Model for end-stage liver disease-sodium underestimates 90-day mortality risk in patients with acute-on-chronic liver failure

Abstract

Background & aims: It is unclear whether the model for end-stage liver disease-sodium (MELD-Na) score captures the clinical severity of acute-on-chronic liver failure (ACLF). We compared observed 90-day mortality in patients with ACLF with expected mortality based on the calculated MELD-Na and examined the consequences of underestimating clinical severity.

Methods: We identified patients with ACLF during hospitalization for cirrhosis in 127 VA hospitals between 01/01/2004 and 12/31/2014. We examined MELD-Na scores by ACLF presence and grade. We used actual and observed 90-day mortality to estimate a standardized mortality ratio (SMR) by ACLF presence and grade. We used transplant center-specific median MELD-Na at transplantation (MMaT) to estimate the proportion likely to receive priority for liver transplantation (LT) based on MELD-Na alone.

Results: Of 71,894 patients hospitalized for decompensated cirrhosis, 18,979 (26.4%) patients met the criteria for ACLF on admission. The median (P25-P75) MELD-Na on admission was 26 (22-30) for ACLF compared to 15 (12-20) for patients without ACLF; it was 24 (21-27), 27 (23-31), and 32 (26-37) for ACLF-1, 2 and 3, respectively. At 90 days, 40.0% of patients with ACLF died (30.8%, 41.6% and 68.8% with ACLF-1, 2 and 3, respectively) compared to 21.3% of patients without ACLF. Compared to the expected death rate based on MELD-Na, mortality risk was higher for patients with ACLF, SMR (95% CI): 1.52 (1.48-1.52), 1.46 (1.41-1.51), 1.50 (1.44-1.55), 1.66 (1.58-1.74) for overall ACLF, ACLF-1, -2 and -3, respectively. Only 9.1% of patients with ACLF reached the national median MELD-Na of 35 and between 17.3% to 35.1% exceeded the MMaT at any center. During index admission, 589 (0.8%) patients with ACLF were considered for LT evaluation and 16 (0.1%) were listed for LT.

Conclusions: In a US cohort of hospitalized patients with decompensated cirrhosis, MELD-Na did not capture 90-day mortality risk in patients with ACLF. Patients with ACLF are at a disadvantage in the current MELD-Na-based system.

Lay summary: Acute-on-chronic liver failure (ACLF) is a condition marked by multiple organ failures in patients with cirrhosis and is associated with a high risk of death. Liver transplantation may be the only curative treatment for these patients. A score called model for end-stage liver disease-sodium (MELD-Na) helps guide donor liver allocation for transplantation in the United States. The higher the MELD-Na score in a patient, the more likely that a patient receives a liver transplant. Our study data showed that MELD-Na score underestimates the risk of dying at 90 days in patients with ACLF. Thus, physicians need to start liver transplant evaluation early instead of waiting for a high MELD-Na number.

Keywords: Cirrhosis; Natural history; Outcomes; Prognosis; Transplant center.

Fig. 1.. Expected vs. observed 90-day mortality by model for end-stage liver disease-sodium score in patients with and without ACLF of several grades.

Expected mortality was based on Kim et al. ACLF, acute-on-chronic liver failure.