Combined Choroidal and Retinal Ischemia During Interferon Therapy: Indocyanine Green Angiographic and Microperimetric Findings

This article presents findings from indocyanine green angiography (ICGA) and microperimetry in a patient with interferon (IFN)-associated combined choroidal and retinal perfusion deficits.

A 40-year-old man was seen for central scotomas and blurred vision during treatment with IFN alfa-2b for metastatic renal cell carcinoma. Findings from retinal examination included cotton-wool spot formation and small hemorrhages. Findings from fluorescein and ICGA demonstrated areas of retinal as well as choroidal nonperfusion. Results of scanning laser ophthalmoscope microperimetry revealed central scotomas corresponding to choroidal perfusion defects. Cotton-wool spots and retinal hemorrhages disappeared during therapy over a 6-month period, but visual impairment and choroidal ischemia persisted. In addition to retinopathy, IFN therapy may also induce persistent choroidal perfusion changes associated with scotomization.

Our patient was seen for blurred vision in both eyes and reading difficulty for several months. Findings from the patient's medical history revealed renal cell carcinoma, for which he had undergone nephrectomy 3 years earlier. Treatment with IFN alfa-2b was administered for metastasis of the lung with an initial dose of 2 × 10⁶ IU subcutaneously and increased to the continuos dose of 10 × 10⁶ IU 5 times per week. His general condition was otherwise good, and he did not take any other medications. Ocular symptoms were noted roughly 1 month after initiation of IFN alfa-2b therapy. On initial examination, visual acuity was 20/20 OU. On ophthalmoscopic examination, cotton-wool spots and small retinal hemorrhages were found scattered throughout the posterior pole in both eyes (Figure 1). The patient was not diabetic, and his blood pressure levels have never been elevated. Other systemic causes of cotton-wool spots were excluded by findings from extensive laboratory examination. Two months later, his visual acuity decreased to 20/50 OD and remained unchanged in his left eye. Results of fluorescein angiography showed a central area of complete absence of choroidal perfusion with overlying intact retinal perfusion and no alteration of the perifoveal capillary net in early-phase fluorescein angiography (Figure 2, A and B). Late-phase fluorescein angiography showed venous dilatation and retinal vessel leakage in both eyes (Figure 2, C). The ICGA revealed a well-demarcated area of confluent and persistent choroidal filling defects (Figure 3, A). The characteristic ICG pattern of the capillary and arteriolar net was missing. Findings from objective testing using scanning laser ophthalmoscope microperimetry revealed central and paracentral scotomas (Figure 4) precisely corresponding to the 2 areas of choroidal hypofluorescence consistent with choroidal ischemia during late-phase ICGA (Figure 3, B). A slight decrease of the Arden ratio was found by whole electroretinography. Within the initial follow-up period of 6 months, the cotton-wool spots and the retinal hemorrhages disappeared. However, there was no recovery of the central choroidal perfusion changes. Functional defects persisted over a 1-year follow-up period, and visual acuity remained 20/63 OD.

Interferon alfa inhibits vascular endothelial cell proliferation and reveals antitumor activity. Therefore, it is used to treat human neoplasms, including metastatic renal cell carcinoma. Interferon-associated retinopathy has been described previously, mainly in the Japanese literature.^1,2 Typical retinal findings include cotton-wool spots, hemorrhages, and microaneurysms. Retinal ischemia may be observed in affected areas by fluorescein angiography. Other ocular findings included conjunctival hemorrhage, central retinal vein occlusion, branch retinal artery occlusion, aggravation of diabetic retinopathy, optic disc edema, oculomotor nerve paralysis, and hypertrichosis.^1,2

To our knowledge, choroidal perfusion changes have not been described in the literature until now and are documented for the first time in a patient undergoing prolonged systemic IFN therapy. The patient had symptoms of focal retinal capillary dropout as manifested by cotton-wool spots. However, there were no retinal infarctions seen within the area of choroidal nonperfusion itself, and the perifoveal capillary perfusion was intact (Figure 2, A). Choroidal filling defects were substantiated by ICGA (Figure 3) and appeared to be responsible for visual impairment based on microperimetry. Electrophysiological findings supported the presence of photoreceptor damage, which is responsible for the central functional defect documented by scanning laser ophthalmoscope microperimetry. The central and paracentral scotomas corresponded in size and location to 2 persisting choriocapillary perfusion defects delineated by ICGA. As reported by other authors, symptoms of retinal vasculopathy subsided spontaneously in our patient; the choroidal defects, however, did not resolve.

The pathogenesis of IFN-associated vasculopathy is unclear. It is known that up to 10% of patients treated with recombinant IFN alfa produce autoantibodies.³ It is therefore postulated that IFN may cause deposition of immune complexes in retinal vessels.² However, capillary-type damage is suspected in IFN retinopathy, while choroidal nonperfusion is more likely owing to occlusion of larger-caliber vessels.

Occlusive choroidopathy was seen in malignant hypertension and Harada disease and was associated histopathologically with fibrin platelet clots in the choriocapillaris.⁴ Another possible origin of choroidal thrombosis could be the underlying renal cell carcinoma, which might have triggered an embolic event by platelet or tumor clots. Venous choroidal thrombosis mostly induced serous retinal detachment, which was not present in our patient. However, embolization of the choriocapillaris by an arterial route did not result in detachments experimentally.⁵ Arterial choroidal occlusion is most likely responsible for intensive functional defects. However, choroidal perfusion changes and photoreceptor function did not recover during long-term follow-up.

Our observation suggests that patients demonstrating retinopathy during IFN therapy should carefully be examined by ICGA and central field testing to recognize additional choroidal nonperfusion. Since there is evidence that vaso-occlusion might not be restricted to retinal capillaries, the risk of ischemia occurring within other organ systems should be taken into account.

The authors have no commercial, proprietary, or financial interest in any research or devices described in this study.

Reprints: Hans Hoerauf, MD, Department of Ophthalmology, Medical University Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany.