphocyte count (/mm3), and total cholesterol (mg/dL); the primary outcome measure was length of stay (LOS) and the secondary one was in-hospital mortality. Results: Out of a total of 203 patients enrolled, 44 (21.7%) patients had a normal status (0–1), 66 (32.5%) had a mild impairment (2–4), 68 (33.5%) had a moderate impairment (5–8), and 25 (12.3%) a severe impairment (9–12). The mean LOS was 8.24 ± 5.75 days; nine patients died. A moderate-severe CONUT correlated with a higher LOS at the univariate [HR 1.86 (95% CI 13.9–3.47); p < 0.0001] and multivariate analysis [HR 1.52 (95% CI 1.10–2.09); p = 0.01]. The CONUT score was also a predictor of mortality, with an AUC of 0.831 (95% CI 0.680–0.982) and with an optimal cut-off at 8.5 points. Nutritional supplementation within 48 h from admission correlated with lower mortality [OR 0.12 (95% CI 0.02–0.56) p = 0.006]. Conclusions: CONUT is a reliable and simple predictor of LOS and in-hospital mortality in medical wards.