Rhegmatogenous Retinal Detachment in Patients With Cytomegalovirus Retinitis: The Foscarnet-Ganciclovir Cytomegalovirus Retinitis Trial THE STUDIES OF OCULAR COMPLICATIONS OF AIDS (SOCA) RESEARCH GROUP IN COLLABORATION WITH THE AIDS CLINICAL TRIALS GROUP (ACTG)*
• PURPOSE: To determine the incidence and risk factors for rhegmatogenous retinal detachment in a population of patients with newly diagnosed cytomegalovirus retinitis. • METHODS: Analysis of selected baseline and timedependent data on patients enrolled in a multicenAccepted for publication Feb 7, 1997. Supported by cooperative agreements from the National Eye Institute to The Johns Hopkins University School of Medicine (U10 EY 08052), School of Hygiene and Public Health (U10 EY 08057), and the University of Wisconsin School of Medicine (U10 EY 08067). Additional support provided by the National Center for Research Resources through General Clinical Research Center grants 5M01 RR 00350 (Baylor College of Medicine), 5M01 RR 00035 and 5M01 RR 00722 (The Johns Hopkins University), 5M01 RR 05096 (Louisiana State University/ Tulane), 5M01 RR 00071 (Mt Sinai Medical Center), 5M01 RR 00047 (New York Hospital-Cornell Medical Center), 5M01 RR 00096 (New York University), 5M01 RR 00048 (Northwestern University), 5M01 RR 000865 (University of California, Los Angeles), 5M01 RR 00083 (University of California, San Francisco), and 5M01 RR 05280 (Univer sity of Miami). Support also provided by the National Institute of Allergy and Infectious Diseases through cooperative agreements UOl AI 27668 (The Johns Hopkins University), UOl AI 27674 (Louisiana State University/Tulane), UOl AI 27669 (Memorial Sloan-Kettering), UOl AI 25917 (New York Hospital-Cornell Medical Center), UOl AI 27667 (Mount Sinai Medical Center), UOl AI 27665 (New York University), UOl AI 25915 (Northwestern University), UOl AI 27660 (University of California, Los Angeles), UOl AI 27670 (University of California, San Diego), UOl AI 27663 (University of California, San Francisco), and UOl AI 25831 (University of Massachusetts). Funding also provided by Astra USA, Inc, Westborough, Massachusetts. Drugs provided by Astra USA, Inc; Burroughs Wellcome Co, Research Triangle Park, North Carolina; and Syntex Research, Palo Alto, California. *AU individuals participating in this study are listed at the end of the article. Financial disclosure statements are on file at the SOCA Coordinating Center, The Johns Hopkins University School of Hygiene and Public Health. Correspondence to William R. Freeman, MD, University of California, San Diego, Shiley Eye Center, 9415 Campus Point Dr, La Jolla, CA 92093. Reprint requests to Douglas A. Jabs, MD, SOCA Chairman's Office, The Wilmer Ophthalmological Institute, The Johns Hopkins University School of Medicine, 550 N Broadway, Ste 700, Baltimore, MD 21205.
VOL.124, No. 1
ter, prospective, randomized, controlled clinical trial of therapy with foscarnet vs ganciclovir. • RESULTS: I n 3 1 6 eyes with cytomegalovirus retinitis at baseline, the risk of rhegmatogenous retinal detachment in an eye involved by cytomeg alovirus retinitis was 1 8 . 9 % at 6 months ( 9 5 % confidence interval [CI], 14-0% to 23.8%) and 3 7 . 9 % at 1 year ( 9 5 % CI, 3 0 . 5 % to 4 5 . 3 % ) . Retinal detachment was not associated with the type of anticytomegalovirus therapy (intravenous foscarnet or ganciclovir) to which the patient was assigned. Extent of retinal involvement by cyto megalovirus retinitis, higher patient age, and lower C D 4 + T-cell counts were associated with an in creased risk of retinal detachment; myopia was not. • CONCLUSIONS: Retinal detachment in patients with cytomegalovirus retinitis is unrelated to the type of intravenous therapy used or to refractive error. T h e median time to retinal detachment in an involved eye with cytomegalovirus retinitis and free of retinal detachment at baseline was 18.2 months. Strategies to reduce the extent of retinitis and possibly the number of reactivations may reduce the incidence of retinal detachment.
