Patient-Reported Outcomes in RLBP1 Retinal Dystrophy: Longitudinal Assessment in a Prospective Natural History Study

Abstract

Purpose: To evaluate the performance of two non-disease–specific patient-reported outcome (PRO) instruments, the National Eye Institute Visual Function Questionnaire-25 (VFQ-25) and the Low Luminance Questionnaire (LLQ), in patients with retinaldehyde-binding protein 1 retinal dystrophy (RLBP1 RD).

Methods: PROs were assessed using the VFQ-25 and LLQ. Rasch analysis was conducted to estimate person and item measures of the VFQ-25 and LLQ questionnaires to determine the association between the two PROs. In addition, the association between these two instruments and their correlations to weighted measures of visual function and disease progression were analyzed in this three-year PRO-focused sub-study of a five-year prospective natural history study.

Results: Forty-two patients participated, with most of them having completed at least two PRO follow-up visits at least one year apart. The mean VFQ-25 scores were lowest for distance activities (39.2–49.0) and peripheral vision (37.5–52.4), with mean LLQ subscale scores generally low (<41), except for the emotional distress subscale. Using Rasch analysis, calibrated item and person measures along with their standard errors were estimated for both ePROs. This indicated that the distribution of the VFQ-25 and LLQ item measures well covered the distribution of person function in this group. This suggests that the item difficulties well cover the person-level performance in this population. As well, the two PROs showed a strong and significant correlation at all assessed time points as assessed with Pearson correlation coefficient (0.81, 0.91, 0.81 and 0.87 at baseline, 1/1.5, 2/2.5 (P < 0.001) and 3/3.5 years (P = 0.002)). The composite scores of both PRO questionnaires strongly correlated with clinical measures of visual function. At 2 to 2.5 years of follow-up, meaningful statistically significant declines in peripheral vision (both VFQ-25 and LLQ), distance vision (VFQ-25), and extreme lighting in dark and bright light (LLQ) subscales were noted.

Conclusions: This study demonstrated a strong association between VFQ-25 and LLQ scores and their association with clinical measures of visual function.

Translational Relevance: PRO instruments can provide insights into the specific disabilities of this unique patient population and help to guide appropriate outcome measures for future clinical trials.

Introduction

Biallelic retinaldehyde-binding protein 1 (RLBP1) mutation–associated retinal dystrophy (RLBP1 RD), a form of retinitis pigmentosa, is a rare disease with limited literature^1–12 and the highest number of patients reported in specific areas of Sweden (Västerbotten County) and Canada (Newfoundland).^{2,6–9,11–14} For RLBP1 RD patients, symptoms usually start with night blindness during early childhood, subsequent progressive visual field restriction, and later loss of central vision, leading to legal blindness by early adulthood or middle age.^13,15,16 These symptoms have a profound impact on the patient's abilities to perform vision-dependent functions of everyday living, thereby affecting their physical, social, and psychological well-being, with important socioeconomic consequences. Considering the progressive nature of the disease, although an ophthalmic examination can provide an objective assessment of a patient's clinical status and disease progression, it can offer little insight into the patient's quality of life (QoL). It is essential to understand the patient's perception of visual function during the course of the disease, in addition to routine structural and functional ocular assessments.¹⁷ 

Patient-reported outcome (PRO) questionnaires are increasingly being used to understand the patient's perspective about the impact of their disease through subjective self-reporting on their experiences, as well as QoL. As yet, no disease-specific PRO instruments have been developed and validated for RLBP1 RD.^2,18 However, non-disease–specific PRO instruments, such as the National Eye Institute Visual Function Questionnaire-25 (VFQ-25)^19,20 and the Low Luminance Questionnaire (LLQ)²¹ have been used to assess QoL in patients with inherited retinal dystrophies (IRDs) and age-related maculopathy, respectively. Recently, a cross-sectional study by Green et al.,² which to our knowledge is the only study using three PRO instruments (VFQ-25, LLQ, and the Visual Activities Questionnaire [VAQ]) for same patients with inherited RD, reported on visual symptoms such as night blindness, difficulty adapting to changes in lighting, and difficulties seeing in bright lighting in patients with RLBP1 RD and showed their substantial impact on patients’ daily lives and physical functioning.² Further studies with additional analysis examining these PRO questionnaires and their association with vision function measures might give different perspectives for patients with RLBP1 RD. 

