A study of the causes of liver enlargement amongst black patients at King Edward VIII Hospital, Durban, South Africa has revealed that congestive cardiac failure (36.7%), amoebic liver abscess (7.1%), hepatocellular carcinoma (5.8%) and cirrhosis (5.4%) are the most common causes in this population. Liver biopsy was needed to determine the cause in 28.7% of patients studied. The diagnostic yield of percutaneous liver biopsy was increased by obtaining 2 or 3 consecutive specimens for histological examination by redirecting the biopsy needle through a single entry site. This benefit was achieved without an increase in morbidity or mortality. Fatalities and complications associated with liver biopsy were more frequent at this hospital than in hospitals in Europe, The United Kingdom and North America. The complication rates after percutaneous or peritoneoscopic biopsy were 2.0% and 2.3% respectively. A total of 6 deaths was recorded. The morbidity and mortality rates were not increased when more than one specimen was taken during percutaneous biopsy. In the majority of patients in whom biopsy was carried out, after-care was either non-existent or inadequate. The "Tru-Cut" needle was used for all percutaneous liver biopsies at King Edward VIII Hospital. Two techniques, including the method recommended by the manufacturer, have been found to be incorrect; the needle must be used correctly if an adequate biopsy specimen is to be obtained for histological examination and if serious complications are to be avoided. Hepatic tuberculosis was diagnosed in 9% of patients with unexplained hepatomegaly who were subjected to liver biopsy. This disease did not yield any consistent clinical findings. In addition, liver function tests were of little diagnostic value and results of hepatic imaging techniques were often normal. Accordingly, a high index of suspicion is needed and liver biopsy is essential in patients with unexplained hepatomegaly or hepatospienomegaly, or pyrexia of unknown origin since biopsy provides the only means of diagnosing hepatic tuberculosis. The accuracy of both ultrasonography and scintigraphy in distinguishing between normal and diseased livers was low (68% and 74% respectively). These techniques performed better at detecting focal than diffuse liver disease; the sensitivity of ultrasonography and scintigraphy in focal and diffuse disease were 88% and 92%, and 27% and 54% respectively. The specificity of both procedures was high for both types of liver disease (range 91-96%). Overlap between the ultrasonographic features of amoebic liver abscess, hepatocellular carcinoma and metastatic carcinoma resulted in a correct final diagnosis being made in only 81% of patients with amoebic liver abscess, 29% with hepatocellular carcinoma and 43% of patients with metastatic carcinoma who had an ultrasound scan. Neither technique was capable of determining the cause of diffuse liver disease. Therefore, when diffuse parenchymal liver disease is suspected, liver biopsy is needed to determine the presence and nature of the disease. In addition, liver biopsy or aspiration is usually required to determine the cause of focal disease in selected patients in whom space-occupying lesions are detected on hepatic imaging studies. / Thesis (M.D.)-University of Natal, Durban, 1990.