Branch Retinal Vein Occlusion and Quadratic Variation in Arteriovenous Crossings Richard M. Feist, M.D., B e n j a m i n H. Ticho, M.D., M i c h a e l J. Shapiro, M.D., and Marilyn Farber, Dr.P.H. To explore further the origin and clinically observed regional variation of branch retinal vein occlusion, we studied fluorescein angiograms of 42 patients (42 eyes) with branch retinal vein occlusion and a control popula­ tion of 126 consecutive patients. In a statisti­ cally significant percentage of crossings, the artery was anterior to the vein in those areas of the retina clinically predisposed to branch retinal vein occlusion. Thirty-nine of the 42 patients with branch retinal vein occlusion sites had artery-anterior-to-vein crossings (P = .002), whereas 183 of all 266 arteriovenous crossings in these same eyes were similarly positioned. The artery lay anterior to the vein in significantly more temporal retinal cross­ ings (337 of 457) than nasal retinal crossings (89 of 149; P = .002). Similarly, significantly more superotemporal quadrant crossings (164 of 209) than inferotemporal quadrant cross­ ings (173 of 248) had the artery anterior to the vein (P = .0045). These results suggested that variation in the pattern of arteriovenous crossings may have a role in the clinical distri­ bution of branch retinal vein occlusion. BRANCH RETINAL VEIN OCCLUSION was first de­ scribed by Leber1 in 1877. Various local factors that determine the location of branch retinal vein occlusions have been suggested, includ-

Accepted for publication March 11, 1992. From the University of Illinois at Chicago, Department of Ophthalmology a n d Visual Sciences, Chicago, Illi­ nois. This study was supported by an unrestricted re­ search grant from Research to Prevent Blindness, Inc., N e w York, New York. Reprint requests to Michael J. Shapiro, M.D., Universi­ ty of Illinois at Chicago, Department of Ophthalmology and Visual Sciences Library, 1905 W. Taylor St., Chica­ go, IL 60612.

664

ing inflammation or abnormalities of blood factors,2'4 angulation and narrowing of the vein, 6 the number of arteriovenous crossings, and presence of a crossing in which the artery is anterior to the vein. 6 Koyanagi 7 is credited for first observing, in 1928, that branch retinal vein occlusion develops predominantly at arteriove­ nous crossings. In 1899, Amman 8 noted that approximately two thirds of occlusions develop superotemporally and one third develops inferotemporally. In 1936, Jensen 6 reviewed the literature, his own series of 61 cases of occlu­ sions, and the ophthalmoscopic findings of 50 patients (100 eyes). In his control population, Jensen tabulated but did not analyze the num­ ber of artery-anterior-to-vein and vein-anterior-to-artery crossings for each quadrant. Jensen confirmed the findings of Koyanagi 7 and Am­ man 8 and also observed the following: (1) in 70% of arteriovenous crossings in the normal fundus, the artery is anterior to the vein (that is, on the vitreous side) and (2) an even higher percentage of crossing sites involved with branch retinal vein occlusion have the artery anterior to the vein. Jensen postulated that the predilection of branch retinal vein occlusion for the superotemporal quadrant was caused by the greater number of crossings in this quadrant and the location of the macula temporal to the disk. Others 910 have recently confirmed Jen­ sen's observation of a greater percentage of artery-anterior-to-vein crossings at branch reti­ nal vein occlusion sites, but no one has modi­ fied the explanation of superotemporal predi­ lection. To examine the relative percentage of artery-anterior-to-vein crossings at branch reti­ nal vein occlusion sites vs nonocclusion cross­ ings, we evaluated fluorescein angiograms in a control population and in a population with branch retinal vein occlusion. A regional varia­ tion in the percentage of artery-anterior-tovein crossings was an incidental finding in this study.

