Abstract
Purpose
The purpose of this study was to characterize the clinical features, electrophysiology, and variant spectrum in ABCA4- and PRPH2-retinopathies and to identify novel electrodiagnostic biomarkers to differentiate between these two genotypes.
Methods
We conducted an international multicenter case-control study of patients with a clinical and genetic diagnosis of ABCA4- or PRPH2-retinopathy. Age at symptom onset, best-corrected visual acuity (BCVA), electroretinography (ERG) components, and fundus autofluorescence (FAF) imaging were compared. Subgroup exploratory analysis was performed on those patients with a phenotype similar to central areolar choroidal dystrophy (“CACD-like”), those with flecks distributed throughout the posterior pole (“fleck-like”) and a healthy control group. Receiver operating characteristic (ROC) analysis was performed to determine the optimal cutoff for ERG parameters to distinguish PRPH2- from ABCA4-retinopathy.
Results
This study included 155 patients with ABCA4-retinopathy, 133 patients with PRPH2-retinopathy, and 52 healthy controls. Significant electrophysiological biomarkers included the light-adapted (LA)30 hertz (Hz) flicker and LA3.0 single flash b-wave peak time (P < 0.001) for the “CACD-like” group with an area under the ROC curve (AUROC) of 0.78 and 0.76 and cutoff thresholds of 27.8 ms or 31.7 ms providing 77% and 85% sensitivity, respectively. Conversely, in the “fleck-like” group, the dark-adapted (DA)0.01 b-wave amplitude and DA3 a-wave amplitude (P < 0.001) had the highest AUROC, namely 0.88 and 0.88, respectively, with cutoff thresholds of 85 µV and 109 µV providing 93% and 88% sensitivity, respectively.
Conclusions
Unique ERG profiles can distinguish PRPH2- from ABCA4-retinopathy. The LA30 Hz peak time and DA0.01 b-wave amplitude may have clinical utility in predicting and interpreting the genotype in patients with overlapping retinal phenotypes.
Keywords: Stargardt disease, pseudo-Stargardt disease, ABCA4, PRPH2, pattern dystrophy, fundus flavimaculatus, electrophysiology, peripherinopathy, peripherin, RDS, imaging biomarkers
ABCA4-retinopathy (STGD1, OMIM #248200), caused by biallelic variants in the ATP-binding cassette (ABC), subfamily A, member 4 (ABCA4) gene, accounts for 12% of inherited retinal disease (IRD)-related blindness.1 Dominant variants in the peripherin-2/retinal degeneration slow (PRPH2) gene were the fourth most common cause of IRDs.2 The phenotype of PRPH2-retinopathy is variable and may manifest as central areolar choroidal dystrophy (CACD2, OMIM #613105), pattern dystrophy (PD, OMIM #169150), retinitis pigmentosa (RP7, OMIM #608133), butterfly-shaped pigment dystrophy (BPD) or a vitelliform macular dystrophy (VMD, OMIM #608161). ABCA4-retinopathy can also present with a wide and sometimes overlapping phenotypic spectrum from a Bull's eye maculopathy (BEM) to Stargardt disease 1 (OMIN #248200) and late onset fundus flavimaculatus (OMIN #248200). ABCA4 can masquerade as a PRPH2-associated CACD or a pattern dystrophy.3–6 Thus, clinically it can be challenging to distinguish PRPH2 from ABCA4-retinopathy.
Despite their similar fundus appearances subtle clues in the outer retina on optical coherence tomography (OCT) may help to distinguish ABCA4 from PRPH2.4 Similarly, quantitative autofluorescence (qAF) signals tend to be higher in ABCA4 as compared to PRPH2.5 Given that the production of the device that measures qAF has been discontinued, such analyses are difficult to implement in clinical practice. Conversely, electrophysiology is routinely performed to classify ABCA4-retinopathy into three groups namely; those with no generalized dysfunction (group 1), cone dysfunction (group 2), and cone and rod dysfunction (group 3).6 While there has been no such electroretinography (ERG) classification system published in PRPH2-retinopathy, animal studies have demonstrated that rods were more vulnerable to reduced and disordered oligomerization of PRPH2 than cones.7 Electrophysiological evidence of additional inner retinal involvement in PRPH2-retinopathy has also been reported.8 Hence, we hypothesize differences in ERG components may help to distinguish PRPH2- from ABCA4-retinopathy in patients with overlapping phenotypes without characteristic outer retinal bands due to macular atrophy. Given ERG is noninvasive, cost-effective, and readily accessible to most IRD clinics, this modality may be a useful adjunct to OCT imaging to distinguish between PRPH2- and ABCA4-retinopathy and the interpretation of variants of uncertain significance. Additionally, differences in ERG may provide insights into their underlying pathophysiology.
This study compared the ERG components in ABCA4- and PRPH2-retinopathy in patients presenting with a similar retinal phenotype. Our aim was to identify novel biomarkers to differentiate between these two commonly occurring IRDs.
Methods
Study Design
This was an international multicenter, case-control study. The study protocol adhered to the tenets of the Declaration of Helsinki and received approval from all local ethics committees of the participating institutions (University of Western Australia [2021/ET000151], Sir Charles Gairdner Osborne Park Health Care Group [RGS04985], Royal Victorian Eye and Ear Hospital/Centre for Eye Research Australia [19/1443H], New Zealand Ministry of Health [NTX/08/12/123], Auckland District Health Board [A+4290], the Save Sight Institute [2022/PID01932], Fondazione Bietti [NEU_01-2014], Vita-Salute San Raffaele University [MIRD2020], Erasmus Medical Center [NL34152.078.10], National Taiwan University Hospital [IRB 201408082RINC], Singapore ethics review board [2015/2766], the Hospital for Sick Children [REB-1000017804], and National Institute of Sensory Organs, National Hospital Organization Tokyo Medical Center [R18-029]). Informed consent was obtained from all patients and controls. Patients were recruited from October 2013 to October 2025.
Patient Identification
Inclusion criteria were molecular and phenotypic confirmation of ABCA4- or PRPH2-retinopathy. Patients were identified from the database of the Western Australian Retinal Degeneration (WARD) study, the Australian Inherited Retinal Disease Registry (AIRDR), and those of participating centers.
