Association of nonstandardized model for end-stage liver disease score exceptions with waitlist mortality in adult liver transplant candidates.

In the US liver allocation system, nonstandardized model for end-stage liver disease (MELD) exceptions (NSEs) increase the waitlist priority of candidates whose MELD scores are felt to underestimate their true medical urgency. We determined whether NSEs accurately depict pretransplant mortality risk by performing mixed-effects Cox proportional hazards models and estimating concordance indices. We also studied the change in frequency of NSEs after the National Liver Review Board's implementation in May 2019. Between June 2016 and April 2022, 60,322 adult candidates were listed, of whom 10,280 (17.0%) received an NSE at least once. The mean allocation MELD was 23.9, an increase of 12.0 points from the mean laboratory MELD of 11.9 (P < .001). A 1-point increase in allocation MELD score due to an NSE was associated with, on average, a 2% reduction in hazard of pretransplant death (cause-specific hazard ratio: 0.98; 95% CI: 0.96, 1.00; P = .02) compared with those with the same laboratory MELD. Laboratory MELD was more accurate than allocation MELD with NSEs in rank-ordering candidates (c-index: 0.889 vs 0.857). The proportion of candidates with NSEs decreased significantly after the National Liver Review Board from 21.5% to 12.8% (P < .001). NSEs substantially increase the waitlist priority of candidates with objectively low medical urgency.