Search

The post-irradiation regression rate of uveal melanomas is a prognostically significant factor for the development of metastases. Other predictive factors for metastases are histological networks which are imagable with confocal ICG angiography. The purpose of this study was to evaluate a possible connection of networks in the ICGA and tumor regression rates.We compared the post-irradiation regression rates (as %) in 20 patients 1 year after brachytherapy with networks identified in pre-treatment indocyanine green angiography (ICGA). The ICG angiography was performed before irradiation, 10 patients were irradiated with Ru-106 and 10 were irradiated with Id-125.The mean preoperative maximum apical height was 5.2 mm [SD: 1.5 mm; Ru106 group: 5.7 mm (SD: 1.0 mm); Id-125 group: 5.0 mm (SD: 1.9 mm)]. In 11 patients (55%) (Ru-106: 5; Id-125: 6) we found networks in the preoperative ICG. The mean regression rate in tumors with networks was 51.3% (SD: 14.7%) and 28.0% (SD: 16.4%) in the group without networks. The difference between both groups was statistically significant (p = 0.003, Mann-Whitney test). No statistically significant difference in the regression rates was found between the two groups of brachytherapy Ru-106 and Id-125 (p = 0.165, Mann-Whitney test).Highly proliferative tumors are known to be more sensitive to irradiation. This may be one reason why tumors with a rapid post-irradiation regression are the more aggressive ones with regard to later development of metastases. Histopathological networks are also known to be a strong indication of more aggressive, metastasizing tumors. These networks are also imagable in ICG angiography. Our observation emphasizes a connection between networks in ICG angiography and regression rates of uveal melanomas after brachytherapy.

Histologically demonstrable microcirculation patterns (microcirculation pattern) of human choroidal melanomas have prognostic significance for the patient. We report on our experience in imaging these microcirculation pattern in vivo using simultaneous confocal Fluorescein (FA)- and Indocyaninegreen (ICG) angiography before and after brachytherapy.The simultaneously procured confocal FA- and ICG angiograms of 50 patients with untreated choroidal melanomas were studied for the visibility of microcirculation pattern. Patients were also followed with simultaneous FA/ICG after brachytherapy.Confocal FA disclosed signs of tumor vascularization in 12 (24%) of the 50 examined patients but microcirculation pattern only in 3 patients (6%). In contrast, simultaneously obtained confocal ICG disclosed microcirculation pattern in 47 patients (94%). In 10 (77%) of the 13 patients the tumor microcirculation changed considerably after brachytherapy: Distortion, thickening, thinning, as well as complete obliteration of vessels could be observed.Histologically demonstrated microcirculation pattern can be imaged in vivo. This offers the possibility to assess the likely biologic behavior of the individual tumor without the need for obtaining a cytologic or histologic specimen via enucleation or fine-needle biopsy. Confocal ICG angiogiography images microcirculation pattern better than FA which can be explained by the different absorption-, fluorescence- and exudation-characteristics ICG. Follow-up with confocal ICG of choroidal melanomas after brachytherapy shows different features and allows for visualization of the microcirculation reaction to the treatment which might be a useful tool for studying the effects of future anti-angiogenesis based tumor therapies.