HEGMATOGENOUS RETINAL DETACHMENT IS A
common complication of cytomegalovirus in fection of the retina, the most frequently encountered intraocular infection in patients with the acquired immunodeficiency syndrome (AIDS). 1 ' 2
© AMERICAN JOURNAL OF OPHTHALMOLOGY i997;i24:6i-70
Two groups have suggested that the extent of retinitis either at baseline (references 4 through 6 and Heinemann MH, Williams RG, Wise-Campbell S, unpublished data, "CMV-associated retinal detach ment in patients with AIDS," presented at the IX International Conference on AIDS, Berlin, June 6-11, 1993) or at a given examination15 is a strong predictor of retinal detachment, particularly when the extent of retinitis is large in the retinal periphery outside the major vascular arcades. The presence of retinitis activity has also been suggested to be a risk factor for cytomegalovirus retinitis-related retinal detachment,15 as has the presence of myopia.12 The Foscarnet-Ganciclovir Cytomegalovirus Reti nitis Trial was designed to compare foscarnet with ganciclovir as initial treatment for previously untreat ed cytomegalovirus retinitis in patients with AIDS.16 The primary outcomes of the trial were death, time to retinitis progression, and change in visual function. A total of 234 patients were randomly assigned to treatment with either foscarnet or ganciclovir, where as six patients elected or were assigned to deferral of treatment. The trial was conducted from March 1990 through October 1991, when the treatment protocol was suspended after a significant survival difference favoring foscarnet Was observed.17 In this report, we present the results of analyses of rhegmatogenous retinal detachment that occurred in 81 of 316 eyes with cytomegalovirus retinitis (and free of retinal detachment at baseline).
Serous macular retinal detachment caused by retinitis in the proximity of the optic nerve or foVea may also cause visual impairment; however, serous retinal de tachments are uncommon, and they resolve with treatment of the retinitis.3 Published reports of large numbers of prospectiVely analyzed patients suggest that the frequency of retinal detachment in either eye among patients receiving anticytomegalovirus treat ment for retinitis may range from 15% to 26% (references 4 through 6 and Heinemann MH, Wil liams RG, Wise-Campbell S, unpublished data, "CMV-associated retinal detachment in patients with AIDS," presented at the IX International Conference on AIDS, Berlin, June 6-11, 1993). The cumulative probability of retinal detachment occurring in either eye 1 year after the diagnosis of cytomegalovirus retinitis has been reported to exceed 50% (references 4 through 6 and Heinemann MH, Williams RG, Wise-Campbell S, unpublished data, "CMV-associ ated retinal detachment in patients with AIDS," presented at the IX International Conference on AIDS, Berlin, June 6-11, 1993). Treatment of eyes with cytomegalovirus-related retinal detachment is difficult. Cytomegalovirus-associated rhegmatogenous retinal detachments are char acterized by widespread areas of necrotic retina, multiple retinal breaks, vitreous traction, and, un commonly, proliferative vitreoretinopathy. Most such retinal detachments can be treated surgically with vitrectomy and silicone oil tamponade, and recent studies have suggested a high initial success rate.4,7'9 Long-term visual outcomes have been disappointing in some studies for a variety of reasons, which include the formation of cataract, difficulties with postopera tive refraction through silicone oil, variability of refraction and hyperopic shift, progression of retinitis, and the possibility of late postoperative optic atrophy in these eyes7,10'14; however, more recent studies have suggested that good visual results can be obtained in a substantial proportion of patients.8,9 Because of the morbidity involved and the need for surgery, it is important to understand the risk factors for retinal detachment. Better understanding of the epidemiolo gy of retinal detachment in cytomegalovirus retinitis should allow for a more informed patient counseling as well as aid in planning clinical trials designed to reduce the incidence of this important visionthreatening complication of cytomegalovirus retinitis.
ed cytomegalovirus retinitis, were aged at least 13 years, and had an absolute neutrophil count of greater than 1,000 cells per |xl and a serum creatinine concentration below 2.0 mg/dl. The diagnosis of AIDS was based upon the Centers for Disease Control 1987 Revised Surveillance Case Definition.18 Cytomegalovirus retinitis was diagnosed when the characteristic necrotizing retinitis was present on examination by an ophthalmologist.13 The treatment protocol and consent procedures were reviewed and approved by the institutional review boards of the individual participating institutions, and all patients gave signed consent before randomization.16
METHODS ELIGIBLE PATIENTS HAD AIDS AND PREVIOUSLY UNTREAT-
