Usher Syndrome on the Island of Ireland: A Genotype-Phenotype Review

Abstract

Purpose: Usher syndrome (USH) is a genetically heterogeneous group of autosomal recessive (AR) syndromic inherited retinal degenerations (IRDs) representing 50% of deaf-blindness. All subtypes include retinitis pigmentosa, sensorineural hearing loss, and vestibular abnormalities. Thorough phenotyping may facilitate genetic diagnosis and intervention. Here we report the clinical/genetic features of an Irish USH cohort.

Methods: USH patients were selected from the Irish IRD registry (Target 5000). Patients were examined clinically (deep-phenotyping) and genetically using a 254 IRD–associated gene target capture sequencing panel, USH2A exon, and whole genome sequencing.

Results: The study identified 145 patients (24.1% USH1 [n = 35], 73.8% USH2 [n = 107], 1.4% USH3 [n = 2], and 0.7% USH4 [n = 1]). A genetic diagnosis was reached in 82.1%, the majority (80.7%) being MYO7A or USH2A genotypes. Mean visual acuity and visual field (VF) were 0.47 ± 0.58 LogMAR and 31.3° ± 32.8°, respectively, at a mean age of 43 years. Legal blindness criteria were met in 40.7%. Cataract was present in 77.4%. ADGRV1 genotypes had the most VF loss, whereas USH2A patients had greater myopia and CDH23 had the most astigmatism. Variants absent from gnomAD non-Finnish Europeans and ClinVar represented more than 20% of the variants identified and were detected in ADGRV1, ARSG, CDH23, MYO7A, and USH2A.

Conclusions: USH is a genetically diverse group of AR IRDs that have a profound impact on affected individuals and their families. The prevalence and phenotype/genotype characteristics of USH in Ireland have, as yet, gone unreported. Understanding the genotype of Irish USH patients may guide clinical and genetic characterization facilitating access to existing/novel therapeutics.

Usher syndrome (USH) is a syndromic inherited retinal degeneration (IRD) comprising retinitis pigmentosa (RP), sensorineural hearing loss (SNHL), and variable vestibular dysfunction.^1–6 The reported prevalence is three to 17:100,000.^2,3,7 It accounts for 18% of all RP^1,8 and 50% of the deaf/blind population⁹; however, USH represents only 3% to 13% of the congenitally deaf population.^8,10 The Usher Syndrome International Consortium⁶ classifies the disease into type 1 (USH1: profound congenital SNHL, vestibular dysfunction, and RP), type 2 (USH2: mild/moderate congenital SNHL, normal vestibular function, and RP), and type 3 (USH3: adolescent-/adult-onset progressive SNHL, progressive vestibular dysfunction, and RP).⁶ Limited cases of late-onset retinopathy with variable SNHL have been described as USH4^11–14 (Table 1). Subtype frequency varies by publication and geography, with USH3 and USH4 being substantially less frequent than USH1 and USH2 in Western populations.^2,6,15,16 

Table 1.

 

Clinical Subtypes of Usher Syndrome: Phenotype, Onset, and Associated Genes/Loci

 

Eleven genes (14 total loci) have been implicated in USH to date, all of which are inherited in an autosomal recessive (AR) fashion (Table 1), although HARS has been reported only in consanguineous Amish populations.¹⁷ Despite this genetic heterogeneity, two genes (MYO7A, USH2A) represent the majority of genetically resolved cases.¹⁸ USH proteins form an interactive unit localizing to ciliated structures including photoreceptors and cochlear ciliated epithelium.^8,19–21 Variable effects on respiratory cilia²² and fertility²³ have been reported. Some hypomorphic genetic variants can cause non-syndromic RP (e.g., USH2A is also the most frequent cause of nonsyndromic AR-RP)^24,25 and may cause isolated deafness (e.g., in MYO7A, USH1C, CDH23, PCDH15) even when monoallelic,^26–28 further compounding the diagnostic complexity. The characteristics and severity of the phenotype depend on the position and degree of deleteriousness of the causative variant. Natural history studies may provide prognostic biomarkers or therapeutic endpoints for novel therapies (e.g., MYO7A gene therapy and USH2A antisense oligonucleotides).^29–33 A detailed review of USH genetics and the role of USH protein complexes in eye and ear can be found in Delmaghani et al.³⁴ 

