The Exponential Constriction Model of the Ellipsoid Zone in Taiwanese Individuals With RPGR-Related X-Linked Retinitis Pigmentosa

Abstract

Purpose

This study documents the natural disease progression and genotype-phenotype correlation in RPGR-related retinitis pigmentosa (RP) in the Taiwanese population.

Methods

A retrospective analysis was conducted on individuals with molecularly confirmed RPGR-related disease-causing variant(s). Demographics, best-corrected visual acuity (BCVA), spherical equivalent (SE), fundus autofluorescence, and optical coherence tomography were assessed.

Results

Fifty-two individuals from 31 families were diagnosed with RPGR-related disease-causing variant(s). Mean follow-up time was 4.2 years. Among 21 genetic variants, 67% involved the open reading frame 15 region (ORF15) variant, and 33% were Exon 1–14 variants. Male patients (69%) had a mean BCVA of 0.9 logMAR and SE of −3.8 D in the right eye and −3.0 D in the left eye, with high myopia in 19% to 20%. BCVA progression was 0.031 logMAR/year in the ORF15 group (P < 0.001) and 0.011 logMAR/year (P = 0.457) in the Exon 1–14 group. An exponential decay model revealed rapid ellipsoid zone (EZ) constriction during childhood in the ORF15 group. Female patients/carriers (31%) had a mean BCVA of 0.3 logMAR and SE of −4.3D, with high myopia of 31% in the right eye and 46% in the left eye. Among symptomatic females, 73% exhibited clinically significant disease. The most common mutation was the c.2592dup variant (15%).

Conclusions

This first longitudinal analysis of RPGR-related RP in Taiwan presents a predictive model of EZ constriction. Findings suggest earlier onset in Exon 1–14 variants and a tendency for faster progression in the ORF15 group, informing insights for genetic therapy development and disease management.

Retinitis pigmentosa (RP) is the most common inherited retinal degeneration (IRD), with a worldwide prevalence of approximately 1:3000–4000.^1,2 It causes progressive visual impairment leading to visual blindness.³ RP results from multiple disease-causing genes and a wide variety of genetic variants, with X-linked (XLRP) inheritance representing 5% to 15% of all RP cases, whereas autosomal recessive and autosomal dominant account for 50% to 60% and 30% to 40% of cases, respectively.⁴ Among the different disease variants, XLRP is the most severe form, characterized by early-onset disease and rapid progression to an advanced stage by the third or fourth decade of life.^5–7 In XLRP, the retinitis pigmentosa GTPase regulator (RPGR) gene is the most common disease-causing gene, accounting for 70% to 80% of all cases, with the rest attributed to variants in the RP2 gene.^6,8,9 Isoform C (NM_001034853) of RPGR gene, also known as the open reading frame 15 region (ORF15), shows abundant expression and plays a role in the pathogenesis of RP in the retina.¹⁰ In RPGR-related XLRP, approximately 60% to 80% of mutation variants are found in the ORF15 region, whereas the remaining variants are distributed in the Exon 1–14 region.¹⁰

The phenotypic presentation of XLRP includes peripheral degeneration, pigment clumping, and accumulation of lipofuscin in the retinal pigment epithelium (RPE), which can be identified through fundus autofluorescence (FAF) in the macular area. This accumulation stems from metabolic stress in the RPE or loss of photoreceptor outer segment loss, resulting in a hyper auto-fluorescence (hyper-AF) presentation.^11–13 A parafoveal hyper-AF ring is a distinctive feature in RP, demarcating the boundary between healthy and degenerative retina.¹⁴ The constriction of hyper-AF ring correlates with disease progression.¹⁵ Because the integrity of the ellipsoid zone (EZ) corresponds to the hyper-AF ring on FAF, it is considered a structural biomarker for visual acuity and disease progression.^16–18

Although RPGR-related IRD is conventionally a X-linked disease affecting the male population, female carriers of a RPGR variant can exhibit variable degrees of interocular asymmetry, ranging from normal appearance to a phenotype similar to affected male patients.¹⁹ The heterogenicity in disease presentation can be attributed to random X chromosome inactivation, a process that balances the presence of two X chromosomes in female carriers.

Recent advancements in genetic therapy targeting photoreceptor degeneration in RPGR-related RP have shown promising results, including restoration of retinal sensitivity and reversal of visual field loss in phase I/II trials.²⁰ Subsequently, phase II/III studies have been initiated to further investigate these findings, aiming to introduce a novel treatment option in the genomic era. However, despite significant progress in gene replacement therapy, our understanding of the disease course and pathophysiology of RPGR-related RP remains incomplete, creating challenges in establishing a useful criterion for therapy initiation and determining optimal therapeutic goals for patients.

This study aims to investigate the natural history of RPGR-related IRD in Taiwan. Our specific objectives are to enhance understanding of disease presentation and progression in both male patients and female carriers and to explore potential genotype-phenotype correlations.

Methods

Study Design

This retrospective, longitudinal, single-center study was conducted on individuals recruited through the Taiwan Inherited Retinal Degeneration Project (TIP). The study received approval from the Research Ethics Committee of National Taiwan University Hospital (IRB No.: 201408082RINC) and conducted following the Tenets of the Declaration of Helsinki. Enrollment spanned from September 2015 to March 2024, with eligibility based on the following inclusion criteria:

• 1)Male or female individuals diagnosed with RPGR-related disease-causing variant(s) confirmed by molecular genetic testing
• 2)Ability to undergo a comprehensive ophthalmic examination, including best-corrected visual acuity (BCVA), color fundus photography (CFP), FAF, optical coherence tomography (OCT), and complete the questionnaire
• 3)Willingness to provide informed consent

Exclusion criteria included:

• 1)Presence of other inherited retinal disease or syndromic conditions that may confound the study findings
• 2)Pre-existing significant ocular condition that may interfere with the interpretation of primary and secondary study outcomes (e.g., advanced diabetic retinopathy, age-related macular degeneration, uveitis, or severe glaucoma)
• 3)Systemic disease or medication known to affect retinal function (e.g., hydroxychloroquine toxicity)
• 4)Inability to comply with study visits and imaging protocol

Next-Generation Sequencing

Details of genetic testing processes have been extensively described in a previous publication from our institution.²¹ In brief, blood samples were collected after obtaining informed consent, and genomic DNA was extracted from peripheral blood leukocytes using the Gentra Puregene Blood Kit (Qiagen, Hilden, Germany). Genetic analysis was performed using a probe capture-based next-generation sequencing (NGS) method to investigate 212 genes associated with IRDs, selected from the RetNet (https://sph.uth.edu/retnet/) and OMIM (https://www.ncbi.nlm.nih.gov/omim) databases. Selected genes can be found in the Supplementary Table S1.

