A new prognostic model to predict dropout from the waiting list in cirrhotic candidates for liver transplantation with MELD score <18.

BACKGROUND & AIMS: The model for end-stage liver disease (MELD) is used for organ allocation in liver transplantation (LT), but its prognostic performance is less accurate in patients with low score. We assess the outcome of patients with MELD < 18 awaiting LT, finding prognostic variables to identify a high dropout risk.
METHODS: Training set consisted of 277 patients and validation cohort of 292 patients. Competing risk regression analysis, taking into account LT, was used for univariate/multivariate analysis.
RESULTS: Ascites, sodium, bilirubin, albumin and glomerular filtration rate were independently associated with a 12-month dropout risk in the training set. Combining these five prognostic parameters, we calculated a new score named liver-renal-risk (LIRER). In the validation set, the 12-month LIRER concordance index showed a discrimination power [0.798, 95% confidence interval (95% CI) 0.793-0.803] better than MELD (0.582, 95% CI 0.575-0.588), Child-Turcotte-Pugh (0.687, 95% CI 0.681-0.693), MELD-sodium (0.721, 95% CI 0.715-0.727) and MELD-ascites-sodium (0.729, 95% CI 0.724-0.735), with a remarkable calibration (Hosmer-Lemeshow test: P = 0.91; R(2) = 0.911). Considering all study patients, the risk of wait list dropout increased with the rise in LIRER. The survival benefit analysis comparing the wait list dropout risk with the mortality of the 216 transplanted patients with same LIRER showed an important benefit for LT in patients with LIRER > 15.9.
CONCLUSIONS: In patients with low MELD (<18), combination of ascites, sodium, albumin, bilirubin and renal function in a new score (LIRER) discriminates patients at high risk of medium-term adverse outcome from those in whom LT may be safely deferred.