Driving with retinitis pigmentosa

ABSTRACT Background To establish the proportion of patients with retinitis pigmentosa (RP) meeting the Australian fitness to drive (FTD) visual standards. Methodology A prospective consecutive case series of patients with a clinical or genetic diagnosis of RP. Data on age at symptom onset, current driving status, inheritance pattern, better eye visual acuity (BEVA), binocular Esterman visual field (BEVF) parameters, genotype and ability to meet the driving standards based on BEVA and BEVF were collected. Outcome measures included the proportion of RP patients overall meeting the standards and clinical predictors for passing. A sub-analysis was performed on those RP patients who reported to drive. Change in BEVA and BEVF parameters across age in specific genotype groups was assessed. Results Overall, 228 patients with RP had a BEVF assessment. Only 39% (89/228) met the driving standards. Younger age at the time of testing was the only significant predictor (p < 0.01) for passing. Of the 55% of RP patients who reported to drive, 52% (65/125) met the standards, decreasing to 14% in the 56- to 65-year-old age group. RP patients harbouring mutations in HK1 or RHO genes may have slower rates of decline in their VF parameters. Conclusion Nearly 40% of RP patients met the driving standards. However, almost 50% of RP drivers were unaware of their failure to meet the current standards. BEVF testing is essential in the assessment of RP patients who are still driving. Phenotype and genotype predictors for passing the standards warrant further investigation. Abbreviation: FTD, fitness to drive; IRD, inherited retinal disease; RP, retinitis pigmentosa; RHO, rhodopsin; HK1, hexokinase 1; PRPF31 pre-mRNA processing factor 31; RPGR, retinitis pigmentosa GTPase regulator; VF, visual field; BEVA, better eye visual acuity; BEVF, binocular Esterman visual field.


Introduction
Retinitis pigmentosa (RP) is known to be caused by over 90 genes with autosomal dominant (AD), autosomal recessive (AR), and X-linked (XL) inheritance patterns (1). RP is the most prevalent form of inherited retinal disease (IRD), which accounts for the most common cause of blindness in workingage adults in both Australia and the United Kingdom (2,3). The ability to retain one's driver license is an important element of patient-reported outcome measures included in the National Eye Institute's Visual Function Questionnaire (VFQ-25).
In 1981, Fishman et al. (4) reported the frequency of car accidents in 42 RP patients with varying degrees of central and peripheral visual field (VF) loss. A higher proportion of accidents was reported in RP patients as compared to their agematched controls (4). Notably, 74% of RP patients in this study voluntarily restricted their driving to daylight hours (4).
Historically, peripheral VF constriction has been considered a driving hazard. However, in 1957 Danielson (5) proposed the quality of the central VF was more important in determining driving eligibility given drivers can adapt by head scanning regularly. Conversely, in 1967 Keeney et al. (6) found a horizontal VF extent of 140° was needed to drive safely. Similarly, Szlyk et al. found the extent of the horizontal VF to be the strongest predictor of car accidents (7,8). To date, unambiguous global visual standards necessary to drive are lacking when assessing individuals with complex and often progressive visual conditions, including RP. Consequently, there exists a preconceived notion that any patient with RP should not drive. In addition to visual function, other factors are crucial for driving including experience, adherence to speed limits, cognition, familiarity with surroundings and physical coordination. Xu et al. (9) assessed VF progression in 52 RP patients over a mean of 12 years. Using survival analysis, they extrapolated 50% of RP patients failed the Canadian visual driving standards by 37 years of age (9). However, this study only included kinetic VF data and did not assess for differences among genetic subtypes.
Given the paucity of current data on driving in those with a genetically confirmed diagnosis of RP, we sought to evaluate the proportion of an unselected RP cohort who pass the Australian fitness to drive (FTD) visual standards as measured by better eye visual acuity (BEVA), presence of central scotomas and horizontal VF extent as measured on the binocular Esterman visual field (BEVF). This work helps to raise awareness as well as to inform clinicians regarding their RP patient's FTD. Potential future genotype-phenotype correlations are also discussed.

