Background/aims Vitreous haemorrhage (VH) and retinal detachment (RD) cause a precipitous decline in vision in a subset of patients with X-linked retinoschisis (XLRS), an otherwise a slowly progressive condition. This study aims to report the frequency of macular and peripheral retinal findings in a large cohort of patients with XLRS and to determine whether peripheral retinal findings are associated with VH and RD.
Methods A retrospective observational case series was performed in 65 patients with XLRS with a pathogenic variant in retinoschisin 1. Chart review included examination notes, fundus photographs and optical coherence tomography (OCT). Fisher exact tests and univariable logistic regression analysis were used to determine the association between peripheral retinal findings (including retinoschisis, metallic sheen, vascular sheathing, pigmentary changes, white spiculations and vitreous veils) and complications (including VH and RD).
Results Seven eyes (8%) showed normal macular structure on OCT. Peripheral retinoschisis was significantly associated with both VH and RD. Out of 10 eyes with complications, 9 (90%) had peripheral retinoschisis, compared with 33 out of 116 eyes (28%) without complications (p=0.0014). In addition, each additional peripheral finding increased the odds of RD by a factor of 4.06 (95% CI 1.58 to 10.39, p=0.028). There were no complications in the 28 eyes with a normal periphery (p=0.84) or in the 35 eyes with metallic sheen (p=0.42).
Conclusion The data suggest that patients with peripheral retinoschisis are at increased risk for VH and RD. Furthermore, patients with additional peripheral retinal findings together with peripheral schisis may carry additional risk for RD.
Statistics from Altmetric.com
X-linked retinoschisis (XLRS) is an inherited retinal degeneration presenting in the first decade of life and characterised by a splitting of the inner retina in the macula with cystic cavities. Approximately 50% of individuals have schisis cavities in the peripheral retina as well.1 Other peripheral retinal findings may include metallic sheen, retinal pigment epithelium (RPE) pigmentary changes, white spiculations, vascular sheathing and vitreous veils.1 The natural history is variable, with some patients experiencing deterioration in the first and second decades (20/60 to 20/120), while others remain relatively stable in childhood; further deterioration can occur later in adulthood secondary to outer retinal atrophy.2 Patients are at increased risk for vitreous haemorrhage (VH) and retinal detachment (RD), and these two complications primarily account for severe visual deterioration at a younger age in affected patients. RD is typically rhegmatogenous, caused by breaks in the inner and outer walls of the schisis cavity, but tractional detachments can also occur. VH is thought to occur in association with retinal breaks and vitreoretinal traction.
XLRS is caused by mutation in retinoschisin 1 (RS1), which has a discoidin domain and is implicated in cell–cell adhesion, consistent with the phenotype of schisis in the setting of RS1 mutations.3 RS1 is expressed in photoreceptors, but the protein is present in both the outer and inner retina.4
Given the potentially devastating visual consequences of VH and RD, a clinically relevant question is whether peripheral retinal findings are associated with these complications. It is biologically plausible that peripheral schisis would predispose patients, and other peripheral retinal changes may play a role as well. Others have noted peripheral schisis in patients with VH5 and RD,6 although prior reports lack statistical analysis. We hypothesise that peripheral schisis and other peripheral retinal changes are associated with an increased risk of VH and RD in patients with XLRS. This question was addressed with a retrospective chart review of 65 patients with molecularly proven XLRS.
This was a retrospective review of medical records of all patients with XLRS seen at Moorfields Eye Hospital, diagnosed based on clinical findings and family history, and confirmed by detection of a disease-causing variant in RS1. Patients without fundus photos on record were excluded. In total, 65 patients were ascertained, with visit dates between 1965 and 2015, and follow-up ranging from a single encounter to 48 years. Although some patients had visit dates as far back as 1965, all patients had visit dates recent enough to have genetic testing. Ethical approval was obtained from Moorfields Eye Hospital for this retrospective observational series.
All patients had undergone a comprehensive ophthalmological examination, including measurement of best-corrected visual acuity, intraocular pressure, slit-lamp examination, dilated fundus examination, colour fundus photography and spectral-domain optical coherence tomography (SD-OCT). In addition, a proportion of patients had measurement of colour vision, kinetic visual field and electrophysiology.
