Abstracts
2002 Digestive Disease Week

# 100572 Abstract ID: 100572 Colorectal Adenocarcinoma in Cirrhotic Patients
Pascal Gervaz, Rattaplee Pak-Art, Santhat Nivatvongs, Bruce Wolff, Dirk Larson, Stacy Klingel, rochester, MN

Background: It has been hypothesized that colorectal adenocarcinoma spread to the liver is infrequent in cirrhotic patients. This was confirmed by meta-analysis of post-mortem series and by studies of Asian population with active chronic hepatitis B. However, the clinical outcome of colorectal carcinoma in western cirrhotic patients remains to be investigated. Hypothesis: The Child classification of liver cirrhosis provides better prognostic discrimination than the TNM classification to determine outcome of cirrhotic patients with colorectal adenocarcinoma. Methods: Retrospective analysis of patients operated for colorectal adenocarcinoma at the Mayo Clinic Rochester, Minnesota between 1975 and 2000, with confirmed liver cirrhosis at the time of abdominal exploration. Results: Seventy-two patients were available for analysis. The median age was 70 (range 42-89) years. Nine patients (12.5%) died in the postoperative period. An elevated level of direct bilirubin, as well as prolonged prothrombin time were significantly associated with postoperative death (Wilcoxon-rank test, p=0.01 and 0.009 respectively). Six patients (8%) developed liver metastases during a median follow-up of 28 months. Most cirrhotic patients were alcoholic (49%) and had compensated liver disease (Child A class = 43%, Child B = 42%, Child C = 15%). For the whole group 1-, 3- and 5-years survival rates were 69%, 35%, and 14% respectively. Causes of death were liver failure (39%), cancer (24%) and others (37%). In univariate analysis, the presence of ascites correlated with poor outcome (p<0.001). Child A patients had a significantly better survival than Child B and C patients (log rank test, p=0.008), while the TNM classification provided no additional prognostic information (log-rank test, p=0.518). Parameters that correlated with decreased survival in multivariate analysis were albumin serum level < 2.5 mmol/l (HR = .575, 95% Confidence Interval .417-.793, p<0.001) and prolonged prothrombin time (HR=1.09, 95% CI 1.037-1.153, p <0.001). Conclusion: Liver metastases from colorectal adenocarcinoma occur in less than 10% of cirrhotic patients. The clinical outcome of cirrhotic patients operated for colorectal carcinoma is primarily correlated with stage of the liver disease. Child classification provides an adequate means to identify patients at risk for immediate postoperative death and poor long-term survival. Transjugular intrahepatic portosystemic shunt (TIPS) should be considered before colorectal resection in Child B and C patients.