AMERICAN JOURNAL OF OPHTHALMOLOGY
namic biomicroscopic examination of the vitre ous. No eyes without tears had pigment gran ules. Therefore, the presence of vitreous pigment granules in this clinical setting has a 93% (15 of 16) sensitivity and 100% (90 of 90) specificity for identifying patients with retinal tears. Since a careful dynamic biomicroscopic ex amination of the vitreous was performed in Hikichi and Trempe's series, it would be inter esting to know how often vitreous pigment granules were noted and if they were indica tive of retinal breaks. Corroboration of our re sults by other investigators would further sub stantiate the utility of this important clinical sign. We believe that any patient with acute symp toms of flashes and floaters, especially with pigment granules in the vitreous, deserves a careful peripheral retinal examination. The presence or absence of any clinical sign or symptom related to vitreous separation does not completely exclude the possibility of a reti nal tear with its attendant risk of retinal de tachment. DAVID A. LIGHTMAN, M.D. Williamsport, Pennsylvania ROY D. BROD, M.D.
References 1. Brod RD, Lightman DA, Packer AJ, Saras HP. Correlation between vitreous pigment granules and retinal breaks in eyes with acute posterior vitreous detachment. Ophthalmology 1991;98:1366-9. 2. Davis MD. Natural history of retinal breaks without detachment. Arch Ophthalmol 1974;92:183. 3. Beyer N. The natural history of asymptomatic retinal breaks. Ophthalmology 1982;89:1033. Reply EDITOR: We appreciate the comments by Drs. Lightman and Brod on our article. In consideration of their response, we re viewed the data of the 785 patients in our study to determine the relationship between retinal breaks and vitreous pigment granules. Of 50 symptomatic eyes (floaters, light flashes, or both) with posterior vitreous detachment with retinal breaks, 45 (90%) eyes had vitre ous pigment granules. No symptomatic eyes without retinal breaks had vitreous pigment granules. These findings support those of Drs.
Lightman and Brod that the presence of pig ment granules are indicative of retinal breaks. In our patient series, four of the five eyes (80%) with retinal breaks without vitreous pigment granules had small breaks (less than l k disk diameter), and one eye had undergone ocular examination one week after the onset of symptoms, which was the shortest time in our series. In the study by Brod and associates, 1 only one eye had a small retinal break without vitreous pigment granules. Machemer and La qua reported 2 that the number of pigment epithelia in the vitreous cavity increased with the duration of retinal detachment in the owl monkey. We agree that vitreous pigment granules are indicative of retinal breaks; however, we also wish to emphasize the importance of the rela tionship between clinical signs or symptoms and the disorder. The first task a clinician un dertakes is evaluation of the initial complaints, that is, flashes or floaters, or both, in this case. On the basis of these complaints, the clinician then conducts an examination to establish the exact diagnosis, the next step of which must be a slit-lamp examination with a dilated pupil. Further examination narrows the differential diagnosis to the exact disorder if the examina tion has been conducted correctly. However, the point we wish to emphasize is that flashes or floaters are the first evidence of a possible pathologic ocular condition that warrants fur ther examination, and we believe it is impor tant to establish the relationship between these signs and the ocular disorder. TAIICHI HIKICHI, M.D. CLEMENT L. TREMPE, M.D. Boston, Massachusetts
References 1. Brod RD, Lightman DA, Packer AJ, Saras HP. Correlation between vitreous pigment granules and retinal breaks in eyes with acute posterior vitreous detachment. Ophthalmology 1991;98:1366-9. 2. Machemer R, Laqua H. Pigment epithelium pro liferation in retinal detachment (massive periretinal proliferation). Am J Ophthalmol 1975;80:1-23.
