- Email: [email protected]
Cystoid Macular Degeneration in Experimental Branch Retinal Vein Occlusion
Cystoid Macular Degeneration in Experimental Branch Retinal Vein Occlusion INGOLF H. L. WALLOW, MD, RONALD P. DANIS, MD, COLLEEN BINDLEY, MICHAEL NEIDER, BA
Abstract: Macular edema and collateral vessels were examined clinically and histopathologically up to 48 months after branch retinal vein occlusion in six eyes of five cynomolgus monkeys. In all six, central macular swelling and fluorescein leakage from the retinal vasculature were confined to the acute stage. However, histopathologically, at the chronic stage, only two maculas were completely recovered and unremarkable, whereas the other four showed variable degrees of cystoid degeneration and photoreceptor cell loss. In the two recovered maculas, six to eight normal-sized capillaries separated the fovea from the nearest cluster of capillary collaterals. In three maculas with cystic degeneration, collaterals incorporated the circumfoveal capillaries. In the fourth macula with cystic degeneration, collaterals were separated from the center by two normal-sized capillaries but were also associated with large areas of capillary nonperfusion partially due to occlusion of the macular arteriole. [Key words: atrophy of photoreceptor cells, branch retinal vein occlusion, collateral vessels, cystoid macular degeneration, foveoschisis, foveal hole, microcystoid changes of outer plexiform layer.] Ophthalmology 95:1371-1379,1988
In humans, branch retinal vein occlusion (BRVO) is second only to diabetic retinopathy in the frequency with which it produces retinal vascular abnormalities_ I In almost half of the cases, these abnormalities cause significant macular edema, reducing visual acuity to a level of 20/50 or less. Visual loss may be minimized by timely laser photocoagulation. 2 Initially macular edema is diffuse and reversible. Later, the edema and degeneration of retinal cells are
Originally received: October 1,1987. Revision accepted: April 13, 1988. From the Department of Ophthalmology, University of Wisconsin School of Medicine, Madison. Presented in part at the Annual Meeting of the Association for Research in Vision and Ophthalmology, April 30, 1986, Sarasota. Supported by NIH grant EY-01634, JDF grant #184502, and a grant from the Miller Foundation, Marshfield, Wisconsin (Dr. Wallow), and by training grant P32-EY07059 (Dr. Danis). Reprint requests to I. H. L. Wallow, MD, Department of Ophthalmology, University of Wisconsin-Madison, 600 Highland Avenue, Madison, WI 53792.
often associated with the formation of cystoid spaces, i.e., cystoid macular edema. An animal model of experimental BRVO may provide the best opportunity to clarify in greater detail the subclassification 3 and etiologl-6 of cystoid macular edema. However, despite several extensive short-term studies of experimental BRVO, 7-9 it has been concluded editorially that". . . macular edema was also never observed after the acute stages." 10 Our study shows that cystoid macular degeneration was frequently present histopathologically when the duration of the experimental BRVO was extended. This degeneration has been overlooked previously and may have a relationship to chronic low-grade edema not detectable clinically.
MATERIALS AND METHODS The major veins of one retinal hemisphere (either superior or inferior) were occluded by argon laser photocoagulation in one eye of four cynomolgus monkeys and in both eyes of a fifth animal. Retinal arterioles were not 1371
1372
WALLOW et aI
•
CYSTOID MACULAR DEGENERATION
Fig 2. A and B, early phase of fluorescein angiogram of monkey 1, left eye, at 9 days (A) and 6 weeks (B). Nonleaking capillaries form collateral brushes which involve temporal half of circumfoveal capillaries. Leakage of dye was present from superior venules (arrowheads) and from some dilated capillaries at venular end (arrows). C, At 3 months, pericentral collaterals contain occasional larger channels (arrow). Dye leakage from microvasculature was absent. D, light micrograph of fovea, 3 months after BRVO. Outer plexiform layer contained small cystic spaces (small arrows); the number of photoreceptor cell nuclei was reduced (open arrow). Capillary collaterals (arrowheads) were located throughout inner nuclear layer. A pigment-laden macrophage (large arrow) can be seen (J.5-l'm epoxy section, toluidine blue; original magnification, X 170).
targeted, except in one case, although incidental bums sometimes hit but did not occlude them. Clinical follow-up was done with stereo color fundus photography and stereo fluorescein angiography, typically at 2 weeks, 1 month, and 3 months after BRVO, and thereafter at
3-month intervals. At 3, 6, 12, 151f2, 40, and 48 months after occlusion, a systematic histopathologic evaluation of the macula was done using light microscopy and was combined with review of our clinical data. All procedures were done on anesthetized animals.
(
Fig 1. Top left. monkey 1, left eye. At 9 days after BRVO, macular edema and hemorrhages were still extensive. Top right. same eye at 6 weeks after BRVO. Macular swelling of the superotemporal area and hemorrhages persisted. Notice lines of hard exudates (arrows) and cotton-wool spots (arrowheads). Second row left. same eye at 3 months after BRVO. All clinical findings persisted, although to a lesser degree. Lines of hard exudates were fragmented . Second row right. monkey 2, right eye. At 3 weeks after BRVO. Residual macular swelling was minimal; lines of hard exudate were conspicuous. Third row left. monkey 5, right eye. At 1 day after laser treatment which occluded superior veins and a segment of superotemporal artery. Macular ischemic whitening and infarction were intense. Third row right. same eye at 39 months after BRVO. The macula appeared flat. Bottom left. monkey 1, right eye. At 2 hours after BRVO, showed massive macular edema and considerable retinal hemorrhage. Bottom right. same eye at 11 months after BRVO. A yellow slit-like depression of center was surrounded by slightly irregular pigmentation of paracentral area and drusen. Notice large caliber ofinferotemporal vein and thin sheathed superotemporal vein.
1373
OPHTHALMOLOGY
•
OCTOBER 1988
•
VOLUME 95
•
NUMBER 10
Fig 3. A. fluorescein angiogram of monkey 2, right eye, 5'h months after BRVO. Six normal-sized circumfoveal capillaries separated the fovea from the nearest collateral brush which did not leak dye and contained several larger channels. B. histopathologically, this fovea was unremarkable (1.5-ILm epoxy section, toluidine blue; original magnification, x21O).
