Clinical versus fundus photographic evaluation of the status of cytomegalovirus retinitis in AIDS patients.

PURPOSE: To evaluate the accuracy of clinical examinations and serial fundus photographic readings in determining the response of cytomegalovirus retinitis to antiviral therapy in patients with acquired immune deficiency syndrome. METHODS: Fifty two consecutive patients with cytomegalovirus retinitis who were prospectively evaluated over a 30-month period for a minimum of 6 months (or until death) were included in this study. There was a total of 708 patients visits. The clinical evaluations included indirect ophthalmoscopy, fundus drawings, 60 degrees fundus photographs, and a comparison of the photographs with those of the previous visit. The fundus photographs were reevaluated in a blinded fashion. Cytomegalovirus retinitis was classified as active (progression of border since last examination) intermediate (border activity without progression), healed (no activity since last visit), or normal (no retinitis). RESULTS: Using the photographic data as the measure of cytomegalovirus retinitis activity, the sensitivity and specificity of clinical assessments were determined. The sensitivity and specificity of clinical versus photographic evaluations varied with retinitis status. In healed retinitis the sensitivity of the clinical examination was 98%, and the specificity was 83%. In cases of border opacification without progression the sensitivity was 80%, and the specificity was 96%. In cases of clinically active retinitis the sensitivity was 63% with a specificity of 100%. Clinical detection of active retinitis and border opacification without progression was reduced when potential problems were present that made visualization of the retinitis border difficult, such as smoldering retinitis, progressive retinal destruction without border opacification, poor media, or fundus pigmentation. CONCLUSIONS: Progressive retinal destruction and visual loss can occur in patients with cytomegalovirus retinitis despite antiviral therapy. Examining the patient through indirect ophthalmoscopy only can result in failure to detect subtle changes.