BACKGROUND: The model for end-stage liver disease (MELD) is used for organ allocation in liver transplantation. The maximal serum creatinine (Cr) level for MELD is set at 4.0 mg/dL; however, there was no outcome data to justify this strategy. METHODS: Ninety-two patients with cirrhosis with Cr level >4 mg/dL were selected from 1438 patients and compared with MELD score-matched controls for three-month and six-month mortality. RESULTS: At three months, patients with Cr level >4 mg/dL had a significantly higher mortality rate than the 184 controls with a lower Cr level (44.6% vs. 29.3%, p = 0.015). This trend was still significant at six months: the mortality rate was 62% in the index group vs. 45.1% in the control group (p = 0.011). The difference between the index and control groups was the smallest (2.5% at three months and 3.4% at six months) when Cr was up-scaled to 5.5 mg/dL. The predictive accuracy of the MELD was estimated by using area under receiver-operating characteristic (AUC) curve. Only the cutoff of 5.5 mg/dL at six months displayed a higher AUC (0.753). CONCLUSIONS: A cutoff at 5.5 mg/dL may be more appropriate for the MELD. The MELD for patients with cirrhosis with advanced renal insufficiency deserves re-evaluation.
BACKGROUND: The model for end-stage liver disease (MELD) is used for organ allocation in liver transplantation. The maximal serum creatinine (Cr) level for MELD is set at 4.0 mg/dL; however, there was no outcome data to justify this strategy. METHODS: Ninety-two patients with cirrhosis with Cr level >4 mg/dL were selected from 1438 patients and compared with MELD score-matched controls for three-month and six-month mortality. RESULTS: At three months, patients with Cr level >4 mg/dL had a significantly higher mortality rate than the 184 controls with a lower Cr level (44.6% vs. 29.3%, p = 0.015). This trend was still significant at six months: the mortality rate was 62% in the index group vs. 45.1% in the control group (p = 0.011). The difference between the index and control groups was the smallest (2.5% at three months and 3.4% at six months) when Cr was up-scaled to 5.5 mg/dL. The predictive accuracy of the MELD was estimated by using area under receiver-operating characteristic (AUC) curve. Only the cutoff of 5.5 mg/dL at six months displayed a higher AUC (0.753). CONCLUSIONS: A cutoff at 5.5 mg/dL may be more appropriate for the MELD. The MELD for patients with cirrhosis with advanced renal insufficiency deserves re-evaluation.