Management of Autoimmune Retinopathies With Immunosuppression FREE

The autoimmune retinopathies (AIRs) are a group of diseases characterized by acute or subacute progressive vision loss, an abnormal finding via electroretinogram (ERG) (either rod-cone or cone-rod patterns), and circulating antibodies directed against retinal antigens. The AIRs encompass the better-studied paraneoplastic syndromes, such as cancer-associated retinopathy (CAR) and melanoma-associated retinopathy (MAR), and a larger group of AIRs that have similar clinical and immunologic features but that are not associated with an underlying malignancy. We coined the term nonparaneoplastic autoimmune retinopathy (npAIR) to describe this latter group, which has previously been called recoverin-associated retinopathy,¹ autoimmune retinopathy in the absence of cancer,² or, simply, autoimmune retinopathy.^{3 - 4} Some npAIR patients have cystoid macular edema (CME) as a prominent feature (npAIR/CME) that appears to be a distinguishing factor. Many npAIR patients are referred with a diagnosis of simplex retinitis pigmentosa (RP), and the 2 disorders have similar clinical features. The challenge is to identify cases of AIR, which can be treated. Complicating the issue is that a few cases of hereditary RP may develop secondary AIR with rapid visual field loss and severe cystoid edema.^{5 - 7}

Cancer-associated retinopathy, MAR, npAIR, and npAIR/CME tend to have common clinical features despite the fact that no uniform set of antiretinal antibodies (ARAs) has been found to be circulating in these patients.^{8 - 9} Patients have a variety of antibody activity, often with 3 to 6 different ARAs found on Western blots, a laboratory technique that detects serum ARAs in patients.⁷ However, there is a commonality in the presentation of patients with AIR; frequently, without a history of visual problems or night blindness, the patient has rapid onset of photopsia, followed by other symptoms, such as night blindness, scotomata, and visual field loss. Some patients also develop diminished central vision. The presentation may be asymmetrical between the eyes. On first examination, there are frequently no obvious retinal changes on ophthalmoscopy, but a standardized ERG will show abnormal responses. Some cases have negative waveforms, consisting of an a-wave that does not return to the isoelectric point on dark-adapted bright flash ERG testing.¹⁰ Kinetic visual fields are better at measuring peripheral losses or scotomata, blind spot enlargements, and pericentral losses.

On questioning, most patients reveal that they have positive autoimmune family histories, with first-degree relatives who have lupus, rheumatoid arthritis, thyroid disease, and other autoimmune disorders. The diagnosis of AIR is made by weighing the available evidence (Table 1) and by finding ARAs on Western blot. Immunohistologic staining of normal retinal tissue with the sera of patients is an additional technique that may pick up ARA activity. The presence of antirecoverin or anti–α-enolase antibodies is further confirmatory evidence.^{11 - 14} However, many patients who fit this clinical picture do not have antibodies against recoverin or α-enolase but have mixtures of antibodies directed against either other unidentified retinal antigens or ones whose pathogenicity is unknown.⁹

Finally, npAIR patients frequently have CME as a prominent feature.^{6 ,14} Many of these patients have been diagnosed by their ophthalmologist as having RP based on clinical findings, but they have no family history of RP. There are also few patients with hereditary RP who develop AIR as a secondary complication that manifests as CME.^{5 - 7 ,14} Patients with RP and ARAs occasionally have visual field loss faster than usual; this last circumstance needs prospective studies to objectively examine whether the presence of ARAs worsens RP.¹⁵ Although optical coherence tomography demonstrates intraretinal cystic spaces or schisislike spaces, many of these cases do not have leakage on fluorescein angiography, and the AIR macular changes are really a form of degenerative schisis.⁷ Prospective masked studies demonstrated that 90% of patients with RP and cysts have circulating ARAs by means of Western blot analysis compared with 13% of patients with RP without macular cysts and 6% of controls.⁷

