Frosted Branch Angiitis Associated With Cytomegalovirus Retinitis Richard F. Spaide, M.D., Albert T. Vitale, M.D., Ilona R. Toth, M.D., and Jorge M. Oliver, M.D.
We examined three patients with acquired immunodeficiency syndrome who had frosted branch angiitis associated with small patches of cytomegalovirus retinitis. Each patient had a low CD4-helper T-lymphocyte count and a T-lymphocyte helper-suppressor ratio of less than 0.1. Treatment with intravenous anticytomegalovirus antibiotics caused the vascular sheathing to resolve within two weeks in all three patients, but each patient continued to have a smoldering retinitis. Retinal biopsy in one of the patients demonstrated virions whose morphologic characteristics were con sistent with cytomegalovirus on electron mi croscopy and the identity of which was con firmed by immunohistochemistry. Although frosted branch angiitis in otherwise healthy patients responds to corticosteroids, similar treatment with corticosteroids for frosted branch angiitis associated with cytomegalovi rus retinitis in patients with AIDS does not seem to be indicated. Before corticosteroid treatment is started for a patient with the clinical signs and symptoms of frosted branch angiitis, careful medical examination of the patient is necessary. was first described in 1976 by Ito and associates. 1 Their patient was a 6-year-old boy who had bilateral uveitis with profound sheathing of the retinal vessels. Since then, 11 additional cases have been de.FROSTED BRANCH ANGIITIS
Accepted for publication March 5, 1992. From the Departments of Ophthalmology (Drs. Spaide and Vitale) a n d Pathology (Drs. Toth a n d Oliver), St. Vincent's Hospital a n d Medical Center of N e w York, N e w York, New York. This study was presented in part at the East Coast Uveitis Club meeting, Jan. 3 1 , 1992, Baltimore, Maryland. Reprint requests to Richard F. Spaide, M.D., Suite 506, O'Toole Bldg., St. Vincent's Hospital and Medical Cen ter of New York, 36 Seventh Ave., New York, NY 10011.
scribed,2"6 but no consistent systemic abnormal ity has been found. We studied three patients with AIDS who had frosted branch angiitis associated with small areas of cytomegalovirus retinitis. The clinical course, proposed cause, and treatment of these cases differed from the previously described cases.
Case Reports Case 1 A 36-year-old homosexual man with AIDS and a history of pneumocystosis had a fourweek history of blurred vision. Ocular examina tion showed a visual acuity of 20/20 in each eye. There were no cells in the anterior chamber of either eye. He had mild vitreitis in both eyes. Nasal to the optic disk in the right eye was a 3 x 4-disk-diameter area of retinitis and retinal hemorrhage consistent with cytomegalovirus retinitis. In the left eye, two foci of active cytomegalovirus retinitis measuring 2 x 3 and 3 x 3 disk diameters, respectively, were located directly inferior and inferotemporal to the inferotemporal arcade, respectively. Cytomegalo virus had not been diagnosed before initial examination. The veins of both eyes had thick perivascular infiltrate and the arteries were also involved in areas close to the retinal veins (Figs. 1 and 2). Fluorescein angiography showed leak age from both the retinal arteries and veins, although the perivenular leakage was more prominent (Fig. 3). Fluorescein also leaked from the optic disk. Laboratory evaluation disclosed a white blood cell count of 4,200/mm 3 . The absolute CD4 helper count was 10/mm 3 and the CD8 suppressor cell count was 340/mm 3 , giving a helper-suppressor ratio of 0.029. The fluores cent treponemal antibody-absorption test and sputum smears for acid-fast bacilli were nega tive.
© A M E R I C A N JOURNAL O F OPHTHALMOLOGY 113:522-528, M A Y , 1992
Vol. 113, No. 5
Frosted Branch Angiitis and Cytomegalovirus Retinitis
Fig. 1 (Spaide and associates). Case 1. Posterior pole of the left eye showing frosted branch angiitis.
Fig. 2 (Spaide and associates). Case 1. Superior retina of the left eye demonstrating the thick perivas cular sheathing, primarily affecting the retinal veins, that was found throughout both eyes.
Induction treatment with intravenous ganciclovir caused complete regression of the perivascular sheathing in 11 days in the right eye and 14 days in the left eye. Induction treatment was continued for three weeks because of con tinuing active retinitis after two weeks of ganciclovir treatment. He did not have discernible retinal atrophy or retinal pigment epithelial scarring in the immediate perivascular areas of either eye in areas not involved with typical cytomegalovirus retinitis. During the next four months, the patient retained a visual acuity of 20/20 in both eyes. The patient switched care to a hospital closer to his home and did not return to our hospital.
