NOT ALL MELD SCORES ARE CREATED EQUAL: MELD DRIVEN BY CREATININE HAS LOWER INTENT TO TREAT SURVIVAL COMPARED TO MELD DRIVEN BY BILIRUBIN OR INR
For the same MELD score at listing, outcomes before and after liver transplantation may vary if the predominant driver of the MELD score is bilirubin, INR or creatinine. For example, a MELD score of 25 driven by creatinine may have different survival as compared to a MELD score of 25 driven by bilirubin. We hypothesized that MELD score subtypes do not have similar intent-to-treat survival.
We examined all adult patients registered for liver transplantation between 2016 – 2020 and excluded patients receiving MELD exceptions or those receiving dual organ transplantation. Using K-Means Clustering Analysis (method to assign primary driver) we classified each patient as either MELD-Br, MELD-INR or MELD-Cr depending on the dominant variable for a given MELD score (Fig A). The primary outcome of interest was intent-to-treat survival defined as survival within 1 year from listing with or without liver transplantation (accounts for waitlist and post-transplant outcomes). AKI and CKD were defined using ADQI and KDIGO criteria, respectively
Overall there were MELD-Br (n=22,286) MELD-INR (n=4,204) and MELD-Cr (n=13,697) registered. ITT Survival within 1 year was lowest for MELD-Cr (62%) as compared to MELD-Br (76%) or MELD-INR (79%). (Fig B) Among registrants MELD 26-35, this difference was even wider (MELD-Cr 49% as compared to MELD-Br 62% and MELD- INR 69%). By sex, ITT survival was 5-10% lower for each of the subtypes for females(Fig C). We explored causes of variation: MELD-Cr had the highest MELD at listing (MELD-Cr 24 vs MELD-Br 20 and MELD-INR MELD 14) but the largest decline in MELD within 3 months of listing (MELD-Cr -25% vs MELD-Br -15% vs MELD-INR +1%). On adjusted analysis, MELD-Cr was associated with higher WL mortality (HR 1.4, 95% CI 1.3-1.4) as well as lower LT rates ( HR 0.7, 95% CI 0.68-0.73). Finally, in the MELD-Cr subgroup, ITT survival was lowest for AKI only (55%) as compared to AKI/CKD (58%), CKD (60%) versus no AKI or CKD (69%).
Not all MELD scores are created equal. For equivalent listing practices, registrants with MELD scores driven by serum creatinine have lower ITT survival and this may be partly explained by changes in kidney function in waitlisted registrants. If continuous distribution is adopted for organ allocation, MELD subtype may serve as (1) a useful variable for assessment of risk of mortality and (2) allow for data driven assignment of additional priority especially among females.