Patients with RP cystoid “edema” may respond to carbonic anhydrase inhibitor agents, such as acetazolamide and topical dorzolamide hydrochloride.^{16 - 18} Rebound failures are common, and immunosuppression therapy may be more effective because the therapeutic action is directed at the cause of the condition.¹⁹

No treatment regimens for the various autoimmune retinopathies have emerged during the past decade.⁸ Immunosuppression with oral or intravenous corticosteroids has shown mixed results.²⁰ Generally, short courses of prednisone are ineffective, although the group that appears to have the best response with corticosteroids seems to be that with CAR, where a combination of chemotherapy and milder immunosuppression variably gives some reversal of visual loss.^{21 - 23} Intravenous immunoglobulin has also been used in patients with CAR and MAR.^{24 - 25}

The prevalence of AIR is unknown, although it is believed to be relatively uncommon. The rarity of AIRs, and the difficulty in firmly establishing the diagnosis, has limited investigations that evaluate treatment outcomes. A review of the literature shows that treatment results are found primarily as case reports and a few small case series. We report our clinical experience using immunosuppression to treat AIR in a series of 30 patients.

Study participants were identified from a cohort of patients with the diagnosis of AIR or CAR syndrome. The diagnostic criteria are listed in Table 1. Except for 1 patient with early CAR, all the patients underwent ARA Western blot testing. The sample included patients from the Jules Stein Eye Institute at the University of California, Los Angeles, but most were from the Kellogg Eye Center. The study was conducted with institutional review board approval from both institutions. Two of us (H.A.F. and J.R.H.) reviewed the medical records to reconfirm that patients met the diagnostic criteria and to evaluate the response to the treatment. There was no standard protocol, partly because no regimens have been established in the literature, and also because individual patients have different severity of disease and different tolerances to these medications, which tend to have adverse effects.

Western blot testing was performed by lysing normal human, mouse, and bovine retinal tissue in a phosphate-buffered saline, 1mM phenylmethylsulfonyl fluoride (PMSF), protease inhibiter cocktail and sonicating the suspension. Thirty micrograms of each lysate was heated for 10 minutes at 80°C in sample buffer with dithiothreitol, followed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Proteins were transferred to nitrocellulose membrane and were incubated overnight with the serum of the patient at 1:100 dilution. Membrane was washed and incubated with anti–human IgG–tagged horseradish peroxidase, and proteins were detected using a chemiluminescent detection kit (SuperSignal West Pico; Thermo Scientific Pierce, Rockford, Illinois) and were exposed to film.

Minimum follow-up for inclusion was 3 months on immunosuppression therapy, and median follow-up was 17 months. Outcome measures for improvement were at least 2 lines of Snellen visual acuity (VA), expansion of the visual field isopters on kinetic perimetry (Goldmann visual fields) by at least 25%, and resolution of CME. Two patients had significant improvement in ERG amplitudes (≥50%). Medical records were reviewed for the following data: initial VA (at the start of treatment), final VA (at the last follow-up visit while on treatment), age at the onset of treatment, sex, the presence of CME, ERG amplitudes at the start of treatment (calculated as the percentage of the mean normal amplitude), the presence of negative waveforms on ERG, ARA activity on Western blot, a personal and family history of autoimmune disease (eg, systemic lupus erythematosus, asthma, Crohn disease, idiopathic thrombocytopenic purpura, multiple sclerosis, psoriasis, rheumatoid arthritis, stiff man syndrome, and thyroid disease), immunosuppressive medications used, dosage, length of treatment, and adverse effects. A digital planimeter (Lasico, Los Angeles, California) was used to measure the area in square centimeters of the visual field (isopters IV-4-e and III-4-e). The percentage change in visual field area was calculated for these isopters, and an increase in visual field size greater than 25% was considered significant. The ERG findings were expressed as a percentage of the mean normal amplitude to show the difference from reference values. The AIRs in this study include CAR, npAIR, and npAIR/CME. Although patients with MAR were evaluated during the medical record review, none were treated with immunosuppression and none were included in this study.

Data obtained were distributed nonparametrically and are expressed using median values, interquartile ranges (IQRs), and ranges.