An Irish survey found that USH comprised 12.4% of all IRDs (n = 189/1522) and estimated the healthcare and social cost of USH to be €10.2 million per year, which did not account for the impact of SNHL (visual-only cost of blindness assessment).³⁵ USH's dual sensory impairment makes it a priority condition for developing interventions, both disease-modifying (e.g., retinal gene therapy, cochlear implants) and supportive (e.g., vision and mobility aids, educational supports).³⁶ Clinical and genetic characterization of USH patients may allow access to novel therapies via clinical trials which may benefit the individual (i.e., sensory restoration/preservation) and society (i.e., participation and economy). Herein, we describe the largest Irish USH population yet published with a focus on (1) clinical and genetic characteristics (including likely novel candidate genetic variants), (2) contrasting clinical features of differing genotypes, and (3) incidence of modifiable disease. 

Methods

Patients recruited to Target 5000^36–40 with a clinical USH diagnosis were included in this retrospective study. All patients signed informed consent to participate with the assistance of sign language translators where required. This study was approved by the institutional review boards of the participating sites (Royal Victoria Eye & Ear Hospital, Dublin, Ireland; Mater Misericordiae University Hospital, Dublin, Ireland; Belfast Health & Social Care Trust, Belfast, Northern Ireland; Trinity College Dublin, Dublin, Ireland) and abides by the Declaration of Helsinki. 

Ophthalmic phenotyping included best corrected logarithm of the minimum angle of resolution (LogMAR) visual acuity (VA), refraction, applanation tonometry and dilated slit lamp biomicroscopy. Refraction was assessed for all phakic eyes (preoperative refraction used when available) and a comparison made between clinical USH subtypes and genotypes. Multimodal imaging including ultra-widefield fundus photography and autofluorescence (“California,” Optos Plc, Dunfermline, UK) and spectral domain optical coherence tomography (OCT; Cirrus 5000; Carl Zeiss Meditec, Jena, Germany). Visual field (VF) to the IV4e target (VF; Goldmann Visual Field, Haag-Streit, Switzerland; or Humphrey Visual Field Analyser; Carl Zeiss MediTec) and International Society for the Clinical Electrophysiology of Vision standard electrophysiology were documented as available. Determination of USH subtype was suggested by the clinical phenotype (Table 1) in accordance with the Usher Syndrome Consortium⁶ and confirmed via genetic sequencing. Audiology was used to help categorize USH subtype; however, quantitative auditory and vestibular data were limited due to the ophthalmic focus of this study. 

For patients who underwent target capture NGS of 254 IRD-associated genes, DNA sequencing was performed as described previously.^37–41 Briefly, sequencing libraries were created and capture performed with the Nimblegen SeqCap EZ kit (Roche Ireland Ltd., Dublin, Ireland), incorporating the exonic regions of 254 IRD-associated genes as per the manufacturer's instructions. Sequencing data were aligned to the human genome (hg38) using BWA version 0.7.15. Downstream analysis and variant calling were performed using Freebayes version 1.1.0. Variants were annotated using SnpEff, dbNSFP, MetaLR and M-CAP.³⁸ USH2A exon sequencing and noncanonical splice site sequencing was performed as previously described.⁴² Briefly, 321 MIPs were designed to cover all coding regions of USH2A (NM_206933.4), as well as 20-nt of intronic sequences upstream and downstream of all exons and four currently known deep-intronic variants. Probes facilitated amplification and inclusion of an 8-nt barcode to each amplicon. Samples were sequenced on a NextSeq 500 system (Illumina, San Diego, CA, USA). Alignment and processing of data were performed using BWA mem and an in-house analysis pipeline. GATK was used to call SNVs and small indels. Two individuals underwent whole genome sequencing as described by Fadaie et al.⁴³ Variant analysis was performed as part of a study by Reurink et al.⁴⁴ The American College of Medical Genetics and Genomics and The Association for Molecular Pathology (ACMG-AMP) criteria for classifying pathogenic variants was utilised for interpretation of candidate variants.⁴⁵ Confirmatory testing and further analysis for unresolved cases was carried out at accredited laboratories (Blueprint Genetics, Helsinki, Finland and the Manchester Centre for Genomic Medicine). 

