OCR Text |
Show Journal of Neuro- Ophthalmology 16( 1): 1- 6, 1996. © 1996 Lippincott- Raven Publishers, Philadelphia Venous Obstruction as the Cause of Retinal/ Choroidal Dysfunction Associated with Arteriovenous Shunts in the Cavernous Sinus Mark J. Kupersmith, M. D., E. Marino Vargas, M. D., Floyd Warren, M. D., and Alejandro Berenstein, M. D. Abstract: Objective: To determine the hemodynamic abnormalities that result in visual acuity loss because of retinal or choroidal dysfunction associated with arteriovenous shunts ( AVS) that involve the cavernous sinus. Methods: Retrospective study of the records of 250 patients with either a carotid cavernous fistula or dural arteriovenous malformation revealed a total of 10 patients with vision loss secondary to retinopathy ( group I) or choroidal effusion ( group II). The cerebral angiograms of these 10 patients and 10 additional patients with visual dysfunction due to elevated intraocular pressure ( group III) were evaluated to determine whether the three groups could be distinguished by a specific vascu-lopathic pattern. Particular attention was given to the extent of thrombosis in the ophthalmic venous system and cavernous sinus. Results: There were angiographic signs of severe thrombosis in the ophthalmic vein in nine patients and in the cavernous sinus in seven patients ipsilateral to the retinal or choroidal lesion. In contrast, in group III, severe thrombosis in the ipsilateral ophthalmic vein and in the cavernous sinus occurred in one and two patients, respectively. Closure of the AVS improved the visual acuity in 1 of 4 patients in group I and in 4 of 5 patients in group II, and normalized the intraocular pressure in all patients. Conclusions: Ophthalmic vein thrombosis, rather than arterialization of the venous system or an arterial " steal," is the principal cause of retinal or choroidal dysfunction associated with AVS to the cavernous sinus. Following AVS closure, visual recovery is more frequent with choroidal effusion or detachment rather than with retinopathy. Key Words: Arteriovenous shunt- Venous obstruction- Retinopathy- Choroidal effusion. Arteriovenous shunts ( AVS) that involve the cavernous sinus frequently cause typical orbital and cavernous sinus dysfunction. The patients complain of blurred or loss of vision, diplopia From the Departments of Ophthalmology ( M. J. K., F. W.), Neurology ( M. J. K.), and Surgical Neuroradiology ( A. B.), New- York University Medical Center; New York Eye and Ear Infirmary, New York, U. S. A.; and Instituto de Oftalmologia ( E. M. V.), Lima, Peru. Supported by the R. L. Kohn Foundation and Research to Prevent Blindness, Inc. Address correspondence and reprint requests to Dr. Mark J. Kupersmith, Professor of Ophthalmology and Neurology, 530 First Avenue, New York, New York 10016, U. S. A. headache, ocular or orbital pain, and a pulsatile bruit. Such AVS typically arise from a carotid cavernous fistula ( CCF) or a dural arterial venous malformation ( DAVM). The clinical findings develop because of the disturbance in the circulation of both the orbit and the cavernous sinus ( 1). Previous reports have described the ophthalmologic abnormalities ( 2- 4) prior to successful treatment ( 5- 7). Attempts to correlate the clinical problems with hemodynamic alterations associated with the AVS ( 8) are rare. The neuro- ophthalmo-logical signs, which include ophthalmoplegia, venous stasis, retinopathy, choroidal effusion, secondary glaucoma, and orbital congestion, are hypothesized to develop as a result of several vascu-lopathic conditions. These clinical abnormalities result from reduced ophthalmic artery perfusion and arterialized venous blood flow induced venous hypertension with a CCF ( 1), and venous hypertension caused by occlusive disease in the cavernous sinus and minor arterialization of venous flow with a DAVM ( 9). In this study we sought to determine if a clinically significant retinopathy or a choroidal effusion or detachment was associated with thrombosis or significant obstruction of the ophthalmic venous system