Study design and population
This prospective study collected data at the Lions Eye Institute, Perth, Australia, from January 2013 to October 2022. The study adhered to the tenets of the Declaration of Helsinki and ethics approval was obtained from the Human Ethics Office of Research Enterprise, the University of Western Australia (RA/4/1/7916, RA/4/20/5454, RA/4/1/8932, and 2021/ET000151), and Sir Charles Gairdner Hospital Human Research Ethics Committee (RGS04985). Informed consent was obtained from all participants. Patients with a diagnosis of RP based on ultra-widefield (UWF) multimodal retinal imaging, electrophysiology confirmation of a rod > cone photoreceptor dysfunction and genetic confirmation of RPcausing variants were included. Clinical data included age at symptom onset, age at BEVF examination, sex, BEVA as measured on the Early Treatment of Diabetic Retinopathy Study (ETDRS) chart, inheritance pattern, phenotype (syndromic or isolated), genetic diagnosis, driving status and BEVF performance parameters.

Imaging procedures and outcome measures
UWF fundus colour and autofluorescence (FAF) images were obtained using the Optomap or California device (Optos PLC, Dunfermline, United Kingdom), elicited by a green excitation laser at 532 nm. BEVF testing was performed with the Humphry field analyser system (Carl Zeiss Dublin, California, United States of America) in a completely darkened room. A horizontal (150°) × vertical (100°) VF was examined using 120 test points arranged in the Esterman grid. Each location was tested once with a suprathreshold algorithm using a size III white stimulus at an intensity of 10 dB against a background luminance of 10 cd/m 2 . Missed points were retested, and a defect was recorded if the repeated test showed a negative response. Fixation monitoring was recorded. BEVF was predominantly performed whilst pupils were dilated (tropicamide 0.5% and phenylephrine 2.5%) and patients were instructed to fixate at the fixation target throughout the testing procedure (i.e. roving was not allowed). Testing was conducted after the patient was given 2-3 minutes to dark adapt. The normal physiological binocular VF extends horizontally from 160° to 200° with significant interindividual variability.

Fitness to drive visual standards
As per the Australian Fitness to Drive (FTD) 2022 guidelines (10), VA should be measured for each eye separately without correction and again with correction if needed. For a private vehicle license, a failure occurred if more than 2 letters were incorrectly read on the 20/40 (6/12) line with both eyes. A conditional private license may be issued if the patient's VA is equal to or better than 20/80 (6/24) if the patient is otherwise alert, has normal reaction times and good physical coordination. All patients with RP require a BEVF assessment given this is a progressive and constricting eye condition. For a private vehicle license, if the horizontal extent is less than 110° but greater than or equal to 90° a conditional license may be obtained. A single cluster of up to 3 adjoining missed points, unattached to any other area of defect, lying on or across the horizontal meridian can be disregarded when assessing the horizontal extent. Further, a vertical defect of only a single-point width but of any length, unattached to any other area of the defect, that touches or cuts the horizontal meridian may be disregarded. There should be no significant defect that encroaches within 20° of fixation above or below the horizontal midline. The horizontal field extent as measured from the last seen point to the next seen point taking into consideration scotoma that may be disregarded was recorded for each patient. A cluster of four or more adjoining points that is completely or partly within the central 20° radius is considered unacceptable. A cluster consisting of three missed points up to and including 20° from fixation and any additional missed points within the central 20° is also unacceptable. As is any central loss, this is an extension of a hemianopia or quadrantanopia and is greater than 3 missed points. Although patients with diplopia are not permitted to drive, none of the RP patients had this symptom. Fields with a false-positive rate of >20% were not considered reliable enough for determining eligibility.
In our study, the visual field span within 10° above and below the horizontal meridian and the number of missed points within the central 20° radius were recorded to determine a patient's eligibility to hold a conditional or unconditional license in accordance with the Australian FTD standards. Each patient was graded dichotomously as a pass (including both conditional and unconditional license holders) or fail by two graders (RCHJ and FKC) with a third grader (MSK) involved in arbitration if there was disagreement.