We have previously shown that histologically described microcirculation patterns (MCP) can be visualized with indocyanine green (ICG) angiography. We have designed a prospective study to evaluate the prognostic value of these angiographically imaged MCP in small choroidal melanocytic lesions. In this report we describe the design of the study, characterize the patient collective, and present the first results.In this prospective nonrandomized observational study, unilateral choroidal melanocytic lesions with 1.5-5.5 mm maximum apical height are observed until growth is determined according to defined criteria. Variables are demographic parameters, subjective symptoms, subretinal fluid, location and dimension of tumor, hemorrhage, color, orange pigment, and MCP determined by ICG angiography: normal, straight, parallel without crosslinking, parallel with crosslinking, arcs without branching, arcs with branching, loop, and network.Seventy patients (22 males, 48 females; age: 33-88 years, median: 64 years) have been included up to now: 19 tumors showed growth so far (time to growth: 51-946 days, median: 127 days). The following parameters were statistically significantly correlated with time to tumor growth: flashes (p = 0.082), orange pigment (p = 0.012), subretinal fluid (p0.001), maximum basal tumor diameter (p = 0.001), maximum apical tumor height (p0.001), parallel with crosslinking (p0.001), arcs with branching (p = 0.006), loop (p0.001), and network (p0.001). Of these, complex MCP (parallel with crosslinking, arcs with branching, loop and/or network) showed the strongest correlation with time to tumor growth in a Cox regression model. Based on our data, the positive predictive value of imaging complex MCP (for growth within 12 months) is 78% and the negative predictive value is 98%.Our patient collective demonstrates comparable prognostic parameters for time to growth as described in the literature. In addition, the ICG angiographic detection of complex MCP is more strongly predictive of the time to growth than other clinically determinable factors. Thus, we recommend this examination for patients with small choroidal melanocytic lesions, if the patient is to be counseled regarding the likely biologic behavior of his tumor.

To report the association between duration of vitrectomy, as well as other risk factors, and the progression of nuclear sclerosis and posterior subcapsular cataract in the Vitrectomy for Macular Hole Study.A cohort study nested within a randomized controlled clinical trial.Using a system similar to the Lens Opacities Classification System II, nuclear sclerosis (NS) and posterior subcapsular cataract (PSC) were scored in the vitrectomy and fellow eye of 74 patients at baseline and at 6, 12, and 24 months postoperatively. Age, baseline blood pressure and refractive power, and duration of surgery were evaluated as risk factors for NS or PSC progression and cataract extraction.The incidence of NS progression in the surgical group of vitrectomy eyes was 81% at 6 months, 98% at 1 year, and 100% at 2 years of follow-up. In contrast, NS progression in the control group of fellow eyes was only 18% at 6 months, 20% at 1 year, and 8% at 2 years. The incidence of PSC progression in the surgical group remained at approximately 11% throughout follow-up, which was not significantly higher than the 3% to 5% incidence in the control group. Vitrectomy was significantly related to progression of NS cataract (P.001) and cataract extraction (P.01). No statistically significant differences were found for NS scores, PSC scores, or progression rates between eyes that had less than median surgical duration (60 min.) or more than the median surgical duration. Additionally, no significant differences were found when eyes that experienced 45 minutes or less surgical duration were compared with eyes that endured more than 75 minutes surgical duration. Age, blood pressure, and refractive power were not found to be predictors for NS and PSC progression.Although vitrectomy is a risk factor for NS progression, the duration of vitrectomy does not increase the risk.

To characterize cytomegalovirus (CMV) retinitis in human immunodeficiency virus (HIV)-infected patients who demonstrate immune recovery while receiving highly active antiretroviral therapy (HAART).Consecutive, noncomparative case series.Twenty-two HIV-positive patients, from two institutions, with a history of CMV retinitis, and with elevated CD4 cell counts after HAART.Duration of healed CMV retinitis without anti-CMV therapy, CD4 cell count, and HIV viral load.Discontinuation of anti-CMV therapy after persistent elevation of CD4 cell count over 50 cell/mm3 (median, 161/mm3; range, 85-408/mm3).The median period of healed CMV retinitis without anti-CMV therapy was 72 weeks (range, 33-116 weeks). Nineteen of 22 patients were still healed without anti-CMV therapy at study end. The three patients with CMV retinitis progression simultaneously had HAART, fail with CD4 cell counts of 37, 35, and 47/mm3.HIV-positive patients with CMV retinitis, who demonstrate a sustained HAART-induced elevation of CD4 cell count on two consecutive counts 3 months apart and whose retinitis remains healed on anti-CMV therapy for greater than 4 months, are likely to remain healed if the anti-CMV therapy is withdrawn. It is important to monitor these patients with indirect ophthalmoscopy because HAART failure may occur and allow CMV retinitis reactivation.