Patients were classified into one of two strata based on the location and the size of the retinal lesions. Lesions were classified according to their location in one of three zones of the retina.16,19 Zone 1 consisted of the area within 1,500 |xm of the edge of the optic nerve or within 3,000 |xm of the center of the fovea. Zone 2 extended from the limits of zone 1 to a circle defined by the ampullae of the vortex veins. Zone 3 extended from the limits of zone 2 to the ora serrata. The size of the lesion was estimated as a percentage of the total retinal area. When both eyes were involved, stratification was based on the extent of disease in the eye with more extensive disease.16 Stratum 1 included subjects with lesions in zone 1 or lesions involving more than 25% of the retina. These patients were considered to be at imminent risk of visual loss and were started on the assigned treatment as soon as possible. Stratum 2 included subjects with lesions in either zone 2 or 3, or both, that involved less than 25% of the retina. Patients in stratum 2 were offered three options: immediate treatment (with ganciclovir or foscarnet, according to random assignment), deferral of treatment (until the retinitis progressed into zone 1 or involved more than 25% of the retina, and then initiation of randomly assigned treatment at the time of progression), or random assignment to either immediate or deferred treatment.1617 Randomization schedules were de signed to provide balance in treatment assignments within clinics and within strata.16,17 The induction dose of foscarnet was 60 mg/kg of body weight given every 8 hours (that is, 180 mg/kg per day) for 14 days. The maintenance dose was 90 mg/kg per day, administered with 1 liter of normal saline solution. In December 1990, the maintenance dose of foscarnet was increased to 120 mg/kg of body weight per day after a repeat induction course for retinitis that had relapsed. The initial maintenance dose remained at 90 mg/kg per day. The induction dose of ganciclovir was 5 mg/kg given every 12 hours (that is, 10 mg/kg per day) for 14 days. The mainte nance dose was 5 mg/kg per day.1617 Patients were seen for data collection at baseline, every 2 weeks for 2 months after randomization, every 4 weeks from 2 months to 6 months after randomiza tion, and every 2 months thereafter. At each visit, patients gave a medical history and had a physical examination, an estimate of Karnofsky performance VOL.124, No. 1
score, best-corrected visual acuity measurement, and ophthalmologic examination, and underwent fundus photography.16 At baseline, 1,3, and 6 months after randomization and every 6 months thereafter, pa tients underwent visual field testing.16 Standardized visual acuity measurements, visual field measurements, wide-angle fundus photography, and a photography reading were performed as previ ously described.16,17,20 Progression was defined as the movement of the border of the cytomegalovirus lesion 750 |xm along a front more than 750 (xm in length or as the occurrence of a new lesion in either eye over 25% of the disk area in size and separated from previous lesions by more than 750 (Jim. The primary evaluation of retinitis progression was performed by the Fundus Photograph Reading Center personnel, who were masked to treatment assignment. At each visit, the ophthalmologist was asked to evaluate whether or not progression had occurred and to estimate the amount of progression. Decisions about reinduction were based on the clinician's estimate of progression. At baseline only, the areas of zones 1 and 2 involved by retinitis were measured at the reading center using planimetry on a mosaic of the photo graphs created by computerized image processing. At baseline and at each follow-up visit, lesion size was also graded by the clinician as within one of four categories: less than 10% of the retinal area, 10% to 24% of the retinal area, 25% to 49% of the retinal area, or 50% or more of the retinal area involved by cytomegalovirus retinitis. At each visit, the ophthal mologist also recorded for each eye whether the retina was detached and, if so, the type of detachment, and whether surgery was planned and, if so, the date of surgery for repair of the retinal detachment. For comparisons between the two treatment groups, this report uses data on the 234 patients randomly assigned to treatment on or before October 7, 1991, the date of the suspension of the treatment protocol. It does not include data from the six patients who either chose or were assigned to deferred treat ment and whose retinitis had not progressed enough by October 7, 1991, to warrant randomization to treatment. Also for these analyses, data are included from all follow-up visits completed as of October 7, 1991, and from all data entered at the coordinating center by June 17, 1992. Comparisons between
RETINAL DETACHMENT IN CYTOMEGALOVIRUS RETINITIS
treatment groups were performed according to each patient's original treatment assignment. For evaluation of risk factors for retinal detach ment, this report uses data on the 240 patients who were randomly assigned to treatment or who chose or were assigned to deferred treatment. For these analy ses, data are included from all follow-up visits com pleted and data entered at the coordinating center as of June 17, 1992. Data were analyzed with respect to eyes rather than patients. All eyes involved with cytomegalovirus retinitis at baseline and without a retinal detachment at baseline were included in the analyses. With respect to a question on follow-up forms, retinal detachments were counted as events. Cox proportional hazards models were used to estimate both crude and adjusted relative risks obtained by means of the sandwich variance estimator, which corrects for the correlation between two involved eyes in the same patient.21,22 Risk factors were defined as either time-indepen dent, when the value of the risk factor did not change during follow-up, or as time-dependent, when the value of the risk factor did change during follow-up. Time-independent risk factors were the following: treatment group to which the patient was randomly assigned, Karnofsky score at baseline (at or above vs below the median of 90), age at baseline (at or above vs below the median of 37 years), race, sex, myopia at baseline (mild myopia defined as refractive error between —0.25 and —6.0 diopters, and severe myopia defined as 6.0 diopters or more), platelet count at baseline (at or above vs under the median of 200 X 103/n
