OBJECTIVE: To compare the clinical efficacy and tolerance of didanosine (ddl) monotherapy with low-dose zidovudine/didanosine (AZT/ddl) therapy among HIV-infected patients previously treated with AZT. METHODS: A randomized controlled trial was carried out of ddl 400 mg daily versus AZT/ddl 300/200 mg daily among patients with CD4 cell counts </=350 mm3 and prior AZT treatment for at least 16 weeks. Fifty eight patients received ddl monotherapy and 66 combined treatment. RESULTS: Patients were similar with respect to demographic, clinical and laboratory characteristics, and prior AZT treatment. Median duration of follow-up was 17.3 months. In the ddl group, 20 patients (34%) discontinued treatment because of toxicity, compared to 19 (29%) in the AZT/ddl group (p=0.38). There was no statistically significant difference in CD4 change between the two groups. In the ddl group, 16 patients (28%) developed a clinical endpoint (death or AIDS-defining opportunistic infection), compared to 33 (50%) in the combined therapy group (relative risk 1.8; 95% confidence interval 1.1-2.9; p=0.01). CONCLUSIONS: For fairly advanced AZT-pretreated patients, monotherapy with ddl was clinically and statistically superior to the low-dose AZT/ddl combination in preventing AIDS-defining illness and death. When access to drugs is limited, e.g. in under-resourced countries, combining available drugs and reducing dosage may be less effective than a single drug at the conventional dosage.