In the five-year natural history study, longitudinal functional and structural retinal measurements were evaluated in patients with RLBP1 RD from Sweden and Canada. The performance of two non-disease–specific PRO questionnaires (VFQ-25 and LLQ) were concurrently evaluated in these patients in a three-year sub-study, to further assess vision-related functioning. The PRO assessments were introduced after most participants had completed their month 24 follow-up visit in the five-year natural history study.²² Here, we report the VFQ-25 and LLQ outcomes, correlation between these two PRO instruments, and their association with clinical measures of visual function endpoints in RLBP1 RD patients over a follow-up period of three years. 

Methods

Study Design and Patient Population

The evaluation of the PROs was conducted as a three-year sub-study, part of a five-year longitudinal, prospective, natural history study of 44 patients aged 17 to 69 with biallelic mutations in the RLBP1 gene (RLBP1 RD patients), at two clinical sites (Umeå, Sweden, and Newfoundland, Canada) between September 2013 and June 2019. The study was conducted in accordance with the principles of the Declaration of Helsinki, International Conference on Harmonization Good Clinical Practice Guideline, and other local regulations such as the Health Research Ethics Authority in Newfoundland, Canada (reference no. OHRP IRB00011348) and Regional Ethic Review Committee in Umeå, Sweden (Dnr 2012-498-31M and 2018-216-32M). In addition, the study was compliant with the Health Insurance Portability and Accountability Act of 1996. The details of this five-year study including the patient population, and inclusion and exclusion criteria are provided in the Supplementary Appendix. Participants in the five-year natural history study had scheduled visits for evaluation of disease progression every six months. 

Study Objectives, Endpoints, and Assessments

This study investigated the longitudinal utility of two established, non-disease–specific PRO questionnaires (VFQ-25 and LLQ) in RLBP1 RD patients. The VFQ-25 included the composite score and 12 subscale components as follows: general vision, distance activities, near activities, peripheral vision, driving, color vision, ocular pain, general health, mental health, role difficulties, dependency, and social functioning.²³ The LLQ included the composite score and six subscale components as follows: extreme lighting, general dim lighting, peripheral vision, mobility, driving, and emotional distress.²¹ For both questionnaires, the range of composite and subscale scores was 0–100, with a higher score representing better functioning (more information on each PRO instrument is given in the Supplementary Appendix). 

Ongoing enrollment of patients between September 2013 and June 2019 and the introduction of PRO assessments after a protocol amendment in September 2015 led to patients having their first PRO assessment at the next available visit for each patient. Thus PRO baseline visits were at different time points, in comparison with the five-year natural history study. To aid the interpretation and creation of summary findings, results have been presented together to reflect the first completed assessment (described as PRO baseline visit), and follow-up visits at 1 to 1.5 (1/1.5) years, 2 to 2.5 (2/2.5) years, and 3 to 3.5 (3/3.5) years after the reference PRO baseline visit, to facilitate comparisons across all patients (Table 1). It was recommended that the PROs be assessed before all clinical measures of visual function assessments, preferably by the same assessor and, where possible, in a quiet, relatively isolated location. As described previously,^21,23,24 the PRO scores were computed based on patient responses ranging from 0 to 100 (100 represents the highest functional level and 0 the lowest) with scores <50 considered low. 

Table 1.

 

PRO Assessment Time Points of Enrolled Patients During PRO Sub-Study

 

Clinical measures of visual functional parameters including visual acuity (VA), Humphrey visual field (HVF) mean deviation (MD), and contrast sensitivity (CS) were assessed for each visit at which the PRO questionnaires were administered. The Early Treatment Diabetic Retinopathy Study best-corrected visual acuity was expressed on a logarithm of the minimum angle of resolution scale. The HVF was measured using the Swedish Interactive Threshold Algorithm standard 30–2 program with foveal thresholds turned on and CS was measured with Pelli-Robson charts (highest score = 2.25; lowest = 0). To account for binocular vision, single weighted averages (w) of visual acuity (wVA), HVF mean deviation (wvMD), and CS (wCS) of individual eye scores were calculated as 0.75 of the value for the eye with better VA and 0.25 of the value for the fellow eye (based on the adjusted World Health Organization categories of visual impairment²⁵ and published data in glaucoma informed visual field index severity ratings).²⁶ The criteria for the selection of the better-seeing eye are described in the Supplementary Appendix. 

