Birdshot Chorioretinopathy

BirdShot ChorioRetinopathy (BSCR)^1–4 is a non-infectious, bilateral, panuveitis associated with distinctive, widely distributed, yellow-white choroiditis spots. Women are affected slightly more often than men and virtually all patients are Caucasian adults. HLA-A29 positivity is essentially disease-defining, although the limited sensitivity of earlier antibody-based HLA typing techniques has been misinterpreted to indicate that roughly 5% of those with BSCR fail to express the HLA-A29 antigen.^4,5 Most patients will have diffuse retinal vasculitis, often producing cystoid macular edema (CME). A minority will have little or no leakage, what might be termed “dry BSCR,” but can still develop diffuse outer retinal atrophy over time without treatment. The signature round or ovoid choroiditis spots of BSCR can be either indistinct or visible only with indocyanine green angiography (ICGA), particularly early in the course of disease, in patients with little or light choroidal pigmentation, following treatment, or in advanced disease where choroidal scarring and atrophy can alter their appearance. While never studied in a randomized, controlled fashion, most support the use of long-term treatment, either with conventional or biologic immunosuppressive agents or with sustaine release corticosteroid implants, in the majority of patients. Monitoring disease activity requires regular multimodal imaging, most notably optical coherence tomography (OCT) and both fluorescein angiography (FA) and ICGA. Periodic visual fields and electroretinography have also been employed. Both sarcoidosis and uveal lymphoma can mimic BSCR and so should be actively excluded when establishing the diagnosis. Five original articles^6-10 in this issue of Ocular Immunology & Inflammation (OII) address important aspects of the pathogenesis, diagnosis, complications, and management of BSCR.

Pohlmann et al⁶ used multimodal imaging with an emphasis on optical coherence tomography angiography (OCTA) in a cross-sectional study of retinal vascular changes in 64 eyes of 32 subjects with BSCR followed in a uveitis referral center in Berlin, Germany. Consistent with their diagnosis, all subjects had bilateral low-grade inflammation in the anterior chamber and vitreous in the setting of multiple choroiditis spots and were both Caucasian and HLA-A29 positive. Seventeen subjects (53%) were female with an age range of 38 to 74 years (mean = 60 years; median = 61 years). All but one (97%) received systemic corticosteroids, a systemic non-corticosteroid immunosuppressive agent, or both at some point during their management. At the time of imaging, four subjects (13%) were on no systemic medications; three (9%) were on low dose prednisone (5.0 – 7.5 mg/kg/day); seven (22%) were on low dose prednisone and a systemic immunosuppressive agent (three on mycophenolate mofetil [MMF]; two on cyclosporine A [CSA]; and one each on methotrexate [MTX] and adalimumab), and 18 (69%) were on a single non-corticosteroid immunosuppressive agent alone - including nine (28%) on MMF, six (19%) on adalimumab, two (6%) on CSA, and 1 (3%) on azathioprine (AZA). Retinal findings included epiretinal membrane (ERM) formation in 19 eyes (30%); vasculitis in 17 eyes (27%); and CME in 8 eyes (13%). All but one patient with a known dye allergy underwent both FA and ICGA. Fluorescein angiography revealed retinal vascular leakage in just over half of all imaged eyes (55%) despite the use of systemic immunosuppressants, and was described as peripheral in 17 (27%), diffuse in 14 (23%), and focal in three eyes (5%). Of the 34 eyes with active vascular leakage, concurrent leakage from the disc occurred in 28 (82%). Hypocyanescent spots were visible on late phase ICGA images in 14 eyes (23%). Spectral Domain-OCT (SD-OCT) of the cohort revealed a mean thickness of 268 microns (median = 265 microns; range 148 – 489 microns), with an inverse correlation between thickness and vision (r = −0.33; nominal p = 0.007), as might be expected in a cohort with more retinal atrophy than CME as summarized below. Not unexpectedly, vision also correlated inversely with vitreous haze (r = −0.32; nominal p = 0.01). Disruption of the outer retinal hyper-reflective bands indicative of compromised photoreceptor morphology and function was noted in 12 eyes (19%). The authors observed multiple hypoautoflourescent lesions in a number of eyes with advanced disease, but did not elaborate on patterns or extent. Sixty-two eyes underwent OCTA examination with automated segmentation of both the superficial (SCP) and deep capillary plexuses (DCP). Vascular abnormalities were common and included the presence of what the authors referred to as capillary loops (SCP = 58%; DCP = 76%), telangiectasis (SCP = 44%; DCP = 63%), increased intercapillary spacing that resembled ischemia in the representative illustrations (SCP = 52%; DCP = 58%), and altered vascular architecture (SCP = 53%; DCP = 40%). Thinning of the automated segments used to capture the SCP and DCP was also common, occurring in 63% and 52%, respectively, of imaged eyes. The authors subdivided the cohort as active, defined as the presence of either FA leakage or hypocyanescent spots on ICGA – all of whom were receiving systemic immunosuppression (18/63; 29%; mean duration 2.2 years); inactive without these FA or ICGA findings, but with advanced disease in the absence of photoreceptor disruption on SD-OCT (29/63 = 46%; mean duration 3.4 years); and inactive and advance with photoreceptor disruption on SD-OCT (16/63 = 25%; mean duration 7.1 years). They also subdivided the cohort based on duration of disease as two years or less (8/64 = 13%), three to four years (14/64 = 22%), five to ten years (34/64 = 52%), or ten years or more (8/64 = 13%). While the authors made no mention of correcting p-values for multiple univariate testing, and so such values should be viewed as nominal, the authors noted that the magnitude of both SCP/DCP thinning and of the observed visual field defects increased with both disease activity and duration. The authors concluded that the extent, progression and response to therapy of BSCR is best monitored using multiple imaging modalities.