Anesthesia consisted of a mixture ofketamine (8 mg/kg) with xylazine (0.6 mg/kg) and acepromazine (0.45 mg/kg) after sedation with ketamine (15 mg/kg). Photographs were taken before and after treatment using either a standard 30° or a 50° fundus camera. Ektachrome 100 film was used for color photographs. Fluorescein angiograms were taken according to a standardized protocol after injection of 10% sodium fluorescein intravenously into an arm or leg vein. The choroidal, arterial, and venous phases were recorded for a predetermined field up to 5 minutes after dye injection. Kodak Tri-X film was used. The color film was processed conventionally, whereas the angiogram films were push-processed using Kodak D-ll developer, diluted 1:l. The animals were euthanatized with an overdose of methohexital sodium.· After enucleation, the eyes were placed in 0.1 M phosphate buffer. The maculas were 1374
excised, fixed in 4% buffered glutaraldehyde, and embedded in an epoxy resin. Serial thick sections, 1.5-JLm thick, were stained with toluidine blue and evaluated by light microscopy.
RESULTS Macular edema beyond the acute stage was not documented in any specimen either by stereo color photographs or stereo fluorescein angiograms. However, four of the six maculas had cystoid spaces of the outer plexiform layer upon histopathologic examination.
CASE REPORTS Monkey 1, left eye. Both eyes were photocoagulated. The right eye will be described later.
WALLOW et al
•
CYSTOID MACULAR DEGENERATION
Fig 4. A and 8, fluorescein angiogram of monkey 3, left eye, at 2 (A) and 12 (8) months after 8RVO. Nonleaking collaterals involved temporal half of the circumfoveal capillaries. At 12 months, larger channels were more frequent, and a second pericentral collateral brush inferonasally was more prominent than at 2 months. C and D, light micrograph of peri fovea (C) and fovea (D) showed large confluent cysts within outer plexiform layer and profound photoreceptor cell loss with central retinal thinning. Notice large collaterals (arrows) corresponding to the fluorescein angiogram (1.5-llm epoxy sections, toluidine blue; original magnification, C, X130; D, X 120).
In the left eye the superior veins were occluded in two laser sessions, 2 weeks apart. Bums were 2 disc diameters (DD) or more from the center. At 9 days, macular edema and hemorrhage was extensive (Fig 1, top left), and pericentral nonleaking collaterals involved the temporal half of the circumfoveal capillaries (Fig 2A). Two superior second-order branch veins leaked fluorescein diffusely. At 6 weeks, considerable edema and hemorrhage persisted, and cotton-wool spots were present between the center and the occluded superior temporal vein. Lines of hard exudates were also noted (Fig 1, top right). Capillary collaterals had the same caliber, course, and extent as before (Fig 2B). Changes at 3 months after occlusion were less conspicuous. The edema was circumscribed to the temporal half of the center; the nasal part had an irregular contour with sharp edges (Fig 1, second row left). Hemorrhages were few, and the lines of hard exudates were broken into clusters of yellow dots. The cotton-wool exudates were fewer in number and smaller in extent. Temporal peri foveal collaterals were slightly more concentrated into fewer, larger channels, but slightly dilated capillaries of the collateral brush seemed to involve the temporal capillaries of the foveal avascular zone (Fig 2C). At this stage, the eye was enucleated, and the macula pro-
cessed for histopathologic evaluation. The contour of the center showed steep rather than gradually sloping edges, which may have been caused by artifacts of tissue processing. Most larger temporal collaterals were located within all levels of the inner nuclear layer (Fig 2D). In serial sections, throughout the center, small cystic spaces of the outer plexiform layer were present. There was considerable nonartifactitious reduction of the number of photoreceptor cell nuclei at the center. Monkey 2. After two laser sessions, 1 month apart, long segments of the inferior veins of the right eye were occluded. Bums were more than 1 DD from the fovea. Three weeks after the second treatment, most of the initial macular edema and hemorrhage had disappeared, but residual edema and part of a yellow-exudate star figure were noted (Fig 1, second row right). The fovea appeared flat with stereo fundus photography. Fluorescein angiography showed several brushes of collaterals temporal to the center; none of which leaked. At 5 months and 1 week, the macula had pigment epithelial mottling and drusen-like spots. The foveola had a diamondshaped appearance but was flat. Fluorescein angiography showed no vascular leakage. The inferior temporal vein was blocked along the entire arcade from the disc to approximately 3 mm temporal to the center. The accompanying artery and its
1375
OPHTHALMOLOGY
•
OCTOBER 1988 •
Fig S. Fluorescein angiogram of monkey 4, left eye, 16 months after BRVO. Several collateral brushes were present along horizontal raphe temporal to macula. Center was separated from the nearest collateral brush by eight or more normal capillaries.
perimacular branches remained open. Temporal to the center, six normal-sized capillaries separated the fovea from the nearest brush of dilated capillary collaterals (Fig 3A). One week later, at 51f2 months after BRVO, the right eye was enucleated and examined histopathologically. The center was unremarkable, and no cystic spaces were seen (Fig 3B). Dilated capillary collaterals were identified in the inner nuclear layer temporal to the center (not included in figure). Monkey 3. After two laser sessions, 6 weeks apart, long segments of the inferior veins of the left eye were occluded with burns 1 DD from the fovea. Two weeks after the second treatment, macular edema and hemorrhage were still prominent, but the fovea appeared to be flat. Fluorescein angiography showed faint leakage from obstructed venules infratemporal to the fovea but no leakage from pericentral capillaries. Two months after occlusion, capillary collaterals touched the temporal border of the foveal avascular zone (FAZ) but did not leak fluorescein (Fig 4A). A second area of pericentral collateralization was present inferonasally. At 12 months, the temporal half of the center showed a slight yellowish discoloration and a subtle pigment irregularity. The retina inferotemporal to the fovea appeared thin and atrophic; no edema was seen. The collaterals at the center were enlarged, involving almost 180 0 of the foveal circumference, but they did not leak (Fig 4B). Results of histopathologic examination at this time showed extensive cystoid spaces of the outer plexiform layer within and around the atrophic center (Fig 4C, D). Monkey 4. Two laser sessions, 6 weeks apart, occluded the inferior veins of the left eye and were 1 DD from the fovea. Macular edema and hemorrhage were prominent 2 weeks after occlusion, and extensive brushes of capillary collaterals appeared temporal to the macula. By week 8, nearly all macular edema and hemorrhage had been absorbed, and in stereo photographs, the fovea appeared to be flat. There was no leakage from the collaterals, which were now separated from the FAZ by several normal-appearing intervening capillaries. The inferior arteries were thin but open.