Data Analysis and Statistics

Descriptive statistics (e.g., mean and standard deviation [SD]) were used to characterize the clinical characteristics and demographics of the patients. All analyses were performed and reported for all the patients combined from Sweden and Canada and by clinical site. PRO scores (composite and subscales) were summarized descriptively at each time point. Box and whisker plots were generated to compare each instrument composite and within subscales. 

To determine whether more modern psychometric methods yield outcomes consistent with the results obtained using the more traditional composite score, a Rasch analysis was conducted to estimate person and item measures of the VFQ-25 and LLQ questionnaires using the method of successive dichotomizations (R package “msd”).^27,28 Estimates of person and item measures from the Rasch analysis are on an equal interval scale, (i.e. a one logit difference represents the same magnitude change in latent trait at every point on the scale). In contrast, composite scores are, in general, not on an equal interval scale (i.e., the difference in composite scores between “N” and “N + 1” does not necessarily represent the same magnitude change in latent traits for all values of N). The LLQ scores were transformed using the recommended scoring system, which suggests how to map all LLQ responses to the same five-point Likert scale (e.g., 0–4).²⁹ A single Rasch analysis was performed after combining data from all time points to estimate a single set of item measures and separate sets of person measures over time. The Rasch analysis corrects for missing item responses, as well as accounts for differences in latent traits between different levels of response (therefore allowing for a more robust comparison between questionnaires). Using this analysis, calibrated item and person measures along with their standard errors were estimated for both ePROs. Rasch person measures were used as outcome variables in subsequent correlation and change from baseline analyses. 

Scatter plots were generated along with a line of best fit obtained from a linear regression analysis to visually demonstrate correlations between VFQ-25 and LLQ Rasch person measures, composite scores and age, wVA, wvMD, and wCS graphically. The association between PROs and clinical measures of visual function endpoints (wVA, wvMD, and wCS) was based on correlation analyses, at all assessed time points (PRO baseline, years 1/1.5, 2/2.5, and 3/3.5). Pairwise Pearson's correlation coefficients and associated P values were estimated to identify the association between the PRO assessments and wVA, wvMD, and wCS. Pearson's correlation coefficient values 1.0 to 0.6 or −1.0 to −0.6 were considered strong, values between 0.4 to 0.59 or −0.59 to −0.4 were considered moderate and <0.4 or <−0.4 were considered low correlations.²⁷ P < 0.05 was considered statistically significant. Only strong and statistically significant correlations were reported as such. 

The change from baseline over time for both PRO instruments (including Rasch person measures) was assessed using two methods: mixed-effects model for repeated measures (MMRM) and linear mixed (random and fixed effect) model. Least squares (LS) mean change from baseline for each PRO (composite and subscales) and 95% confidence intervals (CI) were estimated from MMRM with year and baseline score as fixed effects and subject as random effect. A compound symmetry structure was used to model the variance-covariance structure. A random-effect linear growth model was used to estimate the average yearly change of PRO (composite and subscale scores) as the slope and 95% CI of years from baseline. The linear mixed model included baseline composite PRO score (VFQ-25 and LLQ), years from baseline to each assessment, baseline age, and weighted visual function measures (wVA, wvMD, and wCS) at each assessment as fixed covariates and a random intercept. All analyses were completed using SAS version 9.3. 

Results

Demographics and Ocular Characteristics at PRO Baseline Visit

Forty-two of 44 RLBP1 RD patients enrolled in the natural history study, participated in the PRO sub-study, and completed at least two PRO follow-up visits at least 1 year apart. Specifically, 41, 37, and eight patients completed 1/1.5-, 2/2.5-, and 3/3.5-year PRO follow-up visits, respectively, in this sub-study (Table 1). 

At PRO sub-study baseline visit, all patients were Caucasian with a median age of 48.5 years (range 17–69 years); the majority of patients were females with an equal proportion of males to females in Canada and a higher proportion of females in Sweden. The most common mutations were homozygous p.R234W in 83% of the Swedish patients, and c.141G>A/c.141+2T>C in 53% of the Canadian patients. 

The mean (SD; range) composite score was 54.9 (18.93; 23–97) for VFQ-25 and 37.5 (14.66; 9–78) for LLQ; LLQ scores were generally lower than the VFQ-25 scores (Table 2). Most of the patients had clinical measures of visual function impairment, and the mean values of wVA, wvMD, and wCS were 1, −22.1, and 0.9, respectively (Table 2). 