tively, he or she by the second method would have values that were updated at the fifth follow-up visit. Some time-dependent risk factors used informa tion from baseline as well as from follow-up visits. These were zone 1 involvement, zone 2 involvement, zone 3 involvement, area of retinitis (assessed clinical ly and photographically in four categories: less than 10%, 10% to 24%, 25% to 49%, and 50% or more), activity of retinitis borders (graded centrally as active, intermediate, or atrophic), lesion hemorrhage (clini cally estimated as marked, moderate, minimal, or none), cumulative number of retinitis progressions (assessed clinically or photographically as advance ment of the border by more than 750 |i-m or as development of a new lesion), and retinal detach ment in the fellow eye (13 such retinal detachments). Time-to-event plots were derived using the KaplanMeier method.23 All data analyses were performed using the Statistical Analysis System.24
RESULTS RHEGMATOGENOUS RETINAL DETACHMENT OCCURRED
in 81 (25.6%) of the 316 eyes involved with cytomeg alovirus retinitis at baseline during the study period. Kaplan-Meier analysis of probability of retinal de tachment over time with the eye as unit of analysis showed an 18.9% risk of retinal detachment at 6 months (95% CI, 14.0% to 23.8%) and a 37.9% risk of retinal detachment at 1 year (95% CI, 30.5% to 45.3%). There were no substantial differences in the risk of retinal detachments in involved eyes of patients assigned to ganciclovir vs foscarnet. At the suspen sion of the treatment protocol, there were 31 retinal detachments (18.7%) in 166 involved eyes in patients assigned to ganciclovir and 33 retinal detachments (23.2%) in 142 involved eyes in patients assigned to foscarnet. The crude and adjusted relative risks of retinal detachment comparing ganciclovir vs foscar net were 0.85 (P = .540) and 0.65 (P = .184), respectively. Note the comparison between ganciclo vir vs foscarnet shown in the Table uses all patients, including those randomized to observation, and thus reports slightly different results. Clinical estimates of the area of retina involved at
AMERICAN JOURNAL OF OPHTHALMOLOGY
TABLE. Risk Factors for Rhegmatogenous Retinal Detachment in Involved Eyes Definition of Time-dependent Covariates* Baseline Visit
Time-dependent covariates Zone 1: yes vs no Zone 2: yes vs no Zone 3: yes vs no Area: 10%-24% vs <10% Area: >25% vs <10% Area: >50% vs <10%* Retinal detachment in fellow eye: yes vs no Hemorrhage: marked, moderate vs minimal, none Activity: active vs atrophic Activity: intermediate vs atrophic Number of progressions' Time-independent covariates Myopia: mild vs none Myopia: severe vs none Treatment assignment: ganciclovir vs foscarnet Platelets: &200 x 103 vs <200 x 103 cells/p.1 Hemoglobin: a11 vs <11 g/dl** Predicted creatinine clearance: a1.2 vs <1.2 ml/min/kg
CD4+ T cells: >14 vs <14 cells/jU Karnofsky score: a90 vs <90 Age: >37 vs <37 yrs Race: nonwhite vs white Sex: female vs male Antiretroviral therapy: yes vs no
Visit Before RD
Visit With RD
1.34 1.35 1.73 0.95 1.40 NA 1.12 0.99 NA5 NAS NA
.334 .419 .233 .893 .445 NA .809 .958 NA! NA! NA
1.00 1.20 1.49 1.40 2.50 6.58 1.93 1.16 1.43 1.31 1.01
.992 .638 .274 .450 .064 .0001 .084 .619 .356 .477 .947
0.99 1.13 1.46 3.43 3.68 13.92 1.36 1.63 0.90 1.30 1.31
.980 .751 .303 .031 .055 <.0001 .316 .064 .768 .465 .023
0.98 1.19 0.73 1.34 2.15 0.94 0.36 0.56 1.65 0.90 2.03 1.19
.933 .721 .240 .319 .007 .900 .001 .079 .067 .763 .154 .610
0.88 1.26 0.84 1.36 2.22 0.96 0.47 0.64 1.45 0.94 1.36 1.12
.684 .571 .489 .280 .009 .932 .010 .117 .207 .850 .499 .762
0.85 1.40 0.87 1.34 2.40 1.10 0.48 0.56 1.59 0.92 1.42 1.40
.582 .472 .610 .337 .007 .840 .020 .076 .126 .814 .468 .410
NA = not applicable; RD = retinal detachment; RR = relative risk. Time-dependent covariates defined at the baseline visit are actually time independent because they contain the value of the covariate at baseline and do not vary over time. tRelative risks and P values were estimated by means of Cox regression models with robust variance estimation that included all risk factors as covariates and clinic as a stratification variable. *Area was categorized as <10%, 10%-24%, a=25% at baseline and <10%, 10%-24%, 25%-49%, £50% during follow-up. §Activity was always recorded as high at the baseline visit. UA model substituting number of reinductions for number of progressions had an RR=0.96, P=.743 for the model with time-dependent covariates defined at the visit before the RD, and an RR=1.15, P=.343 for the model with time-dependent covariates defined at the visit with the RD. An additional model substituting number of major (movement of the border a 1,500 (im or a new lesion) progressions for number of progressions had an RR=1.28, P= .186 for the model with time-dependent covariates defined at the visit before the RD, and an RR=1.40, P=.083 for the model with time-dependent covariates defined at the visit with the RD. **An analysis with hemoglobin trichotomized produced similar results.