1376
VOLUME 95
•
NUMBER 10
Little change was noticed up to the final observation at 16 months (Fig 5). The macula appeared normal except for a few small yellowish-white drusen-like spots deep in the retina. Histopathologically, the macula was unremarkable. Monkey 5. The superior veins of the right eye were occluded in three sessions applied over a course of 11'12 months. Heavy photocoagulation burns were placed over segments of the superior temporal artery. Massive ischemic retinal whitening and swelling developed around the compromised superior temporal artery and extended toward the center (Fig 1, third row left). A distal segment of the superior temporal artery later became completely occluded (Fig 6A). Burns were approximately 1 DD from the fovea. Two weeks after the third session, the retina above the center appeared thin. Fluorescein angiography showed extensive capillary nonperfusion superiorly, although the first two macular branches of the superior temporal artery were open. At 6 weeks, dilated collaterals involved the circumfoveal capillaries at the temporal edge of the FAZ. Collaterals and capillary nonperfusion superior to the fovea extended over more than 90 0 of the foveal circumference. There was no vascular leakage. At 3 months, no macular edema or hemorrhage was seen. The retina superior and temporal to the center appeared thin, but a normal foveal reflex was noted. The temporal collaterals were separated from the center by one or two normal-sized capillaries (Fig 6A). Later (39 months after occlusion), collaterals had become more prominent (Fig 1, third row right; Fig 6B). At 40 months the eye was enucleated and showed cystoid spaces of the outer plexiform layer (Fig 6C). In addition, serial sections at different levels showed defects of the inner and outer retinal layers (Fig 6D, E). Monkey 1, right eye. This eye was treated superiorly three times over a period of 11'12 months. Massive transient macular edema and moderate retinal hemorrhages were observed superior to the center after the final treatment (Fig 1, bottom left). At 4 and 6 weeks, central pigmentation was irregular, due to drusen-like white-yellow spots. Numerous nonleaking collaterals were well established by 3 months and included the circumfoveal capillaries (Fig 7A). The center seemed to contain a vertically oriented slit-like yellow depression surrounded by elevated margins (Fig 1, bottom right). The diameter of several lines of circumfoveal capillaries had decreased to normal although other, more remote capillaries had matured into larger-caliber nonleaking collaterals. At 2'12 years the vertical yellow depression in the center remained. Collaterals had further matured. No additional macular changes were noted until 48 months (Fig 7B) when the eye was enucleated. Results of light microscopic examination showed a few pigment epithelial cells of irregular shape and pigmentation and a foveal valley with steep edges and overhanging loops of tissue, as if the foveal rim had contracted (Fig 7C). In addition, there was upward folding of the outer limiting membrane and outer nuclear layer and a cluster of small cystic spaces within the outer plexiform layer (Fig 7C, D). There was mild reduction in the number of central photoreceptor cell nuclei.
DISCUSSION The notion of others lO that "macular edema never occurs after the acute stages of experimental branch retinal vein occlusion (BRVO) in monkeys" was based mainly on clinical observations for up to 63 days7 and
WALLOW et al
• CYSTOID MACULAR DEGENERATION
Fig 6. A and B, fluorescein angiogram of monkey 5, right eye, 3 (A) and 39 (B) months after BRVO. Capillary nonperfusion in areas of obstructed superior macular arterioles was extensive, but two normalsized capillaries separated center from the nearest collateral brush and from the nonperfused zone. C and D, light micrographs of center at 40 months showed confluent cysts of outer plexiform and outer nuclear layers (C). In some sections outer nuclear and outer plexiform layers were discontinuous showing a foveal hole (D). E, from another section near D and showed outer layer hole at higher magnification (1.5-~m, epoxy sections, toluidine blue; original magnification, C, X110; D, X110; E, X21O).
on histopathologic observations, not focused on the macula, for up to 35 days.9 We extended our observations for up to 4 years and added histopathologic evaluations of the macula. We confirm earlier clinical observations that central macular swelling and petaloid angiographic changes signifying fluorescein leakage into cystoid spaces did not occur after the acute stage. Furthermore, no such changes developed even up to 4 years although we might expect a chronic decompensation of the microvascular bed and progressive edema to occur.
Thus, angiographically and biomicroscopically the monkey model is not similar to the typical findings in human macular edema. On the other hand, in four cases results of our histopathologic examination showed microcystoid changes with varying degrees of macular degeneration. Such changes were not shown by fluorescein angiography unless they were accompanied by capillary permeability alteration. Gass ll noted that rapid or marked or prolonged retinal venous obstruction in humans may result 1377
OPHTHALMOLOGY
•
OCTOBER 1988
•
VOLUME 95
•
NUMBER 10
Fig 7. A and B, fluorescein angiogram of monkey I, right eye at 3 months (A) and 4 years (B) after BRVO. At 3 months, collateral brush across the horizontal raphe involved circumfoveal capillaries and contained several larger channels. At 48 months, several lines of circumfoveal capillaries regained normal caliber, and larger channels became centralized into one vessel. C and D, light micrographs of center of 48 months showed cysts of outer plexiform layer and reduced number of photoreceptor cell nuclei (1.5-/Lm epoxy sections, toluidine blue; original magnification, C, X 130; D, XI30).