Retrospective data were collected from physical and electronic clinical and genetic records. Statistical analyses (descriptive statistics and comparison of means) were carried out using SPSS version 28 (IBM Corporation, Armonk, NY, USA) using analysis of variance and independent samples t-tests as appropriate. 

Results

Demographics

USH was identified in 145 probands (117 pedigrees), representing 76.7% of the estimated Irish USH population (Table 2).³⁵ Of them, 51.0% (74/145) were male. All were of white European ethnicity, with 98.6% of Irish ancestry. The clinical majority were USH2 (73.8%) and USH1 (24.1%) with infrequent USH3 (1.4%) and USH4 (0.7%) cases (Fig. 1). Family history of USH was found in 45.5% (n = 66) and of isolated deafness in 6.2% (n = 9). 

Table 2.

 

Demographics by USH Subtype, Genetic Resolution Rate, and Genotype

 

Figure 1.

 

USH subtypes and candidate causative genotype. Candidate genotypes are displayed by clinical USH subtype columns with the largest contributions being MYO7A and USH2A.

Symptoms

Mean age at diagnosis was 26.6 ± 12.0 years. Primary symptoms were night blindness (86.2%, n = 125, mean onset 17.5 ± 12.6 years), subjective VF constriction (75.2%, n = 109, mean onset 28.4 ± 13.4 years), glare (58.6%, n = 85, mean onset 30.4 ± 15.0 years), and reduced VA (55.2%, n = 80, mean onset 29.1 ± 15.1 years) (Supplementary Table S1). Age of visual symptom onset did not significantly differ between subtypes. Age at diagnosis of SNHL was <5 years for USH1/2/3 and early 4^th decade for USH4. Traditional hearing aids were used by 27.6% (n = 40), and cochlear implants by 2.8% (n = 4). 

Genetic Findings

A biallelic genotype was confirmed for 82.1% of cases (n = 119) and 86.3% (n = 101/117) of pedigrees, with 16 patients carrying variants of unknown significance (VUS) (Supplementary Table S2). A single candidate allele was detected in 4.1% (n = 6). Three or more USH-associated variants were detected in 2.8% (n = 4). Predominant genotypes were MYO7A (62.9% of USH1) and USH2A (69.2% of USH2). These two genes represented 80.7% (n = 96/119) of positive genotypes. Compound heterozygous genotypes represented the majority (62.2%, n = 74/119) of resolved cases. The most prevalent variants in MYO7A were c.2904G>T (15.6%) and c.631A>G (13.3%), in USH2A were c.2299delG (17.5%) and c.8981G>A (9.1%), in CDH23 was c.5237G>A (40.0%) and for ADGRV1 was c.6901C>T (40.6%). There were no instances of PCDH15, USH1G, C1B2, WHRN, or HARS genotypes in this cohort. A total of 96 candidate causative variants were identified (Supplementary Tables S2, S3). Of these 40.6% (n = 39) were categorized as missense variants, 24.0% (n = 23) as premature stop variants, 20.8% (n = 20) as frameshift variants, 9.4% (n = 9) as splice-altering variants, 3.1% (n = 3) as structural variants and 2.1% (n = 2) as small indels (Figs. 1, 2, Table 2). Of these variants 44.8% (n = 43) were classified as pathogenic, 38.5% (n = 37) as likely pathogenic, and 16.7% (n = 16) as VUS. VUS, representing 8.5% of candidate variant incidences reported here, none homozygous, have been included as these may be upgraded to (likely) pathogenic or downgraded to (likely) benign status with future evidence. Twenty (∼20.8%) likely novel candidate variants were detected including the largest deletion yet reported in the USH1C gene.³⁸ Of note, a synonymous splice altering variant (USH2A c.8709C>T) was detected in one individual, highlighting the contribution of these potentially overlooked disease causing variants. 