Molecular Diagnosis
A combination of next-generation sequencing, including panels of retinal dystrophy genes, whole exome sequencing (WES), whole genome sequencing (WGS), and Sanger sequencing, was used to identify pathogenic and likely pathogenic variants in ABCA4 and PRPH2. Molecular testing methods varied depending on the protocols of the participating centers. Phase testing was performed for all ABCA4 patients. Pathogenicity of suspected ABCA4 and PRPH2 variants was adjudicated in accordance with the guidelines set out by the American College of Medical Genetics (ACMG) and ClinGen SVI Working Groups, with reference to expert curations.9,10 Patients in the PRPH2-retinopathy group did not carry biallelic pathogenic or likely pathogenic variants in the ABCA4 gene. Conversely patients with ABCA4-retinopathy did not carry pathogenic or likely pathogenic variants in the PRPH2 gene.
Data Collection
Age, sex, genotype, age at ERG acquisition, best corrected visual acuity (BCVA), electrophysiological testing, and fundus autofluorescence (FAF) images were collected and analyzed. BCVA was measured on either a Snellen-style or a logMAR-style chart such as the Early Treatment Diabetic Retinopathy Study (ETDRS) letter chart. Retinal phenotype was characterized using ultrawide field (UWF) FAF imaging (Optos PLC, Dunfermline, UK) where available. Heidelberg HRA + OCT (Heidelberg Engineering, Heidelberg, Germany) 55 degrees or 30 degrees blue wavelength FAF images were graded when UWF images were not available. FAF was graded by two independent examiners (authors FKC and RHJ). A subgroup analysis was then performed on those patients that belonged to the following phenotypes: “CACD-like” or “fleck-like” (Figs. 1, 2). These FAF phenotypes were based on the two most common clinical presentations of ABCA4 and PRPH2, namely (1) circumscribed central macula involvement in the “CACD-like” group and (2) more diffuse hyperautofluorescent (hyperAF) flecks, which may occur anywhere in the posterior pole or extend past the arcades to the equator in the “fleck-like” group. The “fleck-like” group included those with subtle sparse flecks as well as those with more confluent and extensive flecks as long as these flecks were not solely confined to an oval region within the central macula. Atrophy within the foveal region was allowed. Those with extensive atrophy extending beyond the perifoveal region were excluded. Those patients with PRPH2 with other phenotypes including VMD, BPD, and retinitis pigmentosa (RP) were not included in the subgroup analysis. The “CACD-like” phenotype was originally described by Hoyng et al.11 and Boon et al.12 with a well-defined localized oval region of stippled hyper and hypoAF in the central macula with or without hypoAF patches within. The “fleck-like” phenotype resembled a classical Stargardt disease or a pattern dystrophy phenotype. Any patients with diffuse atrophy obliterating most of the posterior pole were not included in the subgroup analysis. Any grading disagreements were resolved by discussion and are discussed in further detail below.
Figure 1.
Similar fundus autofluorescence images of patients in the “CACD-like” group presenting with localized central macular abnormalities with a pathogenic or likely pathogenic variant(s) in the ABCA4 (A–D) or PRPH2 gene (E–H).
Figure 2.
Similar fundus autofluorescence images of patients in the “fleck-like” group presenting with multifocal flecks throughout the posterior pole with a pathogenic or likely pathogenic variant(s) in the ABCA4 (A–D) or PRPH2 gene (E–H).
Electrophysiologic Assessment
Electrophysiology was performed in accordance with the International Clinical Electrophysiology of Vision Society Standards (ISCEV).13–15 Full-field ERG and pattern ERG were performed. Full-field ERG traces were reviewed, component parameters were extracted, and data were compared with a control group of healthy subjects from Sir Charles Gairdner Hospital, Perth, Australia, which included validated recordings for DA0.01, DA3.0, DA10.0, LA3.0, and LA30 hertz (Hz) ERGs (n = 52). Pattern ERG data were available for 24 of 154 patients (16%) with ABCA4-retinopathy and 69 of 133 patients (52%) with PRPH2-retinopathy.
Statistical Analysis
Collected quantitative data were presented in tables as mean, median, and standard deviation (SD). In the case of qualitative variables, data were reported as percentages. Patient demographics were summarized and compared among the three groups. Exploratory analysis was used with receiver operating characteristics (ROC) and the Youden’s J statistic (or Youden’s Index) to establish a cutoff for the LA and DA ERG parameters, which may distinguish PRPH2 from ABCA4. Model sensitivity, specificity, accuracy, and area under the ROC curve (AUROC, 95% confidence interval [CI]) were reported at the optimal cutoff. Significance was achieved if P < 0.05. All analyses were performed using R software version 4.1.3 (R Core Team, Vienna, Austria) and R Studio software version 2024.12.1 (Rstudio Team, Boston, MA, USA). Sensitivity analysis was performed on select ERG components in a subset of patients by calculating a percentage of the lower limit of normal for the amplitude or the upper limit of normal for the peak time to standardize against a site-specific age-adjusted normative database. Correction for multiple statistical testing was not performed given this was an exploratory analysis.
Results
Demographics
In total, the charts of 155 patients with (likely) pathogenic variants in ABCA4 and 133 patients with (likely) pathogenic variants in PRPH2 were ascertained for phenotyping, from 11 tertiary referral centers in 8 countries (Australia, New Zealand, Italy, Netherlands, Canada, Singapore, Taiwan, and Japan; raw data in Supplementary Material S1). The cohort included 164 female patients (57%) and 124 male patients (43%). Electrophysiological data on our patients with PRPH2 was previously reported in part by Heath Jeffery et al.4 (n = 12), Bianco et al.16 (n = 19), Antonelli et al.17 (n = 28), and Heath Jeffery et al.3 (n = 100). The mean ± SD age of symptom onset was 39 ± 14 years (range = 5–78 years) in the PRPH2 group compared to 24 ± 17 years (range = 4–82 years) for the ABCA4 group (Table 1). In addition, 52 healthy controls were included for comparison with ERG performed at a mean ± SD age of 50 ± 19 years (range = 13–77 years).
