totoxic lesion from coaxial light has been estimated to be as short as 7.5 minutes, depending on intensity of illumination, clarity of ocular media, pupil size, and focus of the beam. 3 In contrast to coaxial light, oblique indirect illumination contributes minimally to retinal irradiance. 1 We consider the following factors as playing a perti- nent role in producing the phototoxic lesion in our patient. Coaxial light was used to illuminate the surgical field during the entire procedure. The retrobulbar block, which immobilized the eye, increased the probability that a focused hot spot would remain on a specific area of the retina for an extended period of time. The block also dilated the pupil, and the dilation was not reversed. Because the ocular media were clear and a corneal cover was not used, maximum coaxial light was allowed through the dilated pupil. Supplemental nasal oxygen was administered, which may potentiate phototoxic retinal damage. 4 In patients with a phototoxic lesion that spares the fovea, the long-term prognosis for retention of central visual acuity is good 5 ; however, photic lesions that involve the fovea can lead to notable vision loss. To diminish the likelihood of phototoxic injury, the fol- lowing precautions should be considered. Oblique indi- rect illumination should be used when coaxial light is not necessary. When coaxial light is required, the intensity of the light should be kept as low as is practical, and the microscope should be repositioned intermittently to minimize prolonged illumination of any one area of the retina. If possible, pupillary dilation should be reversed. When visualization posterior to the cornea is not required, opaque material should be placed over the cornea. During surgery, after administration of local anesthesia in patients younger than 40 years, oxygen by nasal canula should be omitted unless medi- cally necessary. 4 By reporting this case, we wish to bring to the attention of ophthalmic surgeons the risk of photic maculopathy associated with pterygium surgery. REFERENCES 1. Michels M, Sternberg P Jr. Operating microscope-induced retinal phototoxicity: pathophysiology, clinical manifestations and prevention. Surv Ophthalmol 1990;34:237–252. 2. Byrnes GA, Antoszyk AN, Mazur DO, Kao T-C, Miller SA. Photic maculopathy after extracapsular cataract surgery: a prospective study. Ophthalmology 1992;99:731–738. 3. Irvine AR, Wood I, Morris BW. Retinal damage from the illumination of the operating microscope: an experimental study in pseudophakic monkeys. Arch Ophthalmol 1984;102: 1358 –1365. 4. Jaffe GJ, Irvine AR, Wood IS, Severinghaus JW, Pino GR, Haugen C. Retinal phototoxicity from the operating micro- scope: the role of inspired oxygen. Ophthalmology 1988;95: 1130 –1141. 5. Postel EA, Pulido JS, Byrnes GA, et al. Long-term follow-up of iatrogenic phototoxicity. Arch Ophthalmol 1998;116:753– 757. Cytomegalovirus Retinitis in HIV-Infected Patients With and Without Highly Active Antiretroviral Therapy Serge Doan, MD, Isabelle Cochereau, MD, Nurhayat Guvenisik, MD, Marie-Christine Diraison, MD, Haı ¨ssam Mousalatti, MD, and Thanh Hoang-Xuan, MD PURPOSE: To assess the impact of highly active antiretro- viral therapy on the epidemiology of cytomegalovirus retinitis in patients infected with the human immunode- ficiency virus (HIV). METHODS: In a study performed in a single center for infectious diseases, we compared the data collected in 1995 (without highly active antiretroviral therapy) with 1997 data (with highly active antiretroviral therapy). RESULTS: In a comparison of 1997 with 1995 data, the mean CD4 cell count of patients with cytomegalovirus (CMV) retinitis was higher (169 150 CD4/l vs 15 47 CD4/l) (P .05), and the relapses of CMV retinitis were less frequent (17% vs 36%) (P .02). Newly diagnosed CMV retinitis decreased from 6.1% (59 of 952 patients) in 1995 to 1.2% (nine of 726 patients) in 1997 (P < .0001). In 1997, patients with newly diag- nosed or relapsing CMV retinitis had a lower mean CD4 (37 42) cell count than patients with no relapsing CMV retinitis (197 160) (P .01). CONCLUSION: The incidence and recurrences of CMV retinitis decreased from 1995 to 1997, probably as a result of restored immunity while the patients were undergoing highly active antiretroviral therapy; how- ever, the increasing frequency of HIV resistance to highly active antiretroviral therapy justifies close ocular follow-up. (Am J Ophthalmol 1999;128:250 –251. © 1999 by Elsevier Science Inc. All rights reserved.) P ROGNOSIS OF HUMAN IMMUNODEFICIENCY VIRUS (HIV) infection has been dramatically improved with the advent of highly active antiretroviral therapy, which con- sists of a combination of nucleoside analogs and HIV protease inhibitors. Highly active antiretroviral therapy reduces the incidence of opportunistic infections and prolongs survival by improving immune status, as reflected Accepted for publication March 11, 1999. From the Service d’Ophtalmologie, Ho ˆpital Bichat, Claude-Bernard 46, rue Henri Huchard, 75018 Paris, France. This study was supported in part by a grant from Ensemble contre le Sida-Sidaction-Fondation pour la Recherche Me ´dicale, Paris, France (I.C.). Inquiries to Isabelle Cochereau, MD, Service d’Ophtalmologie, Ho ˆpi- tal Bichat, Claude-Bernard-46, rue Henri Huchard, 75018 Paris, France; fax: 33 1 40 25 88 46; e-mail: oph.sec@bch.ap-hop-paris.fr AMERICAN JOURNAL OF OPHTHALMOLOGY 250 AUGUST 1999