To delineate the immune parameters associated with reactivation of cytomegalovirus (CMV) retinitis in patients for whom highly active antiretroviral therapy (HAART) was not successful.Prospective, longitudinal observational study of a cohort of 102 patients with CMV retinitis treated with HAART and being followed up at the AIDS Ocular Research Unit of the University of California, San Diego from November 1995 to November 1998. The study included serial clinical and fundus photographic examinations with CD4 T-lymphocyte counts and HIV viral load measurements.Forty-seven of the 102 patients with CMV retinitis responded to HAART. Thirty-five of the patients were successfully withdrawn from anti-CMV therapy. During a median follow-up of 74.71 weeks (range, 4.86-144 weeks) after discontinuation of anti-CMV therapy, four patients experienced a reactivation of CMV retinitis. In each case, the CD4 count decreased before reactivation to a median of 31.5 cells/mm3 (mean, 31.25 cells/mm3; range, 23-39 cells/mm3). The association between the CD4 count decreasing to less than 50 cells/mm3 and reactivation of CMV retinitis was statistically significant (P0.0003).Four patients treated with HAART experienced reactivation of CMV retinitis as their CD4 count decreased. The threshold CD4 count below which reactivation of CMV retinitis occurred in patients for whom HAART was not successful appeared to be 50 cells/mm3. Despite an initial response to HAART, patients are still at risk for reactivation of CMV retinitis if their CD4 count decreases to less than 50 cells/mm3. The HIV viral load did not appear to predict CMV reactivation.

To evaluate and compare the risks and benefits of autologous serum as an adjuvant therapy in macular hole surgery for stage 3 or 4 macular holes.Comparison of 2 consecutive (nonrandomized) cohorts using standardized methods for the determination of hole size and for surgical procedures, and using the same study surgeons. The serum cohort consisted of 106 eyes using autologous serum as an adjuvant, and the no serum cohort consisted of 58 eyes without adjuvants. The primary end point was the closure of the macular hole as determined by the 6-month fundus photographs. Secondary end points included the number and types of postoperative complications. Comparison in outcomes between the 2 cohorts used chi2 and logistic regression procedures, adjusting for preoperative differences between the study cohorts.At 6 months, the (unadjusted) rate of hole closure was significantly greater for the eyes treated with serum than for the eyes not treated with serum (90 [85%] of 106 vs. 40 [69%] of 58, P = .04). However, after adjusting for preoperative differences in hole diameter and the prevalence of epiretinal membranes, no overall difference in hole closure rates due to serum was found (P = .44). In contrast, benefit due to serum for large holes (diameter573 microm) was seen (12 [75%] of 16 vs 13 [57%] of 23, P = .04). No differences in complication rates were found between the cohorts.Any beneficial effect of serum used as an adjuvant to macular hole surgery is small, and, if present, the beneficial effect may be limited to larger holes. A randomized, prospective, controlled study in larger macular holes is needed.

To study the course and visual outcome of Vogt-Koyanagi-Harada (VKH) syndrome in children.The medical records of 97 consecutive patients with VKH syndrome were reviewed. Patients had been seen at King Khaled Eye Specialist Hospital in Riyadh, Saudi Arabia. The records were divided into two groups: those in whom the onset of the disease occurred at the age of 14 years or younger (pediatric group) and those in whom the disease manifested after the age of 14 years (adult group).Eight (61%) of the 13 children required cataract surgery compared to 14(17%) of 84 adult patients. Despite medical and surgical therapy, eight (61%) of the pediatric group had a final visual acuity of 20/200 or worse compared with 22 (26%) of the 84 adult patients (p=0.0103).The course of early-onset VKH (before the age of 14 years) appears to be aggressive. The ocular complications are more severe in children than in adults, leading to rapid deterioration in vision. Despite therapy, the final visual outcome of VKH in children is poor.