Table 1.
Baseline Demographics
| Variable | Level | PRPH2 | ABCA4 | Controls |
|---|---|---|---|---|
| Number of subjects | 133 | 155 | 52 | |
| Age at ERG/VA testing, mean (SD) | 52 (14) | 36 (18) | 50 (19) | |
| Age of onset, mean (SD, range) | 39 (14, 5–78) | 24 (17, 4–82) | ||
| Gender | Female | 66 (50%) | 67 (43%) | 22 (42%) |
| Male | 67 (50%) | 88 (57%) | 30 (58%) | |
| VA (LogMAR), mean (SD) | Right | 0.3 (0.5) | 0.6 (0.6) | |
| Left | 0.2 (0.4) | 0.7 (0.6) | ||
| FAF phenotype | CACD-like | 33 (25%) | 49 (32%) | |
| Fleck-like | 53 (40%) | 44 (28%) |
CACD, central areolar choroidal dystrophy; ERG, electroretinography; ETDRS, Early Treatment of Diabetic Retinopathy Study; FAF, fundus autofluorescence; LogMAR, logarithm of the Minimum Angle of Resolution; SD, standard deviation; VA, visual acuity.
Genetic Results
In total, 60 (likely) pathogenic alleles were identified in the PRPH2 group (see Supplementary Material S1) of which more than one third (36%) were represented by 1 of 5 recurrent variants: (1) c.514C > T p.(Arg172Trp) (n = 13); (2) c.499G > A p.(Gly167Ser) (n = 12); (3) c.515G > A p.(Arg172Gln) (n = 9); (4) c.533A>G p.(Glu178Arg) (n = 7), and (5) c.290G > A p.(Trp97Ter) (n = 7). Most of the PRPH2 variants were missense (n = 29/60) followed by frameshift (n = 14/60) and nonsense (n = 11/60). In total, 123 (likely) pathogenic alleles were identified in the ABCA4 group of which 20 were complex alleles (see Supplementary Material S1). Nearly one third (32%) of all alleles presented as 1 of 5 recurrent variants in compound heterozygous or homozygous form: (1) c.[2588G > C;5603A > T] p.[Gly863Ala;Asn1868Ile] (n = 22); (2) c.5882G > A p.(Gly1961Glu) (n = 21); (3) c.5603A > T p.(Asn1868Ile) (n = 21); (4) c.[5461-10T > C;5603A > T] p.(Thr1821ValfsTer13,Thr1821AspfsTer6;Asn1868Ile) (n = 18), and (5) c.6079C > T p.(Leu2027Phe) (n = 17). Most of the ABCA4 alleles were missense variants (n = 73) followed by variants altering splice sites (n = 22), nonsense (n = 15), and frameshifting (n = 13).
Visual Acuity and Fundus Autofluorescence Findings
BCVA was analyzed cross-sectionally. Two hundred eighty-eight patients had visual acuity (VA) available at one visit when ERG testing and FAF imaging were performed (raw data in Supplementary Material S1). None of the patients had other concurrent vision-limiting disease. The mean ± SD BCVA (logMAR) was 0.3 ± 0.5 and 0.2 ± 0.4 for the right and left eyes, respectively, in the PRPH2 group (see Table 1). The mean BCVA (logMAR) was 0.6 ± 0.6 and 0.7 ± 0.6 for the right and left eyes, respectively, in the ABCA4 group. The mean ± SD age at BCVA/ERG testing was 52 ± 14 years in the PRPH2 group compared with 36 ± 18 years in the ABCA4 group. The mean ± SD age of patients in the subgroups “CACD-like” and “fleck-like” is shown in Table 1. A “fleck-like” FAF phenotype was observed in 44 of 155 (28%) patients with ABCA4 and 53 of 133 (40%) patients with PRPH2, whereas a “CACD-like” phenotype was observed in 49 of 155 (32%) patients with ABCA4 and 33 of 133 (25%) patients with PRPH2. There were eight grading disagreements when classifying the cohort into “fleck-like” and “CACD-like” groups between the two graders, namely, five for the “CACD-like” group and three for the “fleck-like” group. One of these disagreements arose from the FAF phenotype changing over time. Thus, it was decided that only the FAF phenotype at the time of the ERG would be used. The remaining disagreements arose when a few subtle flecks were present outside the central macula region. It was decided that if any flecks were observed outside the central macular region the phenotype would be graded “fleck-like.”
Electrophysiology
Full-field ERG parameters were assessed across the entire ABCA4 and PRPH2 cohort (raw and summary data in Supplementary Material S1). Overall, significant predictors of PRPH2 (for any phenotype) included light-adapted (LA)3 b:a ratio <3.7, LA30 Hz peak time >30 ms and amplitude of <54 µV, and a DA3 a-wave amplitude of <101 µV and b-wave amplitude of <191 µV (P < 0.001 for all 5 variables). The corresponding AUROC ranged from 0.62 to 0.68 with 61% to 70% of sensitivity and 57% to 71% of specificity.
For the subgroup analysis, the largest statistically significant AUROC in the “CACD-like” group phenotype were observed under LA conditions (Table 2; Fig. 3). The AUROC (95% CI) was 0.78 (0.67–0.89) for LA30 Hz flicker peak time (P < 0.001) and 0.76 (0.64–0.87) for LA3 single flash b-wave peak time (P < 0.001) with cutoff thresholds of >27.8 ms and >31.7 ms providing 77% and 85% sensitivity and 73% and 57% specificity for predicting PRPH2, respectively. The AUROC was 0.78 (0.67–0.89) for the LA3 b:a ratio (P < 0.001) with a cutoff threshold <3.7 providing 77% sensitivity and 72% specificity for predicting PRPH2. Note the range for the LA3.0 b:a ratio for the control group was 3.0 to 6.7. Pattern ERG P50 component amplitude reduction to <0.1 µV was also a significant predictor of PRPH2 with 98% sensitivity and 58% specificity (P < 0.001). The difference in LA30 Hz peak time between ABCA4 and PRPH2 cohorts increased with age (Supplementary Material S2).