in extensive atrophy of the inner retinal layers with "fluorescein angiography demonstrating extensive loss of macular capillaries and often little evidence of macular edema." Our monkeys 3 and 5 had inner and outer layer atrophy, but only monkey 5 had an extensive loss of paramacular capillaries. In addition, confluent intraretinal cysts comparable with small schisis cavities were present, resembling the histopathologic appearance of sex-linked juvenile retinoschisis. In this condition, fluorescein angiography of the posterior fundus is frequently normal despite the presence offoveoschisis. 12,13 Another human counterpart is retinitis pigmentosa, where cystoid macular changes may lack or show only minimal fluorescein staining. 14 Thus, these two models might fit a category of macular edema that has also been recognized in human atrophic retinas. The lack of angiographically demonstrable leakage may also explain the paucity of biomicroscopically demonstrable swelling. Our failure to detect cystic spaces clinically might be explained by our method of clinical 1378
examination using solely stereo fundus photographs and by the small size of the cysts, particularly in monkeys 1 and 6. These changes might have been shown through careful slit-lamp examination with high magnification. In our series, the extent of cystoid macular degeneration varied: two eyes had no changes, two had a few small cysts and little photoreceptor cell atrophy, and two had extensive cystoid changes with atrophy. One of the latter (monkey 5) also had foveoschisis and a foveal hole. In the acute phase, a distal segment of the supratemporal artery had become completely occluded, and the initial retinal whitening and swelling had been particularly intense probably causing a degree of inner layer retinal infarction. Even though one or two normal-sized capillaries next to the center persisted and the remaining circumference of the perifoveal capillary bed appeared normal, a sector of parafoveal capillary obliteration developed superiorly over more than 90°. This sector distinguishes monkey 5 from the others. They had no occlusion of paramacular arterioles and no significant
WALLOW et al
•
CYSTOID MACULAR DEGENERATION
areas of capillary nonperfusion close to the center. Macular edema in this case was thus different in nature, representing in its early stage intracellular edema due to acute ischemia. The other five cases were probably secondary to an expansion of the extracellular space due to transient microvascular leakage. Although these other cases shared the same nature of early leakage and edema, their histopathologic outcome varied. We suspect that the location of collaterals in relation to the center may have played a role. In monkeys 2 (inferior vein occlusion) and 4 (superior vein occlusion), six to eight normal-sized capillaries separated the center from the first temporal collateral brush. The maculas were histopathologically unremarkable, indicating that the extensive edema of the acute phase was followed by nearly complete repair. In monkey 1 (both eyes with superior vein occlusion), the most central collateral brush involved the temporal capillaries around the F AZ for 3 months or longer over an extent of approximately 90 0 of the foveal circumference. The macula showed small cysts histopathologically. In monkey 3 (inferior vein occlusion), collaterals incorporated approximately 180 0 of the temporal perifoveal capillaries, and in addition, a second small collateral brush developed approximately 500 ~m infranasal to the center. The histopathologic examination of this case showed many confluent cysts and foveal atrophy. Thus, initial venous occlusion was similar only superficially. Later collateral development showed that the location and extent of obstruction must have varied. The collaterals and macular capillaries in our cases may have shown chronic low-grade insufficiency of the blood retinal barrier not detectable with angiography. Fluorescein angiography seems to be an insensitive indicator of vascular leakage compared with the morphologic tracer, horseradish peroxidase, and subsequent ultrastructural tissue evaluation. Tso and Shih,15 for instance, found peroxidase leakage from macular capillaries in three monkeys after experimental lens extraction although fluorescein angiography was unremarkable. Over time, in our cases where collaterals replaced too many normal capillaries around the F AZ, persistent leaking of serum proteins may have overcome the tissue capacity to resolve fluid accumulation. Recently, we observed that capillary collaterals have an increased endothelial density persisting over more than 1 year,16 suggesting that overall endothelial cell function is altered. If permeability were affected, our cystic degenerative macular changes would fulfill another criterion for qualifying as consequences of macular edema. Factors other than vascular leakage deserve consideration: for instance, the role of Muller cells and the func-
tion of the pigment epithelium, as discussed by Bellhom. 17 The relationship of these cells in our cystic degenerative changes will be addressed in a future study, and detailed evaluations of collaterals regarding their cellularity, wall composition, and permeability are also under way. Thus, although the BRYO monkey model again had not yielded the typical clinical findings of cystoid macular edema in humans, our histopathologic results are new and should revive interest in using this model further.
REFERENCES 1. Orth DH, Patz A. Retinal branch vein occlusion. Surv Ophthalmol 1978; 22:357-76. 2. The Branch Vein Occlusion Study Group. Argon laser photocoagulation for macular edema in branch vein occlusion. Am J Ophthalmol 1984; 98:271-82. 3. Tso MOM. Pathology of cystoid macular edema. Ophthalmology 1982; 89:902-15. 4. Fine BS. Brucker AJ . Macular edema and cystOid macular edema. Am J Ophthalmol1981; 92:466-81. 5. Yanoff M, Fine BS, Brucker A, Eagle RC Jr. Pathology of human cystoid macular edema. Surv Ophthalmol 1984; 28(Suppl):505-11. 6. Gass JDM, Anderson DR, Davis EB. A clinical, fluorescein angiographic, and electron microscopic correlation of cystoid macular edema. Am J Ophthalmol1985; 100:82-6. 7. Hamilton AM, Kohner EM, Rosen D, et al. Experimental retinal branch vein occlusion in rhesus monkeys. I. Clinical appearances. Br J Ophthai mol 1979; 63:377-87. 8. Rosen DA, Marshall J. Kohner EM, et al. Experimental retinal branch vein occlusion in rhesus monkeys. II. Retinal blood flow studies. Br J Ophthalmol 1979; 63:388-92. 9. Hockley DJ, Tripathi RC, Ashton N. Experimental retinal branch vein occlusion in rhesus monkeys. III. Histopathological and electron microscopical studies. Br J Ophthalmol 1979; 63:393-411. 10. Editorial: Retinal vein occlusion [Editorial]. Br J Ophthalmol 1979; 63:375-6. 11. Gass JDM. A fluorescein angiographic study of macular dysfunction secondary to retinal vascular disease. II. Retinal vein obstruction. Arch Ophthalmol 1968; 80:550-68. 12. Burns RP, Lovrien EW, Cibis AB. Juvenile sex-linked retinoschisis : clinical and genetic studies. Trans Am Acad Ophthal Otolaryng 1971 ; 75:1011-21 . 13. Deutman AF. Vitreoretinal dystrophies. In: Krill AE. Krill's Hereditary Retinal and Choroidal Diseases: vol II, Clinical Characteristics. Hagerstown, MD: Harper & Row, 1977; 1044-56. 14. Ffytche TJ. Cystoid maculopathy in retinitis pigmentosa. Trans Ophthalmol Soc UK 1972; 92:265-83. 15. Tso MOM, Shih CoY. Experimental macular edema after lens extraction. Invest Ophthalmol Vis Sci 1977; 16:381-92. 16. Danis RP, Wallow IHL. Microvascular changes in experimental branch retinal vein occlusion . Ophthalmology 1987; 94:1213-21. 17. Bellhorn RW. Analysis of animal models of macular edema. Surv Ophthalmol1984; 28(Suppl):520-4.