Cytomegalovirus retinitis had not been diag nosed in the patient before initial examination. In each eye extensive white exudates sur rounded the vessels; the veins were more prom inently involved than were the arteries (Fig. 4). Fluorescein angiography showed leakage from
Case 2 A 50-year-old homosexual man with AIDS, cutaneous Kaposi's sarcoma, infection of the oral tissues with Candida albicans, and a history of pneumocystosis noted spots in the field of vision in both eyes. Ocular examination three weeks after the onset of symptoms showed a visual acuity of 20/20 in both eyes. There was mild vitreitis in each eye, but no cells were observed in the anterior chambers. In the right eye, two 2 x 3-disk-diameter areas of retinitis, one nasal to the disk and one superotemporal to the fovea, were observed. In the left eye, one 2 x 5-disk-diameter area in the temporal periph ery and a 1 x 2-disk-diameter area of retinitis superonasal to the optic disk were observed.
Fig. 3 (Spaide and associates). Case 1. Fluorescein angiogram of the left eye showing leakage from the retinal arteries and veins. There is prominent stain ing of the disk.
AMERICAN JOURNAL OF OPHTHALMOLOGY
Fig. 4 (Spaide and associates). Case 2. Posterior pole of the right eye showing a wedge-shaped area of cytomegalovirus retinitis, measuring 2 x 3 disk di ameters, superotemporal to the fovea. the disk, subtle leakage from the arteries, and more marked leakage from the veins. Hematologic testing demonstrated a white
blood cell count of 3,500/mm 3 , an absolute CD4 helper cell count of 10/mm 3 , and a CD8 suppressor cell count of 620/mm 3 with a help er-suppressor ratio of 0.016. A fluorescent treponemal antibody-absorption test and spu tum smears for acid-fast bacilli were negative. Induction treatment with intravenous ganciclovir caused complete regression of the vascu lar sheathing in ten days. However, the induc tion dose of intravenous ganciclovir was continued beyond the normal two-week period because the retinitis continued to show moder ate activity. At three weeks of induction, the retinitis was minimally active, but an acute retinal detachment developed in the left eye. Vasculitis was not evident at this time. The retinal detachment was successfully repaired using vitrectomy techniques. A retinal biopsy specimen was taken at the time of vitrectomy from an area of normal retina adjacent to an area of retinitis (Fig. 5). Numerous inclusions were seen by light mi croscopy in the ganglion, and inner and outer nuclear layers. No inclusion bodies were noted in the capillary endothelial cells. Capillaries and one small venule, but no large vessels, were observed in the examined specimen. Intranu clear inclusions containing scattered viral nu-
Fig. 5 (Spaide and associates). Case 2. Photomicrograph of a retinal biopsy specimen demonstrating intracytoplasmic (small arrow) and round intra nuclear (large arrow) viral inclusion bodies consistent with cytomegalovi rus retinitis (hematoxylin and eosin, X400).
Vol. 113, No. 5
Frosted Branch Angiitis and Cytomegalovirus Retinitis
Fig. 6 (Spaide and associates). Case 2. Electron photomicrograph of a retinal biopsy specimen. The intranuclear inclusions containing scattered nucleocapsids measuring 150 to 200 nm, the relative electron lucency of the central core, and the typical envelope are consistent with cytomegalovirus (x 62,000). cleocapsids and nucleolar-associated nucleo capsids, as well as cytoplasmic dense bodies, were found on electron microscopy (Fig. 6). Immunohistochemical staining confirmed cyto megalovirus retinitis, but unfortunately no large blood vessels were present for analysis. Neither cytomegalovirus antigen nor immune complexes were found in the retinal capillaries in this biopsy specimen using immunohisto chemical stains, however. Treatment of the patient was then switched to intravenous foscarnet. Four weeks after initial examination, a small inferonasal rhegmatogenous detachment developed in the right eye at the border of a previously healed area of retini tis. This detachment was delimited by laser photocoagulation. Nine weeks after initial ex amination, the patient had minimal retinitis and a delimited nasal detachment in the right eye. His retina was attached in the left eye and he had no active retinitis. His best-corrected visual acuity was R.E.: 20/25 and L.E.: 20/40. In the regions of the retina not directly involved with typical cytomegalovirus retinitis, no reti nal atrophy or retinal pigment epithelial altera tions were observed around the blood vessels.