Table 2.
ROC Analysis in the “CACD-Like” Group
| Parameter | PRPH2 Vs. ABCA4 | Threshold | Specificity | Sensitivity | Accuracy | Precision | AUROC (95% CI) | P Value |
|---|---|---|---|---|---|---|---|---|
| DA0.01 | ||||||||
| b-wave peak time (ms) | > | 99.20 | 0.36 | 0.88 | 0.68 | 0.70 | 0.56 (0.40, 0.71) | 0.425 |
| b-wave amplitude, µV | < | 97.41 | 0.31 | 0.94 | 0.70 | 0.69 | 0.56 (0.42, 0.70) | 0.380 |
| DA3 | ||||||||
| a-wave peak time (ms) | < | 20.38 | 0.67 | 0.85 | 0.78 | 0.80 | 0.73 (0.60, 0.86) | <0.001 |
| a-wave amplitude, µV | < | 99.62 | 0.32 | 0.94 | 0.69 | 0.68 | 0.53 (0.38, 0.68) | 0.672 |
| b-wave peak time, ms | < | 49.25 | 0.48 | 0.81 | 0.68 | 0.70 | 0.57 (0.43, 0.72) | 0.283 |
| b-wave amplitude, µV | > | 261.20 | 0.60 | 0.51 | 0.55 | 0.67 | 0.45 (0.31, 0.59) | 0.461 |
| b:a ratio | < | 1.44 | 0.39 | 0.89 | 0.71 | 0.71 | 0.62 (0.48, 0.76) | 0.088 |
| DA10 | ||||||||
| a-wave peak time, ms | < | 14.30 | 0.50 | 0.85 | 0.78 | 0.88 | 0.62 (0.37, 0.87) | 0.244 |
| a-wave amplitude, µV | > | 335.38 | 0.40 | 0.95 | 0.84 | 0.87 | 0.53 (0.25, 0.81) | 0.770 |
| b-wave peak time, ms | < | 34.20 | 0.43 | 1.00 | 0.86 | 0.85 | 0.65 (0.34, 0.95) | 0.262 |
| b-wave amplitude, µV | > | 329.05 | 0.50 | 0.68 | 0.65 | 0.85 | 0.45 (0.20, 0.69) | 0.610 |
| b:a ratio | < | 1.53 | 1.00 | 0.46 | 0.57 | 1.00 | 0.77 (0.62, 0.92) | 0.009 |
| a-wave slope | < | 14.70 | 0.75 | 0.96 | 0.93 | 0.96 | 0.86 (0.65, 1.00) | 0.025 |
| LA3 | ||||||||
| a-wave peak time, ms | > | 15.95 | 0.46 | 0.77 | 0.66 | 0.71 | 0.61 (0.47, 0.74) | 0.125 |
| a-wave amplitude, µV | < | 15.70 | 0.40 | 0.85 | 0.68 | 0.69 | 0.53 (0.38, 0.68) | 0.648 |
| b-wave peak time, ms | > | 31.65 | 0.57 | 0.85 | 0.74 | 0.76 | 0.76 (0.64, 0.87) | <0.001 |
| b-wave amplitude, µV | < | 76.15 | 0.60 | 0.81 | 0.73 | 0.76 | 0.66 (0.52, 0.80) | 0.021 |
| b:a ratio | < | 3.65 | 0.72 | 0.77 | 0.75 | 0.82 | 0.78 (0.67, 0.89) | <0.001 |
| LA30 Hz | ||||||||
| Flicker peak time, ms | > | 27.76 | 0.73 | 0.77 | 0.76 | 0.82 | 0.78 (0.67, 0.89) | <0.001 |
| Flicker amplitude, µV | < | 53.82 | 0.47 | 0.90 | 0.73 | 0.73 | 0.66 (0.52, 0.79) | 0.021 |
| On-Off response | ||||||||
| b-wave amplitude, µV | < | 29.61 | 0.75 | 0.73 | 0.73 | 0.94 | 0.76 (0.46, 1.00) | 0.112 |
| d-wave amplitude, µV | < | 23.96 | 1.00 | 0.76 | 0.80 | 1.00 | 0.88 (0.74, 1.00) | 0.015 |
| Pattern ERG | ||||||||
| P50 peak time, ms | < | 54.10 | 0.75 | 0.69 | 0.70 | 0.97 | 0.54 (0.17, 0.91) | 0.800 |
| P50 amplitude, µV | < | 0.10 | 0.58 | 0.98 | 0.86 | 0.84 | 0.79 (0.64, 0.93) | <0.001 |
DA0.01, dark-adapted 0.01 cd/m2; DA3, dark-adapted 3 cd/m2; DA10, dark-adapted 10 cd/m2; LA30 Hz, light-adapted 30 Hz flicker at 3 cd/m2; LA3, light-adapted 3 cd/m2 single flash at 2 Hz.
“<” threshold value below which, and “>” threshold value above which would predict a PRPH2-retinopathy.
Figure 3.
Receiver operating curve (ROC) analysis graphs showing the area under the curve (AUC) for the LA30 Hz flicker peak time (A), LA3 b-wave peak time (B), LA3 b:a ratio (C), and PERG P50 amplitude (D) comparing patients with PRPH2- to ABCA4-retinopathy in the “CACD-like” group.
In contrast, the largest statistically significant AUROC in the “fleck-like” group were observed under DA conditions (Table 3; Fig. 4). The AUROCs were 0.88 (0.81–0.95) and 0.88 (0.81–0.95) for the DA0.01 b-wave amplitude and DA3 a-wave amplitude (P < 0.001) with cutoff thresholds of <85 µV and <109 µV providing 93% and 88% sensitivity and 74% and 75% specificity, respectively, for predicting PRPH2. The AUROC was 0.86 (P < 0.001) for the DA3 b-wave amplitude with a cutoff threshold of <191 µV giving 81% sensitivity and 81% specificity for predicting PRPH2. Similarly, the AUROC were 0.83 and 0.81 for the DA10 a- and b-wave amplitude (P < 0.001) with cutoff thresholds of <127 µV and <204 µV providing 80% and 66% sensitivity and 75% and 85% specificity, respectively, for predicting PRPH2. There was a trend for a more rapid reduction in the a-wave and b-wave amplitudes in the PRPH2 group with increasing age (see Supplementary Material S2).
