Oral-contraceptive-associated liver tumours: occurrence of malignancy and difficulties in diagnosis

Abstract

Seven of ten women with oral-contraceptive-associated liver tumours were found to have hepatocellular carcinoma. The diagnosis was often delayed, although hepatomegaly was always present on examination, and liver-function tests and erythrocyte sedimentation-rates were abnormal in most cases. Other investigations, including routine technetium liver scans and biopsy, were sometimes misleading. There were important differences in alpha-fetoprotein concentration, vascularity on angiography, and survival between liver tumours in pill users and non-users.

PIP: A physician's personal series of 10 women treated from 1970-1979 for oral contraceptive-associated liver tumors is presented. Of the 10 women treated, 7 had hepatocellular carcinoma and 3 had benign adenomas. Symptomatology is described. Problems with diagnosis of liver dysfunctions included misleading biopsies and liver scans. The erythrocyte sedimentation rate was raised in all but 1 woman, and it was above 70 mm/h in 7. Changes in liver function tests were consistent with an intrahepatic tumor, with a striking increase in alkaline phosphatase in 9 (1170 IU/ml), and with only a slight rise in serum aspartate transaminase (mean 55 IU/ml). None of the patients had alpha fetoprotein levels above the upper limit of normal, and all patients were negative for hepatitis B surface antigen and antibody and anticore antibody. The carcinoma characteristics were similar in 7 patients (irregular trabecular arrangement with basophilic and dysplastic cells with nuclear pleomorphism and increased mitotic figures). When these oral contraceptive users were compared with 7 women diagnosed with hepatocellular tumors who had never used oral contraceptives, several striking differences were found. None of the poll users with carcinoma had raised alpha fetoproteins, whereas 4/7 nonpill users did. By arteriography, tumors in nonusers were much less vascular and less well defined. Survival rates also differed, with a 50% survival time of 1-8 years in nonusers compared with 4-8 years in pill users. The striking feature of this series is the delay in reaching a diagnosis in most of the 10 cases treated.