To determine capillary blood flow measurements in eyes with branch retinal vein occlusion using a scanning laser Doppler flowmeter.Retinal capillary blood flow in branch retinal vein occlusion areas and corresponding ipsilateral nonbranch retinal vein occlusion areas, 11 equivalent areas of the contralateral fellow eye of 12 consecutive untreated branch retinal vein occlusion patients, and 16 eyes of 11 age-matched normal control subjects were measured with scanning laser Doppler flowmetry. A template consisting of eight squares, each with a field of 100 x 100 microm (10 x 10 pixel) with space interval of 500 microm equidistant horizontally and vertically was used to obtain blood flow measurements in all subjects. Mean blood volume, flow, and velocity were obtained by averaging the mean values measured in each field. We avoided measurement over large retinal vessels to prevent the aliasing artifact of blood cells from moving faster than the sampling frequency.Branch retinal vein occlusion areas have significantly decreased microvascular blood volume (P = .0009), flow (P = .02), and velocity (P = .016) compared with ipsilateral nonbranch retinal vein occlusion areas in the same eye. Branch retinal vein occlusion areas also have decreased blood volume (P = .001), flow (P = .0042), and velocity (P = .0044) compared with areas of contralateral fellow eyes of branch retinal vein occlusion subjects. Branch retinal vein occlusion areas have significantly decreased blood volume (P = .0012), flow (P = .008), and velocity (P = .02) compared with age-matched normal areas.Average retinal blood volume, flow, and velocity in areas of branch retinal vein occlusion are significantly lower than in healthy retinas. The ability to noninvasively measure hemodynamic changes in the retinal capillary bed may be relevant to development of new therapies for retinovascular disease.

The authors prospectively studied visual outcome, relapse, complications, and survival of patients with acquired immune deficiency syndrome (AIDS)-related cytomegalovirus (CMV) treated with high-dose intravitreous ganciclovir (2 mg/0.1 ml) injections. The outcomes were compared with those of patients treated with standard doses of intravenous ganciclovir in the same institution. The histopathologic and electrophysiologic effects of high-dose intravitreous ganciclovir injections in rabbits also were studied.A nonrandomized case series.A total of 42 patients (74 eyes) were treated with intravitreous injections and 18 patients (27 eyes) were treated with intravenous ganciclovir. Five eyes of three New Zealand white rabbits were injected with ganciclovir, and the sixth eye was a control specimen.Patients treated with intravitreous injections received twice-weekly doses of 2 mg/0.1 ml ganciclovir for 3 weeks, then weekly injections. Patients treated with intravenous ganciclovir received standard doses. Patients were monitored with regular examinations. Rabbit eyes were given intravitreous injections of 1 mg/0.1 ml of ganciclovir weekly for 4 weeks.Assessments of vision, retinal inflammation, and survival were made. Electroretinograms were performed on the rabbit eyes, and they were processed for light and electron microscopy.In the intravitreous group, visual acuity (VA) was stable in 64 of 74 eyes, 5 improved, and 5 deteriorated. Sixty-three (85%) of 74 eyes had final VA of 20/20 to 20/40. Relapse occurred in five eyes (7%; median time, 42 weeks). There were three cases of endophthalmitis. Median survival after diagnosis of CMV retinitis was 36 weeks. In the intravenous group, VA was stable in 18 eyes, 0 improved, and 9 deteriorated. Sixteen (59%) of 27 eyes had final VA of 20/20 to 20/40. Relapse occurred in 15 eyes (56%) at a median time of 21 weeks. Median survival was 21 weeks. The rabbit studies showed no evidence of toxicity.High-dose intravitreous ganciclovir effectively suppressed CMV retinitis, preserved vision, and prevented relapse without deterioration in survival.