Genetic analysis
DNA was collected through the Australian Inherited Retinal Disease Registry and DNA Bank (11). Genomic DNA was analysed using genotyping microarrays, whole-gene sequencing, or disease-specific next-generation sequencing (NGS) panels as appropriate (12). Candidate mutations were confirmed in parents and other affected siblings by Sanger sequencing (Casey Eye Institute Molecular Diagnostics Laboratory, Portland, OR, USA, or Molecular Vision Laboratory, Hillsboro, OR, USA). Variant nomenclature was reported in accordance with the recommendations of the Human Genome Variation Society (12). Pathogenicity was assessed as described previously (13) and interpreted according to the joint guidelines of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) and associated literature (14)(15)(16).

Data and statistical analysis
Mean, standard deviation (SD), median, and interquartile range (IQR) in the form of quartiles 1 (Q1) and 3 (Q3) were reported for all continuous variables. Categorical variables were summarised by frequencies and proportions. Comparisons for age at symptom onset, age at BEVF testing, and BEVA (non-normally distributed) were assessed using the Mann-Whitney U test. The distribution of categorical variables across the passed and failed BEVF groups was examined using the Chi-square test. Logistic regression was utilised to determine whether age at symptom onset, age at BEVF testing, sex, inheritance, and genotype were predictive of FTD. For these models, inheritance was restricted to simplex, AR, and AD, and for genotype, only RHO, PRPF31, and USH2A were included. Others were excluded due to low counts given grouping would be inappropriate. Model estimates were reported as an odds ratio (OR) and respective 95% confidence intervals (CI). A p-value of < .05 was considered statistically significant. The relationship between BEVA and BEVF parameters with age for AD, AR, and XL inheritance was explored and illustrated using scatter plots. General linear modelling was used to associate BEVA, BEVF efficiency, horizontal span, and scotoma in the central 20° with the patient's age at examination for each of the relevant genotypes, and the corresponding slope estimates (95% confidence interval) and their significance were reported. The expected ages at which BEVA reached 20/80 (6/24) and 20/200 (6/60) were estimated for models where p< .001.

Fitness to drive amongst the RP cohort overall
Based on BEVA and BEVF criteria, 39% of our RP cohort (89/ 228) met the Australian visual requirements for FTD. After excluding those <16 years of age, 36% (77/211) met these requirements. The proportion of RP patients who passed the FTD visual requirements decreased from 55% in those aged between 16 and 25 years to 14% in those aged 56-65 years. This proportion increased to 41% in those aged 66 years or older (Figure 1(a), Table S2). A logistic regression model (Table S3) examining age at symptom onset, age at BEVF testing, sex, phenotype, inheritance pattern, and genotype found the main predictors for RP patients to meet the FTD visual requirements were older age at symptom onset, (p = .030), younger age at BEVF testing (p = .004), and presence of a rhodopsin (RHO) mutation (p = .031).

Fitness to drive in current RP drivers
Of those RP patients who underwent BEVF testing, 55% (125/228) reported to still drive at the time of their examination. Of these, only 52% (65/125) met the FTD visual standards for a conditional or unconditional private vehicle license. The proportion of RP drivers who met the FTD requirements decreased from 82% in those aged between 16 and 25 years to 29% in those 56-65 years. This proportion increased to 73% in those aged 66 years or older (Figure 1(b), Table S4). A logistic regression model ( Table 2) examining age at symptom onset, age at BEVF testing, sex, phenotype, inheritance pattern, and genotype found the main predictors for an RP driver to meet the FTD visual standards were older age at onset (p = .038) and younger age at BEVF testing (p = .017). The mean (SD) age at symptom onset for those RP patients who failed and passed the FTD visual standards was 28.8 (15.3) and 32 (13.1) years, respectively. Conversely, the mean (SD) age at BEVF testing for those RP patients who failed and passed the FTD visual standards was 45.8 (9.4) and 34.9 (16.6) years, respectively.
The main reasons for RP drivers failing their FTD visual standards (n = 60) were horizontal VF span <90° degrees in 49 patients (82%) and significant scotomas within the central 20° radius in 50 patients (83%) of which 39 failed both criteria (65% of RP drivers). In our cohort of RP drivers, 76 patients (60.8%) had a BEVF span of 90° or more. The mean (range) BEVA was 20/25 (20/63-20/12); thus, no RP patients failed the FTD visual acuity requirement. The VF span was < 20° (defined as legal blindness in Australia) in 8 of 125 RP patients (6.5%) who reported driving at the time of their BEVF.