Thirty patients who were identified with a clinical diagnosis of CAR (n = 6), npAIR (n = 13), or npAIR/CME (n = 11) underwent treatment with immunosuppression and met the inclusion criteria. For patients with CAR, the distribution of carcinoma was 2 each with breast and ovarian and 1 each with colon and prostate.

All but 4 of 30 patients exhibited diffuse retinal pigment epithelial atrophy without intraretinal bone spicule-like pigment deposits, with only 1 patient with npAIR (patient 16) and 3 with npAIR/CME (patients 22-24) having pigment deposits.

Overall, the combined AIR group (CAR, npAIR, and npAIR/CME) was predominantly female (63% female [19 of 30] vs 37% male [11 of 30]) and had a median age of 51 years (range, 11-85 years; IQR, 35-70 years). Patients 19 and 23, with npAIR/CME, have family histories of RP and may have genetic and autoimmune mechanisms that contribute to their retinopathy. The highest female predominance was in the npAIR/CME subgroup (82% [9 of 11]), followed by the CAR subgroup (67% [4 of 6]), whereas the npAIR subgroup showed no significant sex difference (46% female [6 of 13]). The CAR subgroup was the oldest (median age, 79 years; range, 71-85 years; IQR, 75-84 years), and the npAIR/CME group was the youngest (median age, 36 years; range, 11-66 years; IQR, 16-51 years). The npAIR subgroup had a median age of 51 years (range, 28-78 years; IQR, 43-67 years).

An autoimmune family history, defined as autoimmune diseases that affect the patient and at least 1 first-degree relative, was present in 63% of the patients (19 of 30) in the combined AIR group. An autoimmune family history was most common in the npAIR subgroup (69% [9 of 13]), followed by the npAIR/CME (64% [7 of 11]) and CAR (50% [3 of 6]) subgroups. When the patient's personal autoimmune history was examined, the prevalence was 40% (12 of 30) for the combined AIR group. When patient histories by subgroups were examined, the trends were similar: an autoimmune history was most common in the npAIR group (46% [6 of 13]), followed by the npAIR/CME (36% [4 of 11]) and CAR (33% [2 of 6]) groups.

Immunosuppression therapy was individualized to the patient. Triple therapy using cyclosporine, azathioprine, and prednisone was given to patients with a classic severe presentation. The typical initial doses used were 100 mg/d for cyclosporine, 20 to 40 mg/d for prednisone, and 100 mg/d for azathioprine. In cases with a high clinical suspicion but without a classic presentation or available positive Western blot information, treatment with 1 or 2 subtenon periocular injections of methylprednisolone acetate (40-60 mg) was given to 1 eye as a clinical trial to evaluate for a treatment effect. If there was clear improvement and ARAs were subsequently found on Western blot, immunosuppression therapy was recommended to the patient.

Twenty-nine of the 30 patients received at least 1 systemic immunosuppressive medication during treatment (1 patient with CAR received subtenon methylprednisolone acetate at multiple visits during 2 years). Of these 29 patients, 10 (35%) discontinued at least 1 of their systemic medications due to adverse effects (Table 2). Patients who were unable to tolerate any systemic immunosuppression therapy or chose not to take systemic medications were treated with periocular or intravitreal corticosteroid injections. Some patients received subtenon injections to the worse affected eye as supplemental therapy to the systemic medications. Of 23 patients treated with periocular methylprednisolone, 1 developed elevated intraocular pressure, which was controlled with topical medications. Intravitreal triamcinolone acetate was given to 4 patients with npAIR/CME without complications. Overall, treatment ranged from 3 months to 89 months (median, 17 months; IQR, 6-23.5 months).