The purpose of this study was to evaluate the incidence and characteristics of retinal and choroidal manifestations of toxoplasmosis and/or Mycobacterium avium-intracellulare complex (MAC) in patients with acquired immunodeficiency syndrome (AIDS).The authors analyzed their prospectively collected data and found 120 patients with new retinal lesions (group A) that were diagnosed 3 months or longer following the diagnosis of MAC and/or toxoplasmic encephalitis. The authors also performed a point prevalence study of retinal/choroidal findings in 25 consecutive AIDS patients (group B) without known eye disease who had been recently treated for toxoplasmic encephalitis and/or disseminated MAC infections. In addition, the characteristics of retinochoroidal toxoplasmosis scars in 5 AIDS patients were studied and compared with the characteristics of scars in 18 immunocompetent patients.In this study the incidence of ocular manifestations of MAC was zero (95% confidence interval [CI] 0.0% to 3.8%). Two of 25 patients (8%) (95% CI 1% to 26%) in group A and 2 of 11 patients (18.1%) (95% CI 3.3% to 51.8%) in group B had toxoplasmic retinochoroiditis.In AIDS patients, ocular manifestations of toxoplasmosis are more common than ocular MAC. In addition, when compared with immunocompetent patients, AIDS patients tend to have retinochoroidal scars with less retinal pigment epithelium hyperplasia (1.8+ vs 3+) (P = .03).

To determine the risks and benefits of vitrectomy surgery in eyes with stage 2 macular holes.A multicentered, controlled, randomized clinical trial was performed with participation of 16 community and university-based ophthalmology clinics. Thirty-six eyes with stage 2 macular holes and 12 months of follow-up were studied. Pars plana vitrectomy with separation of the posterior hyaloid membrane and intraocular injection of perfluoropropane (C3F8) was followed by postoperative face-down positioning for two weeks. This protocol was compared with observation alone. Outcome variables included anatomic closure of the macular hole, macular hole size, and four standardized measures of vision.At 12 months, 15 (71%) of 21 eyes randomly assigned to observation progressed to stages 3 or 4, compared with three (20%) of 15 eyes randomly assigned to surgery (P.006). Compared with eyes randomly assigned to observation, eyes randomly assigned to surgery had significantly smaller hole diameters (P.01) and significantly better visual acuity outcomes, as measured by the Word Reading (P = .02) and Potential Acuity Meter (P = .002) charts. No significant differences were found for the Early Treatment Diabetic Retinopathy Study chart and Contrast Sensitivity test.Compared with observation alone, surgical intervention in stage 2 macular holes resulted in a significantly lower incidence of hole enlargement and appeared to be associated with better outcome in some measures of visual acuity.

To describe the echographic characteristics of splitting the outer posterior cortical vitreous in patients with proliferative diabetic retinopathy and vitreous hemorrhage.The authors retrospectively reviewed the echographic findings in 270 patients who were evaluated at the Doheny Eye Institute between January 1983 to December 1989 for proliferative diabetic retinopathy and vitreous hemorrhage. None of the eyes had undergone pars plana vitrectomy before echographic examination.Forty-five patients (17%) had echographic evidence of splitting of the outer posterior vitreous cortex, a finding the authors have termed posterior vitreoschisis. In all patients, differentiation of the posterior vitreoschisis from a true posterior hyaloid detachment was possible, either on the initial or on serial echographic examination, by the separate detachment of the inner wall of the vitreoschisis cavity and the true posterior hyaloid from the retinal surface. The vitreoschisis cavities often were found to contain unclotted blood. In some eyes, the inner wall of the vitreoschisis cavity was adherent to the apex of the most highly elevated area of traction retinal detachment, suggesting that posterior vitreoschisis may itself result in clinically significant vitreoretinal traction, independent of the presence or extent of true posterior hyaloid separation.The authors' finding suggest that spontaneous splitting of the outer posterior vitreous cortex may occur in patients with proliferative diabetic retinopathy and vitreous hemorrhage, which may mimic a true posterior cortical vitreous detachment on echographic examination. Preoperative recognition of posterior vitreoschisis may be important in the surgical management of these patients.