rather than with reduced ophthalmic artery blood flow. For comparison, AVS patients with visual dysfunction from elevated intraocular pressure were also evaluated. SUBJECTS AND METHODS A review of the records of 250 patients with an-giogram- proved DAVM to the cavernous sinus or CCF who were evaluated between 1980 and 1993 by the New York University/ Bellevue Hospital Neuro- Ophthalmology Service revealed a total of 10 patients with vision loss ( visual acuity less than 20/ 40) due to retinal or choroidal dysfunction. None of these patients had an obvious optic neuropathy or glaucomatous cupping. The patients were divided into two groups ( Table 1), five in each, depending on whether the visual disturbance was primarily caused by retinopathy ( hem- 1 2 M. KUPERSMITH ET AL. orrhage and edema involving the macula; group I) or choroidal effusion or detachment ( involving the macula region in addition to peripheral retina; group II). The mean age was 62 years in group I and 71 years in group II. The records of another 10 patients, group III, were selected because of vision loss secondary to elevated intraocular pressure and not to either retinal hemorrhage or choroidal detachment of the posterior pole. The first nine patients meeting this criteria and a tenth, selected because of severe vision loss, composed group III ( mean age 65 years). The tenth case was included as it was the only one in which the patient became blind as a result of high intraocular pressure. All patients had a complete neuro- ophthalmological examination, including perimetry, performed at the time of evaluation for their AVS. None of the patients had iris or retinal neovascularization. All of the patients had arterialized conjunctival vessels of the affected eye. Superselective angiography of both internal and external carotid arteries established the diagnosis and location of a CCF or DAVM. The arterial feeders to each AVS, the cavernous sinus and the connections to the venous system, the ophthalmic vein on each side, and the deep venous system of the brain were evaluated. The arterial injections that optimally demonstrated the cavernous sinus and the ophthalmic venous system of both sides were prospectively analyzed without reference to the nature of the visual abnormality. The time of appearance of the ophthalmic vein ipsilateral to the visual disturbance was compared with the time of appearance of the internal cerebral vein. The internal cerebral vein normally does not opacify early and fills after the superficial cerebral veins. The anterior cavernous sinus on each side was judged to be normal, to have mild thrombosis ( marked irregular appearance of contrast filling), or to have severe thrombosis ( absence of contrast opacification) ( 9). Nonopacification or the abrupt ending of the contrast column was considered a sign of severe obstruction or thrombosis in the ophthalmic vein. The results of analysis of the venous system were organized by clinical grouping. The filling of the ophthalmic artery ipsilateral to TABLE 1. Clinical findings Group I: Case 1 2 3 4 5 Group II: Case 6 7 8 9 10 Group III Case 11 12 13 14 15 16 17 18 19 20 Retinopathy Age 28 67 59 85 73 CCF/ IOP DAVM CCF DAVM DAVM CCF DAVM mmHg 14 16 12 2 30 Choroidal effusion/ detachment Age 72 78 68 73 55 : Secondary gl Age 62 74 63 71 66 58 30 58 74 36 CCF/ IOP DAVM DAVM DAVM DAVM DAVM DAVM aucoma mmHg 25 32 27 18 40 CCF/ IOP DAVM DAVM DAVM DAVM CCF DAVM CCF CCF DAVM DAVM CCF mmHg 33 42 40 35 33 30 60 35 32 30 Proptosis mm 6 4 4 6 7 Proptosis mm 4 5 6 2 6 Proptosis mm 10 7 3 5 3 5 7 1 0 5 Ophthalmoparesis No No No total mild Ophthalmoparesis no no moderate mild total Ophthalmoparesis moderate severe none none none mild severe severe mild moderate Acuity ( initial) FC 20/ 50 FC NLP 20/ 100 Acuity ( initial) 20/ 200 20/ 200 FC 20/ 300 FC Acuity ( initial) 20/ 20 20/ 30 20/ 20 20/ 100* 20/ 70* 20/ 30 blind 20/ 25 20/ 50* 20/ 30 Acuity ( shunt closed) FC 20/ 25 FC NLP - Acuity ( shunt closed) 20/ 30 20/ 200 20/ 70 20/ 70 20/ 30 Acuity ( shunt closed) 20/ 25 20/ 30 20/ 20 not treated