Pathogenic variants in ORF15 were detected using long-range polymerase chain reaction (LR-PCR) followed by NGS, and the candidate variants were confirmed by Sanger sequencing.^22–24 To avoid allele dropout within the LR-PCR stage, two primer sets were used to generate two PCR products, with the length 1786bp and 2079bp. TaKaRa PRIMESTAR GxL DNA Polymerase (TaKaRa Bio. Co., Shiga, Japan) was used to generate the PCR product with the following condition: 30 cycles of 98°C for 30 seconds, 68 °C for 3 min and 30 sec, and hold by 4 °C. The PCR product was then purified using GenepHlow Gel/PCR Kit (Geneaid Biotech Ltd, New Taipei City, Taiwan) or Sera-Mag SpeedBeads (Cytiva, Marlborough, MA, USA). The purified product was then sequenced using NGS (NovaSeq 6000; Illumina, San Diego, CA, USA) and Sanger sequencing (ABI 3730xl; Applied Biosystems, Foster City, CA, USA). All primers, either adopted from previously published studies or modified as needed, are listed in Supplementary Table S2.^22–26

Assessment

All participants in this project underwent a comprehensive series of ophthalmic examinations at the Department of Ophthalmology, National Taiwan University Hospital. Examinations including BCVA assessment, CFP, FAF, and OCT were conducted at the initial visit, with follow-up imaging scheduled every six months. Visual acuity was measured using the Snellen chart and subsequently converted to the logMAR scale for analysis. Specific logMAR values were assigned for visual acuity categories “number of digits,” “hand movement,” “light perception,” and “no light perception,” corresponding to 1.7, 2.3, 2.8, and 2.9, respectively. A questionnaire conducted by our institution was administered during enrollment in the study to gather self-reported age of disease onset based on age range and symptoms including nyctalopia or so. Definitive diagnosis for each subject was established based on the aforementioned ophthalmic examinations, clinical presentations, genetic analyses, and family history.

The classification of retinal phenotypes in female patients/carriers was based on FAF, following the grading criteria established by Nanda et al.²⁷ This includes (1) N-pattern, characterized by a normal or near-normal fundus appearance; (2) R-pattern, marked by radial spoke–shaped reflexes extending from the central macular area; (3) F-pattern, identified by focal pigmentary retinopathy patchy pigmentation and a radial reflex pattern; and (4) M-pattern, corresponding to presentation of male pattern retinitis pigmentosa.

Image Analysis

The purpose of the image analysis was to transform qualitative observations into quantitative data to assess retinal structural changes and objectively measure disease progression in RPGR-related RP. We aimed to quantify the area of hyper-AF ring and hypo-AF atrophy area on FAF and the EZ band length on OCT. The analysis was conducted using ImageJ software (http://imagej.nih.gov/ij/), provided in the public domain by the National Institutes of Health, Bethesda, MD, USA. The hypo-AF atrophy area was delineated by applying color threshold selection with manual adjustments to the area of interest. Since the hypo-AF atrophy areas were mainly distributed at the midperiphery with progression towards the perimacular region, the optic disc area was manually excluded unless the area of interest extends through the disc. For the hyper-AF ring area, the border of the hyper-AF ring was defined and a color threshold was applied to select the area of interest. The optic disc and macular area were included if they fell within the selected hyper-AF ring area. As for central hypo-AF area in cone rod dystrophy (CORD) patients, we delineated the border of the abnormal central area on FAF images and binarized the image using Niblack's method. Subsequently, the image was converted to an 8-bit format with auto-local thresholding.²⁸ Low FAF signals were quantified within the abnormal autofluorescence area in the binarized image.

A digital retinal camera, CR-2 AF device (Canon, Tokyo, Japan) was used to capture color fundus images and a Spectralis HRA and OCT device (Heidelberg Engineering, Heidelberg, Germany) was used to acquire FAF and OCT images. The EZ line was manually measured via ImageJ software for horizontal and vertical scans passing through the fovea. EZ area was calculated by ellipse formula, y = horizontal length (µm) ×  vertical length (µm)  × π, with y as the EZ area (mm²). We calculated the intraclass correlation coefficient, a statistical measure that quantifies the reliability (or consistency) of two graders. The analysis yielded a high intraclass correlation coefficient value of 0.97 (95% confidence interval, 0.97 to 0.98), indicating strong consistency and reliability of two graders.

Statistical Analysis

Subjects who underwent FAF or OCT follow-up for more than six months with a minimum of two imaging visits were included for analysis. The progression rate was calculated using the values obtained as mentioned above in the image analysis process.

Descriptive statistics were performed to summarize the baseline characteristics of the study population. For continuous variables, data were presented as mean ± standard deviation, whereas categorical variables were reported as frequencies and percentages. Group comparisons for continuous variables were conducted using Student's t-test or Mann-Whitney U test, and for categorical variable using χ² test or Fisher's exact test. The Pearson correlation coefficient was calculated to assess the correlation between measurements of the right eye and left eyes. To evaluate interocular symmetrical progression, a paired t-test was performed to compare the means of two measurements taken from the same individual. The primary analysis in this study focuses on the exponential constriction model of the EZ band, and the secondary analysis examines the progression rate of observed characteristics, including BCVA progression, hyper-AF ring constriction, and hypo-AF region. Linear mixed models were employed to examine the rate of progression of BCVA and the constriction trend of the EZ band, both incorporating an exponential decay pattern to accurately capture their relationship with age. In both models, age was treated as a fixed effect, while patient number was included as a random effect to account for intra-patient variability. The use of exponential decay was essential, as both BCVA and the EZ band follow this pattern with increasing age, ensuring a more accurate representation of their progression. Statistical analysis was performed using R version 4.2.0 (https://www.r-project.org). All statistical tests were two-sided, and a significance level of 0.05 was used as the threshold for statistical significance.

Results

Demographics

RPGR-related RP constituted 10% of all RP cases in our institution. A total of 52 individuals from 31 families were enrolled in this study, 34 of them were included from 13 families. 15 individuals had only a single visit, while among those who underwent regular follow-up, the mean clinical follow up duration was 4.2 ± 2.7 years (range 0.7–10.1 years). Among the participants, 36 (69%) were male and 16 (31%) were female. We identified 21 different pathogenic gene variants: 12 (57%) ORF15 variants, seven (33%) Exon 1–14 variants, and two (10%) mutations located in the ORF15 region with CORD phenotype. Eight of these gene variants were identified in multiple families. The variants c.2602_2617dup (n = 6) and c.2237_2238del (n = 6) were each detected in three families, whereas c.2405_2406del (n = 6), c.2592dup (n = 8), c.2032G > T (n = 3), c.614dup (n = 3), c.3160G > T (n = 3), and c.3022G > T (n = 2) were each found in two families. Four novel RPGR variants were identified, including c.2602_2617dup, c.1537_1538del, g.38147250_38147255delins38150129_38150260, and c.3022G > T (Supplementary Table S3). Table 1 presents a demographic comparison between the ORF15 group and the Exon 1–14 group. A detailed discussion and image findings for these two groups will be described in the subsequent paragraphs.