The authors determine the intraocular tolerance of a new widely used liquid perfluorocarbon, perfluoroctylbromide (perflubron).Pars plana vitrectomy was performed on 54 eyes of 54 patients with vitreoretinal disorders at three centers. Diagnoses included giant retinal tears, proliferative vitreoretinopathy, and dislocated intraocular and crystalline lenses. At the conclusion of the vitrectomy, perflubron was removed.Perflubron was efficacious for vitreoretinal manipulation. Of the 45 eyes with retinal detachment, 23 (51.1%) of the retinas were reattached after a single surgery; redetachment occurred in 22 (48.9%) after the initial procedure, and further surgery was necessary to reattach the retina. Final retinal reattachment was achieved in 40 (88.9%) eyes. Mean visual acuity improvement was six lines (P0.0019). Visualization of the water/perfluorocarbon interface was good. There was no evidence of adverse effects from perflubron on the retina, lens, or anterior segment.Findings indicate that perflubron is safe for temporary intraoperative use intravitreally. The absence of adverse effects is consistent with the properties of perflubron that our group has studied in the eyes of animals and in other uses in human patients.

The purpose of this study is (1) to determine baseline characteristics and natural history of immature full-thickness macular holes, (2) to describe progression and resolution, and (3) to present new aspects of pathogenesis of idiopathic macular hole.The authors analyzed 41 eyes with stage 2 macular holes (37 patients) in a multicentered prospective randomized trial; 19 eyes were randomized to observation (versus surgery) and had more than 12 months of follow-up, allowing determination of the natural course. Baseline and subsequent examinations included best-refracted visual acuity (Early Treatment of Diabetic Retinopathy Study, potential acuity meter, Pelli-Robson contrast sensitivity, and Bailey-Lovie reading vision), of clinical examinations, photography, and fluorescein angiography.Mean Snellen visual acuity was 20/66 at baseline. Centric holes usually had a small break (201 microns average mean diameter) with a dark yellow ring and without significant retinal elevation. Eccentric holes had a high maximum/minimum diameter ratio (mean, 1.88 +/- 0.7) and an incomplete cuff of subretinal fluid or yellow ring. Posterior vitreous detachment prevalence was 32% (8/25) in the centric hole group and 0% (0/ 16) in the eccentric hole group (P0.05). For the 19 eyes with 12 months of followup, progression rate to stage 3 (or 4) was 74% (n = 14). The diameter of the stage 2 holes increased significantly between baseline and 12 months (P0.001). Progression rate to stage 3 was 100% (8/8) in the eyes with pericentral hyperfluorescence (PCH) and 55% (6/11) in eyes without PCH (P0.05). Enlargement occurred in 100% of eccentric holes and 60% of centric holes (P = 0.09). Different progression patterns in eccentric and centric holes suggest different mechanisms of pathogenesis.Eccentric and centric stage 2 macular holes may have a different pathogenesis. Most stage 2 macular holes, especially with PCH (P0.05) or eccentric holes, progressed to stage 3 or 4. In addition to purely tangential traction, some component of obliquely oriented anteroposterior vitreous traction component may be important for pathogenesis of senile macular holes, particularly eccentric stage 2 macular holes.

This prospective, randomized, multicenter, controlled trial was designed to evaluate the efficacy and safety of intravenous ganciclovir for the treatment of peripheral cytomegalovirus (CMV) retinitis in patients with AIDS. Patients were randomly assigned to receive either immediate treatment, intravenous ganciclovir, 5 mg/kg twice daily for 14 days followed by 5 mg/kg once daily for 14 weeks, or deferred treatment. Patients randomized to deferred treatment whose retinitis progressed were offered ganciclovir. Of the 22 patients randomized to deferred treatment who were included in the final analysis, 20 were found to have progressive CMV retinitis compared with 10 of the 13 randomized to immediate treatment. The median time to progression in the deferred treatment group, as determined by a masked fundus photography reading center, was 13.5 days compared with 49.5 days in the immediate treatment group (mean +/- SD, 19.3 +/- 4.1 vs. 66.4 +/- 14.0; P = .001, log rank test). These data indicate that ganciclovir delays the progression of CMV peripheral retinitis in persons with AIDS.