Table 3.
ROC Analysis in the “Fleck-Like” Group
| Parameter | PRPH2 Vs. ABCA4 | Threshold | Specificity | Sensitivity | Accuracy | Precision | AUROC (95% CI) | P Value |
|---|---|---|---|---|---|---|---|---|
| DA0.01 | ||||||||
| b-wave peak time, ms | > | 105.94 | 0.34 | 0.88 | 0.64 | 0.62 | 0.55 (0.42, 0.69) | 0.442 |
| b-wave amplitude, µV | < | 84.57 | 0.74 | 0.93 | 0.84 | 0.80 | 0.88 (0.81, 0.95) | <0.001 |
| DA3 | ||||||||
| a-wave peak time, ms | > | 24.41 | 0.35 | 0.93 | 0.61 | 0.54 | 0.62 (0.50, 0.74) | 0.053 |
| a-wave amplitude, µV | < | 108.59 | 0.75 | 0.88 | 0.81 | 0.74 | 0.88 (0.81, 0.95) | <0.001 |
| b-wave peak time, ms | < | 49.40 | 0.50 | 0.95 | 0.70 | 0.61 | 0.65 (0.53, 0.77) | 0.012 |
| b-wave amplitude, µV | < | 191.02 | 0.81 | 0.81 | 0.81 | 0.77 | 0.86 (0.78, 0.93) | <0.001 |
| b:a ratio | > | 1.84 | 0.45 | 0.90 | 0.66 | 0.58 | 0.66 (0.55, 0.77) | 0.010 |
| DA10 | ||||||||
| a-wave peak time, ms | > | 16.45 | 0.95 | 0.49 | 0.64 | 0.95 | 0.76 (0.63, 0.89) | 0.001 |
| a-wave amplitude, µV | < | 127.17 | 0.75 | 0.80 | 0.79 | 0.87 | 0.83 (0.71, 0.94) | <0.001 |
| b-wave peak time, ms | < | 50.75 | 0.57 | 0.67 | 0.64 | 0.77 | 0.52 (0.24, 0.81) | 0.888 |
| b-wave amplitude, µV | < | 204.48 | 0.85 | 0.66 | 0.72 | 0.90 | 0.81 (0.69, 0.93) | <0.001 |
| b:a ratio | < | 1.50 | 0.70 | 0.49 | 0.56 | 0.77 | 0.50 (0.33, 0.66) | 0.982 |
| a-wave slope | < | 15.20 | 0.94 | 0.76 | 0.82 | 0.96 | 0.90 (0.80, 0.99) | <0.001 |
| LA3 | ||||||||
| a-wave peak time, ms | < | 16.92 | 0.37 | 0.93 | 0.68 | 0.65 | 0.61(0.47, 0.74) | 0.112 |
| a-wave amplitude, µV | < | 21.34 | 0.87 | 0.47 | 0.65 | 0.80 | 0.67 (0.55, 0.79) | 0.008 |
| b-wave peak time, ms | > | 32.55 | 0.56 | 0.93 | 0.76 | 0.71 | 0.75 (0.63, 0.86) | <0.001 |
| b-wave amplitude, µV | < | 75.06 | 0.82 | 0.65 | 0.73 | 0.80 | 0.81 (0.72, 0.90) | <0.001 |
| b:a ratio | < | 3.75 | 0.71 | 0.77 | 0.74 | 0.77 | 0.79 (0.69, 0.89) | <0.001 |
| LA30 Hz | ||||||||
| Flicker amplitude, µV | < | 54.03 | 0.77 | 0.84 | 0.80 | 0.75 | 0.86 (0.78, 0.93) | <0.001 |
| Flicker peak time, ms | > | 29.65 | 0.69 | 0.86 | 0.77 | 0.70 | 0.76 (0.65, 0.86) | <0.001 |
| On-Off response | ||||||||
| b-wave amplitude, µV | < | 18.66 | 0.50 | 0.96 | 0.82 | 0.81 | 0.69 (0.49, 0.90) | 0.062 |
| d-wave amplitude, µV | < | 22.45 | 0.58 | 0.77 | 0.71 | 0.80 | 0.67 (0.46, 0.87) | 0.109 |
| Pattern ERG | ||||||||
| P50 peak time, ms | > | 53.85 | 0.88 | 0.29 | 0.48 | 0.83 | 0.53 (0.36, 0.71) | 0.708 |
| P50 amplitude, µV | > | 1.79 | 0.65 | 0.49 | 0.54 | 0.73 | 0.52 (0.36, 0.68) | 0.798 |
DA0.01, dark-adapted 0.01 cd/m2; DA3, dark-adapted 3 cd/m2; DA10, dark-adapted 10 cd/m2; LA30 Hz, light-adapted 30 Hz flicker at 3 cd/m2; LA3, light-adapted 3 cd/m2 single flash at 2 Hz.
“<” threshold value below which, and “>” threshold value above which would predict a PRPH2-retinopathy.
Figure 4.
Receiver operating curve (ROC) analysis graphs showing the area under the curve (AUC) for the DA0.01 b-wave amplitude (A), DA3 a-wave amplitude (B), DA3 b-wave amplitude (C), and DA10 a-wave amplitude (D) comparing patients with PRPH2- to ABCA4-retinopathy in the “fleck-like” group.
Patients with a “CACD-like” phenotype demonstrated a greater reduction in the P50 amplitude for the PERG responses with PRPH2 when compared to ABCA4 despite displaying a similar FAF phenotype (P < 0.001). In the “fleck-like” group, no statistically significant difference was observed in PERG parameters. There was no improvement in the AUROC when ERG components were analyzed as percentages of the lower or upper bounds of the amplitude and peak times, respectively. Figures 5A and 5B show two patients with a “CACD-like” FAF phenotype and contrasting ERG findings. Figures 5C and 5D show an example of two patients with a “fleck-like” FAF phenotype and contrasting ERG findings.