Colour fundus photography was performed with a TRC-50IA retinal fundus camera (Topcon, Tokyo, Japan) or an Optos ultra widefield camera (Optos, Scotland, UK). SD-OCT was performed with the Spectralis OCT (Heidelberg Engineering, Heidelberg, Germany). Electrophysiology included full-field electroretinogram (ffERG) and pattern ERG (pERG), recorded with gold foil electrodes. Protocols incorporated the recommendations of the International Society for Clinical Electrophysiology of Vision.7 8
The main outcome measures were the frequency of peripheral retinal findings, including retinoschisis, metallic sheen, vascular sheathing, pigmentary changes, white spiculations and vitreous veils; and frequency of complications, including VH and RD.
Fisher exact tests were performed on the 2×2 cross-tabulations of peripheral retinal findings and complications, with a correction for multiple testing for seven repeat tests (one for each peripheral finding, including a ‘normal’ periphery). p Values were adjusted based on the multiple-testing correction. Univariable logistic regressions were performed to test the effect of multiple peripheral findings on odds of complications. Fisher exact tests were performed on the cross-tabulations of genotype and peripheral retinal findings, with a correction for multiple testing for 210 repeat tests (35 genotypes × 6 peripheral retinal findings). p Values were adjusted based on the multiple-testing correction.
Patients and clinical findings
Sixty-five patients with molecularly proven XLRS seen between 1965 and 2015 were included. They ranged in age from 3 to 65 years. Best-corrected visual acuity ranged from 6/9 to no perception of light. Colour vision was documented in only four patients, ranging from 0/17 colour plates to 15/17. ffERGs, documented in 16 patients, ranged from normal (one patient) to rod or rod-cone abnormalities with an electronegative waveform. pERGs, documented in eight patients, were variably reduced.
RS1 sequence variants
Thirty-six different RS1 likely disease-causing variants were identified, including 2 large deletions, 3 single nucleotide deletions/duplications, 3 splice site mutations, 1 nonsense mutation and 27 missense mutations. Thirty mutations were previously reported (29 in the literature and 1 in the Leiden Open Variation Database)9–19 (table 1).
Ninety eyes in 47 patients had macular SD-OCT. Sixty-eight eyes (76%) showed macular schisis on SD-OCT, ranging in age from 5 to 36. Thirty-five eyes (39%) showed outer retinal atrophy, ranging in age from 29 to 64. Twenty eyes (22%) showed both macular schisis and outer retinal atrophy, ranging in age from 33 to 64. Seven eyes (8%) in four patients showed no schisis or atrophy, ranging in age from 11 to 45. The 45-year-old was an outlier, with the other patients ranging from 11 to 20 years of age. All seven eyes with a normal macula had peripheral retinal abnormalities. Four eyes had pigmentary changes, three eyes had peripheral schisis, two eyes had white spiculations and two eyes had metallic sheen. All eight patients had missense mutations in RS1. Out of the 43 patients who had SD-OCT in both eyes, 40 (93%) had concordant phenotypes in the two eyes, while 3 (7%) had discordant phenotypes.
Peripheral retinal phenotype
There were 126 eyes with fundus photos. Twenty-eight (24%) had a normal peripheral retina. Peripheral retinal findings, from most to least common, included retinoschisis in 42 eyes (33%), metallic sheen in 35 eyes (28%), vitreous veils in 24 eyes (19%), RPE pigment migration in 20 eyes (16%), white spiculations in 11 eyes (9%) and vascular sheathing in 8 eyes (6%) (table 2). Representative fundus photos of relevant peripheral retinal findings are shown in figure 1.
All peripheral findings were compared across genotypes to define genotype–phenotype correlations. All six eyes in all three patients with the c.214G>A(p.Glu72Lys) mutation had vitreous veils, which was statistically significant (p=0.021). The two most common mutations, c.304C>T(Arg102Trp) and c.305G>A(Arg102Gln), showed 8 out of 12 eyes with peripheral metallic sheen and 3 out of 10 eyes with peripheral white spiculations, respectively. However, this did not reach statistical significance.
