Graefe's Archive

Graefe's Arch Clin Exp Ophthalmol(1990)228:281-296

for Clinical and Experimental

Ophthalmology © Springer-Verlag1990

Argon laser panretinal photocoagulation in ischemic central retinal vein occlusion A 10-year prospective study* Sohan Singh Hayreh 1, Marie R. Klugman 2, Patricia Podhajsky 1, Gary E. Servais 1, and Edward S. Perkins 1 1 Department of Ophthalmology and 2 Department of Preventive Medicine 'Biostatistics), University of Iowa, Iowa City, USA Received June 13, 1989 / Accepted January 12, 1990

Abstract. We conducted a prospective, planned study of argon laser panretinal photocoagulation (PRP) in ischemic central retinal vein occlusion (CRVO) over a 10-year period in 123 eyes. On comparing the lasered eyes versus the nonlasered eyes, there was no statistically significant difference between the two groups in the incidence of development of angle neovascularization (NV), neovascular glaucoma (NVG), retinal and/or optic disc NV, or vitreous hemorrhage, or in visual acuity. Our study, however, did show a statistically significant (P = 0.04) difference in the incidence of iris N V between the two groups, with iris N V less prevalent in the laser group than in the nonlaser group, but only when the P R P was performed within 90 days after the onset of CRVO. The other p a r a m e t e r which showed a statistically significant difference between the two groups was the peripheral visual fields - the laser group suffered a significantly (P_< 0.03) greater loss than the non-laser group. We discuss the implications of these findings in light of the natural history of ischemic C R V O and of ocular NV. Since the original rationale for advocating P R P in ischemic C R V O was the proven beneficial effect of P R P on ocular N V in proliferative diabetic retinopathy, we also discuss the disparities in the disease process between ischemic C R V O and proliferative diabetic retinopathy and in their responses to PRP.

Introduction Currently, panretinal photocoagulation (PRP) is universally considered as the treatment of choice for the prevention of ocular neovascularization (NV), particularly neovascular glaucoma (NVG), in ischemic central retinal * This investigation was supported by grant EY-1151 from the National Institutes of Health, and in part by unrestricted grants from Research to Prevent Blindness, Inc., and from the Alcon Research Institute Offprint requests to." S.S. Hayreh

veto occlusion (CRVO). This practice is based on m a n y reports of the beneficial effects of P R P in preventing and/or treating ocular N V in C R V O [1, 3, 4, 12-23, 25, 26, 28-30, 32] (Table 1). A critical review of all those accounts, however, reveals serious flaws in most of the studies, including the following: 1. M a n y of the reports are based on only a very few cases, so that they cannot be considered any more than anecdotal accounts (Table 1); consequently, no definite conclusions can be drawn f r o m them.

Table 1. Number of eyes with ischemic central retinal vein occlusion (CRVO) treated with panretinal photocoagulation (PRP) and the type of photocoagulator used in the various studies reported on the subject in the literature Authors

Year

Number of eyes

Photocoagulator used

Wetzig and Thatcher [30] Zweng et al. [32] Theodossiadis et al. [26] Freyler and Nichorlis [4] Francois and DeLaey [3] Oosterhuis and Sedney [20] Sedney [23] May et al. [18] May et al. [19] Laatikainen [12] Laatikainen [13] Laatikainen et al. [15]

1974 1974 1974 1974 1974 1975 1976 1976 1979 1977 1977 1977

25" 6" 7" 5a 6" 4~ 22 9a 15" 8 7 12

Laatikainen [14] Coscas and Dhermy [1] Riaskoff [21] Verdaquer et al. [28] Saraux et al. [22] Tasman et al. [25] Magargal et al. [16] Magargal et al. [17] Vidic et al. [29]

1983 15 1978 6 1978 86 1979 30 a 1979 12 1980 3 1981 22 1982 100 1981 15"

