Herpes zoster—associated optic neuritis and necrotizing retinitis are believed to result from viral invasion secondary to varicella reactivation or to dissemination from the dermatomal

exposed to varicella, but her son became secondarily infected with varicella about 2 weeks after her illness started. She had the characteristic varicella rash, most of which had already scabbed. Her visual acuity was 20/20 in her right eye and 20/40 in her left. There was a left relative afferent pupillary defect. Funduscopy showed left optic disk edema with no visible retinal or choroidal lesions. Color vision in the left eye was impaired, and there was a demonstrable visual field defect. She did not have any other neurological deficits. A visual evoked response test revealed a conduction defect of the left visual pathway. The diagnosis of acute varicella infection with left optic neuritis was made. The patient was given iv methylprednisolone (0.5 g q.d. for 5 days) followed by oral prednisolone (45 mg q.d. for 11 days) in tapered doses. At the same time, she was given iv acyclovir (500 mg q8h for 3 days) followed by oral acyclovir (800 mg five times per day for 13 days). Four days after the initial treatment, her visual acuity returned to normal although her color vision was still impaired. Her vision was completely restored after 1 month. She remains well and has not had any further attacks of optic neuritis or other neurological symptoms two years after her admission. Our case is unique in that the visual complications preceded the onset of the varicella rash. To our knowledge, only one other case with a similar presentation has been reported [1]. The pathogenesis of viral-associated optic neuritis has not been well established. The delayed onset of optic neuritis and the frequent bilateral involvement and near-complete recovery in many cases suggest an immune-mediated process with consequent demyelination [2, 3]. Our observation that the ocular symptoms preceded the systemic infection suggests that the optic neuritis may instead be due to direct viral invasion. Herpes zoster—associated optic neuritis and necrotizing retinitis are believed to result from viral invasion secondary to varicella reactivation or to dissemination from the dermatomal lesions. These infections are seen in both immunocompetent and HIVinfected patients and are associated with extensive visual loss [5, 6]. Unlike varicella-associated optic neuritis, the outcome for patients with herpes zoster—associated optic neuritis and necrotizing retinitis is extremely poor despite acyclovir treatment. In most cases of varicella-associated optic neuritis, vision is completely restored, although there may be residual optic disk pallor [2]. The use of steroids to treat optic neuritis is controversial. Other patients [1, 2] have recovered without receiving steroid therapy, although it is believed to hasten recovery [3]. Two patients had severe residual visual loss 6 months after the onset of optic neuritis despite receiving steroid therapy [3, 4]. We administered steroids with acyclovir to our patient since the visual symptoms preceded the rash. The patient's symptoms started to decrease only a few days after the initiation of therapy. The present case illustrates that it is still not understood whether the pathogenesis of varicella-associated complications is due to direct viral invasion, to an immune-mediated process, or to both.