For the combined group (CAR, npAIR, and npAIR/CME) the response to treatment was 70% (21 of 30). When examined by subgroup, the responses were 100% (6 of 6), 54% (7 of 13), and 73% (8 of 11) for CAR, npAIR, and npAIR/CME, respectively. Responses to treatment consisted of improvement in VA in 5 of 21 patients (24%) and expansion of the visual field area of at least 25% in 15 of 21 (71%). Of the 11 patients with CME, 6 (55%) had resolution or significant improvement of the CME. Although not measured routinely, 2 patients (patients 13 and 27) showed significant improvement in the ERG (increase in the rod-isolated or combined b-wave amplitude by >50%). Examination of pretreatment and posttreatment Western blots (in 19 patients who were checked post-treatment) showed that 13 blots were improved (less reaction), 4 were worse, and 2 were unchanged (Figures 1, 2, and 3). Several patients showed improvement on their antibody immunoreactivity but did not show improvement clinically. However, it may be that their progression was slowed or halted.

The AIRs are diseases that besides their clinical and electrophysiologic similarities all exhibit circulating antibodies directed against retinal antigens. The spectrum of AIR includes the paraneoplastic syndromes CAR and MAR and npAIR. The few studies in the literature that have evaluated treatment for AIR have reported conflicting results. We retrospectively reviewed the use of immunosuppression to treat AIR in a group of 30 patients and found that it was effective in all subgroups, although there were differences in the response rates among subgroups. In hindsight, many patients are undertreated because we do not have good markers to evaluate treatment effects.

The paraneoplastic retinopathies have been the most extensively studied of the AIRs. Various investigations^{11 - 12} demonstrate the presence of circulating cross-reactive antibodies that presumably bind to retinal antigens and result in retinal degeneration by currently unknown mechanisms. Several studies have shown that the tumors aberrantly express proteins normally exclusive to retinal tissue, leading to the production of antibodies directed against these retinal antigens. The most common antigen associated with CAR is recoverin,^{26 - 29} but other antigens have been identified, such as α-enolase,³⁰ heat shock cognate protein 70,^{31 - 32} tubby-like protein 1,³³ photoreceptor cell–specific nuclear receptor,³⁴ and neurofilament protein.³⁵ Antibodies directed against retinal bipolar cells,³⁶ a 22-kDa neuronal antigen,³⁷ and transducin have also been identified with MAR.³⁸

Like CAR, npAIR was initially found to be associated with antibodies directed against recoverin.¹ Since then, antibodies directed against the inner plexiform layer,² the inner retinal layer (described as the 35-kDa retinal Muller cell-associated antigen),³⁹ or other as-yet unidentified retinal proteins have been described in npAIR patients.⁴⁰ The pathogenicity of the various unknown ARAs must be determined by future research.

The epidemiologic characteristics of AIR have not been well described in the literature. Adamus et al²⁰ reported the prevalence of ARAs in a group of 193 patients with symptoms resembling paraneoplastic retinopathy or AIR. They found that 58% were women and 42% were men. The mean age of patients with paraneoplastic retinopathy was 62.0 years compared with 55.9 years for npAIR patients. Chan⁸ reviewed the treatment outcomes of 55 patients with CAR from the literature. Meta-analysis of these patients revealed that 51% were women, with a median age of 67 years. In the present series of 30 patients, there was a greater overall female predominance: 63% for the combined group, 67% for CAR, 46% for npAIR, and 82% for npAIR/CME. Likewise, we found a greater difference in median ages: 79 years for CAR, 51 years for npAIR, and 36 years for npAIR/CME.

Autoimmune diseases are regarded as having a multigenic or a complex genetic basis, and this genetic template then variably affects individuals in families, depending on their individual gene makeup. We specifically took family histories for autoimmune diseases to get an indication of the autoimmune tendencies in these patients. This series of patients had a high prevalence of an autoimmune history: 67% for the overall group, 50% for CAR, 69% for npAIR, and 64% for npAIR/CME. The highest prevalence of autoimmune history was found in the npAIR and npAIR/CME groups compared with the CAR group. In some npAIR/CME patients, we hypothesize that development of the AIR/CME component in RP may be a response to retinal antigens released during the degeneration, and these are recognized in the more antigenically sensitive patients, who have a template of autoimmune genes. Some cases may also represent primary idiopathic AIR; histories of head or ocular trauma are occasionally obtained and may represent moments of retinal antigen release.