baseline showed little or no relation to the develop ment of retinal detachment, but larger areas of involvement at the visit when retinal detachment was noted or at the preceding visit were associated with increased risk (Table). Eyes with 50% or more of the retina involved by retinitis at the visit preceding that when retinal detachment was noted had about six
times the risk of detachment compared to eyes with less than 10% of the retina involved. At the visit when retinal detachment was noted, the comparable risk was about 14-fold. When the photographic assessment (using planimetry on a mosaic of photo graphs) of the percent of zones 1 and 2 involved with retinitis at baseline was modeled in place of the
RETINAL DETACHMENT IN CYTOMEGALOVIRUS RETINITIS
clinical assessment, the adjusted relative risk of de tachment for the area between 10% to 24% vs the area less than 10% was 0.84 (P = .695) and, for the area greater than or equal to 25% vs the area less than 10%, was 2.28 (P = .050). The cumulative 6-month risk (unadjusted for covariates) of retinal detachment was 12% for eyes with less than 10% of zones 1 and 2 involved at baseline, 22% for those with 10% to 24% involved, and 37% for those with 25% or more involved (Figure). Patient-specific baseline covariates associated with an approximately 1.5-fold to 2-fold increase in the risk of retinal detachment were the following: older age, higher hemoglobin value, lower CD4+ T-cell count, and lower Karnofsky score (Table). The relative risk of retinal detachment in eyes with more than 6 diopters of myopia was 1.2 to 1.4, but this was not statistically significant. Given the observed number of retinal detachments in the no myopia and mild myopia groups, and with a two-sided type I error level of 5% and 80% power, the minimum detectable relative risk of retinal detachment between the severe myopia and the combined no myopia and mild myopia groups was 2.1. Forty-one eyes had surgery for retinal detachment during follow-up. The outcome of retinal detachment surgery was not analyzed because data were not collected in a way that allowed methodical evaluation of vision at all visits before and after surgery. The median visual acuity in the 30 eyes with data 2 months after surgery was 20/160 + 1 . The median change in visual acuity (ETDRS letters) from before surgery to 2 months after surgery was a loss of 10 letters (two lines).
DISCUSSION THE AIM OF THIS REPORT WAS TO EXAMINE RISK FACTORS
for retinal detachment in patients with AIDS and newly diagnosed cytomegalovirus retinitis. The study database contains information from 240 patients and 316 eyes with cytomegalovirus retinitis studied in 11 centers across the United States. It is important to note that there was no association between treatment assignment to foscarnet vs ganciclovir and retinal
FIGURE. Kaplan-Meier estimates (censored at 12 months) of retinal detachment by reading center-defined area of cytomegalovirus retinitis involvement at baseline.
detachment. The median time to retinal detachment (by eye involved with cytomegalovirus) in our study was 18.2 months. The cumulative probability of development of a retinal detachment in involved eyes was 19% at 6 months and 38% at 1 year. This result agrees with the results of other studies,4,6'15,20 suggest ing that rhegmatogenous retinal detachment is a major problem in eyes with cytomegalovirus retinitis. Although our study on visual outcomes20 reported a 27% to 28% cumulative probability of retinal detach ment at 6 months and a 57% to 66% probability at 1 year after randomization, that report used patients as the unit of analysis, and the reported figure is the probability of a detachment in either eye. In contrast, in this report, we document the probability of a detachment in an eye affected by cytomegalovirus retinitis, a figure expected to be lower. The strongest risk factor for development of retinal detachment was an increased area of retinal involve-
AMERICAN JOURNAL OF OPHTHALMOLOGY
ment by cytomegalovirus retinitis. Other studies also have found that the size of the cytomegalovirus lesion, both at diagnosis4,6 and at the visit before detachment,15 is a risk factor for retinal detachment. Our study used semiquantitative estimates of retinal area, whereas other studies15 quantitated area and zone involvement at each visit. Freeman and associ ates15 reported that zone 2 plus zone 3 involvement was a risk factor when assessed at the visit before detachment.15 The presence of active borders of the cytomegalovirus lesion was not associated with retinal detachment. The results of the analyses of the number of progressions are difficult to interpret in the present study: at the visit before the detachment, there was no association, whereas at the time of retinal detachment, there was an association with the number of progressions (relative risk = 1.31, P = .023). The anecdotal experience of many is that bilateral retinal detachments in patients with AIDS and cytomegalovirus retinitis are common. In this study, 13 (19.1%) of 68 patients had bilateral retinal detach ments. Although the crude (unadjusted) relative risk of retinal detachment when a fellow eye had a detachment was 4.37 (P < .001), this association was weakened when a multivariate analysis was used, which suggests that retinal detachment in the fellow eye is a surrogate marker for other risk factors, such as zone and area of involvement. One previous study suggested that myopia may be a risk factor in cytomegalovirus-related retinal detachments.12 In our study, the relative risks, analyzing either mild or severe myopia, did not demonstrate a statistically significant increase in the relative risk for eyes with myopia. Older age at baseline did have a statistically border line association with an increased risk of retinal detachment; changes in vitreous that occur with age may allow the passage of liquid vitreous through the necrotic area in cytomegalovirus retinitis cases.8,25 Systemic factors associated with increased risk of retinal detachment included a low baseline CD4+ T-cell count and Karnofsky score below 90%. The reason for this result is uncertain. Such patients may have a more rapid retinitis progression, but this was adjusted for. The association of a greater blood hemoglobin at baseline with an increased risk of retinal detachment is unexplained.
