SUMMARY An analysis of 294 patients who died with cirrhosis showed that 24% had developed hepatocellular carcinoma. Haemochromatosis and HBsAg positive chronic active hepatitis were high risk groups (36% and 42% respectively) and the frequency was lowest in primary biliary cirrhosis and HBsAg negative chronic active hepatitis (3 % and 11 % respectively). Those with hepatocellular carcinoma showed a striking male preponderance (11:1) and further analysis has shown that the proportion developing this tumour in each group was closely related to the proportion of males in that group (r=0-97). Age was the only other significant factor, malignant change occurring more commonly in those over the age of 50 years than those below (30 % and 7 % respectively, P < 0.005). The influence of HBsAg was largely accounted for by the known predisposition of males to carry HBsAg. The group of patients who had developed this tumour without cirrhosis were younger (mean age 39 years) and had a lower male to female ratio of 1-1:1 and the place of contraceptive-related tumour within this group is discussed.It is well recognised that hepatocellular carcinoma may develop in longstanding cirrhosis (Stewart, 1931;Berman, 1951;Gall, 1960), though the reported frequency varies both from country to country (Doll et al., 1966) and with the underlying aetiology (Powell et al., 1971). In parts of Africa and the Far East where the incidence is high, environmental factors, such as aflatoxin and the hepatitis B virus, may be important (Vogel et al., 1970;Alpert et al., 1971). Indeed, in these populations up to 80% of patients have HBsAg present in their serum and the most widely held view is that the tumour has developed from a cirrhosis related to the hepatitis B infection (Tong et al., 1971 Davis et al., 1975;Christopherson and Mays, 1976), and males ingesting androgenic sterReceived for publication 8 June 1978 oids (Johnson et al., 1972;Farrell et al., 1975), both of which suggest the additional influence ofhormonal factors.In this paper we analyse the influence of age, sex, HBsAg status, and aetiology of underlying cirrhosis on the frequency of hepatocellular carcinoma development in 294 British patients dying of cirrhosis and the relative importance of these factors in 25 other patients in whom no evidence of an underlying cirrhosis was found.
MethodsThe case notes of all British patients seen at King's College Hospital who were recorded as dying with hepatocellular carcinoma, or as a result of cirrhosis, between 1967 and 1975 were analysed. These comprised 294 patients with cirrhosis, in 71 (24%) of whom hepatocellular carcinoma had been detected either during life or at necropsy, and 25 who had developed carcinoma in an otherwise normal liver. In all cases the diagnosis of cirrhosis and the presence of hepatocellular carcinoma were confirmed by histological examination of liver biopsy tissue. Other types of primary liver tumour, such as cholangiocarcinoma and heamangiosarcoma, were ex-