Occurrences of VH and RD were determined by documentation in the patient notes, and were corroborated by fundus photos when available. There were four eyes with VH (3%) and eight eyes with RD (6%) (table 2). There were no RD or VH in eyes with a normal peripheral retinal exam (p=0.84), or in eyes with metallic sheen (p=0.42)(table 3).
Out of 10 eyes with complications, 9 (90%) had peripheral retinoschisis, compared with only 28% of eyes without complications (p=0.0014). This includes four out of four eyes with VH (p=0.07) and seven out of eight eyes with RD (p=0.014) that showed peripheral schisis (table 3). In addition, the number of peripheral findings was significantly associated with the odds of RD, with each additional peripheral finding increasing a patient’s odds of detachment by a factor of 4.06 (95% CI 1.58 to 10.39, p=0.028). There was no similar effect on the odds of VH.
The natural history of XLRS is highly variable, but if there is progression it is often very slow in childhood and early adulthood, potentially followed by additional gradual decline in later adulthood from macular atrophy. This chronic clinical course can be interrupted by a precipitous drop in vision due to complications of VH or RD, often at a young age. This study describes the frequency of these complications in a cohort of 65 patients from a single centre with molecularly proven XLRS, as well as the frequency of peripheral retinal findings. The study is limited by the small sample size, the retrospective design and the widely variable follow-up of subjects.
We report macular schisis in 76% of eyes in our cohort. Previous reports in the literature range from 68% to 100%.1 5 6 22 The age distribution in the cohort is expected to significantly impact the prevalence of macular schisis, since the natural history of disease is macular schisis followed by macular atrophy and resolution of schisis. Therefore, clinicians typically view macular atrophy as a presumptive sign of prior schisis. George et al and Kellner et al report macular schisis in 68% and 71% of eyes, respectively, but both report that 100% of eyes showed macular pathology on examination, in the form of schisis, blunted fovea, pigment irregularities or RPE atrophy.1 6 In contrast, we report seven eyes in four patients with normal macular structure on SD-OCT (8% of eyes). The previous reports, with the exception of Wang et al, were published before the RS1 gene was identified and the diagnosis was therefore based on clinical examination and family history, with family linkage to the XLRS locus in the case of George et al. In the current study the molecular diagnosis was confirmed in all patients, thus allowing identification of patients with a normal macula despite a pathologic RS1 variant. However, this may not fully explain the discrepancy as all seven eyes had peripheral retina abnormalities and perhaps would have been classified as XLRS even without genetic testing in the setting of a consistent family history. Previous investigations into genotype–phenotype correlations have found that patients with normal macular structure nearly always have missense mutations, although the inverse is not true: patients with those same missense mutations can have macular schisis and variable severity.11 This is consistent with our data. All four patients with normal macular structure had missense mutations. In three out of four cases, there were other patients with the same mutations that did show macular schisis. In the fourth case, there were discordant phenotypes between the two eyes: one eye had a normal macula with peripheral schisis, while the other eye had macular schisis, VH and RD. Prior genotype-phenotype investigations have also revealed worse visual acuity and more frequent ERG abnormalities among patients with non-sense, splice-site and frame-shift mutations compared with those with missense mutations.11 It is well established that an electronegative bright flash dark-adapted ERG is a characteristic finding in XLRS, although it is not universal. The b-wave amplitude is variable and the a-wave amplitude can also be reduced, especially in those with significant peripheral retinoschisis or pigmentary changes.11 Despite the large test–retest variability of ERG recordings and the inherent limitations associated with ERG assessment in children (important given the majority of complications are in childhood), future prospective studies of bright flash dark-adapted a-wave and b-wave amplitudes in relation to VH or RD may suggest these parameters as a functional marker for risk of complication.
