OBJECTIVE To compare the efficacy and toxicity of two doses of didanosine (ddI) in patients with symptomatic HIV disease who are intolerant of zidovudine (ZDV). METHODS The Alpha trial is a randomized double-blind multicentre trial of two doses of ddI. ddI was given as one buffered sachet twice daily in doses adjusted for weight: 750 mg per day for patients > or = 60 kg in the higher-dose group, and 200 per day for the lower-dose group. RESULTS Patients (n = 1775; 907 higher-dose, 868 lower-dose) from nine European countries and Australia were randomized and started trial treatment. Sixty per cent had AIDS, 65% had CD4 cell counts < 50 x 10(6)/l and 55% had received ZDV for more than 12 months. Follow-up was to death or 30 September 1992, and only 33 patients (20 higher-dose, 13 lower-dose) had been lost to follow-up for at least 3 months at that time. The longest follow-up was 28.5 months and the mean (SD) was 12.4 (6.9) months. There was no significant difference in survival between the groups: 67% of patients in each group died, the median survival being 13.0 months in the higher-dose and 12.5 in the lower-dose groups, a difference of about 0.5 months (95% confidence interval, -0.9 to 2.0; log-rank P = 0.7). There was also no significant difference in progression to AIDS or death, development of HIV encephalopathy or death, or development of new AIDS events or death. There were small (but statistically significant) differences in the changes in CD4 cell count and in p24 antigen levels between the groups, with greater increases in CD4 and greater decreases in p24 in the higher-dose group. There were also clear differences in adverse events: pancreatitis developed more frequently in the higher-dose group, 66 patients compared to nine patients in the lower-dose group, of whom 37 and six, respectively, were classified as definite cases. Nine cases (seven higher-dose, two lower-dose) were reported to have died because of or with pancreatitis. Peripheral neuropathy, abnormal liver function and dry mouth were also reported more often in the higher-dose group. CONCLUSIONS The Alpha trial is not able to provide direct evidence for the clinical efficacy of ddI. There was no significant difference between the two doses in mortality or disease progression. However, the higher dose was more toxic.