Correlation between genotype and visual field parameters
We examined changes in BEVA and BEVF parameters across age in common AD, AR, and XL genes associated with RP (Table S5). For AD RP, patients harbouring mutations in HK1 and RHO, increasing age had less impact on visual function (Figure 2(a-d)). In contrast, those with PRPF31 and RP1 had a significant decline in BEVA, with age reaching 20/80 by 44 and 55 years, respectively, and 20/200 by 62 and 69 years, respectively. For AR RP, HGSNAT patients showed relative preservation of BEVA and BEVF efficiency as compared to those AR RP patients with USH2A and EYS mutations (Figure 2 (e-h)). Syndromic RP patients with USH2A mutations had BEVA declining to 20/80 and 20/200 by 46 and 66 years,  respectively. This was at a younger age than those with non-syndromic USH2A RP at 55 and 70 years, respectively. Female XL RP carriers (excluded in our analysis) generally had well-preserved BEVA and BEVF parameters compared to affected males (Figure 2(i-l)). Cases of RPGR were predicted to have their BEVA declining to 20/80 and 20/200 by 32 and 45 years, respectively. Figure 3 illustrates three AD RP patients of varying age with pathogenic variants in RHO. A male in his late teens with a mild phenotype had 97% BEVF efficiency that met the FTD visual standard (Figure 3(a-c)). A male in his late 40s with a sector RP phenotype had a horizontal VF span of 100° with A male in his late teens with temporal hypoautofluorescence and 97% BEVF efficiency meeting the fitness to drive (FTD) visual standard (Figure 3(a-c)). A male in his late 40s with sector RP did not meet the FTD standards despite a horizontal VF span of 100°, due to a cluster of four or more missed points superiorly extending to within the central 20° (Figure 3(d-f)). A male in his late 60s with advanced sector RP failed to meet the FTD visual standard despite sufficient VF span, due to a cluster of four or more missed points within the central 20° (Figure 3(g-i)).  (Figure 4(a-c)). A female in her late 40s with a pericentral phenotype had a 95° horizontal VF span and no scotoma within the central 20° met the FTD visual standard of a conditional licence (Figure 4(d-f)). A male in his 70s had a 20° VF span with 19/24 missed points within the central 20° did not meet the FTD visual standard (Figure 4(g-i)). a cluster of four or more missed points superiorly extending to within the central 20° and thus did not meet the FTD visual standard (Figure 3(d-f)). A male in his late 60s with an advanced sector RP phenotype also had sufficient VF span but failed FTD standard based on a cluster of four or more missed points within the central 20° (Figure 3(g-i)). Figure 4 illustrates three AR RP patients with pathogenic variants in USH2A who were currently driving at the time of testing. A male in his 20s with a mild RP phenotype met the FTD visual standard (Figure 4(a-c)). A female in her late 40s with pericentral phenotype had a 95° horizontal VF span and no scotoma within central 20° and thus met the requirements for a conditional private license (Figure 4(d-f)). A male in his 70s had a 20° VF span with 19/24 missed points within the central 20° (Figure 4(g-i)). Figure 5 illustrates three patients with pathogenic variants in CHM. A male in his late 20s with a 150° horizontal VF span and no scotoma in the central 20° met the FTD visual standard (Figure 5(a-c)). A male in his mid-20s with a horizontal VF span of only 70° and a cluster of four or more missed adjoining points within the central 20°.

Phenotype and genotype correlations
His VF did not satisfy the FTD visual standard for a commercial or private vehicle licence ( Figure 5(d-f)).
A CHM carrier female in her 80s satisfied the FTD visual standard ( Figure 5(g-i)).