not treated 20/ 30 blind 20/ 25 20/ 50 20/ 30 * Acuity reduced secondary to cataract. CCF, carotid cavernous fistula; DAVM, dural arterial venous malformation; FC, finger counting; I0P, intraocular pressure; NLP, no light perception. / Neuro- Ophthalmol, Vol. 16, No. 1, 1996 RETINAL/ CHOROIDAL DYSFUNCTION 3 the affected eye was also assessed. The common or internal carotid artery injection that visualized the ophthalmic artery was evaluated to determine whether the ophthalmic artery filled normally in the arterial phase. Slow emptying was suggested when the visualization of the ophthalmic artery persisted during the capillary or venous phase. Standard intravenous fluorescein angiography was performed in seven patients, all either in group I or II. Treatment of the AVS using embolization ( 1) was attempted in all patients via an intra- arterial ( 6) or intravenous ( 10,11) route. RESULTS Clinical Findings In group I, monocular visual acuity was 20/ 50 in one patient and below 20/ 100 in the other four ( Table 1). Ophthalmoscopy showed diffuse hemorrhagic and ischemic retinopathy ( Fig. 1) in four patients. The fifth, the patient with visual acuity of 20/ 50, had swelling of the inner retinal layers and several punctate hemorrhages, both involving the macula ( Fig. 2). One patient ( case 3) also had a peripheral choroidal detachment. In group II, monocular visual acuity was less than 20/ 100 in all five patients. Ophthalmoscopy showed a choroidal detachment or effusion involving the macula in the five patients, which was confirmed by ultrasonography. Fluorescein angiography showed multiple large areas of capillary nonperfusion ( Fig. IB) in three of four patients in group I and none of three patients studied in group II. The fourth patient investigated in group I had several small areas of capillary nonperfusion. There was no delay in the filling of the FIG. 2. Ophthalmoscopy of the left eye ( case 2, group I) shows faint hemorrhages and edema in the macula. central retinal artery and there were no regions of fluorescein leakage in any patient. In all patients, the fluorescein was seen in the retinal veins within 3 seconds following the appearance of dye in the retinal arteries. In group III, the Snellen acuity test was 20/ 30 or better in six patients and reduced in three patients secondary to cataracts. One patient ( case 17) had no light perception because a central retinal artery occlusion developed when the intraocular pressure became greater than 60 mm Hg. Visual field defects, primarily generalized constriction, were found in nine patients. Varying degrees of ophthalmoplegia were seen in seven patients. No patient had retinopathy or choroidal effusion involving the posterior pole except for the signs of the central retinal artery occlusion in case 17. Angiography Angiography findings ( Table 2) showed the AVS to be ipsilateral in five, contralateral in four, and -?$* ViW FIG. 1. A: Ophthalmoscopy of the left eye ( case 1, group I) shows confluent hemorrhages and cotton wool spots in the retina in all four quadrants and in the macula. B: Fluorescein angiography of the left eye ( case 1, group I) appears to show multiple areas of capillary nonperfusion. J Neuro- Ophthalmol, Vol. 16, No. 1, 1996 4 M. KUPERSM1TH ET AL. bilateral in one patient, in groups I and II. The AVS was ipsilateral in nine and bilateral in one patient, with respect to the involved eye in group III. In groups I and II, the ophthalmic vein ipsilateral to the affected eye filled during the arterial phase in two patients. The vein appeared at the same time as the internal cerebral vein in five patients. In three patients, the ophthalmic vein was not opacified by the injection of either of the internal or external carotid arteries because of thrombosis. In groups I and II, severe thrombosis was noted in the ipsilateral ophthalmic vein in nine and the ipsilateral anterior cavernous sinus in seven patients ( Fig. 3). In the internal carotid artery study, the ipsilateral ophthalmic artery filled early in the arterial phase in all patients. This