Table 2.

 

Patient Demographics, PRO Scores, and Weighted Visual Function Measurements at PRO Sub-Study Baseline Visit

 

PRO Composite and Subscale Scores

During the three-year PRO sub-study, mean VFQ-25 and LLQ composite scores of the subscale scores were relatively stable or gradually declining (Fig. 1). Both the VFQ-25 and LLQ composite scores were similar between Swedish and Canadian patients and relatively stable over time, with the LLQ score consistently lower than the VFQ-25 score (Supplementary Table S1). A wide distribution of individual patient subscale scores was noted for both VFQ-25 and LLQ scores in Swedish and Canadian patients (Supplementary Fig. S1). 

Figure 1.

 

Mean (SD) scores over time (PRO baseline, years 1/1.5, 2/2.5, and 3/3.5 PRO sub-study visits) of all 42 RLBP1 RD patients are represented as bar graphs for composite* and subscale scores for VFQ-25 (A) and LLQ (B). N, number of patients per PRO sub-study visit. *Composite score of all PRO (VFQ-25 or LLQ) subscale scores.

Among the 12 VFQ-25 subscales, mean scores were lowest for peripheral vision (range 37.5–52.4 and gradually declining) and distance activities (range 39.2–49.0) in all patients. As expected, most of the patients had low driving scores (range 14.9–27.1) with only 19 patients evaluable for the driving subscale, consistent with low vision restrictions on driving. The scores for ocular pain, general health, and mental health subscales were relatively high (Fig. 1A). Among all six LLQ subscales, scores were lowest for peripheral vision (range 25.0–37.7 and gradually declining) and extreme lighting (range 28.1–36.9). In general, the driving scores were low (range 6.7–9.8) as most of the patients did not drive (Fig. 1B). 

Rasch Analysis

The Rasch analysis estimated both person and item measures along with corresponding standard errors. To determine how well the item difficulties target the person-level difficulties, histograms of the distribution for each measure were compared (Fig. 2). Overall, the distribution of the VFQ-25 and LLQ item measures well targeted the distribution of person function in this group, with only a small number of person measures beyond 1 logit of the nearest item measure. This suggests that the item difficulties do a good job targeting the person-level performance in this population. 

Figure 2.

 

Wright construct map for VFQ-25C^a,b (A) and LLQ^a (B). ^aPlots the distributions of estimated item and person measures for the VFQ-25/LLQ whereby the axis origin is set to the mean item measure (defined to be 0 logit). ^bHigher person measure has better performance. Higher item measure is an item with greater difficulty.

The item-measure standard errors show a crescent shape distribution, which indicates worse precision for easier and more difficult items (Supplementary Fig. S2). The person-measure standard errors show slightly worse precision for higher performing individuals (Supplementary Fig. S3). To assess uni-dimensionality of the questionnaires, the distribution of infit mean squares were computed. The percent of infits that lie between 0.5 and 1.5 are 77% for VFQ person measures and 69% for LLQ person measures, suggesting reasonable evidence of uni-dimensionality.^28,30 

Correlation Between VFQ-25 and LLQ

Mean LLQ subscale scores were generally lower than VFQ-25 subscale scores and <50 (mean score <41), except for the subscale of emotional distress (range 59.8–65.6; Figs. 1A, 1B). Scatter plot analysis revealed a strong positive correlation between VFQ-25 and LLQ composite scores in all visits of the PRO sub-study (Figs. 3A–D). The Pearson correlation coefficient was 0.81 (P < 0.001) at PRO baseline, 0.91 (P < 0.001) at 1/1.5 years, 0.81 (P < 0.001) at 2/2.5 years, and 0.87 (P = 0.002) at 3/3.5 years visits. There were strong positive correlations between VFQ-25 and LLQ composite scores and Rasch analysis person measures in all visits of the PRO sub-study (Figs. 3E–H). The Pearson correlation coefficient was 0.72 (P < 0.001) at PRO baseline, 0.86 (P < 0.001) at 1/1.5 years, 0.80 (P < 0.001) at 2/2.5 years, and 0.80 (P = 0.018) at 3/3.5 years visits. A strong correlation between Rasch person measures and composite scores between VFQ and LLQ were observed (Supplementary Fig. S4). 

Figure 3.