Risk Factors for the Development of Rhegmatogenous Retinal Detachment in Patients With Cytomegalovirus Retinitis EDITOR: In the article, "Risk factors for the develop-
Vol. 118, No. 5
ment of rhegmatogenous retinal detachment in patients with cytomegalovirus retinitis," by W. R. Freeman, D. N. Friedberg, C. Berry, j . I. Quiceno, M. Behette, S. C. Fullerton, and D. Munguia (Am. J. Ophthalmol. 116:713, De cember 1993), the authors observed that reti nal detachment complicating cytomegalovirus retinitis occurred more frequently when 25% or more of the peripheral retina was involved and when the retinitis was poorly controlled. Our current treatment of cytomegalovirus reti nitis has resulted in few retinal detachments and little loss of vision. We have achieved this by early detection through screening, 1 effective control of the retinitis, 2 and treating those most at risk of retinal detachment with prophylatic laser. We currently screen all AIDS outpatients who have a CD4 count of 0.05 x 109/1 or less every three months. We also screen AIDS pa tients admitted to the hospital for any AIDSrelated illness. Over the last two years we have screened over 270 patients and found 15 (23 eyes) with cytomegalovirus retinitis; five of those patients were asymptomatic at diagnosis. In all, 39 patients were treated for more than six weeks, 22 (35 eyes) with high-dose intravit real ganciclovir and four (seven eyes) with in travenous ganciclovir. The posterior pole was affected in nine eyes, peripheral retina in 23, and both regions in seven. Using high-dose intravitreal ganciclovir (2 mg/0.1 ml), we were able to achieve 100% resolution of the retinitis, with relapse occur ring at a median of 42 weeks in only two pa tients on weekly maintainance intravitreal in jections. The reactivation of the retinitis settled with therapy of two injections a week for three weeks. Of the four patients treated with intravenous ganciclovir, relapse occurred in three within eight weeks. For those patients who had more than 25% of the peripheral retina involved, we adminis tered laser photocoagulation as a band entrap ping the region involved. Using an argon laser and panfundoscope or three-mirror lens, we placed burns three or four deep, one spot apart, to give a full-thickness burn. Nine eyes received prophylactic laser, three of which went on to retinal detachment that was con tained by the band of laser, and visual acuity was not affected. Two patients had retinal detachments requir ing surgery. One had a small area of retinitis and responded well to a scierai buckle proce dure and intravitreal therapy. The other had extensive bilateral disease with subretinal fluid
detaching both maculae. He did poorly despite vitrectomy, silicone oil replacement, and intra venous foscarnet, never achieving good control of the retinitis or vision better than counting fingers. We look forward to a clinical trial confirming our clinical impression that early effective treatment, and prophylatic laser therapy will reduce the 25% incidence of retinal detach ment in cytomegalovirus retinitis. NIGEL MORLET, M.B., F.R.A.C.O. STEPHANIE YOUNG, M.B., F.R.A.C.O. MINAS T. CORNEO, M.S., F.R.A.C.O. Randwick, New South Wales, Australia
References 1. Morlet N, Young SR, Dean R, Gold J. Ophthalmological screening for CMV retinitis in HIV infection. Lancet 1992;340:179. 2. Young SR, Morlet N, Heery S, Hollows FC, Coroneo MT. High dose intravitreal ganciclovir in the treatment of cytomegalovirus retinitis. Med J Aust 1992;157:370-3. Reply EDITOR: We thank Drs. Morlet, Young, and Corneo for their interest in our work on the risk fac tors for the development of retinal detachment in patients with cytomegalovirus retinitis. In deed, it was our intent in performing this study to determine risk factors not only to un derstand the pathophysiologic aspects of these retinal detachments better, but also to help devise strategies to prevent them or limit visu al morbidity because of these retinal detach ments. The risk factors that we identified in clude peripheral retinitis involving greater than 25% of the retina external to the major vascular arcades, activity of retinitis at the most recent examination. Indeed, the amount of peripheral retinitis and the presence of ac tivity of retinitis were so strong that we calcu lated that the presence of retinitis activity in over 25% of the retina external to the major vascular arcades confers a 24-fold increased risk over patients with retinitis but without these risk factors. It is logical to conclude from our study, therefore, that any measures which decrease the amount of retinal involvement with cyto megalovirus retinitis would reduce the risk of retinal detachment greatly. The currently avail able therapies for cytomegalovirus retinitis are