We report 33% of eyes with peripheral retinoschisis, which is lower than previous reports, which have ranged from 43% to 60%.1 5 22 23 Similar to macular schisis, the rate of peripheral schisis may also be impacted by the age distribution of the cohort, since schisis cavities sometimes reabsorb with time. The rates of other peripheral findings have been reported in George et al1 (56 patients), in which reported rates were generally higher than the rates reported herein: metallic sheen in 38% of eyes (vs 28%), vitreous veils in 39% of eyes (vs 19%), pigmentary changes in 29% of eyes (vs 16%), white spiculations in 11% of eyes (vs 9%) and vascular sheathing in 9% of eyes (vs 6%). The complication rates reported here are also on the lower end of the range previously reported. We report VH in 3% of eyes, compared with 4%–21% in previous reports, and RD in 6% of eyes, compared with 5.5%–16% in previous reports.1 5 6 23
We report here that the presence of peripheral retinoschisis is significantly associated with the complications of VH and RD, and furthermore that the odds of RD increase significantly with each additional peripheral retinal abnormality. In addition, there were no complications in eyes with a normal periphery or with metallic sheen, although this did not reach statistical significance. There was one patient with a complication (RD) and no peripheral schisis, although we cannot rule out that he had a previous schisis cavity that spontaneously reabsorbed. Kellner et al6 propose that peripheral schisis is a risk factor for RD because peripheral schisis was present in 75% of fellow eyes in cases of RD, compared with 53% of all eyes without a history of RD. More recently, Wang et al5 reported VH in 4 out of 23 patients with XLRS, and all four had peripheral retinoschisis. This study adds to the previous reports, which lack statistical analysis, by demonstrating a statistically significant association between peripheral schisis and complication rates, as well as an increasing likelihood of RD with each additional peripheral fundus finding. A clinically relevant question is whether these peripheral retinal findings precede the complications or appear subsequently as a result of the complications. Of note, these complications frequently happen in childhood and adolescence. Sometimes patients initially present with VH or RD, and no prior records exist. Kellner et al6 report that all VH in their cohort occurred before the age of 10, and most RD occurred before the age of 11, with the exception of some long-standing detachments of unknown duration found later in life. Patients in our cohort also presented with complications at a young age. Patients with VH ranged from 8 to 21 years old, and patients with RD ranged from 3 to 21 years old, plus one 28-year-old with an old RD of unknown duration. Unfortunately in most cases there were no records prior to the complication, with two exceptions. In one case, a patient had an exudative RD at 21 years of age with previous documentation of vitreous veils in that eye at age 9 and peripheral schisis at age 14. Another patient had VH at age 11 and a tractional RD at age 12 with previously documented peripheral schisis and pigmentary changes in the same eye at age 5. This demonstrates that at least in these cases the peripheral retinal findings preceded the complications. Furthermore, despite having numerous eyes with a normal periphery, we have no documented cases of a patient with normal periphery developing a subsequent complication. Finally, the presence of these peripheral retinal findings in large numbers of patients with XLRS without complications argues that these findings are not a consequence of prior VH or RD. However, we cannot exclude that a patient may have had a VH that resolved without presenting to the ophthalmologist.
This study reports the frequency of both macular and peripheral retinal findings in a large cohort of patients with molecularly proven XLRS. The data suggest that patients with peripheral retinoschisis are at increased risk for VH and RD. Furthermore, patients with additional peripheral retinal findings together with peripheral schisis may carry additional risk for RD. A normal retinal periphery and a retinal metallic sheen may both portend a good prognosis with respect to risk for complications.
Professor Andrew Webster, MB, BChir, at Moorfields Eye Hospital in London, UK, managed many of the patients presented in this study and kindly gave us access to the data. He also provided feedback during manuscript preparation.
Contributors ATF participated in study design, data collection, manuscript preparation and editing.
NA participated in data collection and manuscript editing.
TB participated in statistical analysis and manuscript editing.
MM participated in study design, data collection and manuscript editing.
Funding This work was supported by Moorfields Eye Hospital Special Trustees (grant number ST 1310F), Moorfields Eye Charity (grant number MEC 13 10 A) and the Foundation Fighting Blindness (grant number CD-CL-0711–0518-UCL)
Competing interests ATF owns stock in Ionis Pharmaceuticals (Carlsbad, CA). MM consults for Meira GTx (London, UK), Ora Inc (Andover, MA), Shire (Hampshire, UK) and Editas (Cambridge, MA). NA and TB have no financial disclosures.
Ethics approval Moorfields Eye Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.