We found a higher prevalence of patients with an autoimmune family history who were less responsive to immunosuppression therapy. This trend was most evident in the npAIR subgroup, which had the highest prevalence of an autoimmune family history vs the other subgroups. In the npAIR subgroup, the prevalence of an autoimmune family history was 83% (5 of 6) in nonresponders vs 57% (4 of 7) in responders. This finding suggests that patients with stronger autoimmune tendencies may need more intensive immunosuppression therapy.

Immunosuppression has previously been used to treat AIR with mixed results. Sawyer et al⁴¹ treated 1 of the original 3 patients with CAR with prednisone but saw no improvement. Keltner et al⁴² reported the first patient with CAR responsive to corticosteroid therapy. Since then, there have been numerous case reports in the literature using short-course high-dose intravenous methylprednisolone or oral prednisone that report mild to moderate improvement in VA and visual fields. Plasmapheresis, when used alone, led to no improvement⁴³ ; when used with prednisone, vision improved in 1 patient.⁴⁴ Guy and Aptsiauri²⁴ reported improvement in 2 of 3 patients treated with intravenous immunoglobulin and stabilization in the third. Espandar et al⁴⁵ recently reported stabilization of CAR with alemtuzumab therapy.

In the present series of 30 patients using an ad hoc long-term therapeutic method, immunosuppression therapy was effective in 70%. It was most effective in the CAR and npAIR/CME subgroups (100% and 73%, respectively) compared with the npAIR group (54%). The least responsive subgroup (npAIR) had the highest autoimmune family history (69%). Likewise, there were differences noted between the nonparaneoplastic subgroups: npAIR/CME patients responded significantly better than did those with npAIR (no CME). It may be that the mechanism responsible for the development of CME in these patients is more responsive to immunosuppression therapy. Nonresponders were also less likely to tolerate systemic immunosuppression; 56% of nonresponders (5 of 9) stopped using at least 1 systemic immunosuppressive medication owing to adverse effects compared with 25% of responders (5 of 20). Of the 6 npAIR patients who did not respond, 4 (67%) discontinued taking at least 1 systemic immunosuppressive medication owing to adverse effects compared with 1 of 3 npAIR/CME patients (33%). We found that most AIR patients need immunosuppression therapy for extended periods, with monitoring of visual functions regularly.

Although these results are limited by the limitations inherent in any retrospective series, they demonstrate the potential benefits of using immunosuppression to treat AIR. Patients with earlier disease were more responsive than were severely affected patients (at the time of initial visit). This series of 30 patients represents the largest case series evaluating the treatment of AIR, and the positive results highlight the need for prospective trials to further evaluate which medications are most effective in the treatment of AIR.

Finally, until there are specific diagnostic tests to identify AIR, so that AIR can be clearly distinguished from RP, some cases may be difficult to identify when initially encountered in the clinic. Table 1 gives the current criteria for identifying AIR patients. The most helpful features are the rapid onset and course of photopsia and visual field loss compared with slower losses in typical RP, positive family history of autoimmune disease, lack of pigment deposits in most patients, abnormal ERG findings and visual fields in the face of a relatively quiet fundus appearance, and negative waveforms on ERG (if present). Clinical trials of immunosuppression therapy in uncertain cases for at least 2 to 3 months are usually justified in patients who meet most of the criteria listed in Table 1 and in all patients with severe CME.

Correspondence: John R. Heckenlively, MD, 1000 Wall St, Ann Arbor, MI 48105 (jrheck@umich.edu).

Submitted for Publication: August 27, 2008; final revision received December 7, 2008; accepted December 15, 2008.

Author Contributions: Dr Heckenlively had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Financial Disclosure: None reported.

Funding/Support: This work was supported in part by Foundation Fighting Blindness.

Additional Contributions: Jennifer Burkheiser assisted with illustrations, Kari Branham, MS, obtained family histories, Richard Hackel, MA, provided fundus photography, and Jill Oversier, BS, coordinated patients.