• dialysis;
  • end-stage renal disease;
  • hepatocellular carcinoma;
  • uremia


Background and Aim

The clinical aspects of patients with hepatocellular carcinoma (HCC) undergoing maintenance dialysis are largely unknown. We aimed to investigate the long-term survival and prognostic determinants of dialysis patients with HCC.


A total of 2502 HCC patients, including 30 dialysis patients and 90 age, sex, and treatment-matched controls were retrospectively analyzed.


Dialysis patients more often had dual viral hepatitis B and C, lower serum α-fetoprotein level, worse performance status, higher model for end-stage liver disease (MELD) score than non-dialysis patients and matched controls (P all < 0.05). There was no significant difference in long-term survival between dialysis and non-dialysis patients and matched controls (P = 0.684 and 0.373, respectively). In the Cox proportional hazards model, duration of dialysis < 40 months (hazard ratio [HR]: 6.67, P = 0.019) and ascites (HR: 5.275, P = 0.019) were independent predictors of poor prognosis for dialysis patients with HCC. Survival analysis disclosed that the Child-Turcotte-Pugh (CTP) provided a better prognostic ability than the MELD system. Among the four currently used staging systems, the Japan Integrated Scoring (JIS) system was a more accurate prognostic model for dialysis patients; a JIS score ≥ 2 significantly predicted a worse survival (P = 0.024).


Patients with HCC undergoing maintenance dialysis do not have a worse long-term survival. A longer duration of dialysis and absence of ascites formation are associated with a better outcome in dialysis patients. The CTP classification is a more feasible prognostic marker to indicate the severity of cirrhosis, and the JIS system may be a better staging model for outcome prediction.