To study the safety and effectiveness of foscarnet in the treatment of AIDS patients who have active infection with cytomegalovirus (CMV) that is causing inflammation of the retina (retinitis). In addition, these patients cannot be treated with ganciclovir (DHPG) because of its toxic effect on the body's blood-forming cells or because white blood cell or platelet counts were too low.

CMV is a common virus, which can cause blindness and death in AIDS patients. Previous studies demonstrate that foscarnet has been effective in both AIDS and non-AIDS patients with CMV infection. Although treatment with ganciclovir (DHPG) is also effective, a significant toxicity leading to dose-limiting neutropenia (low white blood cell count) in one third of treated patients has been associated with the drug. Based on the serious nature of CMV retinitis and the lack of alternative drug therapies for DHPG-sensitive patients, the present study will evaluate the safety and efficacy of intravenous (IV) foscarnet in AIDS patients with CMV retinitis.

CMV is a common virus, which can cause blindness and death in AIDS patients. Previous studies demonstrate that foscarnet has been effective in both AIDS and non-AIDS patients with CMV infection. Although treatment with ganciclovir (DHPG) is also effective, a significant toxicity leading to dose-limiting neutropenia (low white blood cell count) in one third of treated patients has been associated with the drug. Based on the serious nature of CMV retinitis and the lack of alternative drug therapies for DHPG-sensitive patients, the present study will evaluate the safety and efficacy of intravenous (IV) foscarnet in AIDS patients with CMV retinitis.

Following routine evaluation studies, patients are randomized to receive foscarnet right away or to delay treatment, as their retinitis has been determined not to be immediately sight-threatening. Patients are hospitalized for the first 3 days and may remain hospitalized for as many as 14 days. Foscarnet is given by vein (IV) in what is called induction therapy, and if the patient's retinitis stabilizes after 2 weeks of treatment, treatment with foscarnet is continued in maintenance therapy for another 8 weeks. During maintenance therapy, patients receive salt solution IV to help prevent any toxic side effect of foscarnet on the kidneys. Patients have regular checkups to monitor their retinitis as well as their general health. Patients taking zidovudine (AZT) prior to entering the study may continue their treatment if they are selected for the delayed treatment group; if they are selected for the immediate treatment group, they begin or resume AZT therapy when they enter the 2nd week of maintenance therapy. Patients are followed as outpatients for at least 10 weeks, with clinic check-ups and lab tests once every week; eye exams are done once a week for the first 2 weeks and then every other week. If clinically indicated, a continued maintenance regimen may be administered after the 10th week; the total duration of therapy plus maintenance is not to exceed 24 weeks. Note: Patients scheduled for the delayed foscarnet treatment are immediately given foscarnet at the first sign that their retinitis is getting worse.

Inclusion Criteria

Concurrent Medication:

Allowed if hematologically stable on that regimen for at least 30 days prior to study entry:

• Oral antibiotics.
• Chemotherapy for Kaposi's sarcoma.
• Acyclovir for outbreaks of herpes simplex or shingles.
• Zidovudine (AZT), either initiated or continued, by patients randomized to both treatment arms. AZT given concurrently with foscarnet may be administered at a dose of 100 or 200 mg every 4 hours (q4h) at the investigator's discretion. Patients randomized to the delayed treatment arm may initiate or continue AZT administration at a dose of 100 or 200 mg q4h at the investigator's discretion. AZT may not be administered during the first 3 weeks of foscarnet therapy. Patients randomized to immediate therapy may begin or resume AZT when they enter the 2nd week of maintenance therapy (week 4 of the 10-week study period), if their hemoglobin is = or > 8 g/dl and absolute neutrophil count is = or > 1000 cells/mm3 at that time. Caution should be used in the concurrent use of foscarnet and ciprofloxacin, as such use has appeared to exacerbate renal failure in one patient.

Patients must have active AIDS-related cytomegalovirus (CMV) retinitis as identified by its characteristic ophthalmoscopic appearance and verified by fundus photography. Patients must also demonstrate one of the following clinical and/or laboratory findings:

• Treatment with ganciclovir (DHPG) resulting in dose-limiting toxicity (absolute neutrophil count (= polymorphonuclear leukocytes plus bands) < 500 cells/mm3 or platelets < 25000 platelets/mm3) occurring on = or > two documented occasions at least 7 days apart while receiving up to a maximum induction regimen of 10 mg/kg/day or a maintenance regimen of up to 5 mg/kg/day. Neutropenia should not be the result of zidovudine (AZT) treatment.
• Ineligibility for DHPG therapy because of baseline neutropenia (absolute neutrophil count < 500 cells/mm3) or thrombocytopenia (platelets < 25000 platelets/mm3) documented on = or > 2 occasions at least 7 days apart. Baseline myelosuppression should not be the result of ongoing therapy with either prescription drugs, including AZT, or over-the-counter medications.

Prior Medication:

Allowed:

• Zidovudine (AZT), according to protocol stipulations.
• Prophylaxis therapy for Pneumocystis carinii pneumonia (PCP).
• Chemotherapy for Kaposi's sarcoma.

Exclusion Criteria

Co-existing Condition:

Patients with any of the following diseases or symptoms are excluded:

• An immediately sight-threatening lesion in a salvageable eye (i.e., patients who have a cytomegalovirus (CMV) lesion that is within 1500 microns of the optic disc or fovea in an eye with correction vision of 20/100 or better).
• Corneal, lens or vitreous opacification which precludes examination of the fundi of either eye.
• Any clinically significant pulmonary or neurologic impairment (i.e., patients who are intubated or comatose), although patients with a history of a seizure disorder or a central nervous system (CNS) mass lesion may be enrolled.

Concurrent Medication:

Excluded:

• Systemic acyclovir as preventive therapy for herpes infection.
• Any nephrotoxic agent.
• Specifically excluded are aminoglycosides, amphotericin B, and parenteral pentamidine. A patient who requires such therapy must be temporarily discontinued from study therapy; if nephrotoxic therapy is given for > 7 days, the patient will be permanently withdrawn from study therapy.
• Other anti-cytomegalovirus (CMV) therapy, specifically ganciclovir, CMV hyperimmune serum/globulin, interferons, and immunomodulators.

Patients will be excluded from the study if they are unwilling or unable to suspend zidovudine (AZT) treatment during the first 3 weeks of the study period (1) if randomized to the immediate treatment arm, or (2) when crossed-over from the delayed treatment arm to foscarnet therapy because of retinitis progression.

Prior Medication:

Excluded:

• Foscarnet for cytomegalovirus (CMV) retinitis.
• Excluded within 7 days of study entry:
• Immunomodulators.
• Investigational agents other than ganciclovir.