VOL.124, No. I
Although this trial was not designed to evaluate the outcomes of surgery for retinal detachment, most studies have suggested that there is a variable loss of vision after the repair of the detachments by vitrectomy with silicone oil injection (references 4, 7 through 14, and 25, and Heinemann MH, Williams RG, Wise-Campbell S, unpublished data, "CMVassociated retinal detachment in patients with AIDS," presented at the IX International Conference on AIDS, Berlin, June 6-11, 1993). Loss of vision associated with retinal detachment repair may be caused by advancement of the retinitis, delay in repair, surgery technique, difficulty obtaining a stable refraction through silicone oil, or cataract. Strengths of this study include the multicentered prospective nature of the study and the high quality of the data collection. However, it must be remembered that patients in this study were selected to be good study candidates and so, for example, had higher Karnofsky scores than the cytomegalovirus retinitis population in general has. The requirements for this study probably resulted in the exclusion of patients who were likely to be noncompliant or poorly compli ant with drug treatment regimens and study proce dures. Extremely ill patients also were excluded. This study used only systemically administered anticytomegalovirus drugs, given intravenously. It did not include patients treated with local intravitreal therapy or with oral anticytomegalovirus agents. Despite these limitations, this study helps to determine the risk factors for the development of rhegmatogenous retinal detachment in eyes with cytomegalovirus retinitis. These results should assist patient counseling, possi bly help in designing risk-reduction strategies, and help to define the nationwide scope of the problem in the United States. Measures that slow the progressive increase in the area of retina involved by retinitis or that improve immune status, or that do both, may reduce the risk of retinal detachment. THE FOSCARNET-GANCICLOVIR CYTOMEGALOVIRUS RETINITIS TRIAL PARTICIPATING CLINICAL CENTERS
Baylor College of Medicine, Cullen Eye Institute, Houston, Texas: Richard A. Lewis, MD, MS (Direc tor), Pamela S. Clogston, COT, Victor Fainstein, MD, Ronald Gross, MD, Tobias Samo, MD, Colette Tuttle, RN.
RETINAL DETACHMENT IN CYTOMEGALOVIRUS RETINITIS
The Johns Hopkins University School of Medi cine, Baltimore, Maryland: Douglas A. Jabs, MD (Director), Linda Apuzzo, John Bartlett, MD, Laura C. Coleson, RN, MPH, J. P. Dunn, MD, Lois Eldred, PA-C, Judith Feinberg, MD, Thomas Flynn, MD, Rosemary King, PA-C, Jo Leslie, PA-C. Louisiana State University Medical Center, New Orleans, Louisiana: Bruce Barren, MD (Director), Deborah Greenspan, RN, Cynthia LeCount, RN, Gholam Peyman, MD. Former director: Rudolph Franklin, MD. Memorial Sloan-Kettering Cancer Center and New York Hospital-Cornell Medical Center, New York, New York: Murk-Hein Heinemann, MD (Di rector), Bruce Polsky, MD, Kathleen Squires, MD, Susanne Wise-Campbell, RN. Mount Sinai School of Medicine, New York, New York: Alan H. Friedman, MD (Director), Tony W. Cheung, MD, Norma Justin, MS, Henry Sacks, MD, PhD, Colette Severin, MS, Steven Teich, MD. New York University Medical Center, New York, New York: Dorothy N. Friedberg, MD, PhD (Direc tor), Adrienne Addessi, MA, RN, Douglas Dieterich, MD, Kevin Frost, Francois Lafleur, MPH. Northwestern University, Chicago, Illinois: David V. Weinberg, MD (Director), Lee Jampol, MD, Robert Murphy, MD, Kathleen Naughton, RN. For mer director: Dale Henderly, MD (deceased). University of California at Los Angeles, Los Ange les, California: Gary N. Holland, MD (Director), Suzette Chafey, RN, NP, Holly Fall, RN, MN, NP, W. David Hardy, MD, Chris Kimbrell, RN, Lesley J. MacArthur, MA. University of California at San Diego, San Diego, California: William R. Freeman, MD (Director), Linda Meixnert, RN, Timothy J. Peterson, MD (US Naval Hospital), Jose I. Quiceno, MD, Leland Rickman, MD, Mary Ann Simanello, RN, Stephen Spector, MD. University of California at San Francisco, San Francisco, California: James O'Donnell, MD (Direc tor), Theodore Bush, RN, Jacqueline Hoffman, Alex ander Irvine, MD, Mark Jacobson, MD, James Larson, COT, Stuart Seiff, MD, Maxine Wanner, Laurel Warner, MD. University of Miami School of Medicine, Miami, Florida: Janet Davis, MD (Director), Elaine Chuang,