 

Scatter plot analysis revealed strong and statistically significant positive correlation between VFQ-25 and LLQ composite* scores during all PRO sub-study visits: (A) PRO baseline (N = 42); (B) Year 1/1.5 (N = 41); (C) Year 2/2.5 (N = 37); and (D) Year 3/3.5 (N = 8) and between VFQ-25 and LLQ Rasch analysis Person Measures scores during all PRO sub-study visits: (E) PRO baseline (N = 42); (F) Year 1/1.5 (N = 41); (G) Year 2/2.5 (N = 37); and (H) Year 3/3.5 (N = 8). *Composite score of all PRO (VFQ-25 or LLQ) subscale scores.

Association of PRO With Clinical Measures of Visual Function Parameters

Overall, strong (Pearson correlation coefficient ≥0.6 or ≤−0.6) and statistically significant correlations were observed between the VFQ-25 composite score and all three clinical measures of visual function at baseline (wVA [N = 42; P < 0.001], wvMD [N = 14; P < 0.05], and wCS [N = 35; P < 0.001]; Figs. 4A–C). Strong (Pearson correlation coefficient ≥0.6 or ≤−0.6) and statistically significant correlations were also observed between the VFQ-25 Rasch analysis person measures and all three clinical measures of visual function at baseline (wVA [N = 42; P < 0.001], wvMD [N = 14; P < 0.05], and wCS [N = 35; P < 0.001]; Figs. 4G–I). At other PRO sub-study visits, VFQ-25 composite score showed strong correlation with wVA and wCS, and mild to strong correlations with wvMD (all P < 0.05; Supplementary Table S2). The LLQ composite score showed a strong and statistically significant correlation with wVA (N = 42; P < 0.001), and mild to strong correlation with wvMD (N = 14; P < 0.05) and wCS (N = 35; P < 0.05) at baseline (Figs. 4D–F). Similarly, the LLQ Rasch person measures showed a strong and statistically significant correlation with wVA (N = 42; P < 0.001), and mild to strong correlations with wvMD (N = 14; P < 0.05) and wCS (N = 35; P < 0.001) at baseline (Figs. 4J–L). At other PRO sub-study visits, a strong correlation of LLQ composite score with wVA was observed and there was mild to strong correlation with wCS (Supplementary Table S2). Last, strong correlations were observed between Rasch person measures and composite scores for both VFQ and LLQ at each time-point (Supplementary Fig. 4). Patient age showed moderate correlation during the initial visits, but at later visits with fewer number of patients, the correlation was no longer significant with the composite scores for either of the PRO questionnaires (i.e., Pearson correlation coefficients not ≥0.6 or ≤−0.6 and P not < 0.05, Supplementary Table S2). RLBP1 RD patients displayed poor VA, high MD, and low CS, which mostly remained consistent during the three years of the PRO assessment period (Supplementary Fig. S5). 

Figure 4.

 

Correlation of PRO composite* scores with visual function measures: (A) VFQ-25 with wVA (N = 42); (B) VFQ-25 with wvMD (N = 14); (C) VFQ-25 with wCS (N = 35); (D) LLQ with wVA (N = 42); (E) LLQ with wvMD (N = 14); and (F) LLQ with wCS (N = 35). Correlation of PRO Rasch analysis Person Measures scores with visual function measures: (G) VFQ-25 with wVA (N = 42); (H) VFQ-25 with wvMD (N = 14); (I) VFQ-25 with wCS (N = 35); (J) LLQ with wVA (N = 42); (K) LLQ with wvMD (N = 14); and (L) LLQ with wCS (N = 35). Scatter plot analysis revealed strong and statistically significant correlation of VFQ-25 and LLQ composite* scores with weighted visual function measures (wVA, wvMD, and wCS) at the PRO baseline visit and with the Rasch analysis Person Measures. *Composite score of all PRO (VFQ-25 or LLQ) subscale scores.