Xenon or argon Argon Xenon Xenon Argon Xenon Xenon Xenon Xenon Xenon Xenon Mostly xenon, some argon Argon Argon Xenon or argon Xenon Argon Argon Argon Argon Argon

a These studies included cases of CRVO, without any evident differentiation into ischemic and nonischemic types

282 2. A l m o s t all the studies are retrospective reviews o f information, n o t systematic, prospective, wellplanned studies (although some claim to be prospective studies). 3. A l m o s t none have a good, c o m p a r a b l e control population, in spite o f frequent claims m a d e to that effect. 4. P R P was performed using either xenon arc p h o t o coagulator, or a r g o n laser, or a mixture o f b o t h (Table 1). 5. There is one strictly r a n d o m i z e d study o f P R P with a r g o n laser, but that is based on only 23 eyes (12 treated and 11 untreated) [15]. 6. In all the earlier studies, all cases o f C R V O were included w i t h o u t differentiating t h e m into the well-established ischemic and nonischemic types [5-7, 9, 10, 24] (Table 1). Since eyes with nonischemic C R V O have no risk o f developing ocular N V [7], including such eyes in a P R P study produces a m a r k e d bias towards a favorable o u t c o m e o f P R P in prevention o f ocular NV. In studies where C R V O was differentiated into ischemic and nonischemic types, the criteria o f differentiation were variable and vague, with retinal capillary nonperfusion at the posterior pole usually described as the m a i n criterion; however, our studies have shown that in early cases o f ischemic C R V O , definite i n f o r m a t i o n on retinal capillary obliteration is n o t available in at least one-third o f the cases, because o f extensive retinal hemorrhages, p o o r quality angiograms, the latent period between onset o f ischemic C R V O and development o f extensive retinal capillary nonperfusion, and multiple other limitations [9]. This casts d o u b t on the reliability o f information on the extent o f retinal capillary nonperfusion, and consequently a b o u t the nature o f the C R V O treated, when consecutive eyes with early C R V O were treated immediately with P R P . 7. M o s t o f the papers do n o t give detailed statistical data in s u p p o r t o f their conclusions. 8. In some studies the data in the paper do n o t fully s u p p o r t the conclusions. F o r example, Laatikainen et al. [15] f o u n d no statistically significant difference in the ocular N V complications between the treated and untreated groups in their r a n d o m i z e d study, and yet they concluded that " d e v e l o p m e n t o f late neovascular complications, particularly rubeosis iridis, neovascular glaucoma, and p r o b a b l y disc neovascularisation, were reduced or prevented by panretinal p h o t o c o a g u l a t i o n . " 9. Some studies h a d m a n y unusual, complicated cases o f C R V O . F o r example, 8 o f the 15 cases with C R V O in the P R P study o f Laatikainen [14] h a d intraocular pressures o f 29-60 m m H g at the initial visit (due to pseudoexfoliation or p r i m a r y open angle g l a u c o m a in 6, angle closure g l a u c o m a in 1, and N V G in 1). A very high intraocular pressure, by itself, can p r o d u c e C R V O and, in o u r experience, m a n a g e m e n t o f high intraocular pressure in such a case is usually the p r i m a r y treatment required. We have seen a g o o d response in some such cases with efficient control o f the high intraocular pressure by a n t i g l a u c o m a therapy alone. In spite o f these and other flaws in the studies reported so far, P R P has come to be generally regarded

as the well-established m e t h o d o f treatment in this condition. F r o m this extremely brief review, it is evident that so far we have h a d no large, randomized, prospective, controlled, well-planned, and systematic study evaluating the role o f P R P treatment in eyes fulfilling the very strict criteria for ischemic C R V O [5, 6, 9, 10, 24]. W i t h these as our m a i n objectives, in 1975 we planned the present study o f P R P with a r g o n laser in ischemic C R V O to find out by a systematic, prospective study, whether P R P prevents or reduces the incidence o f development o f ocular NV, especially N V G , and also if it influences the ultimate visual o u t c o m e in these eyes.