1379
Abstract: Macular edema and collateral vessels were examined clinically and histopathologically up to 48 months after branch retinal vein occlusion in six eyes of five cynomolgus monkeys. In all six, central macular swelling and fluorescein leakage from the retinal vasculature were confined to the acute stage. However, histopathologically, at the chronic stage, only two maculas were completely recovered and unremarkable, whereas the other four showed variable degrees of cystoid degeneration and photoreceptor cell loss. In the two recovered maculas, six to eight normal-sized capillaries separated the fovea from the nearest cluster of capillary collaterals. In three maculas with cystic degeneration, collaterals incorporated the circumfoveal capillaries. In the fourth macula with cystic degeneration, collaterals were separated from the center by two normal-sized capillaries but were also associated with large areas of capillary nonperfusion partially due to occlusion of the macular arteriole. [Key words: atrophy of photoreceptor cells, branch retinal vein occlusion, collateral vessels, cystoid macular degeneration, foveoschisis, foveal hole, microcystoid changes of outer plexiform layer.] Ophthalmology 95:1371-1379,1988
In humans, branch retinal vein occlusion (BRVO) is second only to diabetic retinopathy in the frequency with which it produces retinal vascular abnormalities_ I In almost half of the cases, these abnormalities cause significant macular edema, reducing visual acuity to a level of 20/50 or less. Visual loss may be minimized by timely laser photocoagulation. 2 Initially macular edema is diffuse and reversible. Later, the edema and degeneration of retinal cells are
Originally received: October 1,1987. Revision accepted: April 13, 1988. From the Department of Ophthalmology, University of Wisconsin School of Medicine, Madison. Presented in part at the Annual Meeting of the Association for Research in Vision and Ophthalmology, April 30, 1986, Sarasota. Supported by NIH grant EY-01634, JDF grant #184502, and a grant from the Miller Foundation, Marshfield, Wisconsin (Dr. Wallow), and by training grant P32-EY07059 (Dr. Danis). Reprint requests to I. H. L. Wallow, MD, Department of Ophthalmology, University of Wisconsin-Madison, 600 Highland Avenue, Madison, WI 53792.
often associated with the formation of cystoid spaces, i.e., cystoid macular edema. An animal model of experimental BRVO may provide the best opportunity to clarify in greater detail the subclassification 3 and etiologl-6 of cystoid macular edema. However, despite several extensive short-term studies of experimental BRVO, 7-9 it has been concluded editorially that". . . macular edema was also never observed after the acute stages." 10 Our study shows that cystoid macular degeneration was frequently present histopathologically when the duration of the experimental BRVO was extended. This degeneration has been overlooked previously and may have a relationship to chronic low-grade edema not detectable clinically.
MATERIALS AND METHODS The major veins of one retinal hemisphere (either superior or inferior) were occluded by argon laser photocoagulation in one eye of four cynomolgus monkeys and in both eyes of a fifth animal. Retinal arterioles were not 1371
1372
WALLOW et aI
•
CYSTOID MACULAR DEGENERATION
Fig 2. A and B, early phase of fluorescein angiogram of monkey 1, left eye, at 9 days (A) and 6 weeks (B). Nonleaking capillaries form collateral brushes which involve temporal half of circumfoveal capillaries. Leakage of dye was present from superior venules (arrowheads) and from some dilated capillaries at venular end (arrows). C, At 3 months, pericentral collaterals contain occasional larger channels (arrow). Dye leakage from microvasculature was absent. D, light micrograph of fovea, 3 months after BRVO. Outer plexiform layer contained small cystic spaces (small arrows); the number of photoreceptor cell nuclei was reduced (open arrow). Capillary collaterals (arrowheads) were located throughout inner nuclear layer. A pigment-laden macrophage (large arrow) can be seen (J.5-l'm epoxy section, toluidine blue; original magnification, X 170).
targeted, except in one case, although incidental bums sometimes hit but did not occlude them. Clinical follow-up was done with stereo color fundus photography and stereo fluorescein angiography, typically at 2 weeks, 1 month, and 3 months after BRVO, and thereafter at
3-month intervals. At 3, 6, 12, 151f2, 40, and 48 months after occlusion, a systematic histopathologic evaluation of the macula was done using light microscopy and was combined with review of our clinical data. All procedures were done on anesthetized animals.
(
Fig 1. Top left. monkey 1, left eye. At 9 days after BRVO, macular edema and hemorrhages were still extensive. Top right. same eye at 6 weeks after BRVO. Macular swelling of the superotemporal area and hemorrhages persisted. Notice lines of hard exudates (arrows) and cotton-wool spots (arrowheads). Second row left. same eye at 3 months after BRVO. All clinical findings persisted, although to a lesser degree. Lines of hard exudates were fragmented . Second row right. monkey 2, right eye. At 3 weeks after BRVO. Residual macular swelling was minimal; lines of hard exudate were conspicuous. Third row left. monkey 5, right eye. At 1 day after laser treatment which occluded superior veins and a segment of superotemporal artery. Macular ischemic whitening and infarction were intense. Third row right. same eye at 39 months after BRVO. The macula appeared flat. Bottom left. monkey 1, right eye. At 2 hours after BRVO, showed massive macular edema and considerable retinal hemorrhage. Bottom right. same eye at 11 months after BRVO. A yellow slit-like depression of center was surrounded by slightly irregular pigmentation of paracentral area and drusen. Notice large caliber ofinferotemporal vein and thin sheathed superotemporal vein.
1373
OPHTHALMOLOGY
•
OCTOBER 1988
•
VOLUME 95
•
NUMBER 10
Fig 3. A. fluorescein angiogram of monkey 2, right eye, 5'h months after BRVO. Six normal-sized circumfoveal capillaries separated the fovea from the nearest collateral brush which did not leak dye and contained several larger channels. B. histopathologically, this fovea was unremarkable (1.5-ILm epoxy section, toluidine blue; original magnification, x21O).