Analysis of Change in PRO Scores Over Time

For the VFQ-25 subscales, least squares (LS)mean changes (95% [CI]), estimated by the MMRM model, showed statistically significant (P < 0.05) change for (a) social function subscale from baseline to 1/1.5 years (−6.41 [−11.45; −1.38]); (b) distance activities subscale from baseline to 2/2.5 years (−4.45 [−8.68; −0.21]); and (c) peripheral vision subscale from baseline to 2/2.5 years and 3/3.5 years (−6.67 [−13.12; −0.23] and −12.23 [−24.09; −0.38], respectively). The LS mean change from baseline for all other VFQ-25 scores and time points were not statistically significant (Table 3). For the LLQ subscales, LS mean changes (95% CI) showed significant change (P < 0.05) (a) from baseline to 2/2.5 years for the extreme lighting subscale (−4.27 [−7.81; −0.74]) (b) peripheral vision subscale (−6.56 [−11.63; −1.48]) and (c) Rasch person measures (−0.30 [−0.60; −0.009]). The LS mean change from baseline for all other LLQ subscale scores, and time points were not statistically significant (Table 3). Meaningful declines in peripheral vision subscale were noted with both PRO questionnaires. The average yearly decline of VFQ-25 and LLQ composite and subscale scores, as well as Rasch person measures, from baseline, estimated by linear mixed model, was not statistically significant (P > 0.05; Table 4). 

Table 3.

 

Mean Change From Baseline of VFQ-25 and LLQ Composite^* Scores and Subscale Scores Using MMRM Analysis

 

Table 4.

 

Average Annual Change of VFQ-25 and LLQ Composite^* Scores and Rasch Person Measures From Baseline (Linear Mixed-Model Analysis)

 

Discussion

To the best of our knowledge, this is the first study that concurrently evaluated the performance of two PRO questionnaires (VFQ-25 and LLQ) in patients with IRDs. Although LLQ subscale scores have been observed to relate moderately with VFQ-25 scores in patients with age-related maculopathy,²² they have not been tested together in IRDs. In this study, a strong, positive, statistically significant correlation was noted between the VFQ-25 and LLQ composite scores and for the Rasch analysis person measures. 

The results of this study demonstrated that PRO measures were overall low, reflecting the significant burden of the disease. Evidence suggests that increasing age correlated with a worse QoL in non-inherited eye disease of keratoconus; neovascular age‐related macular degeneration; retinal vein occlusion; and diabetic macular edema.²⁸ Given the wide age range (17–69) of patients with inherited retinal dystrophy in this study, patients may be at varied levels of disease severity, thereby causing wider variability in scoring. In addition, during later visits there were fewer patients available for evaluation. These could have led to no consistent significant correlation with age and composite scores for both VFQ-25 and LLQ. All patients evaluated in this study showed delayed dark adaptation symptoms from an early age, and this vision function attribute has a major contribution to the overall visual function and composite score of the questionnaires. 

Peripheral vision (VFQ-25 and LLQ), distance vision (VFQ-25), and extreme lighting at dark and bright light conditions (LLQ) were the PRO subscales that showed meaningful and statistically significant change between baseline and 2/2.5-year follow-up visit. Statistical analysis confirmed a clinically meaningful worsening of the peripheral vision domains of both VFQ-25 and LLQ at 2/2.5 years and 3/3.5 years from baseline. 

The LLQ captures aspects not covered by the VFQ-25 (responses to bright and dim lighting, and mobility) and was therefore deemed to be an appropriate complementary measure for RLBP1 RD patients in this study. In agreement with an earlier study conducted with the National Eye Institute VFQ-25,¹⁷ the scores for ocular pain, general health, and mental health subscales were relatively high in this study, indicating minimal effect of the disease on these parameters. Most LLQ subscale scores were low; and the scores were lowest for driving-related activities, indicating the substantial effect of visual field impairment on the ability to drive, similar to VFQ-25 driving subscale score, as the majority of these patients do not drive because of limited visual function. 

The QoL measures are multifactorial and dependent on different variables of clinical measures of visual function including visual acuity, visual field, and contrast sensitivity.²⁹ In this regard, this analysis showed a strong correlation of the VFQ-25 composite scores with all clinical measures of visual function (wVA, wvMD, and wCS), whereas LLQ composite scores were strongly correlated with wVA and wCS. Similar conclusions were achieved using more modern psychometric methods (i.e., Rasch analysis). Some previous studies in RLBP1 RP patients reported a similar observation on the positive association between visual function scores and VFQ-25 subscale scores supporting construct validity.^17,30 The use of two PRO questionnaires in association with clinical measures of visual function assessment at each PRO time point over a period of 3 years facilitated the correlation of PROs with clinical measures of visual function parameters. This allowed us to assess whether visual function change over time translated to differences observed in PROs. Based on data from this study, in addition to reports in literature,^17,31 PRO assessments are essential and should complement visual function and clinical examinations. 