Figure 5.
Fundus autofluorescence (FAF) imaging, optical coherence tomography (OCT) line scans through the central macula, ERG tracings for dark-adapted (DA)0.01 cd/m2, 3.0 cd/m2, light-adapted (LA) LA3.0 cd/m2 single flash and LA30 Hz flicker for two patients with a “CACD-like” phenotype (A, B) and 2 patients with a “fleck-like” phenotype (C, D) manifesting PRPH2- and ABCA4-retinopathy respectively as well as a normal control (E) for comparison. In the “CACD-like” phenotype A versus B, PRPH2-retinopathy had a more delayed b-wave peak time (tb) and flicker peak time (tf). For the “fleck-like” phenotype, PRPH2-retinopathy had a lower b-wave amplitude in the DA0.01 and DA3 than ABCA4-retinopathy C versus D. Overall, the b:a ratio for both the DA3 and LA3 were lower in PRPH2- compared to ABCA4-retinopathy. The FAF images for A and B show 30 × 30-degree field. The images for C and D show ultrawide field Optos FAF. The vertical and horizontal scale bars on the OCT mark 200 µm.
Discussion
Despite similar clinical phenotypes on retinal imaging this study found characteristic differences on ERG testing that may help to differentiate PRPH2 from ABCA4-retinopathy. To our knowledge, this is the first use of ERG parameters to distinguish between these two commonly occurring IRDs. Without accounting for the FAF phenotype, a reduced DA3 a- and b-wave amplitude were predictive of a PRPH2-retinopathy as compared to ABCA4. When considering the FAF phenotype, PRPH2-retinopathy was associated with greater cone involvement in the “CACD-like” group and greater rod involvement in the “fleck-like” group as compared to ABCA4-retinopathy.
Most studies comparing ABCA4 to PRPH2 have focused predominately on structural differences using FAF or OCT imaging. We have previously reported an altered outer retinal band profile whereby the thickness of ellipsoid zone/retinal pigment epithelium (RPE; band 2/band 4) ratio was able to discriminate between PRPH2- and ABCA4-retinopathy with an AUC of 0.87.4 Miere et al.18 had retinal experts’ grade FAF images with an accuracy, sensitivity, and specificity of 0.8, 0.8, and 0.8, respectively, in distinguishing ABCA4 from PRPH2. In contrast, retinal fellows had an accuracy, sensitivity, and specificity of 0.7, 0.6, and 0.6, respectively, whereas a deep learning classifier (ResNet50V2) identified 88 of 91 ABCA4 and 10 of 20 PRPH2 correctly with an AUC of 0.9. Duncker et al.19 used qAF to differentiate patients with ABCA4 from PRPH2. Unexpectedly, qAF was not useful in differentiating PRPH2 from ABCA4 given the unexpectedly higher than normal qAF level in some PRPH2 cases. Notably, no studies to date have focused on functional measures to differentiate between these two diseases.
Our work highlights that an LA3 b:a ratio <3.7 is a significant predictor of PRPH2-retinopathy. In our previous study3 of 241 patients with PRPH2-retinopathy, we found 7 (8%) patients who had a reduced b:a ratio (<1.2) for the DA3, whereas 31 (31%) patients had a reduced b:a ratio (<3.0) for the LA3. Hence, a reduced b:a ratio occurred more frequently with the LA3 as compared with DA3. This reduced b:a ratio in the photopic ERG suggests that the dysfunction in PRPH2-retinopathy occurs post-phototransduction in addition to the cone photoreceptors. A reduced b:a ratio (cutoff threshold <1.5) was only a significant predictor of PRPH2-retinopathy in the DA10 for the “CACD-like” group. Similarly, Ba-Abbad et al.8 described an electronegative ERG in both DA and LA recordings in six patients with PRPH2-retinopathy. Interestingly, the On and Off-responses were useful in identifying cone On-bipolar cell dysfunction in all cases. The reduced DA3 b:a ratio in PRPH2-retinopathy may also represent a predominance of DA cones with markedly reduced rods (manifesting as a photopic hill phenomenon in the dark). Of those PRPH2 patients with a reduced DA3.0 b:a ratio <1.5 and <1.0, the mean b-wave latency was 50 and 47 ms, respectively. Whereas a reduced b:a ratio has also been reported in ABCA4-retinopathy with Khan et al.20 describing a case series of eight children, three with a reduced b:a ratio and two with an electronegative ERG, there have been no such reports in the adult population. This reduction in the b:a ratio may relate to trans-synaptic degeneration of the bipolar cells. However, further human retinal histological analysis would be required to confirm this hypothesis.
We have previously reported the CACD phenotype to be associated with 13 specific missense variants in PRPH2.3 Other papers have also attributed certain missense variants to a CACD retinal phenotype.21,22 Another localized central macular FAF phenotype – BEM – has also been associated with PRPH2, although much less commonly.23 ABCA4-retinopathy may present with foveal sparing in late-onset Stargardt disease.24–26 Kurz-Levin et al.27 found no clear structure function correlation in patients presenting with a BEM. Our previous study3 found patients with a CACD phenotype secondary to PRPH2 showed a trend towards greater cone than rod involvement as reported previously by Hoyng and Deutman.11 In contrast work by Ikelle et al.,7 they found cones were more resilient to pathogenic PRPH2 variants as opposed to rods which were highly susceptible to PRPH2 haploinsufficiency.
In contrast to CACD, we have previously found loss of function alleles, including truncating, splice, or start-loss variants in PRPH2 to be associated with a “fleck-like” phenotype.3 Vasireddy et al.28 reported a fleck phenotype to occur in association with variants in three main genes namely ABCA4, PRPH2, and less commonly ELOVL4. More recently, WDR19 has also been added to this list of genes.29 We found more severe rod-mediated dysfunction in those with a PRPH2- as compared to ABCA4-retinopathy. We hypothesize loss of function alleles in PRPH2, manifesting as a “fleck-like” phenotype, likely impact rod and cone function equally. This is consistent with work by Ikelle et al.7 in which structural, functional, and molecular studies demonstrated that rods were highly susceptible to PRPH2 pathogenic variants. Work by Conley et al.30 helps to explain why ABCA4-retinopathy demonstrates less rod involvement. They found ABCA4-deficient rods simultaneously generate less A2E than cones. Their work suggested that primary cone toxicity may be the driving force behind ABCA4-retinopathy.