Discussion
To the best of our knowledge, this is the first report on the proportion of a large cohort of Australian RP patients passing the FTD visual standards. Furthermore, we evaluated the proportion of current drivers with RP who satisfied the FTD visual standards based on BEVA and BEVF. Overall, 39% of our RP cohort (89/228) met the Australian visual requirements for FTD. Of those RP patients who reported to be driving, only 52% (65/125) met the requirements for a conditional or unconditional private vehicle license. The proportion of RP patients who met these requirements decreased with increasing age. A notable exception to this was observed with an increase in those RP patients passing their FTD requirements in those aged > 66 years of age, both overall and in those RP patients who reported to still drive. This could be attributed to a small sample size within this older age group, which artificially skewed the results or a bias towards avoiding BEVF testing among this older age group (Table 1). Our data demonstrated that the strongest predictors for RP patients to meet the FTD visual standard were younger age at BEVF testing or the presence of a RHO mutation, although there was a non-significant trend in those with an older age at symptom onset. Xu et al. (9) retrospectively assessed 275 Goldmann VFs over a mean period of 12 years (range 2-29) from 52 RP patients. They found that the average rates of VF loss were less for AD (2.7%) as compared to AR (10.3%) inheritance although this trend did not reach statistical significance given there were a limited number of patients with AD inheritance (9). Given that RHO is a well-known causal gene for AD RP, this is consistent with our findings that those RP patients harbouring mutations in RHO were more likely to meet the FTD visual requirements (1). Unlike our study, however, Xu et al. (9) did not analyse individual genotypes within their RP cohort and did not discuss how they grouped RP patients into different inheritance patterns. Sandberg et al. (17) also compared the VF decline in AD RP due to RHO mutations to those with AR RP attributed to USH2A and XL RP secondary to RPGR mutations. They found the rate of VF decline to be 2.6% in RHO as compared to 7.0% in USH2A and Figure 5. Ultra-widefield pseudocolor and fundus autofluorescence images and binocular Esterman visual field (BEVF) of three patients with pathogenic variants in CHM. A male in his late 20s with a 150° horizontal VF span and no scotoma in the central 20° met the fitness to drive (FTD) visual standard ( Figure 5(a-c)). A male in his 20s with a horizontal VF span of only 70° and a cluster of four or more adjoining missed points within the central 20° did not satisfy the FTD visual standards for a commercial or a private vehicle licence ( Figure 5(d-f)). A CHM carrier female in her 80s is also shown for comparison ( Figure 5(g-i)).