Figure 2.

 

Mutation types for candidate causative USH genotypes. The majority of USH-associated variants are missense, frameshift and nonsense with a small contribution from splice-altering, structural and copy number variants. CNV, copy number variation.

Visual Function

At mean age of 43.3 ± 15.8 years, mean BCVA was LogMAR 0.47 ± 0.58 with no statistically significant difference between genetic subtypes (ANOVA, P = 0.223) or genders (P = 0.102) (Table 3). BCVA was ≤0.3 LogMAR in 56.6% (n = 82), and ≥1.0 LogMAR in 9.7% (n = 14). VA was compared between MYO7A and USH2A genotypes by age in decade-wide ranges, showing a worsening trend for MYO7A from the sixth decade (increasing proportion of ≥1.0 LogMAR eyes and decreasing proportion of ≤0.3 LogMAR eyes (Fig. 3). Mean diameter of VF was 31.3° ± 32.8°. There was no statistically significant difference in VF diameter between genders (P = 0.608), the USH1 and USH2 groups (P = 0.468), nor between MYO7A and USH2A (P = 0.762). However, VF was statistically significantly worse for ADGRV1 (16.0° ± 12.7°) when compared with USH2A (31.5° ± 30.5°; P < 0.001), MYO7A (31.0° ± 35.8°; P = 0.027), and CDH23 (39.0 °± 35.7°; P = 0.038). When contrasting variant status (i.e., null/null, null/missense, null/splice altering, missense/missense, and missense/splice altering), there were no significant differences in VA or VF (Figs. 4C–F). Blind registration criteria were reached due to VF constriction in 45.8% (n = 44/96), compared with 11.5% (n = 15/130) for VA loss. 

Table 3.

 

Visual Acuity, Visual Field, and Legal Blindness Data

 

Figure 3.

 

( A ) Boxplot contrasting VA for USH2A and MYOYA genotypes demonstrating worsening median VA with age, more notably with MYO7A. Number of eyes with recorded VA per genotype for each age range is numerically displayed at the top of the table above each bar. (B) Proportion of eyes by decade age ranges with VA ≥1.0 LogMAR (i.e., legal blindness) for USH2A and MYO7A genotypes. (C) Proportion of eyes by decade age ranges with VA ≤0.3 LogMAR for USH2A and MYO7A genotypes.

Figure 4.

 

(A) VA box plots for ADGRV1, CDH23, MYO7A, and USH2A genotypes. (B) VF (diameter) box plots for ADGRV1, CDH23, MYO7A, and USH2A genotypes. (C) VA box plots by variant type for USH2A. (D) VF (diameter) box plots by variant type for USH2A. (E) VA box plots by variant type for MYO7A. (F) VF (diameter) box plots by variant type for MYO7A.

Modifiable Clinical Features

Anatomically modifiable diseases (i.e., cataract or cystoid macular lesions [CML]) were detected in 76.6% (n = 111/145) (Table 4). Visually significant cataract (i.e., cataract or pseudophakia) was present in 77.4% (n = 209/270) of eyes, being bilateral in 96.2% (n = 101/105) of these. VA was LogMAR 0.65 ± 0.70 and 0.46 ± 0.57 for pseudophakic and cataract eyes (P = 0.047), respectively. Pseudophakic patients were significantly older (53.7 ± 12.0 vs. 40.7 ± 14.8 years; P < 0.001), suggesting greater linear progression of maculopathy. Due to the retrospective nature of this study, we cannot suggest that cataract surgery causes worse vision in USH, but rather visual loss is multifactorial (i.e., cataract, CML, and macular atrophy). 

