Studies have indicated that maintenance therapy with AZT over extended periods may be limited by dose-dependent toxicity, primarily myelosuppression, and by the emergence of drug-resistant HIV strains. It is anticipated that the combination of AZT with either ddI or ddC may promote higher antiviral efficacy, with acceptable toxicity and less likelihood of development of drug-resistant strains, than AZT alone.

Approximately 1200 patients are randomized in a 2:1:1:2 ratio to one of the following four treatment arms: AZT plus ddI, AZT plus ddI placebo, AZT plus ddC placebo, and AZT plus ddC. Average follow-up is 2 years.

Inclusion Criteria

Required:

• Documented HIV infection OR working diagnosis of HIV OR evidence of idiopathic suppression with an AIDS-defining opportunistic infection or malignancy (except Kaposi's sarcoma).
• CD4+ cell count = or < 200/mm3 or = or < 15 percent of total lymphocyte count within previous 90 days OR history of AIDS-defining opportunistic infection.
• Current PCP prophylaxis.

Exclusion Criteria

Co-existing Condition:

Patients with the following symptoms and conditions are excluded:

• Symptoms of pancreatitis or peripheral neuropathy of grade 2 or worse.
• Requirement for acute therapy for any active AIDS-defining opportunistic infection or systemic chemotherapy for malignancy.
• Stage 2 or worse (moderate) AIDS Dementia Complex.
• Other disorders or conditions for which the study drugs are contraindicated or that may prevent adequate compliance with study therapy.

Concurrent Medication:

Excluded:

• Acute therapy for active AIDS-defining opportunistic infection.
• Systemic chemotherapy for malignancy.
• Antiretroviral therapy other than that provided by this study.

Patients with the following prior conditions are excluded:

• History of pancreatitis or peripheral neuropathy of grade 2 or worse.
• History of intolerance to the study drugs at entry doses and/or frequencies.
• History of phenylketonuria.