Table 1.

Demographics of X-Linked Retinitis Pigmentosa

	ORF15 Group (N = 26)	Exon 1–14 Group (N = 6)	P Value*	Male Group (N = 32)	Female Group (N = 15)	P Value†	Total (N = 47)
Age, mean (range)	40.6 ± 21.3 (7–81)	37.7 ± 17.8 (8–65)	0.479	39.9 ± 20.7 (7–81)	39.8 ± 15.9 (11–65)	0.202	39.6 ± 19.3 (7–81)
Gene variant type			0.6856‡		ORF15: exon1–14 = 10:5	0.698‡	ORF15: exon1–14 = 36:11
Deletion	13	3		16	7		23
Insertion	9	2		11	6		17
Nonsense	3	0		3	1		4
Missense	1	1		2	0		2
Structural variant	0	0		0	1		1
Age at first exam, mean (range)	35.3 ± 21.8, (4–76)	29.8 ± 18.7, (5–60)	0.659	34.3 ± 21.3, (4–76)	34.7 ± 15.3, (3–55)	0.08	34.9 ± 19.7 (3–76)
Onset age, mean (range)	8.2 ± 6.3 (0–21)	5.3 ± 4.7 (0–13)	0.354	7.6 ± 6.2 (0–21)	17.4 ± 14.1 (0–50)	0.006	11.4 ± 11.7 (0–50)
Clinical follow up time year, mean (range)	n = 19	n = 4	0.646	n = 23	n = 9	0.688	n = 32
	4.2 ± 2.6 (0.7–9.9)	4.4 ± 2.8 (2.3–9.0)		4.3 ± 2.6 (0.7–9.9)	3.9 ± 2.9 (0.8–10.1)		4.2 ± 2.7 (0.7–10.1)
Baseline BCVA (logMAR)	OD: 1.0 ± 0.9 (0.1–2.8)	OD: 0.9 ± 0.8 (0.2–2.0)	0.344§	OD: 1.0 ± 0.9 (0.1–2.8)	OD: 0.3 ± 0.5 (0.0–2.3)	0.002§	OD: 0.8 ± 0.8 (0.0–2.8)
	OS: 0.9 ± 0.8 (0.1–2.8)	OS: 0.9 ± 0.8 (0.2–1.7)	0.401§	OS: 0.9 ± 0.8 (0.1–2.8)	OS: 0.3 ± 0.2 (0.0–1.0)	0.001§	OS: 0.7 ± 0.7 (0.0–2.8)
Refractive error (D)	OD: −4.2 ± 3.7 (−12 to 0.5)	OD: −1.8 ± 1.7 (−4.6 to 0)	0.358	OD: −3.8 ± 3.6 (−12 to 0.5)	OD: −4.3 ± 2.9 (−9.6 to 0.8)	0.933§	OD: −4.0 ± 3.6 (−12 to 0.8)
	OS: −3.2 ± 3.8 (−10 to 2.1)	OS: −1.8 ± 1.5 (−4.3 to −0.6)	0.504§	OS: −3.0 ± 3.6 (−10 to 2.1)	OS: −4.3 ± 3.2 (−9.5 to 0.8)	0.455§	OS: −3.5 ± 3.5 (−10 to 2.1)
Image
FAF image
Follow up time year, mean (range)	3.0 ± 2.4 (0.4–9.8)	2.5 ± 0.2 (2.2–2.8)	0.633§	3.0 ± 1.4 (0.4–9.8)	2.8 ± 1.6 (0.7–5.2)	0.576§	2.9 ± 2.0 (0.4–9.8)
Peripheral hypo AF	12	2	1.00‡	14	6 (ORF15: exon 1–14 = 4:2)	0.03║	20
Hyper-AF ring	5	1		6	2 (ORF15: exon 1–14 = 1:1)		8
OCT image
Follow up time year, mean (range)	3.9 ± 2.8 (0.7–11.8)	3.7 ± 1.9 (1.7–6.3)	0.866	3.9 ± 2.7 (0.7–11.8)	2.6 ± 2.9 (0.7–10.1)	0.193	3.8 ± 2.7 (0.7–11.8)
Visible EZ band	11	3	0.53	14	14	0.001	18

P value for group comparisons were conducted by Student's t-test or Mann-Whitney U test for continuous variables, and χ² test or Fisher's exact test for categorical variables.

^*P values for comparison between ORF15 group and Exon 1–14 group.

^†P values for comparison between male group and female group.

^‡P values conducted by Fisher's exact test.

^§P values conducted by Mann-Whitney U test.

^║P values conducted by χ² test.

Molecular Assessment of RPGR Genetic Variant

Among the 47 individuals with diagnosed RPGR-related RP, mean age was 39.6 ± 19.3 years (range 7–81 years). Of the male patients, 26 (81%) carried an ORF15 variant and 6 (19%) had an Exon 1–14 variant. For the female patients/carriers, 10 (67%) carried an ORF15 variant and five (33%) had an Exon 1–14 variant. The most commonly identified variants were deletions, accounting for 49% (23/47), followed by insertions at 36% (17/47), nonsense variants at 9% (4/47), missense variants at 4% (2/47), and structural variants at 2% (1/47). The location of identified RPGR variants are presented in Figure 1A.

Figure 1.

(A) RPGR mutation location in our study population. (B) Visual progression of male patients in the right eye. (C) Visual progression of male patients in the left eye. (D) Visual progression of female patients in the right eye. (E) Visual progression of female patients in the left eye. The red line represents the overall trend between logMAR and age, using a population-averaged (marginal regression line) across all individuals derived from a linear mixed model. Male OD: logMAR = −0.215 + 0.034 × age; Male OS: logMAR = −0.122 + 0.031 × age; Female OD: logMAR = 0.217 + 0.003 × age; Female OS: logMAR = −0.268 + 0.0004 × age.

Age of Onset and Symptom Presentation

Among male patients, 19 individuals (59%) reported symptom onset occurring between 0–10 years of age, whereas eight (25%) individuals reported onset between 11–20 years of age. Nyctalopia was reported in almost all patients (94%) at their first visit. Furthermore, decreased visual field was found in 69% of patients, and blurred vision was reported by 59%.