Table 4.

 

Modifiable Clinical Features of USH Phenotypes and Genotypes (Cataract, Refraction, Cystoid Macular Lesions)

 

CML were identified in 26.9% (n = 56 eyes, 34 patients) of 208 eyes with OCT, being bilateral in 64.7% (n = 22/34 patients). Mean VA of eyes with (0.50 ± 0.47 LogMAR) and without (0.48 ± 0.63) CML was not significantly different (P = 0.889), with this VA difference being confounded by variable severity of CML, lack of treatment naivete, and concomitant outer retinal atrophy. Sixteen of 34 (47.1%) CML patients were male, and CML prevalence was similar between USH subtypes. Age of patients with CML (42.9 ± 14.5 years) and those without (46.2 ± 16.2 years) was not significantly different (P = 0.318). CML was more severe in MYO7A eyes (334.2 ± 114.0 µm) than in USH2A eyes (246.1 ± 81.3 µm) (P = 0.035); however, VA was not significantly different between CML patients in these two genotypes (0.41 ± 0.26 vs. 0.49 ± 0.55; P = 0.513). Only 26.8% (n = 15/56) of CML eyes were pseudophakic, suggesting that postsurgical inflammation may not be a major driver of CML. Macular atrophy (subfoveal ellipsoid zone loss) was present in 41.8% (n = 87/208 eyes in 44 patients), 97.7% (n = 43/44) of which were bilateral (mean age 49.6 ± 16.2 years). Mean VA in atrophic eyes without CML was 0.82 ± 0.74 LogMAR, significantly worse than eyes with CML (P = 0.004). Mean central retinal thickness (CRT) for CML eyes was 287.3 ± 96.5 µm versus 226.7 ± 51.3 µm for non-CML eyes. Non-CML eyes with subfoveal atrophy had a mean CRT of 214.0 ± 52.0 µm. These differences were statistically significant for CML versus non-CML eyes (P < 0.001) and for CML versus non-CML atrophic eyes (P < 0.001). Outer retinal atrophy was noted in conjunction with CML in 28 eyes of 14 patients (50.0%), limiting the clinical utility of isolated CRT measurement in discriminating between atrophy, CML, or epiretinal membrane. 

Mean spherical equivalent refractive error for phakic eyes was −1.27 ± 6.02 diopters sphere (DS) with mean astigmatism of 1.47 ± 1.07 diopters cylinder (DC). There was no statistically significant difference in phakic refractive error between the USH1 (−1.53 ± 6.07) and USH2 (−2.56 ± 3.07) groups (P = 0.412). However, there was a statistically significant trend toward greater myopia in the USH2A group (−3.0 ± 2.39 DS) when compared with MYO7A (0.16 ± 2.44; P < 0.001), and ADGRV1 (−0.66 ± 1.13; P < 0.001) and approaching significance for CDH23 (−0.91 ± 2.14; P = 0.055). The only significant difference in astigmatism was between CDH23 (2.19 ± 1.37 DC) and ADGRV1 (1.07 ± 0.99 DC) (P = 0.023). 

Discussion

We describe the largest reported Irish USH cohort (145 patients) with 24.1% USH1, 73.8% USH2, 1.4% USH3, and 0.7% USH4. Biallelic genetic resolution rate was 82.1%, with a single USH-implicated allele found in 4.1%. VA ≤0.3 LogMAR was maintained in ≥1 eye in 63.1% at a mean age of 43.31 years. ADGRV1 patients had the greatest VF constriction. USH2A patients were significantly more myopic than other genotypes highlighting both a point of intervention and phenotypic guidance for genetic testing. CML (26.9%) and cataract (77.4%) were prevalent; however, pseudophakic individuals had worse VA, likely due to the more advanced retinopathy of older subjects. 

