CMV Medical Article

Methods
Patients
Patients with AIDS and newly diagnosed, active cytomegalovirus retinitis were enrolled at 18 clinical sites in the United States. Eligible patients had to be at least 18 yearsold, and their best corrected visual acuity had to be 20/200 or better in at least one affected eye. Exclusion criteria included opacities that would prevent visualization of the fundus,contraindications to intraocular surgery or to therapy with intravenous ganciclovir, the presence of overt signs or symptoms of extraocular cytomegalovirus infection, an absolute neutrophil count below 500 cells percubic millimeter, a platelet count below 25,000 cells per cubic millimeter, a serum creatinine concentration above 1.5 mg per deciliter (133 µmol per liter), and a Karnofsky score below 60. Informedconsent was obtained from all patients.
Base-Line Evaluation and Randomization
Before randomization, patients underwent a complete base-line examination, which included nine-field fundusphotography. The randomization was stratified so as to distribute patients with unilateral and bilateral cytomegalovirus retinitis equally among the three treatment groups, and blocking was used to assignpatients equally to the treatment groups over time.
Treatment and Follow-Up
Eligible patients were assigned with equal probability to receive one of three treatments: an intraocular implant (Vitrasert,Chiron Vision, Irvine, Calif.) with a release rate of 1 µg of ganciclovir (Cytovene, Roche Laboratories, Nutley, N.J.) per hour, an intraocular implant with a release rate of 2 µg of ganciclovir perhour, or intravenous ganciclovir. A 2-µg-per-hour implant was included to determine whether a release rate higher than that used in a previous study37 (1 µg per hour) would have greater efficacy.Investigators and patients were unaware of the release rate of the ganciclovir implant. The surgical procedure to insert the implant has been described elsewhere.35,36 Patients assigned to receive...