4.7% in RPGR.
Given the vast heterogeneity of genotypes implicated in RP, our logistic regression model was limited in its inclusion of three genotype groups with sufficient numbers namely, RHO, PRPF31, and USH2A.
Limited literature, both within Australia and internationally, has been published regarding the proportion of RP drivers passing the FTD standards. As far back as 1981 Fishman et al. (4) suggested RP patients were involved in more car accidents than their age-matched controls. A decade later Szlyk et al. (8,9) found that the extent of the horizontal VF was the strongest predictor of car accidents. Since then, however, there has been a paucity of published data regarding RP patients and their ability to meet FTD standards. There is a need to increase awareness among ophthalmic clinicians, clinical geneticists, genetic counsellors, and visual rehabilitation providers, involved in the care of RP patients, to consider whether their individual RP patients are fit to drive. We observed that many of our RP patients had never had prior BEVF assessment despite years of RP-related visual symptoms and active driving history. Given the vast phenotypic spectrum, which encompasses RP, these patients should each be considered individually regarding their FTD. It is often difficult for clinicians to initiate conversations with their RP patients, which may prompt a discussion of their cessation of driving. Xu et al. (9) reported the median age of survival for RP patients to meet the Canadian driving standards was 37 years. Although this survival analysis was based on Goldmann perimetry alone, we found a similar pattern where RP patients aged 26-35 years had a 50% chance of meeting the FTD visual standards using BEVF. In addition, Xu et al. (9) reported a statistically significant (p < .001) more rapid decline for smaller targets (I4e) as compared to larger targets (V4e). This highlights the need for clinicians to perform early and repeated BEVF assessments given RP patients may lose their ability to see smaller targets earlier on. We also found that VA was not responsible for RP drivers failing the FTD standards. Using regression modelling, we estimated that the age at which BEVA would drop to below 20/80, the minimum requirement for a conditional licence, was 40-55 years for RP genotypes PRPF31 and USH2A, nearly two decades after RP drivers had failed to meet the VF FTD requirement. The only exception was RPGR, which progressed to 20/80 by 32 years, but none of our RPGR cohort (age 18-61 years) reported to drive.
In this study, we found that the horizontal VF span was < 20°, meeting the requirements for legal blindness, in 6.5% of RP patients who reported to be driving at the time of testing. Given that some RP patients may not be willing to disclose their true driving status, this is likely an underestimate. The median age of survival for legal blindness in studies for RP patients with a specific genetic subtype namely, RHO, RPGR, USH2A, or CRB1 varies from 44 to 77 years (17)(18)(19)(20). None of these studies, however, assessed the current driving status and the proportion of their RP patients able to meet the FTD visual standards. Interestingly, Xu et al. (9) reported a significant effect of sex on the survival distribution for legal blindness, where male sex had a hazard ratio of 3 when compared to female sex (p = .022). This effect, however, was not significant with respect to the driving standard. Similarly, we observed that sex was not a significant predictor for meeting the FTD visual standard. Given nearly 50% of RP drivers did not realize they failed the FTD visual standard, it is perhaps appropriate that the Michigan Retinal Degeneration Questionnaire, unlike the VFQ-25, does not include a question regarding the patient's own perception of their driving ability (21).
Our study has several limitations. Although each RP patient was assessed with the same Humphry visual field analyser device, different ophthalmic technicians were responsible for administering the BEVF test with various levels of experience and possibly providing variable levels of encouragement and instruction. In the future, it would be useful to conduct repeat VF testing across different days to assess for inter-visit and inter-examiner variability. This is unlikely to have been a major issue, however, given the median false-positive error rate was only 4%. Secondly, we relied upon RP patient's selfreporting their driving status. It is indeed possible some patients may not have divulged their true circumstances for fear of losing their licence. Given that the ability to drive in Western society confers a level of autonomy and independence, it is understandable that many RP patients find this topic stressful and may have been weary of performing VF testing voluntarily.
Thirdly, BEVFs were only tested at one point in time, and further data collection over the coming years would be required to determine individual BEVF rates of progression in this RP cohort. A future study with a larger sample size, ideally including RP patients from multiple IRD clinics, would be helpful in further elucidating any statistically significant differences in RP patients' likelihood of passing their FTD requirements across different genetic subtypes and inheritance patterns. Fourthly, our patients were dilated during their BEVF testing. Future studies are required to determine the impact of pupil dilatation on suprathreshold VF testing. Finally, the Australian FTD standard allows for roving BEVF where fixation is not monitored. Some patients with a borderline VF may have been able to pass if allowed to rove during the procedure. Further studies are required to determine the reclassification rate from failure to pass if RP patients were allowed to rove during testing.
We prospectively evaluated the proportion of an unselected RP cohort who passed the driving standards as based on BEVA and BEVF. Despite misperceptions within the medical community and general public that RP patients should not drive, we found nearly 40% of our RP cohort met the current standard. Of those RP patients who reported to still drive, just over 50% met the requirements for a conditional or unconditional private vehicle license. Clinicians should assess all their RP patient's FTD with annual Estermann visual fields. RP patients with a visual acuity of less than 20/80, horizontal extent less than 90°, or a significant defect that encroaches within 20° fixation should be advised they do not meet the driving standard.

Disclosure statement
The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this article.

Funding
The author(s) reported that there is no funding associated with the work featured in this article.

Brief summary statement
To investigate the proportion of patients with retinitis pigmentosa meeting the Australian driving standards. In those who reported to drive, predictors for meeting the fitness to drive standards were explored.