USH Subtype

This distribution of USH subtypes correlates well with publications from the UK,¹⁵ with MYO7A representing 30% to 50% of USH1 cases (65.7% in Ireland).⁴⁶ USH3 is more prevalent in certain populations (e.g., rural Finland, Ashkenazi Jews)^7,16,47 possibly due to the geographic isolation or genetic founder effects. Although similar geographical constraints exist in Ireland, few people with USH3 (n = 2) have been recruited to the study, perhaps due to the national history of mass emigration and few people of Finnish or Ashkenazi Jewish ethnicity recruited to Target 5000. Additionally, there may be poor engagement of specific insulated genetic groups with the health service (e.g., Irish Travellers, although USH-associated genetic variants are not reported in this group) or a true lack of USH3 and USH4 within the genetic landscape of the Irish people.⁴⁸ USH4 may be underdiagnosed clinically due to late-onset SNHL and retinal degeneration. 

Diagnosis/Phenotype

USH may be suggested by a failed hearing test in childhood as 3% to 13% of congenitally deaf children have USH.⁸ However, many USH2 patients attend mainstream schools as hearing loss is missed.⁵ Delayed walking or poor balance in the first years of life may suggest USH1,⁴⁹ but RP may not manifest until the late first to second decades. The RP of all USH subtypes is similar in clinical appearance⁶ and similarly recorded in this study. The differences are in age of onset and rate of progression (MYO7A reaching legal blindness 15 years before USH2A).²⁹ Audiology and genotype are more helpful in USH subtype determination than RP features.³⁰ 

Genetics

Of this total cohort 82.1% was fully resolved (i.e., biallelic USH-associated variants), with 23.1% (n = 6/26) of the unresolved remainder having one allele identified, helping to guide further focused genetic analysis such as single gene or long-read sequencing (i.e., introns and exons).^41,50 The c.2299delG variant is the most common pathogenic USH2A mutation associated with USH,⁵¹ which accounted for 17.5% of USH2A cases in our cohort (10.6% of candidate variants in any USH gene). The most commonly detected USH2A variant associated with nonsyndromic RP is the c.2276G>T variant. This was found in only 7.0% (n = 10) of our USH2A cohort in trans with null mutations (e.g., large deletion, nonsense) in patients with a consistent USH phenotype⁵² as previously reported in USH2.⁵³ USH2A variants causing USH have been reported as more deleterious than those associated with nonsyndromic RP.²⁵ 

The most common genotypes in this cohort were MYO7A and USH2A, together comprising 80.7% of resolved cases, consistent with international findings.²⁹ Although CLRN1 is reported as the most frequent USH3 genotype,¹⁶ only two cases were detected in our entire cohort. More deleterious mutations (e.g., frameshift, large deletions resulting in nonsense outcomes) in USH genes have been associated with a more aggressive phenotype both in terms of RP severity/progression and syndromic nature (i.e., USH rather than isolated AR-RP),²⁵ although 10% of missense/missense genotypes manifest USH phenotypes.⁴² 

Over 20% of candidate variants identified in this cohort were absent in non-Finnish Europeans in gnomAD or ClinVar, indicating a high degree of variants specific to Ireland in this cohort. Therefore this study contributes significantly to the repertoire of USH variants identified globally. Without an open-source population genome sequencing project in Ireland, the frequency of these variants in the general population could not be determined. However, the findings of this study are highly relevant to geneticists and clinicians globally particularly for USH patients of Irish descent. Homozygous candidate pathogenic variants were seen in 37.8% (n = 45/119) of resolved cases. This is higher than the 26.0% (n = 111/427) reported in a large European cohort⁵⁴ or the French data (11.3%, n = 26/231).⁵² As Ireland is an island with a relatively low population density, a higher rate of homozygosity may be expected. However, detecting the presence and frequency of superfamilies is not possible without more comprehensive sequencing techniques allowing haplotyping. 