Our exploratory study is subject to the limitations of retrospective clinical data, such as missing data and variable protocols, as well as devices used for VA, FAF, and ERG acquisition. A prior study has described discordance in VA derived from a Snellen chart as compared with a logMAR style chart.31 Thus, any observed differences in VA between the ABCA4 and PRPH2 groups should be interpreted with caution. Given our ERG data was collected from eight different countries and electrode types were not recorded we used local normative ranges to perform a sensitivity analysis using the percentage of lower bound of the normative amplitude range or the upper bound of the normative peak time range. Interestingly, we found no improvement in the AUROC when this was incorporated. However, differences in ERG parameters should still be interpreted with caution given age was not incorporated as a covariate. Given multiple statistical testing, additional studies are required to confirm high AUROC values were not false positives. Furthermore, we did not collect data on which subjects were related nor additional variants in other genes that may modify the retinal phenotype. We have proposed several cutoff thresholds without validation in a separate cohort of patients. Future studies with an independent sample source are required to determine the clinical utility, and positive and negative predictive values of these cutoff thresholds.
Our study demonstrated unique ERG profiles in ABCA4- and PRPH2-retinopathy when grouping by two common FAF phenotypes namely, “fleck-like” or “CACD-like”. Whereas differences in the “fleck-like” group were mostly observed under DA conditions, differences in the “CACD-like” group were mostly observed under LA conditions. Given genetic testing can have up to a 30% to 40% failure rate32,33 electrophysiological biomarkers may be of particular importance when genetic testing is delayed, not feasible or when variants of unknown significance arise.
Supplementary Material
Acknowledgments
Supported by Australian Government Research Training Program Fees Offset Scholarship (RCHJ); Retina Australia (JAT); Retina Australia (TML); Retina Australia (TLM); Macular Disease Foundation Australia, Future Health Research and Innovation Fund, the McCusker Charitable Foundation, Channel 7 Telethon Trust, Retina Australia, National Health & Medical Research Council of Australia (project and fellowship grant no.: GNT1116360, GNT1188694, GNT1054712, and MRF1142962, FKC); the Henry Brent Chair in Innovative Pediatric Ophthalmology Research (EH); Foundation Fighting Blindness USA enhanced career development award CD-CMM-0224-0873-HSC (AV); grants from Grant-in-Aid for Young Scientists (A) of the Ministry of Education, Culture, Sports, Science and Technology, Japan (16H06269); Grant-in-Aid for Scientists to support international collaborative studies of the Ministry of Education, Culture, Sports, Science and Technology, Japan (16KK01930002); the National Hospital Organization Network Research Fund, Japan (H30-NHO-Sensory Organs-03); Foundation Fighting Blindness Alan Laties Career Development Program (CF-CL-0416-0696-UCL), USA; the Health Labour Sciences Research Grant, AMED (23EK0109634H0001 and 23EK0109632H0001), the Ministry of Health Labour and Welfare, Japan (201711107A and 23809955); Great Britain Sasakawa Foundation Butterfield Awards, UK (KF); the Italian Ministry of Health and Fondazione Roma through the work at IRCCS- Fondazione Bietti (LZ).
Disclosure: R.C. Heath Jeffery, None; J.A. Thompson, None; J. Lo, None; A.L. Vincent, None; M. Patil, None; L. Bianco, None; M. Battaglia Parodi, None; L. Ziccardi, None; C. Dell'Aquila, None; L. Barbano, None; W.C. Tang, None; C.M. Chan, None; C.J.F. Boon, None; J. Hensman, None; T.-C. Chen, None; C.-Y. Lin, None; P.-L. Chen, None; A. Vincent, None; A. Tumber, None; E. Heon, None; J.R. Grigg, None; R.V. Jamieson, None; E.E. Cornish, None; B.M. Nash, None; J. Chou, None; T.M. Lamey, None; S. McLenachan, None; D. Roshandel, None; K. Fujinami, None; E. Chelva, None; T.L. McLaren, None; F.K. Chen, None
References
- 1. De Roach JN, McLaren TL, Thompson JA, et al.. The Australian Inherited Retinal Disease Registry and DNA Bank. Tasman Med J. 2020; 2: 60–67. [Google Scholar]
- 2. Pontikos N, Arno G, Jurkute N, et al.. Genetic basis of inherited retinal disease in a molecularly characterized cohort of more than 3000 families from the United Kingdom. Ophthalmology. 2020; 127(10): 1384–1394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Heath Jeffery RC, Thompson JA, Lo J, et al.. Retinal dystrophies associated with peripherin-2: genetic spectrum and novel clinical observations in 241 patients. Invest Ophthalmol Vis Sci. 2024; 65(5): 22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Heath Jeffery RC, Lo J, Thompson JA, et al.. Analysis of the outer retinal bands in ABCA4 and PRPH2-associated retinopathy using OCT. Ophthalmol Retina. 2024; 8(2): 174–183. [DOI] [PubMed] [Google Scholar]
- 5. Duncker T, Tsang SH, Lee W, et al.. Quantitative fundus autofluorescence distinguishes ABCA4-associated and non-ABCA4-associated bull's-eye maculopathy. Ophthalmology. 2014; 122: 345–355. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Lois N, Holder GE, Bunce C, Fitzke FW, Bird AC.. Phenotypic subtypes of Stargardt macular dystrophy-fundus flavimaculatus. Arch Ophthalmol. 2001; 119(3): 359–369. [DOI] [PubMed] [Google Scholar]