Anesthesia consisted of a mixture ofketamine (8 mg/kg) with xylazine (0.6 mg/kg) and acepromazine (0.45 mg/kg) after sedation with ketamine (15 mg/kg). Photographs were taken before and after treatment using either a standard 30° or a 50° fundus camera. Ektachrome 100 film was used for color photographs. Fluorescein angiograms were taken according to a standardized protocol after injection of 10% sodium fluorescein intravenously into an arm or leg vein. The choroidal, arterial, and venous phases were recorded for a predetermined field up to 5 minutes after dye injection. Kodak Tri-X film was used. The color film was processed conventionally, whereas the angiogram films were push-processed using Kodak D-ll developer, diluted 1:l. The animals were euthanatized with an overdose of methohexital sodium.· After enucleation, the eyes were placed in 0.1 M phosphate buffer. The maculas were 1374
excised, fixed in 4% buffered glutaraldehyde, and embedded in an epoxy resin. Serial thick sections, 1.5-JLm thick, were stained with toluidine blue and evaluated by light microscopy.
RESULTS Macular edema beyond the acute stage was not documented in any specimen either by stereo color photographs or stereo fluorescein angiograms. However, four of the six maculas had cystoid spaces of the outer plexiform layer upon histopathologic examination.
CASE REPORTS Monkey 1, left eye. Both eyes were photocoagulated. The right eye will be described later.
WALLOW et al
•
CYSTOID MACULAR DEGENERATION
Fig 4. A and 8, fluorescein angiogram of monkey 3, left eye, at 2 (A) and 12 (8) months after 8RVO. Nonleaking collaterals involved temporal half of the circumfoveal capillaries. At 12 months, larger channels were more frequent, and a second pericentral collateral brush inferonasally was more prominent than at 2 months. C and D, light micrograph of peri fovea (C) and fovea (D) showed large confluent cysts within outer plexiform layer and profound photoreceptor cell loss with central retinal thinning. Notice large collaterals (arrows) corresponding to the fluorescein angiogram (1.5-llm epoxy sections, toluidine blue; original magnification, C, X130; D, X 120).
In the left eye the superior veins were occluded in two laser sessions, 2 weeks apart. Bums were 2 disc diameters (DD) or more from the center. At 9 days, macular edema and hemorrhage was extensive (Fig 1, top left), and pericentral nonleaking collaterals involved the temporal half of the circumfoveal capillaries (Fig 2A). Two superior second-order branch veins leaked fluorescein diffusely. At 6 weeks, considerable edema and hemorrhage persisted, and cotton-wool spots were present between the center and the occluded superior temporal vein. Lines of hard exudates were also noted (Fig 1, top right). Capillary collaterals had the same caliber, course, and extent as before (Fig 2B). Changes at 3 months after occlusion were less conspicuous. The edema was circumscribed to the temporal half of the center; the nasal part had an irregular contour with sharp edges (Fig 1, second row left). Hemorrhages were few, and the lines of hard exudates were broken into clusters of yellow dots. The cotton-wool exudates were fewer in number and smaller in extent. Temporal peri foveal collaterals were slightly more concentrated into fewer, larger channels, but slightly dilated capillaries of the collateral brush seemed to involve the temporal capillaries of the foveal avascular zone (Fig 2C). At this stage, the eye was enucleated, and the macula pro-
cessed for histopathologic evaluation. The contour of the center showed steep rather than gradually sloping edges, which may have been caused by artifacts of tissue processing. Most larger temporal collaterals were located within all levels of the inner nuclear layer (Fig 2D). In serial sections, throughout the center, small cystic spaces of the outer plexiform layer were present. There was considerable nonartifactitious reduction of the number of photoreceptor cell nuclei at the center. Monkey 2. After two laser sessions, 1 month apart, long segments of the inferior veins of the right eye were occluded. Bums were more than 1 DD from the fovea. Three weeks after the second treatment, most of the initial macular edema and hemorrhage had disappeared, but residual edema and part of a yellow-exudate star figure were noted (Fig 1, second row right). The fovea appeared flat with stereo fundus photography. Fluorescein angiography showed several brushes of collaterals temporal to the center; none of which leaked. At 5 months and 1 week, the macula had pigment epithelial mottling and drusen-like spots. The foveola had a diamondshaped appearance but was flat. Fluorescein angiography showed no vascular leakage. The inferior temporal vein was blocked along the entire arcade from the disc to approximately 3 mm temporal to the center. The accompanying artery and its
1375
OPHTHALMOLOGY
•
OCTOBER 1988 •
Fig S. Fluorescein angiogram of monkey 4, left eye, 16 months after BRVO. Several collateral brushes were present along horizontal raphe temporal to macula. Center was separated from the nearest collateral brush by eight or more normal capillaries.
perimacular branches remained open. Temporal to the center, six normal-sized capillaries separated the fovea from the nearest brush of dilated capillary collaterals (Fig 3A). One week later, at 51f2 months after BRVO, the right eye was enucleated and examined histopathologically. The center was unremarkable, and no cystic spaces were seen (Fig 3B). Dilated capillary collaterals were identified in the inner nuclear layer temporal to the center (not included in figure). Monkey 3. After two laser sessions, 6 weeks apart, long segments of the inferior veins of the left eye were occluded with burns 1 DD from the fovea. Two weeks after the second treatment, macular edema and hemorrhage were still prominent, but the fovea appeared to be flat. Fluorescein angiography showed faint leakage from obstructed venules infratemporal to the fovea but no leakage from pericentral capillaries. Two months after occlusion, capillary collaterals touched the temporal border of the foveal avascular zone (FAZ) but did not leak fluorescein (Fig 4A). A second area of pericentral collateralization was present inferonasally. At 12 months, the temporal half of the center showed a slight yellowish discoloration and a subtle pigment irregularity. The retina inferotemporal to the fovea appeared thin and atrophic; no edema was seen. The collaterals at the center were enlarged, involving almost 180 0 of the foveal circumference, but they did not leak (Fig 4B). Results of histopathologic examination at this time showed extensive cystoid spaces of the outer plexiform layer within and around the atrophic center (Fig 4C, D). Monkey 4. Two laser sessions, 6 weeks apart, occluded the inferior veins of the left eye and were 1 DD from the fovea. Macular edema and hemorrhage were prominent 2 weeks after occlusion, and extensive brushes of capillary collaterals appeared temporal to the macula. By week 8, nearly all macular edema and hemorrhage had been absorbed, and in stereo photographs, the fovea appeared to be flat. There was no leakage from the collaterals, which were now separated from the FAZ by several normal-appearing intervening capillaries. The inferior arteries were thin but open.