Case 3 A 28-year-old human immunodeficiency vi rus-positive transsexual (genetic male) had de creased vision in the right eye at initial exami nation; the visual acuity was R.E.: hand motions and L.E.: 20/20. No cells were ob served in the anterior chamber of either eye, but there was a trace of vitreous cells in the right eye. The patient had cytomegalovirus retinitis affecting the macula, optic disk, and temporal arcades of his right eye. Peripheral to these areas of retinitis, the retinal veins were encased with prominent white sheathing (Fig. 7). The left eye had two cotton-wool spots, but no cytomegalovirus retinitis and no vascular sheathing. The patient was not known to have a previous diagnosis of either cytomegalovirus retinitis or AIDS before the initial examination. The absolute CD4 count was 20 cells/ mm 3 and the CD8 count was 180 cells/mm 3 , resulting in a helper-suppressor ratio of 0.09. The fluorescent treponemal antibody-absorp tion test and sputum smears for acid-fast bacilli were negative. Treatment with intravenous foscarnet was initiated, but the patient showed clinical signs
AMERICAN JOURNAL OF OPHTHALMOLOGY
Fig. 7 (Spaide and associates). Case 3, right eye. Pronounced sheathing of the vessels is evident pe ripheral to a patch of cytomegalovirus retinitis. of increasing renal failure while being treated. Retinal sheathing in the right eye was almost completely resolved after nine days of foscarnet treatment, but there were still areas of active retinitis (Fig. 8). On the tenth day of intrave nous foscarnet treatment, the patient had no sheathing of the vessels, but treatment had to be discontinued because he developed extreme fatigue, neutropenia, and worsening renal fail ure. His renal function and neutrophil count improved during the next week, but he refused additional treatment with foscarnet.
Discussion The first cases of frosted branch angiitis were described in Japanese children who had a rapid decrease in their visual acuity associated with the frosted branch angiitis. 13 Medical examina tion of these patients did not disclose an etiologic agent. Treatment with corticosteroids was thought to be beneficial. 3 Kleiner and associates 4 described three pa tients from the United States who ranged in age from 23 to 29 years and who had frosted branch angiitis and decreased visual acuity. One pa tient had an increased anti-streptolysin O titer that was normal on repeat testing six days later; otherwise no systemic abnormality was found in any patient. Of special interest, results of all viral studies in their patients were negative; the
Fig. 8 (Spaide and associates). Case 3. After nine days of treatment with intravenous foscarnet, the retinitis appeared more granular, but was still opaque. The vein near the top of the photograph is the same vein illustrated in the pretreatment photo graph (Figure 7). first patient had negative urine and blood cul tures for virus, and was seronegative for anti bodies against HIV. The second patient had negative acute and convalescent viral antibody titers, while the third had negative cytomegalo virus titers. 4 All patients had rapid improve ment of clinical signs and symptoms when treated with oral corticosteroids. 4 Sugin and associates 6 described one 32-yearold patient and one 26-year-old patient who had unilateral frosted branch angiitis associat ed with ipsilateral decreased visual acuity. The first patient was hospitalized five weeks before his loss of vision for what was thought to be a viral syndrome. His white blood cell count at the time of admission was 20,000 cells/mm 3 with 80% polymorphonuclear leukocytes, how ever, which is a degree of granulocytosis not commonly seen in viral infections. The second patient had no systemic abnormality and was seronegative for antibodies against HIV. Both patients seemed to have an improvement in their clinical signs and symptoms after treat ment with oral corticosteroids. The authors thought early diagnosis and treatment with corticosteroids was imperative. 6 Vander and Masciulli 6 described one case of bilateral frosted branch angiitis in a 33-yearold patient who had decreased visual acuity. The patient had unremarkable results of a sys-
Vol. 113, No. 5
Frosted Branch Angiitis and Cytomegalovirus Retinitis