- 7. Ikelle L, Makia M, Lewis T, et al.. Comparative study of PRPH2 D2 loop mutants reveals divergent disease mechanism in rods and cones. Cell Mol Life Sci. 2023; 80: 214. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Ba-Abbad R, Robson AG, Yap YC, et al.. Prph2 mutations as a cause of electronegative ERG. Retina. 2014; 34(6): 1235–1243. [DOI] [PubMed] [Google Scholar]
- 9. Richards S, Aziz N, Bale S, et al.. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genet Med. 2015; 17(5): 405–424. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Abou Tayoun AN, Pesaran T, DiStefano MT, et al.. Recommendations for interpreting the loss of function PVS1 ACMG/AMP variant criterion. Hum Mutat. 2018; 39(11): 1517–1524. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Hoyng CB, Deutman AF.. The development of central areolar choroidal dystrophy. Graefes Arch Clin Exp Ophthalmol. 1996; 234(2): 87–93. [DOI] [PubMed] [Google Scholar]
- 12. Boon CF, Klevering JB, Cremers FPM, et al.. Central areolar choroidal dystrophy. Ophthalmology. 2009; 116(4): 771–782. [DOI] [PubMed] [Google Scholar]
- 13. Robson AG, Frishman LJ, Grigg J, et al.. ISCEV Standard for full-field clinical electroretinography (2022 update). Doc Ophthalmol. 2022; 144: 165–177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Bach M, Brigell M, Hawlina M, et al.. ISCEV standard for clinical pattern electroretinography (PERG) - 2012 update. Doc Ophthalmol. 2013; 126: 1–7. [DOI] [PubMed] [Google Scholar]
- 15. Hood D, Bach M, Brigell M, et al.. ISCEV Standard for clinical multifocal electroretinography (2011 edition). Doc Ophthalmol. 2012; 124: 1–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Bianco L, Arrigo A, Antropoli A, et al.. PRPH2-associated retinopathy: novel variants and genotype-phenotype correlations. Ophthalmol Retina. 2023; 7(5): 450–461. [DOI] [PubMed] [Google Scholar]
- 17. Antonelli G, Parravano M, Barbano L, et al.. Multimodal study of PRPH2 gene-related retinal phenotypes. Diagnostics (Basel). 2022; 12(8): 1851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Miere A, Zambrowski O, Kessler A, et al.. Deep learning to distinguish ABCA4-related Stargardt disease from PRPH2-related pseudo-Stargardt pattern dystrophy. J Clin Med. 2021; 10(24): 5742. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Duncker T, Tsang SH, Woods RL, et al.. Quantitative fundus autofluorescence and optical coherence tomography in PRPH2/RDS- and ABCA4-associated disease exhibiting phenotypic overlap. Invest Ophthalmol Vis Sci. 2015; 56(5): 3159–3170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Khan KN, Kasilian M, Mahroo OA, et al.. Early patterns of macular degeneration in ABCA4-associated retinopathy. Ophthalmology. 2018; 125(5): 735–746. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Boon CJF, den Hollander AI, Hoyng CB, et al.. The spectrum of retinal dystrophies caused by mutations in the peripherin/RDS gene. Prog Retin Eye Res. 2008; 27: 213–235. [DOI] [PubMed] [Google Scholar]
- 22. Boon CJF, van Schooneveld MJ, den Hollander AI, et al.. Mutations in the peripherin/RDS gene are an important cause of multifocal pattern dystrophy simulating STGD1/fundus flavimaculatus. Br J Ophthalmol. 2007; 91(11): 1504–1511. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Essilfie J, Sanfilippo CJ, Sarraf D.. Bull's eye maculopathy with mutations in RDS/PRPH2 and ROM-1. Retin Cases Brief Rep. 2018; 12(Suppl 1): S55–S58. [DOI] [PubMed] [Google Scholar]
- 24. Westeneng-van Haaften SC, Boon CJF, Cremers FPM, et al.. Clinical and genetic characteristics of late-onset Stargardt's disease. Ophthalmology. 2012; 119(6): 1199–1210. [DOI] [PubMed] [Google Scholar]
- 25. Fujinami K, Waheed N, Laich Y, et al.. Stargardt macular dystrophy and therapeutic approaches. Stargardt macular dystrophy and therapeutic approaches. Br J Ophthalmol. 2024; 108(4): 495–505. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Fujinami K, Segouniotis PI, Davidson AE, et al.. Clinical and macular analysis of Stargardt disease with preserved foveal structure and function. Am J Ophthalmol. 2013; 156(3): 487–501. [DOI] [PubMed] [Google Scholar]
- 27. Kurz-Levin MM, Halfyard AS, Bunce C.. Clinical variations in assessment of Bull's eye maculopathy. Arch Ophthalmol. 2002; 120(5): 567–575. [DOI] [PubMed] [Google Scholar]
- 28. Vasireddy V, Wong P, Ayyagari R.. Genetics and molecular pathology of Stargardt-like macular degeneration. Prog Ret Eye Res. 2010; 29(3): 191–207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Nunziata A, Antropoli A, Bianco L, et al.. WDR19-associated retinopathy presenting with adult-onset Stargardt-like phenotype. Ophthalmic Genet. 2025; 46: 281–284. [DOI] [PubMed] [Google Scholar]
- 30. Conley SM, Cai X, Makkia R, et al.. Increased cone sensitivity to ABCA4 deficiency provides insight into macular vision loss in Stargardt's dystrophy. Biochim Biophys Acta. 2012; 1822(7): 1169–1179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Chen FK, Agelis L, Peh KK, et al.. Visual acuity fractions derived from a Snellen chart and letter scores on the early treatment diabetic retinopathy study chart. Asia Pac J Ophthalmol. 2014; 3(5): 277–285. [DOI] [PubMed] [Google Scholar]
- 32. Britten-Jones AC, Hickey DG, Edwards TL, et al.. Panel-based genetic testing in a consecutive series of individuals with inherited retinal diseases in Australia: identifying predictors of a diagnosis. Genes (Basel). 2025; 16(8): 888. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Britten-Jones AC, Gocuk SA, Goh KL, et al.. The diagnostic yield of next generation sequencing in inherited retinal diseases: a systematic review and meta-analysis. Am J Ophthalmol. 2022; 249: 57–73. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.