artery remained opacified during the capillary or venous phase in only one patient ( case 1). In group III, in relation to the internal cerebral vein, the ophthalmic vein ipsilateral to the affected eye filled from the cavernous sinus at the same time in one, earlier in five, and not at all in one patient. In the three other patients the ophthalmic vein was only opacified by injection of the ipsilateral external carotid artery so the filling relationship to the internal cerebral vein could not be determined. Nine patients had minor or no evidence of thrombosis in the ipsilateral ophthalmic vein ( Fig. 4). Two patients had severe thrombosis in the ipsilateral anterior cavernous sinus. In eight patients, the ophthalmic artery was visualized and filled during the early arterial phase. Each of the two other patients had a large fistula and the ophthalmic artery was not demonstrated. No patient had delayed emptying of the ophthalmic artery. Clinical Results of Treatment Nine patients, four in group I and five in group II, had AVS closure, six after treatment by embolization and three spontaneously. After shunt closure, the vision in one patient in group I and four patients in group II improved within 3 months ( Ta- TABLE 2. Angiography findings 1 II III G Ipsi 3 2 9 roup AVS location Contra 2 2 Bi 1 1 Ipsilateral ophthalmic vein occlusion Severe 5 4 1 Mild 1 9 Ipsilateral cavernous sinus occlusion Severe 3 4 2 Mild 2 1 8 Ophthalmic vein filling compared to internal cerebral vein Early 2 5 Same Late 3 2* - 4 * * Ipsi'. ipsilateral; Contra: contralateral; Bhbilateral. * Ophthalmic vein not opacified ( thrombosed) in three cases by any injection. ** Ophthalmic vein not opacified in one case by any injection; ophthalmic vein opacified by external carotid injection only, so this relationship could not be determined in three cases. / Neuro- Ophthalmol, Vol. 16, No. 1, 1996 ble 1). In group I, case 2 also had resolution of a paracentral scotoma, the retinal hemorrhages, and edema. Ophthalmoscopy did not demonstrate improvement in the other four patients in group I. In group II, the choroidal detachment resolved within 3 months in all cases. Following spontaneous thrombosis of the DAVM after a plane flight, one patient ( case 8, Group II) also had resolution of a third nerve paresis. Another ( case 10, Group II) had resolution of complete ophthalmoplegia after embolization therapy. Eight of the patients in group III had AVS closure by embolization. The intraocular pressure normalized within 1 week in all eight patients, five of whom had improvement in visual field abnormalities. All eight patients had pretreatment ocular motility and eyelid dysfunction, which remitted. DISCUSSION Vision loss from retinopathy or choroidal detachment in patients with cavernous sinus region AVS was associated with stasis and increased pressure in the ophthalmic venous system caused by severe thrombosis in the ipsilateral ophthalmic vein. Significant thrombosis was also commonly found in the ipsilateral anterior cavernous sinus, which would normally drains this ophthalmic vein. Previous reports hypothesized that transmission of the ophthalmic venous system hypertension to the vortex veins is the cause of congestion, ocular edema, or detachment of the choroid ( 12- 16). It appears that the venous hypertension caused by arterialized venous flow in the presence of a patent ophthalmic vein has a different effect than when venous thrombosis occurs. The findings of severe orbital congestion and intraocular pressure elevation without clinical signs of retinopathy in group III suggest that the episcleral veins may be more sensitive than the retinal veins to an elevation of ophthalmic venous pressure or that another pathological mechanism, ophthalmic vein obstruction, is necessary to cause clinical retinal dysfunction. RETINAL/ CHOROIDAL DYSFUNCTION 5 FIG. 3. A: The capillary phase of the lateral internal carotid artery angiogram in case 1 ( group I) showed persistence of contrast media in the cavernous sinus ( broad arrow) and a failure ( narrow arrow) of