Roberts et al¹¹ recently reported on the use OCT and OCTA to access retinal thickness, retinal capillary density and foveal avascular zone (FAZ) integrity in 37 eyes of 21 subjects with BSCR for an average of 6 years (±3.5SD) seen in a referral clinic in Chicago, USA. The same number of eyes from 21 age-matched, healthy subjects served as controls. While the researchers found that both foveal thickness and FAZ area were similar in patients with BSRC and controls, parafoveal retinal thickness and SCP, DCP, and full retina vascular densities were significantly diminished in those with BSRC. Although these vascular densities were each inversely correlated with best-corrected vision, other potentially confounding factors may have contributed to vision loss, including ERM formation (65%), CME (22%), CNV (5%) and subretinal fluid (3%). A relatively high proportion of patients in the study (62%) received a fluocinolone acetonide implant.

Hafidi et al⁷ described their experience treating 21 patients with BSCR seen between 2003 and 2013 in a referral clinic in Lyon, France. All were Caucasian and HLA-A29 positive, and 11 were female. Fourteen received initial pulse methylprednisolone for three days (250–750 mg/day) followed by oral prednisone therapy; four were treated initially with oral prednisone (up to 1 mg/kg/day followed by taper); two received CSA (3–4 mg/kg/day); and one who was treated initially with interferon alpha-2b, but was quickly switched to CSA. The authors noted that 12 of 14 (86%) patients who received initial pulse methylprednisolone achieved control of their inflammation at 12 months, compared to four of seven (57%) who did not. The mean follow-up was just over four years (50.8 months; range 13–177 months) and all patients ultimately achieved quiescence, both clinically and angiographically. Twelve patients required long-term, low-dose prednisone (mean 5.9 mg/day; range 4-10 mg/day); seven required both low-dose prednisone and cyclosporine and one was controlled with interferon alpha-2b. Nine patients received intravitreal dexamethasone as adjunctive therapy for CME. While the number of subjects was small, follow-up variable, and the potential for referral, selection, and treatment bias very real, the authors suggested that initial pulse methylprednisolone therapy may allow faster control of inflammation, with less need for non-corticosteroid immunosuppressive therapy. Perhaps not unexpectedly given the authors’ preferred practices, treatment-related side effects were common and included corticosteroid-induced Cushingoid changes (52%) and systemic hypertension (29%), CSA-induced nephrotoxicity in three patients, and interferon associated asthenia in two patients. To minimize the risks of such complications, current guidelines recommend limiting corticosteroid exposure through routine use of non-corticosteroid immunosuppressive agents, such as MTX, MMF and/or a tumor necrosis factor (TNF) inhibitor.^12,13 If not otherwise contraindicated, 0.59 mg fluocinolone acetonide intravitreal implants (Retisert®) can also be used to achieve lasting control of the inflammation, but are associated with a high rate of ocular hypertension and glaucoma.^14–16 Although employed in the current study, the popularity of CSA as a corticosteroid-sparing immunosuppressive agent in patients with uveitis has decreased over time^17,18 for reasons related to both tolerability and efficacy.^19,20

