To determine the relationship of viral susceptibility to zidovudine (AZT) and baseline viral load (as determined by plasma viremia and quantitative endpoint dilution). To determine the relationship between viral load and susceptibility during different antiretroviral therapy strategies. To correlate measures of viral load and short term clinical and laboratory markers (such as weight, CD4 count, p24 antigenemia, and beta2 microglobulin) on the different therapy arms.

High-grade resistance to AZT has been detected in HIV isolates from approximately 25 percent of individuals with AIDS who received AZT for at least 1 year. To elucidate the clinical significance of in vitro AZT resistance, it is necessary to distinguish between clinical failure caused by AZT resistance and clinical decompensation caused by other factors.

High-grade resistance to AZT has been detected in HIV isolates from approximately 25 percent of individuals with AIDS who received AZT for at least 1 year. To elucidate the clinical significance of in vitro AZT resistance, it is necessary to distinguish between clinical failure caused by AZT resistance and clinical decompensation caused by other factors.

One hundred-twenty patients who have been receiving AZT for at least 1 year are randomized to 1) continue with AZT, 2) switch to treatment with didanosine at 1 of 2 doses, or 3) receive both AZT and ddI. Treatment is given for 16 weeks, with a possible extension to 32 weeks. Patients are followed at weeks 2, 4, 8, 12, and 16. For analysis purposes only, patients are stratified according to degree of susceptibility of HIV isolates to AZT.

• Drug: Zidovudine
• Drug: Didanosine

Inclusion Criteria

Concurrent Medication:

Allowed:

• Chemoprophylaxis against Pneumocystis carinii pneumonia (PCP), Mycobacterium tuberculosis, or Herpes simplex virus, or against other opportunistic infections as indicated.
• Corticosteroids for no longer than 21 days (only as part of PCP therapy).
• Erythropoietin and G-CSF.

Patients must have:

• Documented HIV-seropositivity.
• CD4 count 100 - 300 cells/mm3.
• Prior continuous AZT dose = or > 300 mg/day for 1 year or longer.

Prior Medication: Required:

• AZT for at least 1 year prior to study entry.

Exclusion Criteria

Co-existing Condition:

Patients with the following symptoms and conditions are excluded:

• Medical contraindication or is considered noncompliant in the opinion of the investigator.
• Peripheral neuropathy = or > grade 2.

Concurrent Medication:

Excluded:

• Anti-HIV agents other than study drugs.
• Biologic response modifiers (other than erythropoietin or G-CSF).
• Systemic cytotoxic chemotherapy.
• Regularly prescribed medications (such as antipyretics, analgesics, allergy medications) that are associated with an increased risk of pancreatitis, peripheral neuropathy, or bone marrow suppression.

Concurrent Treatment:

Excluded:

• Radiation therapy.

Patients with the following prior conditions are excluded:

• History of acute or chronic pancreatitis, gout, or uric acid nephropathy.

Prior Medication:

Excluded:

• Other antiretrovirals besides AZT.
• ddI or ddC for more than 30 days within the past year or any time within 3 months prior to study entry.
• Acute therapy for an infection or other medical illness within 14 days prior to study entry.

• Bristol-Myers Squibb
• Glaxo Wellcome