temic examination, including negative results of studies for cytomegalovirus and HIV. The visual acuity in this patient improved to 20/30 without corticosteroid treatment. In contrast, our three patients had frosted branch angiitis in association with cytomegalo virus retinitis in the setting of profound reduc tion of absolute helper T-cell counts secondary to AIDS. The four eyes that did not have direct infection of the macula with cytomegalovirus had a visual acuity of 20/20. In all three pa tients, the frosted branch angiitis was seen only in association with the cytomegalovirus retini tis. One patient had unilateral involvement in an eye with typical cytomegalovirus retinitis, whereas his fellow eye without cytomegalovi rus retinitis had no vascular sheathing. Our patients did not appear to have syphilis 7 or tuberculosis 8 as possible explanations of the retinal vasculitis. Each one had areas of retini tis that had the clinical appearance of cytomeg alovirus retinitis. Treatment with anticytomegalovirus antiviral medications such as ganciclovir and foscarnet resulted in rapid reso lution of the perivascular infiltrate and slower resolution of the retinitis. Cytomegalovirus ret initis was confirmed by retinal biopsy in one of our patients. Although corticosteroids seemed to be previously effective, none of the previous ly described patients had evidence of cytomeg alovirus retinitis or AIDS. Our patients already had severe immunosuppression from AIDS and cytomegalovirus infection, which contraindicated the use of corticosteroids. Retinal vasculitis was previously described in patients with AIDS who live in Africa,910 but these patients had perivascular infiltration in the retinal periphery only. Cytomegalovirus retinitis is uncommon in African patients with AIDS and was not seen in any patient in either study. 910 Those authors thought the vasculitis was a noninfectious retinopathy associated with AIDS. 910 In contrast, our patients had im pressive perivascular deposition of inflammato ry exudate throughout widespread areas of their retinas and had typical findings of cyto megalovirus retinitis. Additionally, treatment with anticytomegalovirus antiviral medication was associated with rapid resolution of the perivascular infiltrate in our patients. The perivascular infiltrate seen in our pa tients may have resulted from direct cytomega lovirus infection of the retinal vessels or the retina immediately surrounding the vessels. Cytomegalovirus has a certain tropism for vas cular endothelial cells. 1112 The more rapid reso lution of the perivascular infiltrate may have
resulted from higher tissue levels of antiviral in the vascular and perivascular regions. Some aspects of our patients' findings did not seem to be completely explained by this hypothesis, however. The vessels of the retina seemed to be globally involved, not just in areas contiguous with cytomegalovirus retinitis. Although the vessels were involved, resolution of the vasculi tis was not accompanied by any discernible retinal atrophy or retinal pigment epithelial scarring one might expect if the perivascular areas were directly infected with cytomegalovi rus. The inflammatory exudate around the retinal vessels in conventional frosted branch angiitis may be related to antigen-antibody complex deposition. 6 Such immune-complex deposition leads to vasculitis and local tissue inflamma tion. 13 If this were true, one might suppose that altering either the intraocular local antigen or antibody concentrations might favorably re duce the amount of immune-complex forma tion and thereby reduce the amount of perivas cular infiltrate. In our patients, the antigen presumably was derived from the intraocular cytomegalovirus infection. The rapid resolu tion of the perivascular infiltrate with intrave nous ganciclovir may have resulted from lower ing the intraocular antigen load through antiviral treatment. The rapid resolution of frosted branch angiitis with corticosteroid treatment in patients without AIDS might be the result of dampening a specific antibody response, or more generally, by reducing the inflammatory response to the immune-complex deposition. In frosted branch angiitis associated with cytomegalovirus retinitis, the hypothesis of di rect infection of the blood vessels and sur rounding retina by the cytomegalovirus and the hypothesis of immune-complex deposition are not mutually exclusive; both may occur simul taneously. The retinal biopsy specimen in one of our patients was taken after the frosted branch angiitis resolved and did not include any large retinal vessels, and thus did not answer the question. The cases of frosted branch angiitis previous ly described in the literature did not seem to have clinical signs of a systemic vasculitis, but rather had clinical signs and symptoms limited to the eye; some cases were unilateral. 6 This suggested that the production of the inciting antigen in conventional frosted branch angiitis might occur within the eye. Although not every case described in the literature was tested, no previously described case of frosted branch
