The study protocol was reviewed and approved by the Institutional Review Board of Seoul National University Hospital (No. 2201-113-1293). The need for written informed consent was waived due to the retrospective design of the study and the use of deidentified patient information. All study procedures were conducted in accordance with the tenets of the Declaration of Helsinki.

This was a retrospective cohort study of the medical records of consecutive patients with pericentral RP. All study procedures were performed at Seoul National University Hospital (Seoul, Korea) between August 2023 and May 2024.

The diagnosis of pericentral RP was made when fundus examination showed bony-spicule pigmentation or retinal pigment epithelium atrophy in the near mid-periphery (5°-30°) within and beyond the vascular arcades and when Goldmann visual field (GVF) testing revealed pericentral scotomas in the corresponding areas surrounded by the far retinal periphery with preserved appearances. Full-field ERG showed subnormal though detectable responses. Patients were excluded if they had a history of ocular trauma or intraocular inflammation, such as uveitis in the affected eye, or if retinal changes due to secondary causes, such as prior hydroxychloroquine usage, could not be ruled out [13]. Additionally, individuals with a history of intraocular interventions that could affect retinal status, including vitrectomy, intravitreal injections, or panretinal photocoagulation laser, before the baseline evaluation were also excluded from the study.

All patients underwent comprehensive ophthalmological examinations including best-corrected visual acuity (BCVA), slit-lamp biomicroscopy, indirect fundus examination, ultra-widefield fundus photography, full-field ERG, GVF, optical coherence tomography (OCT), and fundus autofluorescence (FAF). BCVA was converted to logarithm of the minimum angle of resolution (logMAR) units for statistical analyses. Full-field ERG was recorded using the RETI-port system (Roland Consult) according to the recommendations of the International Society for Clinical Electrophysiology of Vision [14]. GVF was measured by moving the stimulus target from I4e to V4e on a calibrated standard Goldmann perimeter and 43 eyes were available for the II4e target. The acquired field images were scanned and converted into area measurements (in degrees²) through software-based quantification using Adobe Photoshop ver. 24.7 (Adobe), following the same validated procedure as previously published [15].

Spectral-domain OCT images were obtained using Heidelberg Spectralis (Heidelberg Engineering) or Zeiss Cirrus 4000 (Carl Zeiss Meditec). Cross-sectional OCT images were captured along the horizontal or vertical meridian through the fovea. The preserved ellipsoid zone (EZ) width was measured as the distance between the temporal and nasal borders on the horizontal image and between the superior and inferior borders on the vertical image of the EZ, where the EZ line disappeared. If the entire length of the EZ line exceeded the scan length of the OCT image, the borders of the EZ were set to be those of the scanned OCT image [16,17]. The same process was used to measure the preserved external limiting membrane width [18]. EZ and external limiting membrane widths were measured using the built-in calipers of the Heidelberg or Zeiss Cirrus software [17]. The OCT findings were categorized into four groups, as described by Lupo et al. [19]: (1) no specific macular abnormalities; (2) presence of cystoid macular edema (CME); (3) presence of vitreomacular traction (VMT) or epiretinal membrane (ERM); and (4) macular retinal thinning. FAF images with a 30° field of view were acquired using the Heidelberg Spectralis instrument. FAF was classified according to the patterns proposed by Fakin et al. [18] as follows: (1) normal macular autofluorescence (AF); (2) abnormal central hypo-AF; (3) a typical ring of perimacular hyper-AF; and (4) a hyper-AF foveal patch.

Molecular genetic tests were conducted on peripheral blood samples collected from patients who provided informed consent. Samples were analyzed with a next-generation sequencing (NGS)-based gene panel or by whole exome sequencing (WES). Specifically, the NGS-based gene panel targeted 248 genes associated with inherited retinal diseases. WES was performed by Macrogen. Genomic DNA samples were prepared using an Agilent SureSelect Human All Exon Kit V6 array (Agilent), and sequencing was performed using an Illumina NovaSeq 6000 System (Illumina). Each variant’s clinical significance was determined based on the recent guidelines from the American College of Medical Genetics and Genomics on standards for interpreting and reporting sequence variations [20,21]. Further functional impact predication was performed using in silico tools for variants classified as variants of uncertain significance (VUS) in the NGS or WES analysis. The gene status was categorized as “solved” when the causative gene could be determined given the gene-specific inheritance patterns, and as “inconclusive” when only one pathogenic or likely pathogenic variant was identified in an autosomal recessive gene and the pathogenicity could not be fully explained.