© A M E R I C A N JOURNAL OF OPHTHALMOLOGY 113:664-668, JUNE, 1992

Vol. 113, No. 6

Branch Retinal Vein Occlusion

665

Patients and Methods

Results

A control population of 126 consecutive pa­ tients with fluorescein angiograms meeting inclusion criteria was obtained from the pho­ tographic files of 1988 at our institution. Inclusion criteria were clear media, arteriove­ nous phase frames sufficient to determine rela­ tive location of vessels at crossings, and ab­ sence of disk swelling or retinal arterial or venous occlusive disease. Some angiograms contained frames of both eyes, and others con­ tained frames of only one eye; because this might potentially produce variability, only the eye photographed first during the early stages of each angiogram was included in the study. Each angiogram was reviewed independently by two of us (R.M.F. and B.H.T.), and the location, order, and anterior vessel (artery or vein) of each visible crossing site were deter­ mined. First-order crossings were denned as those involving major trunk veins (such as the superotemporal vein). Second-order crossings were denned as those involving veins resulting from a bifurcation of a first-order vein, and third-order crossings were defined as those involving veins resulting from a bifurcation of a second-order vein. Crossings involving smaller veins that joined major veins at right angles were not included in the study. Disagree­ ments between the examiners were discussed and a consensus was reached when possible. Those crossings that were not adequately photographed to allow a consensus were not included in the study. Of this control popula­ tion, 43 patients were male and 83 were female; 49 were black, 55 were white, 21 were Hispanic, and one was a native of India. Primary diag­ noses in these patients included age-related macular degeneration (19 patients) and back­ ground or proliferative diabetic retinopathy (22 patients). A study group of 42 consecutive eyes with branch retinal vein occlusion examined with fluorescein angiography at our institution was studied in a similar manner to the control popu­ lation. Occlusion was superotemporal in 20 of these eyes and inferotemporal in 22. Statistical analysis, comparing the percent­ age of artery-anterior-to-vein crossings be­ tween groups and between quadrants, was per­ formed by chi-square analysis.

Overall, in 70% (426 of 607) of crossings, the artery lay anterior to the vein. Although no significant difference was found between vari­ ous orders of crossing within a quadrant, a significant difference was found between quad­ rants. A significantly higher number of arteryanterior-to-vein crossings were found in the superotemporal quadrant (164 of 209) than in the inferotemporal quadrant (173 of 248; P = .045), and in the two temporal quadrants (337 of 457) when compared to the nasal quadrants (89 of 149; P = .002). The percentage of artery-anterior-to-vein crossings at occlusion sites (39 of 42, 93%) was statistically significant when compared to the control group (426 of 606, 70%; P = .003), to all crossings in eyes with occlusions (183 of 266, 69%; P = .002), and to the combined group of all eyes in this study (609 of 872, 70%; P = .002). In review of Jensen's 6 control group of 50 patients, there were a higher number of arteryanterior-to-vein crossings superotemporally (242 of 330) vs inferotemporally (187 of 270; P = .3) and temporally (429 of 600) vs nasally (252 of 378; P = .13). This variation in Jensen's control patients was not statistically significant.

Discussion Although this study confirmed the previous observation 6,910 that branch retinal vein occlu­ sion crossing sites are artery-anterior-to-vein more often than would be predicted by chance alone, it raised the question of whether regional variation in the retinal circulation may have a role in branch retinal vein occlusion location. Possible explanations for our finding of varia­ tion in arteriovenous crossings included a true regional variation for all retinal arteriovenous crossings, a regional variation for only those crossings central enough to be recorded in a standard 30-degree fluorescein angiogram, a regional variation only for those veins large enough to fit the criteria used in this study, an insufficient sample size, or observer bias. Al­ though regional variation in vessel crossings might exist only inside the region normally included in a 30-degree fluorescein angiogram,

666

AMERICAN JOURNAL OF OPHTHALMOLOGY

i

h

r

June, 1992

/--

i »

V

-A*"

it is here that most clinically relevant branch retinal vein occlusions develop. The finding of statistically significant regional variation in our study but not in Jensen's might be explained, at least partially, by the different methods used. Jensen 6 examined the fundi of his patients using ophthalmoscopy, without the additional clues of crossing anatomy offered by fluorescein angiography. Possible explanations for the clinical predom­ inance of temporal branch retinal vein occlu­ sions include the location of the macula tem­

