A Phase II Randomized, Double-Blind, Placebo-Controlled Trial of the Virologic Effect of Two Different Nucleoside Treatment Strategies (Zidovudine Versus Zidovudine in Combination With Didanosine) for HIV Infection in Subjects With CD4+ Counts >= 550 Cells/mm3

Inclusion Criteria Concurrent Medication: Allowed: Antibiotics for bacterial infections as clinically indicated. Recombinant erythropoietin (EPO) and G-CSF as clinically indicated for grade 3 or worse anemia and neutropenia, respectively. Antipyretics. Analgesics. Allergy medications. Oral contraceptives. Nonprescription medications such as vitamins or herbal therapies. Concurrent Treatment: Allowed: Radiation therapy to local lesion only. Acupuncture. Patients must have: HIV seropositivity. CD4 count >= 550 cells/mm3. No ARC or AIDS conditions by CDC criteria. Consent of parent or guardian if less than 18 years of age. Exclusion Criteria Co-existing Condition: Patients with the following symptoms or conditions are excluded: Presence of factors predisposing to pancreatitis such as active alcoholism. Other medical conditions that would interfere with study compliance. Concurrent Medication: Excluded: Other antiretrovirals or systemic immunomodulators. Systemic corticosteroids. Systemic cytotoxic chemotherapy. Intravenous pentamidine. Concurrent Treatment: Excluded: Radiation therapy except to local lesion. Patients with the following prior conditions are excluded: History of chronic diarrhea, defined as more than four loose or watery stools on average daily for the past month. History of grade 2 or worse peripheral neuropathy. History of pancreatitis. Bacterial infection requiring antibiotics within 14 days prior to study entry. Prior Medication: Excluded: Prior HIV therapy with antiretrovirals or systemic immunomodulators. Prior Treatment: Excluded within 2 weeks prior to study entry: Transfusion. Active substance abuse or alcoholism.