THE LIVER FRAILTY INDEX (LFI) IMPROVES MORTALITY RISK PREDICTION OVER MELD 3.0 ALONE FOR WOMEN, OLDER ADULTS, AND OBESE PATIENTS WITH CIRRHOSIS

Background: The Model for End-Stage Liver Disease-Sodium (MELDNa) is an established metric of 3-month waitlist mortality (WLM) in cirrhosis patients but underestimates mortality risk in certain subgroups. The Liver Frailty Index (LFI), a metric of physical frailty not traditionally captured by MELDNa, enhances WLM prediction over MELDNa alone, particularly among women, older, and obese adults. Given the recent update of MELDNa to MELD 3.0, we evaluated the relationship between MELD 3.0 and LFI in these subgroups.
Methods: Data from ambulatory adults with cirrhosis awaiting LT without MELD exceptions between 2012-22 in the 10-center Functional Assessment in Liver Transplantation (FrAILT) study were analyzed. The primary outcome was WLM with deceased donor LT (DDLT) as the competing event. Fine-Gray subdistribution hazard model performance of MELD 3.0+LFI vs. MELD 3.0 was assessed using bootstrap cross-validated Wolbers’ concordance index, continuous net reclassification improvement (NRI), and integrated discrimination improvement (IDI) at 3, 6, and 12 months, and in subgroups of women, older (≥65y), and obese (BMI≥30 kg/m²) adults.
Results: Among 2,259 patients, median (IQR) LFI was 4.0 (3.5-4.5), MELD-Na was 19 (15-23), and MELD 3.0 was 18 (14-22). 42% were women, 18% were older, and 37% were obese. With a median (IQR) follow-up of 8.4 (3.2-20.5) months, WLM with DDLT competing risk was 3%, 5%, and 10% at 3, 6, and 12 months. Figure shows the predicted probabilities of survival among four individual cirrhosis patients with varying levels of MELD 3.0 and LFI. MELD 3.0+LFI also improved prediction among key subgroups, including women, older, and obese patients (Table).
Conclusion: In this multi-center ambulatory cohort of cirrhosis patients, the addition of LFI to MELD 3.0 improved WLM risk prediction compared to MELD 3.0 alone overall and in women, older, and obese adults. Our data support the use of LFI to identify cirrhosis patients who are particularly vulnerable to WLM and may benefit most from intensive prehabilitation and/or accelerated paths to LT.