Cotton-wool spots and cytomegalovirus (CMV) retinitis are seen frequently in AIDS patients. Human immunodeficiency virus (HIV) infection of the retina has been proposed as a mechanism for the high incidence of retinal pathology. An autopsy study of the eyes from 25 consecutive cases of AIDS was performed using gross examination, light microscopy, trypsin digestion of retinal vasculatures, and immunohistochemistry to evaluate the possible role of HIV, as well as CMV, in the pathogenesis of retinitis and retinal vasculopathy. Brain tissue was studied in the first 20 of these cases to evaluate any correlation between retinal and central nervous system pathology. CMV retinitis was observed in 15 cases (60%). Cotton-wool spots were seen in nine cases (36%). CMV encephalitis was detected in four cases, whereas HIV encephalitis was noted in five cases. We were unable to demonstrate a correlation between CMV retinitis and CMV encephalitis. However, the number of cases studied was small, and the frequency of CMV encephalitis was low. On the other hand, bilateral CMV retinitis demonstrated a correlation to HIV encephalitis (P less than 0.005, Fisher's exact test). HIV infection of the retina was not detected by typical morphologic changes or immunohistochemistry. Immunohistochemistry localized CMV infection solely to areas of active retinitis. These findings suggest that bilateral CMV may serve as a marker of HIV encephalitis, possibly indicating a severely immunodepressed state.(ABSTRACT TRUNCATED AT 250 WORDS)

The authors report four cases in which retinal biopsy findings yielded unexpected or previously unreported diagnoses in patients with inflammatory retinitis. The tissue diagnoses included Wegener's retinal vasculitis in an immunosuppressed patient with a clinical diagnosis of cytomegalovirus retinitis, a novel viral form in the retina of a patient with cytomegalovirus retinitis, a case of acute retinal necrosis due to cytomegalovirus infection in an immunologically normal adult, and a case of ganciclovir-resistant herpes family viral retinitis. These cases illustrate the use of retinal biopsy in obtaining tissue for diagnosis and guiding treatment in selected cases of retinitis.

Objective: To describe the clinical course of varicellazoster optic neuropathy preceding acute retinal necrosis in patients with acquired immunodeficiency syndrome. Design: Case series. Setting: Two tertiary care centers in San Diego, Calif, and London, England. Patients: Three human immunodeficiency virus-positive men with previous cutaneous zoster infection, optic neuropathy, and necrotizing retinitis. Results: All patients had an episode of zoster dermatitis treated with acyclovir. Visual loss consistent with an optic neuropathy ensued, followed by typical herpetic retinitis. The cause of visual loss was not suspected to be varicella-zoster until after the retinitis occurred. Despite aggressive medical treatment, 4 of 6 eyes progressed to retinal detachment. Conclusions: Varicella-zoster may cause an optic neuropathy in patients with acquired immunodeficiency syndrome, especially in those with previous shingles. A high index of suspicion is necessary to establish the diagnosis and begin early antizoster treatment.

We studied eight patients in whom cytomegalovirus retinitis was an initial manifestation of AIDS. Six of these patients had cytomegalovirus retinitis alone at the time of diagnosis and two patients had simultaneous cytomegalovirus retinitis with other AIDS associated diseases (Kaposi's sarcoma in one patient and cytomegalovirus colitis in one patient). Treatment with ganciclovir (dihydroxy propoxymethyl guanine), a new antiviral medication related to acyclovir, resulted in regression of retinitis in all five patients treated with this drug. Our study shows that cytomegalovirus retinitis may be the initial manifestation of AIDS and that ganciclovir is a clinically effective antiviral agent for cytomegalovirus.