In the female group, four (27%) individuals were asymptomatic. Among those exhibiting phenotypic symptoms, four (36%) individuals experienced symptom onset between 0–10 years of age, two (18%) individuals between 11–20 years of age, and four (36%) individuals between 21–40 years of age. Nyctalopia was reported by more than half of female patients/carriers (64%), with other commonly reported symptoms including photophobia (46%) and decreased visual field (27%).

Disease Progression in Male Patients

Analysis of Visual Acuity Progression

Mean baseline visual acuity was 0.9 ± 0.84 logMAR in the right eye and 0.9 ± 0.74 logMAR in the left eye. Spherical equivalent (SE) was −3.8 ± 3.5 D (range −12.0 to 0.5D) in the right eye and −3.0 ± 3.6 D (range −10.0 to 2.1D) in the left eye after excluding patients who had undergone posterior chamber intraocular lens implantation. High myopia was present in 19% and 20% of the right and left eyes, respectively. Baseline BCVA (r = 0.767, P < 0.001) and refractive error (r = 0.624, P = 0.002) between the right and left eyes showed high correlation. The progression rate of BCVA was 0.034 ± 0.005 and 0.031 ± 0.004 logMAR/year (P < 0.001) in the right and left eyes for all male patients (Figs. 1B, 1C). However, pooled correlation analysis between BCVA and SE in both eyes demonstrated a very low correlation (r = 0.165, P = 0.274)

Patient Characteristics in Different Disease Status

Patients were classified into early and late disease stage based on their FAF and OCT findings. FAF was available in 25 patients, with a mean follow-up of 2.95 ± 2.19 years (range 0.4–9.8 years, n = 20), and OCT was obtained from 24 patients, with 14 of them presenting visible EZ width. Hyper-AF ring and remaining EZ band were identified as early disease characteristics, whereas those presenting with hypo-AF atrophy and loss of EZ band were classified as being in a late disease stage.

In patients with visible EZ band, mean age was 17.8 ± 9.9 years old (range 5–32 years). The mean EZ band horizontal width was 1827 ± 1418 µm in the right eye and 1814 ± 1441 µm in the left eye. Constriction rate of EZ width was −90.9 ± 26.3 µm/year (P = 0.004) for the right eye and −89.9 ± 24.5 µm/year (P = 0.003) for the left eye. All metrics showed statistically significant differences in mean EZ at baseline for the right and left eye, with details provided in Table 2.

Table 2.

Baseline Value and Progression Rate for Biomarkers

				Pearson's Correlation Coefficient
	OD Mean ± SD	OS Mean ± SD	P Value	r	P
Hyper-AF baseline (mm2)	58.6 ± 59.8	47.8 ± 50.0	0.277	0.951	0.004
Hypo-AF baseline (mm2)	171.0 ± 90.8	166.4 ± 93.4	0.639	0.957	<0.000
EZ horizontal width (µm)	1827.2 ± 1418.5	1813.9 ± 1441.5	0.536	0.988	<0.000
EZ vertical width (µm)	1536.0 ± 1449.0	1518 ± 1439.6	0.445	0.995	<0.000
EZ area (mm2)	3.8 ± 7.8	3.8 ± 8.2	0.347	0.999	<0.000
Hyper-AF rate (LMM) (mm2/year)	−4.8 ± 2.6	−3.8 ± 2.2	0.281	0.903	0.014
Hypo-AF rate (mm2/year)	3.6 ± 1.2	2.7 ± 1.3	0.061	0.973	<0.000
EZ horizontal rate (µm/year)	−90.9 ± 26.3	−89.9 ± 24.5	0.979	0.833	0.001
EZ vertical rate (µm/year)	−67.4 ± 23.3	−73.7 ± 22.9	0.323	0.883	<0.000
EZ area rate (mm2/year)	−0.3 ± 0.2	−0.3 ± 0.1	0.444	0.994	<0.000

LMM, linear mix model regression; OD, right eye; OS, left eye.

P value was calculated by simple paired t-test for inter-ocular symmetry.

^*After taking the natural logarithm of the raw values of hyper-AF ring, linear mixed model regression was conducted.

Hyper-AF ring was observed in six patients with a mean age of 12.3 ± 5.8 years old (range 7–24 years), which presented as a wide hyper-AF ring with a border near the outer periphery, constricting toward the posterior pole as patients aged from 4 to 10 years old. Patients aged 7 to 20 years exhibited a slower constriction rate of hyper-AF ring from the posterior pole to near the macular area, and the hyper-AF ring remained relatively stable when reaching 20 years old. Mean area of hyper-AF ring was 58.6 ± 59.8 mm² in the right eye and 47.8 ± 50.0 mm² in the left eye. Constriction rate of ring area based on linear regression was −4.1 ± 1.8 mm²/year (P = 0.032) when data from both eyes were recruited. Regarding assumed exponential changes in hyper-AF ring, the exponential decay model was y = 335.59e^{−0.26  × t} in the right eye and y = 224.40e^{−0.24  × t} in the left eye, where y represents the ring area (mm²) and t denotes age in years. The decay constant was 0.26 for the right eye and 0.24 for the left eye, respectively.

As for patients with advanced disease characteristics, mean age was 39.7 ± 17.8 years old (range 15–81 years) for those with peripheral hypo-AF atrophy. Generalized peripheral atrophy gradually converged toward the retinal vessels region and merged into larger atrophic spots, eventually evolving into definitely decreased autofluorescence lesions (DDAF) mostly in their third to fourth decade. Mean hypo-AF region was 171.0 ± 90.8 mm² in the right eye and 166.4 ± 93.4 mm² in the left eye. Progression rate of hypo-AF region was 3.6 ± 1.2 mm²/year (P = 0.00882) in the right eye and 2.7 ± 1.3 mm²/year (P = 0.0513) in the left eye. A relatively small increase in hypo-AF lesions was observed, which could be attributed to nearly half of the patients already presenting with advanced disease at initial assessment.

To further clarify our findings, we stratified individuals with hypo-AF atrophy into two groups based on preservation of macular region, as determined by the presentation of the EZ band on OCT. In the preserved macular group (n = 7), the mean hypo-AF region measured 95.4 ± 76.1 mm² in the right eye and 92.1 ± 88.9 mm² in the left eye, with a mean age of 23.4 ± 8.1 years. Conversely, in the affected macular group (n = 7), the mean hypo-AF region was 236.8 ± 34.5 mm² in the right eye and 220.9 ± 44.1 mm² in the left eye, with a mean age of 46.3 ± 17.1 years. Furthermore, we observed a relatively preserved baseline BCVA in the preserved macular group (0.6 ± 0.4 logMAR in the right eye and 0.5 ± 0.3 logMAR in the left eye) with a slower BCVA progression rate (0.03 ± 0.01 logMAR/year; P = 0.005 in the right eye and 0.02 ± 0.01 logMAR/year; P = 0.088 the left eye) compared to the affected macular group. The latter group demonstrated baseline BCVA of 0.9 ± 0.3 and 1.2 ± 0.6 logMAR in the right and left eyes, respectively, with a more rapid BCVA progression rate of 0.09 ± 0.02 logMAR/year (P < 0.001) in the right eye and 0.04 ± 0.01 logMAR/year (P = 0.009) in the left eye.