Interventions

The SNHL of USH1 is treatable with pre-lingual cochlear implants, improving safety, communication and social interaction while the milder SNHL of USH2 may benefit from traditional amplifying hearing aids.^55,56 Although people with acquired blindness may develop enhancement of other senses via cortical modulation (e.g., absolute pitch in blind musicians),⁵⁷ the dual sensory impairment of USH cruelly limits this compensatory process. The RP associated with USH is not currently treatable; however, novel therapies have been approved (e.g., Luxturna in biallelic RPE65-retinopathy) or are in advanced human clinical trials for other IRDs including MYO7A and USH2A-associated USH.^58–62 Access to these novel therapies may be aided by accurate phenotype and genotype. Although disease-modifying techniques (e.g., gene therapies) are on the horizon, practical benefit can be gained from treatment of currently modifiable disease (e.g., refractive error, cataract, CML) and relevant supports can be accessed from first clinical contact.^36,63,64 

Refractive Error

Past studies have shown USH caused by MYO7A (−1.00DSE) or USH2A (−1.50DSE) to be associated with low myopia.^65,66 Our cohort concurred with this (mean −1.27 ± 6.02 DS) but interestingly found that USH2A patients, who have slower retinopathy progression,²⁹ were significantly more myopic than other genotypes (−3.00 ± 2.39 DS), and CDH23 patients had the most astigmatism (+2.19 ± 1.37 DC). Despite the limitation of RP, early refractive correction may encourage maximal visual development as demonstrated with other IRDs (e.g., RPGR-RP or RS1-retinoschisis).⁶⁶ Accurate documentation of extraretinal features including refraction may prove a useful composite tool to guide genetic testing choice (e.g., single gene testing/long-read sequencing) or aid variant interpretation (e.g., milder hearing loss, RP, higher myopia would suggest that an USH2A genotype is more likely than MYO7A). 

Cataract

Cataract was prevalent in all USH subtypes. VA was actually worse in pseudophakic eyes (0.65 ± 0.70) than in cataractous eyes (0.46 ± 0.57; P = 0.047). Although VA after cataract surgery is less predictable in the presence of macular pathology,^67,68 small retrospective studies have demonstrated no statistically significant difference in RP progression in phakic versus pseudophakic eyes.⁶⁹ Cataract surgery in RP is considered beneficial considering advanced VF constriction but this may be limited by macular dysfunction.⁷⁰ Contrast sensitivity may improve with cataract surgery, which was not formally assessed in this study. Clear crystalline lenses were documented in 22.6% (n = 61/270 eyes), younger (32.3 ± 15.4 years) than the group mean (43.31 ± 15.76 years). Following on from Piazza et al.³⁰ finding that VA of ≤0.3 LogMAR was maintained in at least one eye of 69% of USH1 and 94% of USH2 by age 29 years, we found that, at mean age of 43 years, the same level of vision was maintained in 59% of USH1 and 66% of USH2, showing a progressive diminution of VA with age, in keeping with the findings of Stingl et al.⁷¹ 

CML

CML may be (1) a factor of abnormal outer retinal metabolic fluid homeostasis (i.e., RPE pump/outer blood retinal barrier dysfunction),⁷² (2) inflammatory-mediated with glial activation (localizes to the INL/ONL),⁷³ and/or (3) representative of neuroretinal tissue loss/death.⁷⁴ Prevalence of CML varies in the literature from 13–49% of RP^75,76 and 15.7% of all USH subtypes⁷⁷ (26% of USH2 eyes).⁷⁸ Our figures indicate 26.9% CML across all subtypes (Table 4); however, actual figures may be higher as not all cases were treatment naïve and not all patients had OCT.⁶⁴ CML can be detected and monitored with OCT,⁷⁹ and severity can be graded to assist in monitoring of treatment response.⁸⁰ Younger patients have been reported to have CML more often⁸⁰; however, this was not borne out in our cohort with no significant age difference noted between patients with and without CML. Intuitively, end-stage macular atrophy is more likely in older adults with longer disease duration. 