the contrast media to pass into the orbital portion of the ophthalmic vein. B: The capillary phase of the lateral internal carotid artery ( open arrow) angiogram in case 2 ( group I) shows persistence of contrast media in the cavernous sinus ( broad arrow) and a failure of the contrast media ( smaller arrow) to pass into the orbital portion of the ophthalmic vein. It has been suggested that vision loss can also result if a high flow AVS, supplied by the ophthalmic artery ( 17,18), shunts arterial blood away from the normal arteries directly into the veins, bypassing the intraocular capillaries. In patients with AVS, not directly supplied by the ophthalmic artery, an arterial steal has also been hypothesized to be an important factor that pathologically narrows the arterial or venous blood and oxygen gradient in the ocular tissues ( 8,19). The resulting ischemia and hypoxia affects the retina and choroid, anterior segment, eye muscles, pupillary sphincter, and possibly the intracranial optic nerve. However, similar retinopathy has been seen in patients FIG. 4. In contrast to the angiograms of patients in groups I and II, in group III severe thrombosis of the ophthalmic venous system was seen in only one case. A: Lateral internal carotid artery ( broad arrow) angiogram shows a high flow CCF ( case 14) draining into the cavernous sinus ( arrow) with complete opacification of the ophthalmic venous system ( asterisk). B: Lateral external carotid injection shows a DAVM ( case 19) supplied by the middle meningeal artery ( white arrow) with opacification of a portion of the cavernous sinus ( broad arrow) and the orbital ophthalmic vein ( arrow). / Neuro- Ophthalmol, Vol. 16, No. 1, 1996 6 M. KUPERSMITH ET AL, with DAMVs who had only minor amounts of arteriovenous shunting ( 19). Prior studies have not correlated the clinical retinal or choroidal disorder with the angiographic findings at the time of the visual loss before treatment. One study documented the association of thrombosis in the cavernous sinus, particularly in the posterior compartment, and signs of orbital congestion, but the retinal or choroidal disturbances were not considered ( 9). In two reports, the published angiograms of two patients ( 14,20) appear to demonstrate thrombosis of the ophthalmic vein. These two cases support a prominent role for ophthalmic venous obstruction in the pathophysiology of vision loss from retinal and choroidal dysfunction. Our results suggest that the retina is more susceptible to severe reduction of outflow from the ophthalmic venous system than to a hypothesized arterial steal. Delayed or absent filling of the ophthalmic artery was not seen in any patient with retinal or choroidal dysfunction. Also, in our patients with severe retinopathy, fluorescein angiography showed normal flow into the central retinal artery. Thus, except in the previously reported patients who had surgical procedures to occlude the internal carotid artery ( 8,21), the role of an arterial steal probably is minor in comparison to the effects of stagnant hypoxia caused by venous outflow obstruction in inducing retinal damage. In addition, none of our patients had signs of ocular arterial hypoperfusion, retinal microaneurysms, and neovascularization, which have been infrequently reported ( 22,23). Hypertension in the ophthalmic venous system, which raises the episcleral venous pressure, is considered to cause elevation of intraocular pressure ( 1,24,25). In our patients without major thrombosis in the ophthalmic venous system, arterialized flow was the principal cause of the venous hypertension. Although we randomly chose patients from a pool of 225 patients with elevated intraocular pressure, our sample might be biased. It is possible that other cases without retinopathy or • choroidal detachment might have ophthalmic venous system obstruction or thrombosis as the cause of the ocular hypertension. We would not predict nor expect that all cases with ophthalmic venous system obstruction will develop retinopathy or choroidal effusion. However, when these clinical abnormalities are found, thrombosis in the ophthalmic venous system should be suspected. In this setting, closure of the pathological AVS is more likely to result in restoration of vision in patients with choroidal detachment than with retinopathy. REFERENCES 1. Kupersmith, MJ. Neurovascular Neuro- ophthalmology. Heidelberg: Springer Verlag, 1993; 69- 140. 