Geno-Phenotype Comparison Between ORF15 Group and Exon 1–14 Group

Among the Exon 1–14 group (n = 6), only one individual reported symptom onset between 11–20 years old, whereas the others experienced symptom onset between 0–10 years old. In the ORF15 group (n = 26), half of the individuals had symptom onset between 0–10 years old, and 27% (n = 7) had onset between 11–20 years old. The Exon 1–14 group had a mean onset age of 5.3 ± 4.7 years, whereas the ORF15 group had a mean onset age of 8.2 ± 6.3 years, indicating a trend of earlier disease onset in the Exon 1–14 group, although no statistical significance could be reached (t-test, P = 0.14). The BCVA progression rate in the right eye and left eye was 0.038 ± 0.006 and 0.031 ± 0.004 logMAR/year (P < 0.001), respectively, in the ORF15 group, but was 0.019 ± 0.011 and 0.011 ± 0.013 logMAR/year in the Exon 1–14 group, without no significance (P = 0.115 and 0.457, respectively).

In our comparative analysis of disease characteristics between the two groups, we observed a relatively constricted hyper-AF ring in the single patient from the Exon 1–14 group when compared age-matched patients in the ORF15 group. All other patients in the Exon 1–14 group with available FAF were categorized into the hypo-AF atrophy with preserved macular group. Notably, the mean age of this subgroup was higher (30.8 years) compared to their counterparts in the ORF15 group (mean age 20.5 years). Further analysis of patients aged 30–40 years revealed distinct differences between the Exon 1–14 group and the ORF15 group. Only 20% of patients in the ORF15 group demonstrated preservation of the macular region on both FAF and OCT, whereas all patients in the Exon 1–14 group had an intact macula. Overall, the Exon 1–14 group exhibited a trend of earlier disease onset, but the ORF15 group were correlated with more severe disease progression. However, no statistical evidence was available, most likely because of the small number of patients.

Prediction Model of EZ Band Length in the ORF15 Group

To clarify the constriction trend of the EZ band, we applied a model demonstrating that its variation follows an exponential decay pattern, supporting the hypothesis of early disease progression in the ORF15 group. The linear mixed models incorporating exponential decay were y = 11440.18e^{-0.16  × t} for the right eye and y = 18197.22e^{-0.19  × t} for the left eye, where y represents the horizontal length (µm) of EZ band and t denotes age in years. As depicted in Figure 2, the EZ band in the ORF15 group rapidly condensed during childhood, with horizontal lengths of approximately 2309 and 2721 µm remaining in the right and left eyes, respectively, by the age of 10 years. All patients older than 40 years exhibited a complete loss of the EZ band. We observed a 5-year-old patient presenting with EZ band length of approximately 10,000 µm. Detailed clinical findings and the impact of this case on the EZ band constriction model are provided in Supplementary Data 4.

Figure 2.

Exponential prediction model of horizontal EZ band in ORF15 group. (A) The right eye. (B) The left eye.

Disease Progression in Female Patients/Carriers

Baseline visual acuity was 0.3 ± 0.3 logMAR (range 0.0–2.3) in the right eye and 0.3 ± 0.2 logMAR (range 0.0–1.0) in the left eye. All recorded BCVA were better than 0.5 logMAR, except for one 60-year-old patient who presented with a visual acuity exceeding 1.0 logMAR. Mean SE was −4.3 ± 2.8 D (range −9.6 to 0.75D) in the right eye and −4.3 ± 3.2 D (range −9.5 to 0.75D) in the left eye, after excluding patients/carriers who underwent posterior chamber intraocular lens implantation. High myopia accounted for 31% and 46% in the right eye and left eye, respectively. Statistical analysis revealed no significant correlation between age and BCVA (r = 0.476, P = 0.073 in the right eye and r = 0.227, P = 0.417 in the left eye) (Figs. 1D, 1E). Likewise, SE and BCVA demonstrated a very weak correlation in both eyes (r = −0.13, P = 0.537).

Phenotype Classification According to Fundus Autofluorescence Scan

According to the classification system by Nanda et al, our female patients/carriers were classified as below: 7% (1/15) as N pattern, 46% (7/15) as R pattern, 20% (3/15) as F pattern, and 27% (4/15) as M pattern (Fig. 3). If patients had interocular asymmetry, categorization was based on the eye with worse presentation.

Figure 3.

Different grade of female X-linked retinitis pigmentosa based on fundus autofluorescence Phenotype heterogeneity in female patients. Five examples of CFP, FAF, and macular OCT are presented. (A) A 52-year-old asymptomatic patient, exhibiting normal presentation on CFP, FAF, and OCT. (B) A 45-year-old patient, demonstrating relatively normal appearance on CFP, with a wedge-shaped reflex extending from central to peripheral regions with macular sparing on FAF, and normal macular OCT. (C) RPE atrophy with foveal sparing evident on CFP, accompanied by radial reflex pattern and asymmetrical pigmentary changes on FAF, limited EZ layer on horizontal and vertical view in the right eye on OCT. (D) Presence of bone spicule pigmentation with generalized RPE atrophy in the right eye, extensive hypo-AF atrophy in the right eye and pigmentation localized to inferonasal and supertemporal region in the left eye, total loss of EZ layer in the right eye with remaining EZ band in the left eye. (E) Characteristic hyper-AF ring observed on FAF, corresponding to a limited EZ band on OCT.

Patients/carriers with F pattern was older (range 40–55 years). All individuals demonstrated interocular asymmetry in their presentations. Two patients/carriers exhibited R pattern in their contralateral eye, while another displayed asymmetric hypo-AF lesions area in the inferonasal quadrate of both eyes. During follow up, these lesions progressed to lesions without encroaching upon previously unaffected areas. Among the four M patterned patients/carriers, ages were 12, 36, 49, and 60 years, respectively. Two individuals showed symmetric M pattern with hyper-AF ring in both eyes, whereas the 36-year-old patient spontaneously developed peripheral hypo-AF atrophy. Notably, both the constriction of the hyper-AF ring and the progression of hypo-AF remained stable during follow up period. The remaining two patients/carriers with male pattern demonstrated asymmetry between eyes. One individual presented with pericentral hypo-AF in the right eye and lower quadrant semicircle pigmentation in the fellow eye. The second patient exhibited generalized hypo-AF with macular involvement in the right eye, contrast by multifocal hypo-AF areas in the fellow eye.