The predominance of mild CML in all USH subtypes and the overlap with outer retinal atrophy in 50% obscures the contribution of each factor to vision loss and is a result of the retrospective nature of this analysis, which includes a wide age-range and diverse USH genotypes. Landmark studies excluded those previously treated⁸⁰ whereas our cohort had a mix of treated and treatment naïve patients. A review of RP-associated CML management demonstrated benefit from a practical stepwise approach of (1) topical carbonic anhydrase inhibitor (CAI, e.g., dorzolamide three times daily), (2) oral CAI, and (3) intravitreal steroid (e.g., triamcinolone).^63,64 

Blind Registration

The criteria for blind registration in Ireland are BCVA ≥1.0 LogMAR in the better seeing eye or widest monocular VF diameter ≤20°,⁸¹ which were met in 34.5% of USH patients in this study, mostly due to VF constriction. This is compounded by the restrictions of SNHL, making USH a severe obstacle to normal life (e.g., work, social integration, activities of daily living). 

Limitations

Assessment of visual function was from the most recent visit thus metrics of progression are not available, although Table 3 attempts to address this. Due to the retrospective multicenter nature of this dataset, treatment algorithms were not necessarily standardized between all clinical centers (e.g., CML treatment naïve). Indeed, CML may be under-represented in this cohort as older patients may have had significant CML before progression to macular atrophy. The focus of this study was ophthalmic, and thus quantitative data on SNHL/balance is limited. Symptom onset was patient reported and thus is subject to recall bias (i.e., hearing loss). Genotyping was performed via multiple sites with varying IRD-panel content thus true genetic resolve rate may be higher than reported here. 

Relevance

Irish descent is claimed by 11.4% of the global population.⁸² Thus genetic variants in Irish populations may have founder effects on IRD and systemic disease in the Western world. Irish communities abroad may not be assimilated, and concentrations of Irish parents may lead to rare AR inherited disease in their offspring.⁸³ USH is a devastating dual sensory impairment with an estimated economic impact of €10.2 million/year in Ireland.³⁵ As progress is made in gene-specific treatments for MYO7A⁸⁴ and USH2A (NCT01505062; NCT03780257), clinically and genetically characterizing the USH populations may facilitate access to these novel therapies for Irish patients and their international diaspora. 

Conclusion

Although no definitive cure for USH retinopathy is yet available, modifiable auditory and visual treatments and practical/social supports can be implemented now. Accurate clinical phenotyping and genetic characterization enables genetic counseling and determines prognosis and suitability for upcoming therapeutic clinical trials. A coordinated service in Ireland³⁶ has enabled genetic characterization of 63.0% of estimated USH cases with further testing possible for the unresolved population. This is the first comprehensive publication of clinical and genetic data on Usher Syndrome in Ireland, which has implications for clinical care, quality of life, and cost-effective use of novel treatments as they become available in Ireland and globally. 

Acknowledgments

The authors thank the Usher syndrome patients and families of Ireland; Fighting Blindness Ireland; the School of Genetics & Microbiology, Trinity College Dublin, D02 VF25. Mater Retinal Research Group. Retina Centre, Mater Private Hospital, Dublin, Ireland; Mater Misericordiae University Hospital, Dublin, Ireland; the Research Foundation, Royal Victoria Eye & Ear Hospital, Dublin, Ireland; and Belfast Trust Hospitals, Belfast, Northern Ireland. 

Supported by grants from Fighting Blindness Ireland (FB16FAR, FB18CRE), The Health Research Board of Ireland (POR/2010/97); Health Research Charities Ireland (MRCG-2013-8, MRCG-2016-14), the Irish Research Council (GOIPG/2017/1631), and Science Foundation Ireland (16/1A/4452). 

Disclosure: K.A.J. Stephenson, None; L. Whelan, None; J. Zhu, None; A. Dockery, None; N.C. Wynne, None; R.M. Cairns, None; C. Kirk, None; J. Turner, None; E.S. Duignan, None; J.J. O'Byrne, None; G. Silvestri, None; P.F. Kenna, None; G.J. Farrar, None; D.J. Keegan, None