2. deSchweinitz GA, Holloway TB. Pulsating Exophthalmos. Philadelphia: WB Saunders, 1908; 11- 120. 3. Henderson JW, Schneider RC. The ocular findings in carotid cavernous fistula in a series of 17 cases. Am } Ophthalmol 1959; 48: 585- 97. 4. Sattler CH. Beitrag zur kenntnis des pulsierenden ex-opohthalmus. Ztschr f Augenh 1920; 43: 534- 52. 5. Debrun G, Lacour P, Vinuela F. Treatment of 54 traumatic carotid- cavernous fistulas. / Neurosurg 1981; 55: 678- 92. 6. Kupersmith MJ, Berenstein A, Choi IS, et al. Management of nontraumatic vascular shunts involving the cavernous sinus. Ophthalmology 1988; 95: 121- 30. 7. Kupersmith MJ, Berenstein A, Flamm E, Ransohoff J. Neu-roophthalmologic abnormalities and intravascular therapy of traumatic carotid cavernous fistulas. Ophthalmology 1986; 93: 906- 12. 8. Sanders MD, Hoyt WF. Hypoxic ocular sequelae of carotid-cavernous fistulae. Study of the causes and failure before and after neuro- surgical treatment in a series of 25 cases. Br J Ophthalmol 1969; 53: 82- 97. 9. Grove AS. The dural shunt syndrome. Pathophysiology and clinical course. Ophthalmology 1983; 90: 31^ 44. 10. Manelfe G, Berenstein A. Treatment of carotid cavernous fistulas by venous approach. / Neuroradiol 1980; 7: 13- 21. 11. Hanneken AM, Miller NR, Debrun GM, Nauta HJW. Treatment of carotid- cavernous sinus fistulas using a detachable balloon catheter through the superior ophthalmic vein. Arch Ophthalmol 1989; 107: 87- 92. 12. Costin JA, Weinstein MA, Berlin AJ, et al. Dural arteriovenous malformations involving the cavernous sinus: A case report. Br J Ophthalmol 1978; 62: 478- 82. 13. Guerry D, Harbison JW. Bilateral choroidal detachment and fluctuating proptosis secondary to bilateral dural arteriovenous fistulas treated with transcranial orbital decompression with resolution: Report of a case. Trans Am Ophthalmol Soc 1975; 73: 64- 73. 14. Harbison JW, Guerry D, Weisinger H. Dural arteriovenous fistula and spontaneous choroidal detachment: New cause of an old disease. Br J Ophthalmol 1978; 62: 483- 90. 15. Jorgensen JS, Gutthoff RF. 24 cases of carotid cavernous fistulas: Frequency, symptoms, diagnosis and treatment. Acta Ophthalmologica 1985; 63: 67- 71. 16. Klein R, Meyers SM, Smith JL, et al. Abnormal choroidal circulation. Association with arteriovenous fistula in the cavernous sinus area. Arch Ophthalmol 1978; 96: 1370- 73. 17. Bogousslavsky J, Vinuela F, Barnett HJM, Drake CG. Amaurosis fugax as the presenting manifestation of dural arteriovenous malformation. Stroke 1985; 10: 891- 3. 18. Halbach VV, Higashida RT, Hieshima GB, et al. Dural arteriovenous fistulas supplied by ethmoidal arteries. Neurosurgery 1990; 26: 816- 23. 19. Newton TH, Hoyt WF. Dural arteriovenous shunts in the region of the cavernous sinus. Neuroradiology 1970; 1: 71- 81. 20. Sergott RC, Grossman RI, Savino PJ, et al. The syndrome of paradoxical worsening of dural- cavernous sinus arteriovenous malformations. Ophthalmology 1987; 94: 205- 12. 21. Kalina RE, Kelly WA. Proliferative retinopathy after treatment of carotid- cavernous fistulas. Arch Ophthalmol 1978; 96: 2058- 60. 22. Harris MJ, Fine SL, Miller NR. Photocoagulation treatment of proliferative retinopathy secondary to carotid- cavernous fistula. Am ] Ophthalmol 1980; 90: 515- 18. 23. Weiss DI, Shaffer RN. Neovascular glaucoma complicating carotid- cavernous fistula. Arch Ophthalmol 1963; 69: 304- 7. 24. Nordmann J, Lobstein A, Gerhard JP, Levy JP. A propos de 14 cas de glaucome par hypertension veineuse d'origine extraoculaire. Ophthalmologica 1981; 142( suppl): 501- 5. 25. Phelps CD, Armaly MF. Measurement of episcleral venous pressure. Am ] Ophthalmol 1978; 85: 35- 42. / Neuro- Ophthalmol, Vol. 16, No. 1, 1996 |