Twenty patients with chronic type B hepatitis were entered into a randomized, controlled study of adenine arabinoside monophosphate. Before entry, all patients were documented to have stable levels of hepatitis B surface antigen, hepatitis B e antigen, serum hepatitis B virus deoxyribonucleic acid, and deoxyribonucleic acid polymerase activity. Ten patients received adenine arabinoside monophosphate and 10 received no treatment. The two groups were well matched with respect to age, sex, known duration of hepatitis B surface antigen, presence of symptoms, serum aminotransferase levels, and hepatic histopathology. During the 4 wk of therapy, serum levels of hepatitis B virus fell dramatically. However, serum hepatitis B virus-deoxynbonucleic acid or deoxyribonucleic acid polymerase activity, or both, remained detectable, and levels of hepatitis B virus invariably rose once therapy was stopped. From 2 to 9 mo after therapy, 4 of the 10 treated patients became hepatitis B e antigen or hepatitis B virus-deoxyribonucleic acid and deoxyribonucleic acid polymerase negative, or both, and the results of routine serum biochemical tests improved. However, 2 of these 4 patients later relapsed. In the control group, 2 patients became seronegative for hepatitis B virus-deoxyribonucleic acid and deoxyribonucleic acid polymerase and manifested improvement in serum biochemical results by 18-24 mo after randomization. Thus, long-term improvements in clinical and serologic features of disease occurred in 20% of both treated and control patients. Side effects of adenine arabinoside monophosphate therapy were common, and 3 patients developed a severe and prolonged neuropathic pain syndrome. These results suggest that a 4-wk course of adenine arabinoside monophosphate therapy does not induce an increased rate of long-term remissions in chronic type B hepatitis.