Mean annual exponential rates of decline in visual acuity for RP have been reported to be as high as 8.2% to 8.6% [9,22-24]. Therefore, to discriminate between cases with unusually rapid progression, this study set the criterion for rapid visual acuity loss at 10% per year. To identify clinical biomarkers capable of predicting the risk of future rapid decline in visual acuity, the enrolled eyes were divided into two groups based on annual changes in visual acuity in each eye: group 1, comprised with eyes that exhibited a rapid a nnual rate of visual a cuity decline of ≥10%; and group 2, comprised with eyes that showed a slower annual rate of change, <10%, during the follow-up period.

To calculate the annual change in visual acuity in each eye, we converted visual acuities to logMARs to provide a good fit for describing the disease progression of RP [9,24], which has been used in previous longitudinal studies of RP [9,24,25]. We considered only the data from subjects with ≥2 visits, and 55 eyes of 29 patients were available for longitudinal analysis. Linear regression analyses were performed with visual acuity as the dependent variable and age as the independent variable to derive the slope (rate of change) for each eye during follow-up. A mixed linear regression model accounting for repeated measures was used to estimate the mean rate of change in visual acuities for all available subjects and each group [9,24,25].

All statistical analyses were performed using IBM SPSS ver. 26 (IBM Corp). Continuous variables are presented as mean ± standard deviation, and the level of statistical significance was set at p < 0.05. Univariate logistic regression analysis and backward stepwise multivariate regression analysis were used to detect significant variables, and the Hosmer-Lemeshow test was performed to assess the fit of the logistic regression model.

The aim of this study was to perform a detailed analysis of the clinical characteristics of South Korean patients with pericentral RP and to identify clinical biomarkers capable of predicting the risk of future rapid decline in visual acuity based on initial multimodal measurements. We compared the clinical characteristics of pericentral RP with those of typical RP described in previous literature. Among studies on typical RP patients in South Korea, Lee et al. [26] analyzed the clinical characteristics of a significant number of 365 patients with an average age of 38 years and average logMAR visual acuity of 0.90. In the present study, the average age of the patients was higher and the average visual acuity was better, indicating a favorable trend in pericentral RP compared to typical RP. The older age at symptom onset and diagnosis in our study compared to the average age reported in Lee et al. [26] was likely due to the presence of mild or asymptomatic cases in patients with pericentral RP, resulting in delayed symptom recognition and diagnosis. In a study by Hwang et al. [27], which analyzed the clinical characteristics of 492 eyes of 246 South Korean patients with typical RP, the average age of the patients (48 years) was comparable, while the average logMAR visual acuity (0.31) was worse than that of the present study. In our study of pericentral RP patients, the GVF area was better preserved than that of a large cohort of 601 typical RP patients from a clinical trial with baseline average GVF area of about 1,500°² to 1,800°² by subgroups [28]. Considering the use of a smaller stimulus in the present study of pericentral RP patients (II4e vs. V4e), the difference in displaying a wider preserved GVF area in the pericentral type than in typical RP can be significant. In terms of OCT-based classification, our study had a higher percentage of cases without macular abnormalities and a lower incidence of CME and VMT/ERM than previous studies on typical RP [19,27,29,30]. These differences suggest that in pericentral RP, the macula may be better preserved, leading to a lower incidence of macular complications such as CME and VMT/ERM. This preservation may also account for the milder symptoms and favorable BCVA observed in patients with pericentral RP. In summary, this study suggests that patients with pericentral RP have a more favorable average visual acuity, wider preserved visual field, and fewer macular complications than patients with typical RP from previous studies, indicating a relatively mild phenotype. However, it is important to note that this comparison is indirect, as we did not conduct a direct age-matched or controlled comparison between our cohort and those in other studies.

In other pericentral RP cohorts abroad, Karali et al. [8] conducted an analysis of the clinical and genetic characteristics in a cohort of 54 European patients with pericentral RP. The study reported a mean BCVA of 0.7 decimal (0.15 logMAR), with 85.2% of patients showing a preserved BCVA of 0.5 or better in decimal, consistent with the results of our study. Additionally, multiple studies of American pericentral RP cohorts have shown similar outcomes in visual acuity [9,11,12]. Karali et al. [8] reported a relatively less preserved visual field area in their European cohort compared to that in the present study, whereas Sandberg et al. [9] reported a more favorable visual field in their American cohort, similar to the findings of our study. Karali et al. [8] reported the prevalence of CME and ERM to be 14% and 15%, respectively, which are similar to the findings of the present study and slightly lower than those in a typical RP cohort. The frequencies of the abnormal central hypo-and hyper-AF foveal patch in FAF were similar to those in the present study. This study showed a lower frequency of the perimacular hyper-AF ring compared with the European cohort of Karali et al. [8], but it remained the most identifiable abnormal pattern.