Silpa-archa et al⁸ retrospectively analyzed 144 consecutive patients with typical BSCR and at least 12 months of follow up seen at a tertiary referral center in Boston, Massachusetts, USA, between 2005 and 2014. For comparison, subjects were categorized as early (< 50 years of age; n = 68) vs late-onset (≥ 50 years of age; n = 76) BSCR. Patients with less than 12-month follow-up (n = 73) or atypical features were excluded. The demographics of the cohort were as expected with BSCR, with a mean age of 51.9 +/- 10.8 years (range 21 – 77 years), a slight majority being female (63%), and essentially all Caucasian (99%). A frequency plot of age at presentation revealed a more or less normal distribution, illustrating the fact that half or more of all patients with BSCR are under 50 years of age. Only two-thirds of patients (67%) had active anterior chamber or vitreous inflammation on initial examination. The authors used univariate analyses to compare a number of independent demographic and clinical features at baseline between the early and late onset sub-cohorts, and only one – the presence of more than trace (> 0.5) anterior chamber (AC) cell, was differed between the two groups, with all four of those identified with more than trace AC cell occurring in subjects less than 50 years of age. However, such small numbers required the use of the nonparametric Fisher’s exact test and the authors made no mention of performing p-value corrections for multiple comparisons, so the clinical significance and generalizability of this result remain unclear. The authors also compared the rate of ocular complications. Overall, 98% of patients with BSCR had at least one complication, the most common of which were cataract (93%), ERM formation (33%), CME (30%), ocular hypertension/glaucoma (13%), and optic atrophy (9%). Only two subjects (1%) developed CNV. There were no observed differences in the rates of these complications based on age. Of note, non-corticosteroid immunosuppressive or biologic (NSIB) agents were ultimately used in 98% of subjects, 61% of whom were on three or more agents – the most common of which were MMF (86%), CSA (81%), infliximab (30%), and daclizumab (17%). The authors characterized the response to treatment as ‘good,’ which they defined as occurring within six months, in 61% of patients. Drug-free remission was achieved in 24% of subjects, following a mean treatment time of just over seven years (86 months). Univariate analyses suggested that NSIB agents were initiated within four months of presentation more frequently in younger patients (49% vs 26%; p = 0.01) and that younger patients tended to require more NSIB agents overall (2.6+/- 1.0 vs 2.3+/-1.0; p = 0.04), but, again, p-values as presented appeared not to have been corrected for multiple testing and so should be viewed as nominal. The authors noted that seven of the 217 patients (3%) with BSCR seen in their clinic over the period of the study had a positive family history of the condition. Testing revealed HLA-A29 positivity in only 95% of patients, which raises the question of how HLA-A29 testing was performed in study participants.

Calvo-Río et al⁹ described two patients with BSCR treated in a referral clinic in Madrid, Spain, who were refractory to conventional immunosuppressive therapy together with a TNF inhibitor, despite 40 mg/day of prednisone. Both were switched from their TNF inhibitor to the IL-6 inhibitor tocilizumab (TCZ; 8 mg/kg every 4 weeks). Prior to initiating TCZ, both patients had bilateral vitritis, diffuse vasculitis and CME associated with decreased vision. One-month after initiating TCZ, the patients experienced increased vision and improvement in their uveitis with resolution of their CME. Both were ultimately weaned from their prednisone and controlled for 10 and 18 months, respectively, with TCZ, which was well-tolerated with no associated adverse events. Others have reported similarly positive responses to TCZ in patients with BSCR and macular edema,^21–25 and other non-infectious forms of uveitis,²⁴ including promising results in a cohort of children with treatment refractory juvenile idiopathic arthritis (JIA).²⁶

Khanna et al¹⁰ utilized Adaptive Optics Scanning Laser Ophthalmoscopy (AOSLO) to study photoreceptors in 16 consecutive patients with HLA-A29 positive BSCR seen in a tertiary referral clinic in Chicago, Illinois, USA. Images of four eyes of three subjects were of adequate quality to allow alignment with infrared reflectance photos and correlated with spectral domain optical coherence tomography (SD-OCT) findings. In each case, inner segment/outer segment junction and interdigitation zone disruption or discontinuity on SD-OCT corresponded to reduced waveguiding ability of photoreceptors on en face AOSLO imaging. The authors suggested that in eyes with BSCR, discontinuities on SD-OCT in the outer retinal hyper-reflective bands normally associated with photoreceptor outer segments represent an absence of AOSLO waveguiding cones. They suggested further that that AOSLO might one day be useful tool for detecting photoreceptor change in patients with BSCR. Currently, however, AOSLO imaging of photoreceptors is available at relatively few academic centers, and its use is limited to analysis of the central macula and by a requirement for clear media and relatively normal overlying retinal anatomy.

Together, these studies highlight important issues related to the clinical presentation, diagnosis, and management of BSCR. While the central retina thickness is often spared until late in the disorder, diffuse retinal vasculitis, widespread outer retinal atrophy, CME and ERM formation are all common and can produce significant functional vision loss in the setting of relatively preserved Snellen visual acuity. Multimodal imaging with regular OCT, FA and ICGA is important, therefore, for assessing both foveal and extra-foveal disease activity and response to therapy. Many also perform periodic visual fields and electroretinography to follow retinal function. Although there is little controversy regarding the treatment of active CME or visually significant vitritis, broader determinants of therapy have not been well-studied and considerable center-to-center variability exists regarding the use of imaging, visual field, and electroretinography-based endpoints to either initiate or modify treatment. For those patients who require long-term therapy, most would recommend either non-corticosteroid immunosuppressive agents or sustained release corticosteroid implants. Shorter acting regional corticosteroids, in contrast, should be reserved for adjunctive control of disease recurrences or exacerbations.

Declaration of interest

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

Funding

Supported in part by The Pacific Vision Foundation (ETC) and The San Francisco Retina Foundation (ETC).

Additional information

Funding

Supported in part by The Pacific Vision Foundation (ETC) and The San Francisco Retina Foundation (ETC).