Materials and methods This prospective, planned study was performed in the Ocular Vascular Clinic at the University of Iowa, from 1977 to 1986, in 123 eyes with ischemic CRVO.

Study protocol We performed the following two types of prospective studies.

Controlled randomized study. In this study, randomization was done using a random number table, and serially numbered sealed envelopes were provided for consecutive patients who decided to join the study. Each envelope contained a card indicating whether the eye would have PRP treatment or no treatment. No other treatment for CRVO was given in any eye. We originally started with such a study of PRP. After about 2 years, however, we found out that although we were seeing enough patients with ischemic CRVO in our clinic, the number of patients willing to sign up for the controlled randomized study was very low. The major reason for this was the problem of travel, since patients often travel 100 300 miles one way to come to the University Hospitals, the long winter is very severe, and travel can be hazardous; there is little public transport, and most patients with ischemic CRVO are elderly. Many patients, though willing, felt that they could not adhere to the strict time schedule of the protocol for follow-up. Another important factor was lack of continued financial support to conduct the controlled randomized study, because the National Institutes of Health supported the study for only 2 years, and they did not renew funds, on the ground that it was "inappropriate to conduct a major study on vein occlusion that addresses a well-known fact that neovascularization can be eliminated" by PRP. Patient choice study. In view of the problems mentioned above, we had to alter our strategy from the strict controlled randomization to a "patient choice" scheme (or what could be called "patient randomization study"). All patients with ischemic CRVO seen in the Ocular Vascular Clinic at Iowa City were given the same information as for the controlled randomized study and the choice to have PRP or not was left entirely to them. The patients were encouraged to consult their local physician or ophthalmologist to get more information and to help them decide. We did not influence their decision in any way, other than to provide all the same relevant information on the subject as was given to those who were enrolled in the controlled randomized study. These patients were treated, evaluated, and followed according to the same protocol as the controlled randomized study, except that for follow-up the patients were allowed a certain amount of latitude, a few days to weeks, from the times indicated by the study protocol, so that the follow-up schedule was tailored to suit the convenience of the individual patient.

283

Patient population Only patients with a definite diagnosis of ischemic CRVO were included. The diagnosis was based on criteria discussed at length elsewhere [5, 6, 9, 10, 24]: in earlier cases, based on combined information from visual acuity, visual fields, ophthalmoscopy, and fluorescein fundus angiography [5, 6], with the later addition of information obtained from relative afferent pupillary defect [24] and electroretinography [10]. Only patients with active retinopathy, irrespective of the duration since onset, were included in the study. Patients with any possibility of having nonischemic CRVO were excluded from the study, because PRP is contraindicated in those cases. Patients with any evidence of diabetic retinopathy were also excluded. Patients who had had previous PRP at any time, or neovascular glaucoma or vitreous hemorrhage secondary to retinal or optic disc NV were excluded from the study, but patients with iris, angle, retinal, or optic disc NV were not. There were 21 eyes in the controlled randomized study and 102 eyes in the patient choice study.

I2e, I4e and V4e) in all cases. To prevent artifacts, appropriate refraction was used while plotting the visual fields: (a) for central 30 ° visual fields a manifest refraction with appropriate presbyopic correction and (b) for visual fields peripheral to the central 30 ° only manifest refraction for I2e and no corrective lens for I4e and V4e (unless the eye had high myopia or hyperopia, i.e., of 10 diopters or more). Anterior segment, lens, and vitreous were examined by a detailed slit-lamp examination. In the anterior segment, we specifically looked for any NV, apart from any other abnormality. Intraocular pressure was measured by a Goldmann applanation tonometer in both eyes.

Ophthalmoscopic andfluorescein fundus angiographic evaluations were andertaken. A complete ophthalmoscopic examination, with fully dilated pupil, using indirect and direct ophthalmoscopes and biomicroscopy, was conducted to record all the fundus changes in detail. Color fundus photographs were taken. At the baseline visit and at selected intervals during follow-up, fluorescein fundus angiography was performed. The area covered by fundus photography and angiography usually included posterior pole and midperiphery.