1376
VOLUME 95
•
NUMBER 10
Little change was noticed up to the final observation at 16 months (Fig 5). The macula appeared normal except for a few small yellowish-white drusen-like spots deep in the retina. Histopathologically, the macula was unremarkable. Monkey 5. The superior veins of the right eye were occluded in three sessions applied over a course of 11'12 months. Heavy photocoagulation burns were placed over segments of the superior temporal artery. Massive ischemic retinal whitening and swelling developed around the compromised superior temporal artery and extended toward the center (Fig 1, third row left). A distal segment of the superior temporal artery later became completely occluded (Fig 6A). Burns were approximately 1 DD from the fovea. Two weeks after the third session, the retina above the center appeared thin. Fluorescein angiography showed extensive capillary nonperfusion superiorly, although the first two macular branches of the superior temporal artery were open. At 6 weeks, dilated collaterals involved the circumfoveal capillaries at the temporal edge of the FAZ. Collaterals and capillary nonperfusion superior to the fovea extended over more than 90 0 of the foveal circumference. There was no vascular leakage. At 3 months, no macular edema or hemorrhage was seen. The retina superior and temporal to the center appeared thin, but a normal foveal reflex was noted. The temporal collaterals were separated from the center by one or two normal-sized capillaries (Fig 6A). Later (39 months after occlusion), collaterals had become more prominent (Fig 1, third row right; Fig 6B). At 40 months the eye was enucleated and showed cystoid spaces of the outer plexiform layer (Fig 6C). In addition, serial sections at different levels showed defects of the inner and outer retinal layers (Fig 6D, E). Monkey 1, right eye. This eye was treated superiorly three times over a period of 11'12 months. Massive transient macular edema and moderate retinal hemorrhages were observed superior to the center after the final treatment (Fig 1, bottom left). At 4 and 6 weeks, central pigmentation was irregular, due to drusen-like white-yellow spots. Numerous nonleaking collaterals were well established by 3 months and included the circumfoveal capillaries (Fig 7A). The center seemed to contain a vertically oriented slit-like yellow depression surrounded by elevated margins (Fig 1, bottom right). The diameter of several lines of circumfoveal capillaries had decreased to normal although other, more remote capillaries had matured into larger-caliber nonleaking collaterals. At 2'12 years the vertical yellow depression in the center remained. Collaterals had further matured. No additional macular changes were noted until 48 months (Fig 7B) when the eye was enucleated. Results of light microscopic examination showed a few pigment epithelial cells of irregular shape and pigmentation and a foveal valley with steep edges and overhanging loops of tissue, as if the foveal rim had contracted (Fig 7C). In addition, there was upward folding of the outer limiting membrane and outer nuclear layer and a cluster of small cystic spaces within the outer plexiform layer (Fig 7C, D). There was mild reduction in the number of central photoreceptor cell nuclei.
DISCUSSION The notion of others lO that "macular edema never occurs after the acute stages of experimental branch retinal vein occlusion (BRVO) in monkeys" was based mainly on clinical observations for up to 63 days7 and
WALLOW et al
• CYSTOID MACULAR DEGENERATION
Fig 6. A and B, fluorescein angiogram of monkey 5, right eye, 3 (A) and 39 (B) months after BRVO. Capillary nonperfusion in areas of obstructed superior macular arterioles was extensive, but two normalsized capillaries separated center from the nearest collateral brush and from the nonperfused zone. C and D, light micrographs of center at 40 months showed confluent cysts of outer plexiform and outer nuclear layers (C). In some sections outer nuclear and outer plexiform layers were discontinuous showing a foveal hole (D). E, from another section near D and showed outer layer hole at higher magnification (1.5-~m, epoxy sections, toluidine blue; original magnification, C, X110; D, X110; E, X21O).
on histopathologic observations, not focused on the macula, for up to 35 days.9 We extended our observations for up to 4 years and added histopathologic evaluations of the macula. We confirm earlier clinical observations that central macular swelling and petaloid angiographic changes signifying fluorescein leakage into cystoid spaces did not occur after the acute stage. Furthermore, no such changes developed even up to 4 years although we might expect a chronic decompensation of the microvascular bed and progressive edema to occur.
Thus, angiographically and biomicroscopically the monkey model is not similar to the typical findings in human macular edema. On the other hand, in four cases results of our histopathologic examination showed microcystoid changes with varying degrees of macular degeneration. Such changes were not shown by fluorescein angiography unless they were accompanied by capillary permeability alteration. Gass ll noted that rapid or marked or prolonged retinal venous obstruction in humans may result 1377
OPHTHALMOLOGY
•
OCTOBER 1988
•
VOLUME 95
•
NUMBER 10
Fig 7. A and B, fluorescein angiogram of monkey I, right eye at 3 months (A) and 4 years (B) after BRVO. At 3 months, collateral brush across the horizontal raphe involved circumfoveal capillaries and contained several larger channels. At 48 months, several lines of circumfoveal capillaries regained normal caliber, and larger channels became centralized into one vessel. C and D, light micrographs of center of 48 months showed cysts of outer plexiform layer and reduced number of photoreceptor cell nuclei (1.5-/Lm epoxy sections, toluidine blue; original magnification, C, X 130; D, XI30).
