EXPLORE XLRP-2 was a non-interventional study conducted in 23 centres across 10 countries (Austria, Belgium, England, Finland, France, Germany, Israel, Italy, the Netherlands and Spain). Each participating site identified male and female patients who: (1) were aged ≥12 years at screening; (2) had XLRP confirmed by a retina specialist and had a predicted disease-causing sequence variant in RPGR confirmed by genetic testing; (3) were able and willing to give informed consent or assent (with the guidance of a legally acceptable representative, as applicable); and (4) had no participation (currently or previously) in a gene therapy trial. No formal sample size calculation was performed. The sample size of approximately 150 to 200 patients was primarily based on pragmatic considerations, such as disease rarity and ability to enrol enough eligible patients per disease stage for this descriptive and exploratory study.

Insights from clinical experts and patient experts were collected via advisory board meetings to guide development of the study design.

The study received local ethical committee approval at each study site (Austria: 1088/2022; Belgium: B3222022000832; England: 307690; Finland: 48/2022; France: 22.02349.000114; Germany: 839/2021BO1, 2022-200343-BO-bet; Israel: 0391-21-SOR, 9186-22-SMC; Italy: 1026/2021/Oss/AOUFe; the Netherlands: 21.177/VS; Spain: PI2022051) and adhered to the tenets of the Declaration of Helsinki. All participants gave informed consent.

Staging of vision was performed at sites based on information from the last visit (vision status of the better-performing eye). Patients were rated ‘mild’, ‘moderate’ or ‘severe’ based on visual acuity and visual field diameter measurements (Supplementary Table 1) in accordance with the ‘Visual Standards – Aspects and Ranges of Vision Loss’ report from the International Council of Ophthalmology [13].

Data were collected from two sources. One source was cross-sectional self-administered surveys. Feedback on survey questions was collected via interviews by qualified personnel at a call centre, or at the participating site if local regulations did not allow remote survey interviews via call centre. The decision on the collection approach was based on suggestions from patient experts and clinical experts and aimed to provide equal circumstances for the participants irrespective of the level of their visual impairment. Although we collected data from both patients and caregivers, participants could not enrol as both a patient and a caregiver to avoid potential bias in the results. Caregiver outcomes will be published separately.

The other source was retrospective data collected from patients’ medical records available at sites involved in the EXPLORE XLRP-2 study (i.e., ophthalmic centres routinely managing patients with IRDs). Retrospective data included patients’ socio-demographics, comorbidities, clinical parameters, date of first visit to participating site, diagnostic tools used, outcomes of the assessment, consultations, medical resource use, and up to 5 years of retrospective data, where available.

The cross-sectional patient surveys consisted of five patient-reported outcome (PRO) tools: 1. modified Low Luminance Questionnaire (mLLQ); 2. Work Productivity and Activity Impairment General Health v2.0 (WPAI-GH2); 3. Hospital Anxiety and Depression Scale (HADS); 4. Patient Global Impression (PGI)-Mobility; 5. PGI-Daily Activity; and a long-term impact patient questionnaire developed for the study (the outcomes of which will be published separately).

The mLLQ was a modified version of the original 32-item LLQ, an eye disease-specific questionnaire for assessing self-reported visual problems under low luminance and at night [14]. The mLLQ used in this study is the same as the one used in the “Phase 3 Randomized, Controlled Study of AAV5-hRKp.RPGR for the Treatment of X-linked Retinitis Pigmentosa Associated With Variants in the RPGR Gene” (https://clinicaltrials.gov/study/NCT04671433). Modifications were made to the LLQ based on patient feedback, with input from a clinical expert and instrument development experts. The reasons for the modifications included difficulty with interpretation, irrelevance, poor fit between response choice and the item, items assessing more than a single concept, difficulty selecting a response, and difficulty selecting between two response choices. The adult mLLQ, administered to individuals ≥18 years of age, consisted of 30 items across six domains: driving (4 items), extreme lighting (7 items), mobility (6 items), emotional distress (4 items), general dim lighting (6 items) and peripheral vision (3 items). The adolescent mLLQ, administered to individuals 12–17 years of age, comprised 22 items across five domains: extreme lighting (4 items), mobility (6 items), emotional distress (4 items), general dim lighting (5 items) and peripheral vision (3 items). Scores range from 0 (maximal difficulty) to 100 (no difficulty) for both adults and adolescents.

The WPAI-GH2 measured effects of general health and symptom severity on work productivity and regular activities [15]. It consisted of six questions: i. working for pay status (yes/no; if no, only question 6 was answered); ii. hours missed due to health problems; iii. hours missed for other reasons; iv. hours actually worked; v. degree to which health affected productivity while working (measured on a visual analogue scale from 0 to 10); and vi. degree to which health affected productivity in regular unpaid activities (visual analogue scale from 0 to 10). For the last two questions, the lowest score (0) represents no impact. The recall period for questions 2–6 was 7 days. Responses were used to derive four scores (absenteeism, presenteeism, work productivity loss and activity impairment). These four scores were expressed as impairment percentages (0–100%), with higher numbers indicating greater impairment/less productivity.

The HADS measured symptoms of anxiety and depression and comprised seven items, each with depression and anxiety subscales. Scoring for each item ranged from 0 to 3, with 3 denoting the highest anxiety or depression level. A total subscale score of >7 points out of a possible 21 denoted considerable symptoms of anxiety or depression [16, 17]. The recall period was 7 days.

The PGI-Mobility measured the impact of patients’ vision status on their mobility (e.g., walking outside, travelling, using stairs or curbs). The PGI-Daily Activity measured the impact of vision status on patients’ daily activities (e.g., seeing the television, recognising or meeting people, or reading) [18]. Each consisted of one item, and scoring ranged from ‘not at all’ to ‘very much’ on a 5-point scale. The recall period was 7 days.

Data were summarised using descriptive statistics. Continuous/ordinal variables were summarised using number of patients (n), mean, standard deviation (SD), median, minimum, maximum, and 95% confidence interval (CI). Categorical variables were summarised with n, percent and 95% CI. No truncation of negative lower CI value to 0 has taken place. Correlations between level of disease stage and level of burden were analysed exploratively using appropriate correlation methods, and all reported p-values were non-adjusted, nominal and exploratory in nature.

Analysis sets considered in this study included: (1) full analysis set (FAS): patients enrolled with data on at least one PRO and not enrolled as both patient and caregiver; (2) modified FAS (mFAS): a subset of FAS participants who did not discontinue the study for withdrawal reasons (withdrawal by legally authorised representative, withdrawal by parent or guardian, withdrawal by subject); and (3) modified per-protocol (mPP): a subset of mFAS participants who satisfied all inclusion and exclusion criteria.

Primary analysis was based on the mPP analysis set. Descriptive statistics are provided for the overall population as well as for each severity stage of XLRP. For ordinal variables, in addition to descriptive statistics, frequency distribution (number and percentage) is provided for each level of the ordinal variable by stage of XLRP. Correlation between stage of XLRP and continuous/ordinal variable was analysed and Kendall τb or τc rank correlation coefficient with corresponding p-values are provided, depending on the variable.

For the relationship of stage of XLRP with nominal variables (e.g., type of employment: full-time paid, part-time paid, etc.), frequency distributions (number and percentage with corresponding 95% CI) are provided for each category of categorical variable by stage of XLRP. Correlations between stage of XLRP and binomial variables (e.g., yes/no responses, such as ‘XLRP had impact on level of education’) were analysed and rank-biserial correlation coefficient (a measure of association estimated by Goodman–Kruskal’s gamma) with corresponding p-value is provided.