In order to obtain insight in attitudes towards detection and management of hepatic metastases of colorectal origin, a questionnaire was sent to hospitals in 13 Western countries. Response rate was 98.0% (n = 284). In almost all hospitals (98%) some method of follow-up was employed. Carcino-embryonic antigen (CEA) determinations were performed in 84% of all hospitals: most frequently in Germany and the U.S.A., but only in 50% of the British hospitals. Hepatic resection for liver metastases was performed in 95% of all hospitals. Resectability criteria varied considerably among the countries. In the majority of German and American hospitals multiple hepatic metastases were considered resectable (including bilobar disease in 58% of German hospitals). In the majority of British and Dutch hospitals only solitary metastases were considered resectable, or liver resections were not performed at all. The mean reported number of liver resections annually per hospital, reflecting these attitudes, was 11.2 and 7.2 for German and American hospitals, and 2.1 and 1.8 for British and Dutch hospitals respectively. When irresectable hepatic metastases were diagnosed, some form of chemotherapy was applied in 74% of hospitals. Hepatic artery infusion of chemotherapeutics was performed most frequently. The mean reported number of medically treated patients annually per hospital was 34 for Germany, 18 for the U.S.A., and 12 and 9 for Great Britain and the Netherlands respectively. Adjuvant chemotherapy was performed after liver resection in 30% of all hospitals, most frequently in German and American hospitals. Considerable disparity was observed in attitudes towards detection and in management of hepatic metastases among Western countries. On the basis of the reported 1421 liver resections and 3590 medically treated patients (annually) it is concluded that selection of the best detection and treatment policies is obviously hampered by insufficient clinical data and inconclusive evidence of purported optimal approach. To determine the optimal policy useful information can only be provided by inclusion of patients in prospective randomized trials.