AMERICAN JOURNAL OF OPHTHALMOLOGY
angiitis h a d a b n o r m a l c y t o m e g a l o v i r u s titers, positive c y t o m e g a l o v i r u s c u l t u r e s , or findings suggestive of c y t o m e g a l o v i r u s r e t i n i t i s . This s u g g e s t s that frosted b r a n c h angiitis in p a t i e n t s w i t h o u t AIDS may b e c a u s e d b y a g e n t s or factors other t h a n c y t o m e g a l o v i r u s . Frosted b r a n c h angiitis is a n u n c o m m o n s y n d r o m e that h a s b e e n p r e v i o u s l y d e s c r i b e d as a p r i m a r y d i s o r d e r in i m m u n o c o m p e t e n t p a t i e n t s . In our series, frosted b r a n c h angiitis a p p e a r s to h a v e o c c u r r e d as a s e c o n d a r y p h e n o m e n o n in p a t i e n t s w i t h AIDS a n d c y t o m e g a lovirus r e t i n i t i s . O t h e r forms of s e c o n d a r y frosted b r a n c h angiitis may well be e l u c i d a t e d in the future. T r e a t m e n t of the p r i m a r y form w i t h corticosteroids a p p e a r s to b e effective. C o r t i c o s t e r o i d s w o u l d s e e m to be c o n t r a i n d i cated in the s e c o n d a r y form of frosted b r a n c h angiitis t h a t is associated w i t h c y t o m e g a l o v i r u s retinitis in p a t i e n t s w i t h A I D S . S u c h p a t i e n t s , as d e s c r i b e d in this s t u d y , a l r e a d y h a v e p r o f o u n d i m m u n o s u p p r e s s i o n a n d t h e y r e s p o n d to a n t i c y t o m e g a l o v i r u s t r e a t m e n t . Before corticosteroid t r e a t m e n t is b e g u n for a p a t i e n t w i t h t h e clinical signs a n d s y m p t o m s of frosted b r a n c h angiitis, the p a t i e n t ' s m e d i c a l s t a t u s s h o u l d b e carefully c o n s i d e r e d . If t h e p a t i e n t is i m m u n o compromised, additional medical examination for infectious causes of r e t i n a l vasculitis is essential.
References 1. Ito, Y., Nakano, M., Kyu, N., and Takeuchi, M.: Frosted-branch angiitis in a child. Jpn. J. Ophthalmol. 30:797, 1976.
2. Yamane, S., Nishiuchi, T., Nakagawa, Y., Iwase, T., and Hama, M.: A case of frosted-branch angiitis of the retina. Folia Ophthalmol. Jpn. 6:1822, 1985. 3. Watanabe, Y., Takeda, N., and Adachi-Usami, E.: A case of frosted branch angiitis. Br. J. Ophthal mol. 71:553, 1987. 4. Kleiner, R., Kaplan, H. J., Shakin, J. L., Yannuzzi, L. A., Crosswell, H. H., and McLean, W. C : Acute frosted retinal periphlebitis. Am. J. Ophthalmol. 106:27, 1988. 5. Sugin, S. L., Henderly, D. E., Friedman, S. M., Jampol, L. M., and Doyle, J. W.: Unilateral frosted branch angiitis. Am. J. Ophthalmol. 111:682, 1991. 6. Vander, J. F., and Masciulli, L.: Unilateral frost ed branch angiitis (correspondence). Am. J. Ophthal mol. 112:477, 1991. 7. Lobes, L. A., Jr., and Folk, J. C : Syphilitic phle bitis simulating branch vein occlusion. Ann. Oph thalmol. 13:825, 1981. 8. Fountain, J. A., and Werner, R. B.: Tuberculous retinal vasculitis. Retina 4:48, 1984. 9. Kestelyn, P., Van de Perre, P., Rouvroy, D., Lepage, P., Bogaerts, J., Nzaramba, D., and Clumeck, N.: A prospective study of the ophthalmologic findings in the acquired immune deficiency syn drome in Africa. Am. J. Ophthalmol. 100:230, 1985. 10. Kestelyn, P., Lepage, P., and Van de Perre, P.: Perivasculitis of the retinal vessels as an important sign in children with AIDS-related complex. Am. J. Ophthalmol. 100:614, 1985. 11. Ho, D. D., Rota, T. R., Andrews, C. A., and Hirsh, M. S.: Replication of human cytomegalovirus in endothelial cells. J. Infect. Dis. 150:956, 1984. 12. van Dorp, W. T., Jonges, E., Bruggeman, C. A., Daha, M. R., van Es, L. A., and van der Wonde, F. J.: Direct induction of MHC Class I but not Class II expression on endothelial cells by cytomegalovirus infection. Transplantation 48:467, 1989. 13. Fujikawa, L. S.: Advances in immunology and uveitis. Ophthalmology 96:1115, 1989.
"Saint N o n n a , " said Cadfael didactically, threading the woodland belt that spread more than a mile inland from the strait, " w a s the mother of Saint David. She has many sacred wells about the country, that give healing, especially to eyes, even to curing blindness. This holy woman must have chosen to name herself after the saint." Ellis Peters, The Summer of Danes New York, Mysterious Press Books, 1991, p. 109