Fig. 1 (Feist and associates). The central retinal artery is located nasal to the central retinal vein in the optic disk of the 4-month embryo. As the retinal vessels extend through the retina, the arter­ ies would be more likely to reach crossing sites temporal to the disk after veins.

poral to the disk, the greater area of retina temporal to the disk, and the possibility of a regional variation in arteriovenous crossings. We developed a theoretical model of the ori­ gins of regionally variable retinal arteriovenous crossings (Figs. 1 through 3). Two factors essen­ tial to this model are the relatively superficial location of the major retinal vessels and the embryologic development of the retinal vasculature. The central retinal artery is typically located nasal to the central retinal vein at the optic disk. During embryologic development,

Fig. 2 (Feist and associates). The inferonasal retinal vein is often small or absent and the inferotemporal vein is often more nasal than the superotemporal vein. This might predispose to a lower percentage of artery-anterior-to-vein crossings in the in­ ferotemporal quadrant than in the superotemporal quad­ rant.

Vol. 113, No. 6

„,*>.. ^^prjr^OTCT^-^■f"' & W H * p ? ■'-'

^

^T*-M^" ^ t e ^ i m y ^ ■v^P'Px *

u * ^

K-

T

L^I-

■7 . /

667

Branch Retinal Vein Occlusion

Fig. 3 (Feist and associates). The major retinal vessels are located in the superficial nerve fiber layer. During embryologic development, less resistance to cellular migra­ tion might be present ante­ rior to the first vessel to reach a future crossing site.

'7

the retinal vessels extend from the optic disk by the extension of noncanalized endothelial tips. 1113 There is no unanimous agreement on the process of retinal vascular modeling, but the central retinal artery and vein are present in the optic disk by the fourth month of gestation and may extend peripherally either by budding of the veins as suggested by Michaelson, 12 or by remodeling of capillaries derived from vascular mesenchyme as suggested by Ashton. 13 The anterior location of the major retinal vessels is clinically manifest by spontaneous vitreous hemorrhage after posterior vitreous detachment and pathologically by the lack of internal limiting membrane in some areas over­ lying the major retinal vessels. 1418 Because of their more nasal location, arterial tips may be less likely to arrive first at sites in temporal retina than venous tips. If there is less compact tissue anterior to the initial vessel, the second vessel to reach the future site of a vessel crossing may be more likely to be anterior. In this model, arterial vessels would more likely be anterior in areas of the retina that were temporal to the optic disk. This corresponds with our findings. In comparing the venous patterns of the four quadrants, Jensen 6 noted that the inferonasal vein is often absent and the inferotemporal vein often is more nasal than the superotemporal vein. In this model, if the embryonic tips of the inferotemporal vein are farther from their respective quadrant than in the superotemporal quadrant, the arterial tips could have an advantage in the inferotemporal quadrant as compared to the superotemporal

quadrant. This would translate into a slight­ ly lower percentage of artery-anterior-to-vein crossings in the inferotemporal quadrant. This also corresponds with our findings. The origin of branch retinal vein occlusion undoubtedly includes both systemic factors such as hypertension and local anatomic factors such as arteriovenous crossings. Ascertainment of the clinical distribution of occlusions is prob­ ably biased, as temporal occlusions would be more likely to become symptomatic than would nasal occlusions distant from the macula. Be­ cause of the complexity of this issue, the ques­ tion of regional variation in the retinal vasculature deserves greater attention, both in relation to branch retinal vein occlusion and to other retinal vascular entities.

References 1. Leber, T.: Die Krankheite der Netzhaut und des sehnerven. In Graefe, A., and Saemisch, T. (eds.): Handbuch der Gesammten Augenheilkunde, Pathologie und Therapie. Leipzig, Verlag Von Wilhelm Engelmann, 1877, pp. 521-535. 2. Greenwood, A.: Thrombosis of the central reti­ nal vein and its branches. JAMA 82:92, 1924. 3. Koyanagi, Y.: Die Pathologisch Anatomie und Pathogenese de Kruzungsphanomens der Netzhautgefasze bei Hochdruck. Graefes Arch. Ophthalmol. 135:526, 1936. 4. : Veranderungen an der Netzhaut bei hochdruck. Pathol. Anat. 15 Concilium Ophthalmol. 1:145, 1938.