Our findings identified some overlapping causative genes in South Korean pericentral RP cohort compared to those reported in previous typical and pericentral RP studies. In a large South Korean cohort of typical RP cases (279 cases) [31], the most common causative mutations were EYS (8.2%), USH2A (6.8%), and PDE6B (4.7%). Similarly, another study on the genetic profiles in South Korean patients with inherited retinal disease identified EYS (22%), PDE6B (17%), USH2A (11%), PDE6A (11%), and RHO (11%) as the most common causative gene in typical RP [4]. In typical RP, EYS variants have been reported to constitute the largest portion in East Asia, whereas USH2A variants are the most prevalent in Western and European populations [4,32-34]. In contrast, in the South Korean pericentral RP population analyzed in this study, USH2A mutations were detected, while no solved cases with EYS mutations were identified. Although the small number of patients who underwent genetic analysis limits our ability to comprehensively determine the distribution of causative genes, this observation suggests that pericentral RP may has a distinct genetic landscape. Further studies with larger sample sizes are needed to confirm these difference and to explore the underlying genetic mechanisms specific to pericentral RP.

Several studies have been conducted to identify causative genes in Western or European patients with pericentral RP. In the European pericentral RP cohort studied by Karali et al. [8], variants in USH2A were detected in 41% of patients with a high prevalence of pericentral RP in Usher syndrome. Comander et al. [11] reported that RHO was the most commonly identified gene, consistent with the findings of Grondahl et al. [35] and Selmer et al. [36], followed by USH2A, HGSNAT, PDE6B, and other genes. The results of the study suggested that pericentral RP shares causative genes with other mild forms of RP. In contrast, Matsui et al. [12] reported that ABCA4 was the most commonly identified gene, followed by CERKL, CRX, DHDDS, NR2E3, PRPH2, PROM1, and RHO. Our South Korean cohort in this study also demonstrated an overlapping causative gene profile with Western or European pericentral RP cohorts, with the identification of mutations in ABCA4, USH2A, HGSNAT, and CRX.

This study also identified genes that have been previously reported to be associated with mild phenotypes of RP. Karali et al. [8] demonstrated that pericentral RP patients with mutations in USH2A exhibited a milder phenotype compared to a previous RP cohort. Similarly, Comander et al. [11] reported USH2A and HGSNAT mutations in patients with mild forms of RP. Additionally, mutations in CRX and GUCY2D also have been reported in some mild cases of RP in previous studies [12,37,38]. These findings suggest a potential link between the milder phenotype observed in pericentral RP and specific genetic mutations.

The diagnostic yield of 23.5% observed in this study reflects the challenges in identifying causative genes for pericentral RP. Several gene mutations identified in this study remained inconclusive, as they involved only a single heterozygous autosomal recessive variant or compound heterozygous mutations including VUS. It is possible that a second pathogenic variant was not detected due to technical limitations in our sequencing approach. In patients with single heterozygous mutations, there may be other potential pathogenic variants that require further investigation and validation. This suggests that many genes responsible for pericentral phenotype of RP remain unknown, highlighting the need for advanced genetic testing approaches and broader studies to uncover the full spectrum of genetic landscape.

This study identified genes that include those previously detected in the South Korean RP population [4,31,39], as well as those associated with a pericentral or milder phenotype compared to those previously described in patients with RP [8,11,12,37,38]. To our knowledge, this is the first study to investigate the genetic mutation profile of a South Korean population with pericentral RP. Although the size of the cohort limited an in-depth analysis of gene distribution, it is notable that pericentral RP demonstrated genetic heterogeneity. Consequently, further research is essential to gain a more comprehensive understanding of the genetic characteristics of pericentral RP.

In multivariate logistic analysis, “sex” and “b-wave latency of dark-adapted 0.01 response” in full-field ERG showed significant potential to discriminate eyes with high risk of rapid progression in pericentral RP. Therefore, monitoring initial b-wave latency is particularly important for predicting visual decline, especially in male patients.

Consistent with the findings of this study, previous studies have reported that the rate of RP progression is lower in women. The study by Ito et al. [40], which included 235 eyes from 121 patients with RP, showed slower progression rates of visual acuity and average retinal sensitivity of the central four points of the Humphrey visual field 10-2 program in women than in men in a mean follow-up of 6.71 years. The authors proposed that there may be sex-based differences in the progression of central visual function. Some animal studies have reported the expression of estrogen receptors in various retinal layers and have recognized the effects of sex hormone profiles as significant factors contributing to functional differences in the retina [41,42]. An animal study using ERG to investigate retinal function in rats demonstrated better preservation of ERG parameters in females of reproductive age than in males or menopausal females, suggesting a beneficial effect of estrogen and the estrous cycle on retinal function [43].

