The nutritional index 'CONUT' is useful for predicting long-term prognosis of patients with end-stage liver diseases.

Organ allocation in Japan remains difficult due to the shortage of deceased-donor livers. The screening tool for controlling nutritional status (CONUT) has been considered to be an established assessment model for evaluating nutritional aspects in surgical patients. However, the application of this CONUT for evaluating the prognosis of patients with end-stage liver diseases has not been evaluated. We assessed the predictability of the prognoses of 58 patients with end-stage liver disease using various prognostic models. The patients registered at the transplantation center of Tohoku University Hospital for the waiting list of Japan Organ Transplant Network for liver transplantation were retrospectively analyzed. The prognoses of the patients were evaluated using the following 5 models: CONUT, the model for ELD with incorporation of sodium (MELD-Na), Child-Turcotte-Pugh score (CTP), prognostic nutritional indices (Onodera: PNI-O), and the Japan Medical Urgency criteria of the liver (JMU). Cox's proportional hazard model, log-rank test and concordance(c)-static were used for the statistics. The indices were 17.74 ± 5.80 for MELD-Na, 9.21 ± 2.19 for CTP, 33.92 ± 11.16 for PNI-O, and 7.57 ± 3.09 for CONUT. Univariate analysis revealed the significance of CONUT (p = 0.017, Odds: 1.325) but not MELD-Na, CTP, JMU or PNI-O for prediction. The cumulative survival rate was clearly discriminated at CONUT point 7. The c-static was 0.081 for the 6-month (M) survival rate, 0.172 for 12M, 0.517 for 36M, 0.821 for 48M, and 0.938 for 60M for CONUT. In conclusion, CONUT shows best predictability for the distant prognoses of patients with ELD.