668

AMERICAN JOURNAL OF OPHTHALMOLOGY

5. Ennema, M. C : Over Venesluiting in het Netvlies. Amsterdam, the Netherlands, University of Assen, 1940. Dissertation. 6. Jensen, V. A.: Clinical studies of tributary thrombosis in the central retinal vein. Acta Ophthalmol. l(suppl. 10):1, 1936. 7. Koyanagi, Y.: Die Bedeutung der gefasskreuzung fur die Entschung der Astthrombose der Retinal en Zentalvenen. Klin. Monatsbl. Augenheilkd. 81:219, 1928. 8. Amman, E.: Die Netzhautblungen bei Blut und Gefasser Krankhumgen. Beitrage Augenheilkd. 4:211, 1899. 9. Duker, J. S., and Brown, G. C : Anterior loca­ tion of the crossing artery in branch retinal vein occlusion. Arch. Ophthalmol. 107:998, 1989. 10. Weinberg, D., Dodwell, D. G., and Fern, S. A.: Anatomy of arteriovenous crossings in branch retinal vein occlusion. Am. J. Ophthalmol. 109:298, 1990. 11. Ozanics, V., and Jakobiec, F. A.: Prenatal de­ velopment of the eye and its adnexa. In Tasman, W., and Jaeger, E. A. (eds.): Foundations of Clinical

June, 1992

Ophthalmology, vol. 1. Philadelphia, J. B. Lippincott, 1990, pp. 55-58. 12. Michaelson, I. C : The mode of development of the vascular system of the retina with some observa­ tions on its significance for certain retinal diseases. Trans. Ophthalmol. Soc. U.K. 68:137, 1948. 13. Ashton, N.: Retinal angiogenesis in the human embryo. Br. Med. Bull. 26:103, 1970. 14. Green, W. R.: Vitreoretinal junction. In Ryan, S. J., Glaser, B. M., and Michels, R. G. (eds.): Retina, vol. 3. St. Louis, C. V. Mosby Co., 1989, p. 29. 15. Foos, R. Y.: Vitreoretinal juncture over reti­ nal vessels. Graefes Arch. Clin. Exp. Ophthalmol. 204:223, 1977. 16. Green, W. R.: Vitreoretinal juncture. In Ryan, S. J. (ed.): Retina, vol. 3. St. Louis, C. V. Mosby, 1989, p. 29. 17. Wolfe, E.: The Anatomy of the Eye and Orbit. New York, McGraw-Hill Inc., 1955, p. 417. 18. Wolter, J. R.: Pores in the internal limiting membrane of the human retina. Acta Ophthalmol. 42:971, 1964.

OPHTHALMIC MINIATURE

He murmured reassuringly in public-school English as he squeezed a clear and cooling drop of a n o d y n e into each of Eliott's troubled eyes. He had him sit forward in a comfortable chair and look into a little black machine that rested on a highly polished Chippendale table. H u n t e r - H y d e looked into it from the opposite side and conducted a responsible and leisured search of Eliott's eyes. Then he invited him to sit at his own desk and sketch upon a little pad the things he saw in his eyes. Together they studied the drawings, joking at resemblances to melting airplanes, badly carpentered crucifixes, and a little uroboros w h o couldn't quite make his tail. Finally, with a charming stutter that was worth every one of the ten guineas, he told Eliott he h a d something in his eyes called muscae volitantes, "Which is Latin, you see, for f-flying flies, or, to put it another way, f-flies in flight." Gail Godwin, Dream Children New York, Avon Books, 1983, p. 124

Branch retinal vein occlusion and quadratic variation in arteriovenous crossings.

To explore further the origin and clinically observed regional variation of branch retinal vein occlusion, we studied fluorescein angiograms of 42 pat...
2MB Sizes 0 Downloads 0 Views