A recent cross-sectional study used semistructured concept elicitation and cognitive debriefing interviews (qualitative interviews) in 21 RLBP1 RD patients in Sweden and Canada to understand the patient experience and evaluate the content validity of three existing PRO tools (VFQ-25, LLQ, and VAQ).² VAQ assesses light-dark adaptation questions that are not evaluated by VFQ-25 and LLQ. It was observed that the participants had difficulties interpreting item wording and response options for many of the items, even struggling to choose a response for some. There was a high overlap in the questions and PRO domains evaluated between the instruments, and item redundancy if all instruments were used together.^2,32 Notably, as discussed by Green et al.,² neither of these two questionnaires (VFQ-25 and LLQ) assess the light-dark adaptation, a key characteristic in RLBP1 RD patients. In this study using MMRM analysis, we observed meaningful declines in peripheral vision with both VFQ-25 and LLQ at 2/2.5 years. The results of this sub-study are consistent with the findings of the associated five-year natural history study, which also showed a slow progression of the disease in patients with RLBP1 RD. 

The mean changes in the composite scores and the Rasch analysis person measures for both VFQ-25 and LLQ were generally stable during the assessment period with any changes in scores observed reflecting the possible random fluctuations in the data or slight declines/improvements in PRO subscale score. No statistically significant change was noted in the mean annual change with linear mixed model in any of the VFQ-25 and LLQ subscale scores. 

Overall, this study provided more insights on the correlation of VFQ-25 and LLQ with each other and with clinical measures of visual function, with certain limitations to be acknowledged as follows: (i) the observational and descriptive nature of the study; (ii) the application of non-disease–specific PRO questionnaires; (iii) some of the patients exhibiting a floor effect with no possibility for further decline; (iv) although this study was not aimed to explore the genotype-phenotype correlation, phenotypic variabilities that are commonly seen with pathogenic variants of IRDs³² may have affected the measured PROs; (v) lack of complete follow-up PRO measures; (vi) less PRO measures in younger patients compared to the older patients; (vii) lack of disease progression analysis with the PRO measures; and (viii) small sample size especially at the last follow-up visit. Future studies evaluating PRO measures using VFQ-25 and LLQ over a longer time interval would help to understand the disease progression in younger and older patients and its impact on their QoL. 

Conclusions

A strong, positive, and statistically significant correlation was noted between VFQ-25 and LLQ composite scores and Rasch analysis person measures in RLBP1 RD patients in this 3-year observational study. The composite scores of both PRO questionnaires strongly correlated with clinical measures of visual function (VFQ-25 with wVA, wvMD and wCS and LLQ with wVA and wCS), as did the Rasch analysis person measures. At the 2-/2.5-year follow-up, meaningful statistically significant declines in peripheral vision (both VFQ-25 and LLQ), distance vision (VFQ-25), and extreme lighting in dark and bright light (LLQ) subscales were noted. Further research can lead to PRO instruments that comprehensively capture the appropriate QoL issues in patients with RLBP1 RD and inherited retinal dystrophies. 

Acknowledgments

The authors thank Jean Lecot (Novartis Pharma AG, Basel) for his support in conducting the Rasch Analysis. 

Supported by Biomedical Research. 

Medical writing and editorial support for this manuscript were provided by Nilesh Kumar Jain, PhD, and Shaswati Khan, PhD (Novartis Healthcare Pvt. Ltd., Hyderabad, India). Writing support was funded by Novartis Pharma in accordance with Good Publication Practice (GPP3) guidelines. 

Disclosure: J. Whelan, Novartis (F, R), Bayer (R), Alcon (R), Sentrex (R), Roche (C); J. Green, Novartis (C), Memorial University, Newfoundland, Canada (R); M. Burstedt, Novartis (F, R); E. Greco, Novartis (E), PHASTAR (E); X. Ni, Novartis (E), Sarepta Therapeutics (E); C. Spera, Novartis (E); A. Mullins, Takeda Pharmaceuticals (E), Novartis (E); J.-Y. Deslandes, Blue Companion (E), Novartis (E); Z. Su, Shanghai Luoxin Pharmaceutical Co. Ltd. (E), Novartis (E); M. Wald, Biogen (F), Novartis (E, F); C.L. Grosskreutz, Novartis (E); G. Normand, Novartis (E); A. Charil, Eisai Inc. (E), Novartis (E, I); D. Lu, None; K. Stasi, Saliogen Therapeutics (E); K. Holopigian, Novartis (E)