MD, Millie Espinal, Paul Mendez, MD, Ruth Vandenbroucke. University of Massachusetts Medical Center, Worcester, Massachusetts: Sarah H. Cheeseman, MD (Director), John Gittinger, MD, Richard Haubrich, MD, Harry Kachadoorian, CRA, Karen Tolson, BS. Resource Centers Chairman's Office, The Johns Hopkins University School of Medicine, Baltimore, Maryland: Douglas A. Jabs, MD (Study Chairman), Amy C. Klemm, Joan M. Kline, Maria Stevens, Renee M. Webb. Former Members: Jude BrownBellamy, MAS, Jan A. Markowitz, PhD. Coordinating Center, The Johns Hopkins Univer sity School of Hygiene and Public Health, Baltimore, Maryland: Curtis L. Meinert, PhD (Director), Karen L. Binder, Ronald Brookmeyer, PhD, Vivian E. Brown, William H. Burns, MD, Betty J. Collison, John Dodge, Michele Donithan, MHS, Nancy Fink, MPH, Charlotte Gerczak, MLA, Janet T. Holbrook, MS, MPH, Milana R. Isaacson, BS, Camara P. Jones, MD, MPH, Charlene R. Levine, BS, Yuan-I Min, MPH, Jill L. Meinert, Deborah J. Nowakowski, Rosetta M. Owens, Alfred Saah, MD, MPH, Gordon R. Sandford, MD, Stephen Singer, Michael Smith, Alice L. Sternberg, ScM, James Tonascia, PhD, Mark L. Van Natta, MHS. Fundus Photograph Reading Center, University of Wisconsin, Madison, Wisconsin: Matthew D. Davis, MD (Director), Maria Agres-Segal, Jane Armstrong, Rosemary Brothers, Marika Chop, Ginger Freitag, Larry Hubbard, MA, Dolores Hurlbert, Yvonne Magli, Kathleen Miner, Michael Neider, Suzanne Thomas, Marilyn Vanderhoof-Young. Drug Distribution Center, ERC Bioservices Corpo ration, Rockville, Maryland: Gary Stewart, MS, RPh, Robert Hughes. Former Member: Lisa Welch, RPh, MAS. Serum and Cytomegalovirus Culture Repository, Biomedical Research Institute, Rockville, Maryland: James Leef, PhD, Sandy Palmer. National Eye Institute, Bethesda, Maryland: Rich ard L. Mowery, PhD. National Institute of Allergy and Infectious Diseas es, Bethesda, Maryland: Susan Ellenberg, PhD, Joyce Korvick, MD. Study Committees Writing Committee: William R. Freeman, MD (Chair), Janet Davis, MD, Matthew
AMERICAN JOURNAL OF OPHTHALMOLOGY
D. Davis, MD, M. H. Heinemann, MD, Douglas A. Jabs, MD, Mark C. Van Natta, MHS. Officers of the Study: Douglas A. Jabs, MD (Chair), Matthew D. Davis, MD, Curtis L. Meinert, PhD, Richard L. Mowery, PhD. Steering Committee: Douglas A. Jabs, MD (Chair), Bruce Barron, MD, Tom Clark, CRA, Pamela S. Clogston, Matthew D. Davis, MD, Douglas Dieterich, MD, Millie Espinal, William R. Freeman, MD, W. David Hardy, MD, Janet T. Holbrook, MS, MPH, Gary N. Holland, MD, Larry Hubbard, MA, Joyce Korvick, MD, Richard A. Lewis, MD, MS, Jan A. Markowitz, PhD, Curtis L. Meinert, PhD, Law rence Merin, RPB, Richard Mowery, PhD, Bruce Polsky, MD, James Tonascia, PhD. SOCA-ACTG Joint Executive Committee: Doug las A. Jabs, MD (Chair), Matthew D. Davis, MD, Susan Ellenberg, PhD, Judith Feinberg, MD, Mark Jacobson, MD, Joyce Korvick, MD, Curtis L. Meinert, PhD, Richard L. Mowery, PhD, Fred Sattler, MD, James Tonascia, PhD. Former members: John Mills, MD, Colin Jordan, MD. Policy and Data Monitoring Board: Byron W. Brown, Jr, PhD (Chair), Brian Conway, MD, James Grizzle, PhD, Robert Nussenblatt, MD, John P. Phair, MD, Harmon Smith, PhD. Nonvoting mem bers: Matthew D. Davis, MD, Susan Ellenberg, PhD, Douglas A. Jabs, MD, Curtis L. Meinert, PhD, Richard L. Mowery, PhD, James Tonascia, PhD.