ERG enables the evaluation of photoreceptor function in patients with RP and can identify early stages of RP by detecting a delay in b-wave latency [44]. While the prolongation of b-waves in dark-adapted 0.01 ERG is derived from ON-bipolar cells of the rod system [45], it depends on the input from functional rod photoreceptors and is considered to reflect rod dysfunction clinically [14,46]. Hood [47] suggested that damage to the outer plexiform layer is responsible for the relatively large but markedly delayed responses, and the delays occur relatively early in the disease process. Although rods are primarily responsible for peripheral and dim light vision, numerous studies have explained the mechanism of central cone cell death, which results in central vision loss in relation to preceding rod dysfunction in RP as follows: (1) cone survival relies on trophic factors from rods; (2) cones experience nutrient deprivation following rod loss; and (3) oxidative stress, pro-inflammatory microglial activation, or toxins derived from rod cell death [48-52]. Campochiaro and Mir [49] suggested that the rate of rod degeneration is an important prognostic feature of RP, explaining that rod cell death induces superoxide radicals that contribute to cone cell death. Jones and Marc [53] reported that in the context of neural retinal remodeling, microneuroma formation and rewiring due to rod cell death affect visual acuity by disrupting normal retinal circuitry and creating aberrant neural connections. The authors proposed that microneuromas formed during retinal remodeling exhibit an unusual synaptic architecture with highly unstable and resonant visual circuitry. Moreover, dysfunctional rewiring leads to persistent self-activating signals that interfere with normal visual information processing and severely affects visual acuity. As suggested in previous studies, rod cell dysfunction leads to a cascade of events that adversely affects cone cells and central vision. The results of this study suggest that in pericentral RP, which initially shows a subnormal but detectable rod response on ERG, early loss of the rod response could serve as an early indicator for predicting future deterioration of central vision as the disease advances.

This study has several limitations. First, this was a retrospective analysis with a relatively small sample size. Despite the small sample size, this is the largest study to date on the overall characteristics of pericentral RP in the South Korean population, including genetic and prognostic factor analyses. Secondly, longitudinal data for each clinical finding, aside from visual acuity, were not evaluated. Future research focusing on the longitudinal changes in clinical biomarkers may provide new methods for predicting the prognosis of eyes with pericentral RP. Third, if both eyes of the patients met the inclusion criteria, both were included in the study. Because RP is an inherited degenerative disease that typically affects both eyes, one might expect similar characteristics in each eye. However, the bilateral involvement is not always symmetrical. Although both eyes of the same patient exhibited pericentral RP, they did not necessarily share identical structural abnormalities or visual function. This suggests that the detailed condition of each eye can vary, highlighting the importance of evaluating each eye individually. Fourth, genetic testing was not performed in all the patient cohorts, which may have provided more evidence regarding the underlying pathophysiology of the disease. Due to the limited number of patients who underwent genetic analysis and the small number of genes identified with mutations, we were unable to perform further analysis of the gene distribution or genotype-phenotype correlation in the cohort. The identified mutations in the study were spread across various genes. The study primarily focused on identifying the types of genetic mutations within the cohort. Future research could address this gap by focusing on comprehensive genetic analyses, potentially providing valuable insights into the disease. Finally, the cutoff rate of 10% per year decline in visual acuity used to categorize into groups 1 and 2 for logistic analysis in this study might be arbitrary. The 10% threshold for annual decline in visual acuity was selected based on previously reported progression rates [9,22-24]. However, we acknowledge that this threshold may not fully reflect the natural course of disease. Further studies with larger sample sizes, extended follow-up periods, and comprehensive longitudinal analyses are required to establish the appropriate standards for identifying patients with high-risk pericentral RP.

In conclusion, the clinical characteristics of pericentral RP in the South Korean population exhibited a milder phenotype, including better visual acuity, a larger visual field area, and fewer macular abnormalities, compared to those of typical RP patients reported in previous studies, and displayed substantial genetic heterogeneity. Additionally, male sex and prolonged b-wave latency in dark-adapted 0.01 ERG emerged as valuable biomarkers for predicting the risk of rapid disease progression in subsequent follow-ups. Comprehensive multimodal assessment of clinical features in patients with pericentral RP can yield valuable insights, potentially improving the diagnosis, classification, follow-up strategies, and patient counselling for this condition.