Treatment protocol PRP was performed using argon laser, in a scatter fashion similar to the protocol advocated in the Diabetic Retinopathy Study [2], but we used a Rodenstock lens. The presence of marked retinal hemorrhages and edema in these eyes (not a problem in diabetic retinopathy) posed serious technical problems and made it hard to obtain good laser burns in sufficient amounts at one sitting. Ouite often, the area of the retina without dense homorrhages was sparse to begin with. Where there was marked retinal edema, it was like trying to burn a soaking wet filter paper; therefore, it was sometimes hard to obtain an adequate reaction even with maximum power, particularly in eyes with nuclear sclerosis of any importance. In view of these problems, the treatment had to be divided into multiple sessions to obtain an adequate number of burns. The treatment sessions were scheduled as close together as logistically possible (usually at intervals of 1-2 weeks) until adequate treatment was given. The spot size used was usually 500 gm. The power and the exposure time were regulated to obtain an adequate burn. PRP was done outside the macular and optic disc region, extending as far to the periphery as possible within the limits of view of the Rodenstock lens. In the nasal retina, the laser burns were applied from the nasal peripapillary retina to the periphery. In the superior and inferior retina, the burns were applied outside the main temporal vascular arcades, and in the temporal retina from approximately two disc diameter from the fovea temporally.

Protocol f o r baseline and follow-up evaluations At the initial baseline visit, the patients had a complete, detailed review of ocular and systemic history as well as a detailed ocular evaluation required for the differential diagnosis of ischemic CRVO from nonischemic CRVO [9]. The patients were seen 1, 2, 5, 8, 11, 14, 17, 20, 23, 26, 32, 38, 44, and 50 months after the initial visit for follow-up evaluation. Thereafter, the follow-up was every 6 months, or yearly if the retinopathy showed no activity. At the baseline visit and at each follow-up visit the eyes had the following parameters recorded : Visual acuity was recorded, with undilated pupils, using a regular Snellen chart, with the lighting condition and the chart identical throughout the study. Best corrected visual acuity was recorded, urging the patients to fixate eccentrically in view of the presence of central scotoma, making a maximum effect to read the chart, and including change in head position that would make them see best. Visual fields were recorded by a trained perimetrist, using a Goldmann perimeter, and three object sizes and intensities (i.e.,

Data analyses The data were analysed in two ways: (a) for the entire sample of patients (123 eyes) researched prospectively in this study, i.e., by combining the data of patients from both the controlled randomized and patient choice studies, because of the small sample size (21 eyes) in the controlled randomized study itself; and (b) for only the controlled randomized patients.

Data analyses o f entire sample o f patients In both lasered and nonlasered eyes, the data were included up to either the last follow-up visit, or until the eye had cyclocryotherapy for NVG, or until a previously nonlasered-eye patient opted to have PRP, in which case the visit just before PRP was considered the last visit for nonlaser data. We did not use NVG or vitreous hemorrhage as the end-point in this part of the study because we wanted to know the long-term outcome in the laser and nonlaser groups. Thus, the data give information on the long-term outcome, as well as on NVG and vitreous hemorrhage, in the lasered and nonlasered eyes. Our previous prospective natural history studies [71 on the incidence of ocular NV in ischemic CRVO revealed that, using a survival curve analysis, the incidence of anterior segment NV had a linear course from onset to about 200 days after onset, and thereafter the curve became almost flat (Fig. 1). This clearly indicates that the maximum risk of developing anterior segment NV is from the time of onset until approximately 200 days thereafter, and the risk then diminishes markedly (Fig. 1). In view of that, we divided our cases into the following three categories for the purposes of data analysis: Category I: This included patients in whom time from onset of ischemic CRVO to entry into the study was