REFERENCES 1. Jabs DA, Green WR, Fox R, Polk BF, Bartlett JG. Ocular manifestations of acquired immune deficiency syndrome. Ophthalmology 1989;96:1092-1099. 2. Gallant JE, Moore RD, Richman DD, Keruly J, Chaisson RE, Zidovudine Epidemiology Group. Incidence and natural history of cytomegalovirus disease in patients with advanced human immunodeficiency virus disease treated with zidovudine: the Zidovudine Epidemiology Study Group. J Infect Dis 1992;166:1223-1227. 3. Gangan PA, Besen G, Munguia D, Freeman WR. Macular serous exudation in patients with acquired immunodeficiency syndrome and cytomegalovirus retinitis. Am ] Ophthalmol 1994;118:212-219. 4. Jabs DA, Enger C, Haller J, de Bustros S. Retinal detach ments in patients with cytomegalovirus retinitis. Arch Oph thalmol 1991;109:794-799. 5. Gross JG, Bozette SA, Mathews WC, et al. Longitudinal
study of cytomegalovirus retinitis in acquired immune defi ciency syndrome. Ophthalmology 1990;97:681-686. 6. Jabs DA. Ocular manifestations of HIV infection. Trans Am Ophthalmol Soc 1995;93:623-683. 7. Freeman WR, Henderly DE, Wan WL, et al. Prevalence, pathophysiology, and treatment of rhegmatogenous retinal detachment in treated cytomegalovirus retinitis. Am J Oph thalmol 1987;103:527-536. 8. Garcia RF, Flores-Aguilar M, Quiceno JI, et al. Results of rhegmatogenous retinal detachment repair in cytomegalovi rus retinitis with and without scleral buckling. Ophthalmolo gy 1995;102:236-245. 9. Lim JI, Enger C, Haller JA, et al. Improved visual results after surgical repair of CMV-related retinal detachments. Oph thalmology 1994;101:264-269. 10. Dugel PU, Liggett PE, Lee MB, et al. Repair of retinal detachment caused by cytomegalovirus retinitis in patients with acquired immunodeficiency syndrome. Am J Ophthal mol 1991;112:235-242. 11. Davis J, Chuang E. Repair of retinal detachment caused by cytomegalovirus retinitis in patients with the acquired immu nodeficiency syndrome, [letter]. Am J Ophthalmol 1992;113: 468-470. 12. Irvine AR. Treatment of retinal detachment due to cytomeg alovirus retinitis in patients with AIDS. Trans Am Ophthal mol Soc 1991;89:349-367. 13. Kupperman BD, Flores-Aguilar M, Quiceno JI, et al. A masked prospective evaluation of outcome parameters for cytomegalovirus-related retinal detachment surgery in pa tients with acquired immune deficiency syndrome. Ophthal mology 1994;101:46-55. 14- Chuang EL, Davis JL. Management of retinal detachment associated with CMV retinitis in AIDS patients. Eye 1992; 6:28-34. 15. Freeman WR, Friedberg DN, Berry C, et al. Risk factors for development of rhegmatogenous retinal detachment in pa tients with cytomegalovirus retinitis. Am J Ophthalmol 1993;116:713-720. 16. Studies of Ocular Complications of AIDS Research Group in collaboration with the AIDS Clinical Trials Group. Studies of the ocular complications of AIDS foscarnet-ganciclovir cytomegalovirus retinitis trial, 1: rationale, design, and meth ods. Controlled Clin Trials 1992;13:22-39. 17. Studies of Ocular Complications of AIDS Research Group in collaboration with the AIDS Clinical Trials Group. Mortality in patients with the acquired immunodeficiency syndrome treated with either foscarnet or ganciclovir for cytomegalovi rus retinitis. N Engl J Med 1992;326:213-220. 18. Centers for Disease Control and Prevention. Revision of the CDC surveillance case definition for the acquired immuno deficiency syndrome. MMWR Morb Mortal Wkly Rep 1987;36:1S-19S. 19. Holland GN, Buhles WC Jr, Mastre B, Kaplan HE, UCLA CMV Retinopathy Study Group. A controlled retrospective study of ganciclovir treatment for cytomegalovirus retinopa thy: use of a standardized system for the assessment of disease outcome. Arch Ophthalmol 1989;107:1759-1766. 20. Studies of Ocular Complications of AIDS (SOCA) Research Group in collaboration with the AIDS Clinical Trials Group (ACTG). Foscarnet-ganciclovir cytomegalovirus retinitis tri al, 4: visual outcomes. Ophthalmology 1994;101:1250-1261. 21. Cox DR. Regression model and life-tables. J R Stat Soc [B]
RETINAL DETACHMENT IN CYTOMEGALOVIRUS RETINITIS
1972;34:187-220. 22. Lin DY, Wei LJ. The robust inference for the Cox proportion al hazards model. J Am Stat Assoc 1989;84:1074-1079. 23. Kaplan EL, Meier P. Nonparametric estimation from incom plete observations. J Am Stat Assoc 1958;53:457-481. 24- SAS/STAT software: the PHREG procedure. Version 6.
Cary, North Carolina: SAS Institute, 1991 (SAS tech rep; P-127). 25. Freeman WR, QuicenoJI, CrapottaJA, Listhaus A, Munguia D, Flores-Aguilar M. Surgical repair of rhegmatogenous retinal detachment in immunosuppressed patients with cytomegalovirus retinitis. Ophthalmology 1992;99:466-474-
Authors Interactive™ Comments and questions regarding this article may be addressed to authors via THE JOURNAL'S Web Site. Selected comments, questions, and author responses are posted on Authors Interactive™ at http://www.ajo.com/
AMERICAN JOURNAL OF OPHTHALMOLOGY