Patients/carriers with N and R pattern showed normal macular structure on OCT. F pattern exhibited a mix of hypo-AF areas with narrowed EZ band width and non-pigmented areas which remained intact. Most patients/carriers with M pattern had visible EZ band width in both eyes, but one patient presented with complete EZ band loss in one eye.

Common Genotypes and Presentative Cases

Eight genetic variants were identified in multiple families within this cohort: c.3160G > T, c.3022G > T, c.2405_2406del, c.2602_2617dup, c.2592dup, c.2237_2238del, c.2032G > T, and c.614dup. The most prevalent genetic variant was c.2592dup, observed in 15% of the study population. The variants c.2602_2617dup and c.2237_2238del both accounted for 12% of cases, whereas the remaining variants were each presented in 5% of cases.

RPGP Variant c.2602_2617dup

This subgroup comprises six patients from three distinct families, designated as patients A through F. Patients B and D are siblings, whereas patient A is their cousin within the same family pedigree. Patients C and E are siblings from a separate family. The onset age for all patients was between 0 and 10 years, with the exception of patient E, who reported onset at 21 years of age. Nyctalopia was reported in all male patients and in patient F, a female who presented with a male type phenotype. CFP, FAF, and macular OCT for all patients are presented in Figure 4.

Figure 4.

The 2602_2617dup-related phenotypes. (A) A 5-year-old male presenting with near-normal appearance on CFP, exhibiting a hyper-AF ring with boarder in the midperiphery on FAF, and thinning of the outer retinal layer outside the macular region. (B) A 10-year-old male presenting arteriolar attenuation with atrophic RPE on CFP, a hyper-AF ring in the parafoveal region with a peripheral hypo-AF region in the far periphery on FAF, and remaining EZ width in the macular region with outer retinal layer thinning. (C) A 33-year-old male with impaired BCVA (0.8logMAR in both eyes). Bone spicule pigmentation, atrophic RPE, arteriolar attenuation, and disc pallor on CFP, general DDAF lesions with macular involvement on FAF. Severely attenuated outer retinal layer, with scattered remnants of the ellipsoid zone. (D) A 6-year-old female, exhibiting relatively normal appearance on CFP, a wedge-shaped reflex extending from central to peripheral regions with macular sparing on FAF, and normal macular appearance on OCT. (E) A 40-year-old female. Bone spiculate pigmentation with atrophic RPE in the lower hemispherical region on CFP, corresponding with DDAF lesions around the inferior vascular arcade on FAF in the left eye. The remaining region appears relatively healthy with a tapedal reflex, as dose the right eye. Evident EZ band loss on both horizontal and vertical views of the left eye, correlating with FAF features. (F) A 48-year-old female. Atrophic RPE with arteriolar attenuation on CFP, mottled hypo-AF in the mid periphery with temporal upper quadrate DDAF lesion in the right eye and lower hemispherical atrophy in the left eye on FAF. Limited remaining EZ band in the right eye and wider EZ width in the left eye were correspondent with the presentation on FAF.

In this subgroup, three male patients exhibited similar disease phenotypes irrespective of age, but displayed distinct features corresponding to different stages of disease progression (Figs. 4A–C). Conversely, three female patients/carriers, each from different families and of varying ages, presented notably diverse FAF patterns, ranging from R type to M type (Figs. 4D–F). Furthermore, advanced FAF patterns were observed in the older patients/carriers.

RPGP Variant c.2237_2238del

This subgroup comprises six patients from three distinct families, designated as patients G through L. Patients G and H are siblings. Patients J, K, and L form a two-generation pedigree, with J in the first generation and K and L as siblings in the second. Age of onset ranged from 10 to 21 years for male patients, except patient G, who reported symptoms since birth. Female patients/carriers had onset at 20 and 25 years, respectively. All patients reported nyctalopia with blurred vision. Visual field loss progressed to tunnel vision in patients I and J, while patients G and H experienced a decrease in visual field.

In this subgroup, all male patients exhibited an atypical phenotype, characterized by atrophic RPE on CFP, diffusive midperiphery atrophy sparing the macula on FAF, and generalized outer retinal layer thinning on OCT (Figs. 5G, 5H). Advanced cases displayed bone spiculation and vascular attenuation (Figs. 5I, 5J). Two female patients/carriers were classified as R pattern on FAF, exhibiting normal findings on CFP and OCT (Figs. 5K, 5L). This subgroup exhibited homogeneous presentation associated with the same genetic variant, albeit manifesting distinct phenotypically from the previously described c.2602_2617dup variant.

Figure 5.

c.2237_2238del related phenotypes. (A) A 24-year-old male exhibiting atrophic RPE on CFP, diffusive scattered hypo-AF regions in the midperiphery without evolving into DDAF or merging into larger atrophic lesions, with a preserved macular region on FAF. Thinning of the outer retinal layer with preserved EZ width in the macula on OCT. (B) A 25-year-old male presenting with atrophic RPE on CFP, and borderless general hypo-AF with arteriolar attenuation on FAF. Thinning of the outer retinal layer with preserved EZ width at macula on OCT. (C) A 41-year-old male with visual acuity impairment (BCVA 1.0 logMAR in the right eye and 1.3 logMAR in the left eye). Color fundus showed peripheral bone spiculation, vascular attenuation, and atrophic RPE, hypo-AF lesions evolving into DDAF on FAF. No visible ellipsoid zone on OCT. (D) A 53-year-old male. Generalized bone spiculation, atrophic RPE, and optic disc pallor on CFP, generalized hypo-AF on FAF, and no visible ellipsoid zone on OCT. (E) A 28-year-old female displaying near-normal appearance on CFP and OCT, with an R pattern on FAF. (F) A 22-year-old female exhibiting near-normal appearance on CFP and OCT, with ab R pattern on FAF.

The atypical phenotype observed in male patients with c.2237_2238del, which was characterized by diffusive midperiphery atrophy sparing the macula, was also observed in a 35-year-old male individual with the c.2032G > T variant. Interestingly, individuals exhibiting this phenotypic presentation maintained relatively preserved macular regions and measurable EZ width even at 30–40 years old, unlike those with convergence of hypo-AF atrophic lesions with age, who generally experienced earlier macular involvement in our observation.

RPGR-Related Cone Rod Dystrophy

A total of five individuals from four families were diagnosed with cone-rod dystrophy in our cohort. Four (80%) were male, and one (20%) was female. Genetic variants were located in c.3160G > T and c.3022G > T, both of which were nonsense variants in the ORF15 region.