in extensive atrophy of the inner retinal layers with "fluorescein angiography demonstrating extensive loss of macular capillaries and often little evidence of macular edema." Our monkeys 3 and 5 had inner and outer layer atrophy, but only monkey 5 had an extensive loss of paramacular capillaries. In addition, confluent intraretinal cysts comparable with small schisis cavities were present, resembling the histopathologic appearance of sex-linked juvenile retinoschisis. In this condition, fluorescein angiography of the posterior fundus is frequently normal despite the presence offoveoschisis. 12,13 Another human counterpart is retinitis pigmentosa, where cystoid macular changes may lack or show only minimal fluorescein staining. 14 Thus, these two models might fit a category of macular edema that has also been recognized in human atrophic retinas. The lack of angiographically demonstrable leakage may also explain the paucity of biomicroscopically demonstrable swelling. Our failure to detect cystic spaces clinically might be explained by our method of clinical 1378
examination using solely stereo fundus photographs and by the small size of the cysts, particularly in monkeys 1 and 6. These changes might have been shown through careful slit-lamp examination with high magnification. In our series, the extent of cystoid macular degeneration varied: two eyes had no changes, two had a few small cysts and little photoreceptor cell atrophy, and two had extensive cystoid changes with atrophy. One of the latter (monkey 5) also had foveoschisis and a foveal hole. In the acute phase, a distal segment of the supratemporal artery had become completely occluded, and the initial retinal whitening and swelling had been particularly intense probably causing a degree of inner layer retinal infarction. Even though one or two normal-sized capillaries next to the center persisted and the remaining circumference of the perifoveal capillary bed appeared normal, a sector of parafoveal capillary obliteration developed superiorly over more than 90°. This sector distinguishes monkey 5 from the others. They had no occlusion of paramacular arterioles and no significant
WALLOW et al
•
CYSTOID MACULAR DEGENERATION
areas of capillary nonperfusion close to the center. Macular edema in this case was thus different in nature, representing in its early stage intracellular edema due to acute ischemia. The other five cases were probably secondary to an expansion of the extracellular space due to transient microvascular leakage. Although these other cases shared the same nature of early leakage and edema, their histopathologic outcome varied. We suspect that the location of collaterals in relation to the center may have played a role. In monkeys 2 (inferior vein occlusion) and 4 (superior vein occlusion), six to eight normal-sized capillaries separated the center from the first temporal collateral brush. The maculas were histopathologically unremarkable, indicating that the extensive edema of the acute phase was followed by nearly complete repair. In monkey 1 (both eyes with superior vein occlusion), the most central collateral brush involved the temporal capillaries around the F AZ for 3 months or longer over an extent of approximately 90 0 of the foveal circumference. The macula showed small cysts histopathologically. In monkey 3 (inferior vein occlusion), collaterals incorporated approximately 180 0 of the temporal perifoveal capillaries, and in addition, a second small collateral brush developed approximately 500 ~m infranasal to the center. The histopathologic examination of this case showed many confluent cysts and foveal atrophy. Thus, initial venous occlusion was similar only superficially. Later collateral development showed that the location and extent of obstruction must have varied. The collaterals and macular capillaries in our cases may have shown chronic low-grade insufficiency of the blood retinal barrier not detectable with angiography. Fluorescein angiography seems to be an insensitive indicator of vascular leakage compared with the morphologic tracer, horseradish peroxidase, and subsequent ultrastructural tissue evaluation. Tso and Shih,15 for instance, found peroxidase leakage from macular capillaries in three monkeys after experimental lens extraction although fluorescein angiography was unremarkable. Over time, in our cases where collaterals replaced too many normal capillaries around the F AZ, persistent leaking of serum proteins may have overcome the tissue capacity to resolve fluid accumulation. Recently, we observed that capillary collaterals have an increased endothelial density persisting over more than 1 year,16 suggesting that overall endothelial cell function is altered. If permeability were affected, our cystic degenerative macular changes would fulfill another criterion for qualifying as consequences of macular edema. Factors other than vascular leakage deserve consideration: for instance, the role of Muller cells and the func-
tion of the pigment epithelium, as discussed by Bellhom. 17 The relationship of these cells in our cystic degenerative changes will be addressed in a future study, and detailed evaluations of collaterals regarding their cellularity, wall composition, and permeability are also under way. Thus, although the BRYO monkey model again had not yielded the typical clinical findings of cystoid macular edema in humans, our histopathologic results are new and should revive interest in using this model further.
REFERENCES 1. Orth DH, Patz A. Retinal branch vein occlusion. Surv Ophthalmol 1978; 22:357-76. 2. The Branch Vein Occlusion Study Group. Argon laser photocoagulation for macular edema in branch vein occlusion. Am J Ophthalmol 1984; 98:271-82. 3. Tso MOM. Pathology of cystoid macular edema. Ophthalmology 1982; 89:902-15. 4. Fine BS. Brucker AJ . Macular edema and cystOid macular edema. Am J Ophthalmol1981; 92:466-81. 5. Yanoff M, Fine BS, Brucker A, Eagle RC Jr. Pathology of human cystoid macular edema. Surv Ophthalmol 1984; 28(Suppl):505-11. 6. Gass JDM, Anderson DR, Davis EB. A clinical, fluorescein angiographic, and electron microscopic correlation of cystoid macular edema. Am J Ophthalmol1985; 100:82-6. 7. Hamilton AM, Kohner EM, Rosen D, et al. Experimental retinal branch vein occlusion in rhesus monkeys. I. Clinical appearances. Br J Ophthai mol 1979; 63:377-87. 8. Rosen DA, Marshall J. Kohner EM, et al. Experimental retinal branch vein occlusion in rhesus monkeys. II. Retinal blood flow studies. Br J Ophthalmol 1979; 63:388-92. 9. Hockley DJ, Tripathi RC, Ashton N. Experimental retinal branch vein occlusion in rhesus monkeys. III. Histopathological and electron microscopical studies. Br J Ophthalmol 1979; 63:393-411. 10. Editorial: Retinal vein occlusion [Editorial]. Br J Ophthalmol 1979; 63:375-6. 11. Gass JDM. A fluorescein angiographic study of macular dysfunction secondary to retinal vascular disease. II. Retinal vein obstruction. Arch Ophthalmol 1968; 80:550-68. 12. Burns RP, Lovrien EW, Cibis AB. Juvenile sex-linked retinoschisis : clinical and genetic studies. Trans Am Acad Ophthal Otolaryng 1971 ; 75:1011-21 . 13. Deutman AF. Vitreoretinal dystrophies. In: Krill AE. Krill's Hereditary Retinal and Choroidal Diseases: vol II, Clinical Characteristics. Hagerstown, MD: Harper & Row, 1977; 1044-56. 14. Ffytche TJ. Cystoid maculopathy in retinitis pigmentosa. Trans Ophthalmol Soc UK 1972; 92:265-83. 15. Tso MOM, Shih CoY. Experimental macular edema after lens extraction. Invest Ophthalmol Vis Sci 1977; 16:381-92. 16. Danis RP, Wallow IHL. Microvascular changes in experimental branch retinal vein occlusion . Ophthalmology 1987; 94:1213-21. 17. Bellhorn RW. Analysis of animal models of macular edema. Surv Ophthalmol1984; 28(Suppl):520-4.
1379