The only female patient/carrier was 73 years old, who was asymptomatic with BCVA 0.2 logMAR in the right eye and 0.1 logMAR in the left eye. Mean age for male patients was 46.2 ± 13.3 years old (range 36.8–68.9 years). The age of onset was 37 ± 13.9 years-old (range 27–61 years). Three (75%) reported blurred vision as initial symptom, as well as photophobia, and one (25%) reported nyctalopia. No visual field defect or central visual blind spot was reported.

Baseline BCVA was 0.9 ± 0.5 logMAR (range 0.2–1.3) in the right eye, and 0.9 ± 0.5 logMAR (range 0.2–1.3) in the left eye. BCVA progression rate was 0.02 ± 0.01 logMAR (P = 0.246) in the right eye and 0.03 ± 0.01 logMAR (P = 0.139) in the left eye. Mean clinical follow-up time was 2.3 ± 2.1 years (range 0.7–5.3, n = 3). All patients demonstrated central hypo-AF atrophy at the macular region with peripheral hyper-AF ring on FAF, and loss of central EZ band was evident on OCT. Mean central hypo-AF atrophy area was 3.8 ± 1.7mm² in the right eye and 5.6 ± 2.7mm² in the left eye. Mean central EZ area loss was 3.1 ± 1.97 mm² in the right eye and 3.2 ± 2.3 mm² in the left eye.

We observed expansion of the central atrophic area, as well as an increased DDAF lesion within the central atrophic area in one 41-year-old patient with four years of longitudinal follow up (Figs. 6A, 6B). Concomitantly, expansion of the vanishing EZ band at the macular region was observed on OCT. Another 37-year-old patient with an identical variant presented with a smaller central atrophic area on FAF (Fig. 6C), highlighting late progression of CORD.

Figure 6.

Cone-rod dystrophy presentation. (A) A 41-year-old male presenting with BCVA 1.3 logMAR in both eyes. Central atrophic RPE on CFP, macular hypo-AF atrophic region with multiple DDAF spots encircled by a peripheral hyper-AF ring on FAF, loss of the EZ band in the central region on OCT. (B) The same patient as in (A) after four years follow-up, demonstrating expansion of the DDAF region within the hypo-AF area, with lesions merging into a larger size. (C) A 37-year-old male with BCVA 0.7 logMAR in both eyes. Central atrophic RPE on CFP, relatively small hypo-AF region compared to patient A, notably without DDAF spot development. A merely visible remaining EZ band in the central macular region on OCT.

Discussion

In this retrospective longitudinal cohort study, we elucidate the natural disease course of 52 individuals with RPGR-related disease-causing variant(s) in Taiwan, and present the first predictive model of the exponential constriction rate of the EZ band in the male patients of the ORF15 group. In our institution, RPGR-related RP constituted 10% of all RP cases.²¹ We identified 21 distinct genetic variants in the RPGR region: 14 were in the ORF15 region, and seven were in the Exon 1–14 region. Among male individuals, 81% carried genetic variants in the ORF15 region, whereas 19% had variants in the Exon 1–14 region. Our findings demonstrated a higher percentage of variants in the ORF15 region compared to previous studies,^5,29–32 most of which reported frequencies ranging from 50% to 73% across genders. Notably, Kuruvilla et al.³³ reported the lowest frequency of the ORF15 variants (34%) in New Zealand, including patients of Māori ancestry and Samoan families. Given that our study population comprises Asian individuals, we postulate that regional or ethnic differences might influence the prevalence of ORF15 variants.

Despite numerous studies on the natural disease course of RPGR-related RP, the contribution of genetic variant location to disease severity remains controversial. Our findings indicate that disease progression tends to be more severe in the ORF15 group compared to the Exon 1–14 group. We observed an earlier mean age of onset (5.3 years) in the Exon 1–14 group compared to the ORF15 group (8.2 years). Similarly, Bellingrath et al.³⁴ reported a trend toward younger disease onset in the Exon 1–14 group. Additionally, the annual BCVA decline rate was 0.031logMAR in the ORF15 group, whereas the Exon 1–14 group exhibited a slower decline rate (0.01 logMAR/year), indicating more gradual progression in the Exon 1–14 group, although the disparity did not reach statistical significance. The BCVA progression rate in the present study aligns with previous publications, ranging from 0.013 to 0.043 logMAR/year.^5,18,31,35 Di Iorio et al.⁵ reported an annual BCVA decline rate of 0.011logMAR in patients with the Exon 1–14 variant compared to 0.044logMAR in those with the ORF15 variant (P < 0.001). Talib et al.³¹ documented a decline rate of 0.015 logMAR/year in the Exon 1–14 group, whereas in the ORF15 group it was 0.022 logMAR/year. Conversely, another study³⁰ reported a faster BCVA progression rate (0.035logMAR/year) in the Exon 1–14 group than in the ORF15 group (0.022 logMAR/year). Our prediction model for the EZ band in the ORF15 group demonstrated an exponential constriction rate with increasing age, representing the first model to support the hypothesis of EZ band constriction. This finding highlights early disease progression and reinforces the hypothesis that the RPGR-related RP patients harboring the ORF15 variant are more susceptible to rapid disease progression at a young age.

Discrepancies in image characteristics between the two groups also corroborated slower decline in the Exon1–14 group. Individuals with Exon 1–14 variants were, on average, 10 years older than those with ORF15 variants—30.8 years versus 20.5 years—among patients exhibiting hypo-AF atrophy with preserved macular region. Moreover, when comparing groups of individuals of the same age, all patients in the Exon 1–14 group exhibited a preserved macular region, whereas only 20% of those in the ORF15 group retained an intact macular region. Zou et al.³⁰ found a higher prevalence of hypo-AF with affected macula in the ORF15 group in contrast to the predominantly observed hyper-AF ring and hypo-AF with preserved macula in the Exon 1–14 group. Age distribution differences were cited as the primary reason for these distinct presentation patterns, with a median age of 14 years for the Exon 1–14 group and 32.5 years for the ORF15 group. Nevertheless, our study did not observe significant age differences between groups, suggesting that the Exon 1–14 variant may exert a milder influence on disease progression, although patient numbers were too low to reach statistical significance.

In our study, 73% of female patients/carriers were affected. This high disease burden contrasts with previous findings, which reported symptomatic disease onset rates ranging from 20% to 70%.^19,32,36 Only one patient in our study had a visual acuity of 1.0 logMAR in the better-seeing eye, whereas previous studies reported 13% to 17% of individuals with BCVA > 1.0 logMAR.^19,32 Additionally, previous studies demonstrated a positive correlation between age and decreased visual acuity, with blindness occurring in the sixth decade of life.³² This observation was absent in our cohort, most likely because of the younger age distribution of our patients. Nevertheless, we observed higher disease onset percentage and severity in older patients/carriers, supporting age as a significant factor in disease progression. Among female patients/carriers, 46% exhibited R pattern, followed by 20% with F pattern and 27% with M pattern. In comparison, Nanda et al.²⁷ reported 13% of severe phenotypes without interocular heterogeneity. Contrary to Talib et al.,³² who discovered 23% of heterozygotes with RP or CORD phenotype and 9% exhibiting asymmetry on FAF, our study identified a higher percentage of M pattern (27%) and interocular asymmetric presentation (33%). Our results were similar to those of Marques et al.,¹⁹ who reported that 29% of individuals displayed male pattern and 25% showed asymmetry. Phenotypic heterogeneity and interocular asymmetry in female patients/carriers are believed to result from X-chromosomal inactivation, where one X chromosome is preferentially expressed over the other.³⁷ Consequently, the mutant X chromosome may be selected or inactivated individually, resulting in mosaic presentation on FAF and manifesting as a wide range of phenotypes. Our study supports the hypothesis that severe phenotypes are not rare within heterozygous female patients/carriers. However, selection bias remains a concern because individuals without symptoms are less likely to seek medical attention.

Our study revealed a mean SE of −3.8D in the right eye and −3.0D in the left eye in male patients, with 20% individuals exhibiting high myopia. Female patients/carriers presented a mean of SE −4.3D in both eyes, with 31% and 46% showing high myopia in the right eye and the left eye, respectively. In IRD, the RPGR gene locus has been postulated as one of the critical determinants for refractive error development, particularly associated with increased risk of high myopia.³⁸ Previous studies supported this hypothesis by reporting that approximately one-third of males exhibit high myopia, with mean SE ranging from −3D to −4.9D.^30,31 In a Japanese cohort studied by Mawatari et al.,³⁹ the prevalence of high myopia was 50% in male patients, whereas female patients/carriers showed a mean SE of −8.25D and −10.25D in the right and left eyes, respectively. Other studies identified a frequency of high myopia ranging from 33% to 66% with mean SE from −5 to −8D in female carriers,^32,33,40,41 with an exception of the study by Marques et al,¹⁹ who reported only 17% of high myopia. Higher percentages of high myopia in female patients/carriers compared to male individuals were also detected by Yang et al.,⁴⁰ although refractive error was unrelated to age or genetic variant type of RPGR variants. Additionally, higher refractive error was shown to be correlated with faster BCVA progression.³¹ According to Di Iorio et al,⁵ high myopia was correlated with faster BCVA decline (0.081 logMAR/year in patients with high myopia). However, this association was not observed in our cohort. Given the higher proportion of adults in our cohort, refractive error was presumed to be less variable because of maturity. Despite a higher prevalence of high myopia in young Asian adults compared with the non-Asian population,⁴² our study did not detect a higher frequency of high myopia in RPGR-related RP individuals. On the contrary, we observed a lower percentage of high myopia among male patients. A previous publication from our institution also supported this finding, reporting no significant differences in refractive error between different types of IRDs.⁴³ The disparity of high myopia frequency between genders may be attributed to X-chromosomal inactivation in females, which could cause large variabilities in disease severity.³⁷ Currently, the association between myopia and RPGR variant remains unclear, necessitating further investigation to elucidate their relationship.

Characteristics of RP were quantified and analyzed to deepen our understanding of disease progression in the present study. In younger patients, we detected a hyper-AF ring with a statistically significant constriction rate of y = 335.59e^{−0.26  × t} in the right eye and y = 224.40e^{−0.24  × t} in the left eye. In comparison to Tee et al.,¹⁵ our study found a higher initial value with faster decay of hyper-AF ring, with half-lives of 2.67 and 2.89 years. We observed an EZ width decline rate of −90 µm/year from a mean baseline of 1827µm, with a mean age of 17.3 years. Most studies^{7,17,29,44,45} reported annual EZ width decline from 200 to 300µm/year, with Tee et al.⁴⁶ suggesting that EZ width constriction correlates with both baseline length (r = 0.714, P < 0.001) and age (r = −0.523, P < 0.001). They reported a constriction rate of 176.6 µm/year from a baseline of 2438 µm, with mean age 19.1 years. In addition, Birtel et al.¹⁸ observed larger EZs in younger patients and pronounced EZ decline in patients under 20 years old, reporting a decline of horizontal EZ width at −45.37 µm/year from a baseline EZ of 1091 µm (median age 21 years). Our study population exhibited relatively restricted baseline EZ and slower decline rates despite the younger age of patients, suggesting a more advanced disease stage compared to the abovementioned studies.

In our individual population, the c.2592dup variant was the most prevalent genetic alteration, accounting for 15% of all individuals. The c.2237_2238del and c.2602_2617dup variants were each found in 12% of cases, with c.2405_2406del observed in only 5% of individuals. Lin et al⁴⁷ reported c.2405_2406del as the most frequently encountered RPGR variants in a UK cohort of 228 individuals with RPGR-related RP, affecting 10% of the cohort. In the Chinese population, the most frequently detected variants were the c.2405_2406del and c.2236_2237del variants, present in 15% and 12% of individuals, respectively.⁴⁸ Despite consistent findings of commonly detected gene variants in previous literature,^29,35 our study identified unique variant locations prevalent in the Taiwanese population, highlighting regional genetic variations in individuals with RPGR-related RP.

Our study has several limitations. To begin with, missing data and images may have introduced bias in this retrospective study. In addition, individuals of varying ages presented clinically at different disease stages, complicating comprehensive observation of disease progression. Furthermore, the small sample size in the Exon 1–14 genetic variant group and hyper-AF ring group may have limited our ability to achieve statistically significant results. Additionally, the Snellen scale has limitations in measuring vision changes between visual acuity levels, especially in patients presenting with low vision. Lens status with nuclear sclerosis may also introduce bias in visual acuity assessment. Penultimately, selection bias among female carriers is possible, because those without symptoms may not have sought evaluation at an ophthalmology clinic. Finally, initial visual field findings were based on subjective responses collected through the questionnaire rather than formal visual field tests such as Humphrey field analyzer or Goldmann kinetic perimetry.

In conclusion, this longitudinal study provides a comprehensive analysis of disease progression in individuals with RPGR-related RP in the Taiwanese population. Notably, it is the first study to present a predictive model for EZ band constriction in patients with the ORF15 variants. Male individuals carrying the ORF15 genetic variant exhibited a more severe phenotype, whereas those with the Exon 1–14 variant tended to experience earlier symptom onset. Additionally, a lower percentage of high myopia was observed in male patients. Among female carriers, a higher frequency of symptomatic onset was noted, with 27% exhibiting a male phenotype. These findings contribute to a better understanding of the disease course and could contribute to identifying new therapeutic windows for upcoming gene therapies.