Ophthalmic findings in the KIF1A-associated neurological disorder (KAND)

Introduction

KIF1A-associated neurological disorder (KAND) is a recently identified genetic disorder caused by pathogenic variants in the KIF1A gene. KIF1A encodes a member of the kinesin-3 protein family, and functions in anterograde transport of cargo along microtubules in an ATP-dependent fashion. The phenotypic spectrum in KAND patients is variable, but stems from the degeneration of neuronal cells, and can result in spasticity, neurodevelopmental delay and regression, intellectual disability, autism, microcephaly, progressive spastic paraplegia, seizures, cerebellar atrophy and peripheral neuropathy.

Prior to the umbrella term KAND, which encompasses all syndromes caused by pathogenic variants in KIF1A, this syndromic entity was referred to by many other names, reflecting the heterogeneous manifestations of this entity. Previous nomenclature for the syndrome included hereditary spastic paraplegia, progressive encephalopathy with edema, hypsarrhythmia and optic atrophy (PEHO) syndrome, spastic paraplegia type 30, ataxia telangiectasia type disorder and hereditary sensory and autonomic neuropathy type III, among others (^1–6, Supplemental Table 1 and references therein). Some of these previous studies documented the presence of optic nerve atrophy in study participants with pathogenic variants in KIF1A; however, many other studies did not assess ocular features. Of the studies that did look at ocular findings, the majority of the data were based on review of clinical charts and patient report, and did not include standardized, quantitative ophthalmologic assessment (Supplemental Table 1). Notably absent from these studies are descriptions of the degree of visual disability experienced by patients with KAND, how variable the disability may be with regards to age, genotype correlation with the visual phenotype, and essential elements of the ophthalmic examination including visual acuity, ocular motility, and the severity of optic atrophy. The lack of information regarding the expected ophthalmic findings and outcomes in individuals with KAND thus makes it difficult for ophthalmologists to properly counsel these patients with regards to their ophthalmic health and prognosis.

In a study by Lee et al. ⁷, optic nerve atrophy was reported in 10/14 individuals based on retrospective chart review; however, this was corroborated poorly by magnetic resonance imaging (MRI), with only 4/14 individuals showing signs of optic nerve atrophy on MRI. Nemani et al. reported optic atrophy in 7/12 participants, also based on retrospective chart review ⁸. A study by Boyle et al. based on patient telephone interview of 100 participants reported 50% of participants with optic nerve atrophy, 26% with strabismus, and 8% with cataracts ⁹. A number of other studies with smaller groups of participants report variable prevalence of optic nerve atrophy in KAND participants ^10–13. In the ophthalmic literature, only one case report of the ophthalmic findings in a KAND patient has thus far been reported ¹⁴. The case reported a five-month-old boy, who was documented to have poor eye contact, optic nerve hypoplasia and atrophy on dilated fundus examination and abnormal response on flash visual evoked potentials.

The KIF1A Outcome measures, Assessments, Longitudinal And endpoints (KOALA) Study is a longitudinal study that combines coordinated standardized clinical assessments of KAND participants, including assessment of motor function, neurological examination, neurocognitive and neuropsychological status, and ophthalmic examination ¹⁵. Given the relatively large number of participants enrolled in this study, a more complete picture of the ophthalmic manifestations of KAND is being documented, with the aim of understanding what to expect in terms of ophthalmic findings, prognostication and genotype-phenotype correlation with respect to severity of the ophthalmic presentation. In this study, we report comprehensive ophthalmic examination findings in 24 participants with genetically confirmed KAND using clinical evaluation, fundus photography, optical coherence tomography and in select cases, visual evoked potentials.

Methods

KOALA participants were recruited for in-person assessments through KIF1A.org and our ongoing online natural history study. The study was cross-sectional in nature. Approval for the study was provided by the Columbia University institutional review board. Guardians or independent adults provided informed consent. Minors who were capable provided assent. All participants had a pathogenic/likely pathogenic variant in KIF1A on genetic testing (exome/genome sequencing or panel gene sequencing) in a clinical diagnostic laboratory. All laboratory results were reviewed by a certified genetic counselor.

Clinical examination elements

A standard ophthalmic examination for participants enrolled in this study consisted of best corrected visual acuity, intraocular pressure measurement with the iCare TA01i tonometer or by palpation of the globes, cycloplegic autorefraction or retinoscopy, color testing with Hardy-Rand-Rittler (HRR) pseudoisochromatic screening color plates, visual field testing to confrontation, extraocular motility examination, stereopsis testing with the Stereo Optical Stereo Fly Test booklet, slit lamp biomicroscopy and dilated fundoscopy. In participants able to report visual acuity, measurements were collected in Snellen acuity, and converted to the log of the minimum angle of resolution (logMAR) for statistical analysis. “Fix and follow” vision was noted for preverbal children who demonstrated vision through the ability to fix and follow a penlight or object of interest. Allen optotypes were used for participants who were unable to recognize alphanumeric characters. Humphrey automated perimeter visual fields were attempted when it was deemed that participants had the capacity. Not all participants were able to cooperate with all sections of the standard examination, owing to neurocognitive ability or age limitations. Optical coherence tomography imaging was obtained using swept-source OCT on the Cirrus 6000 device (Carl Zeiss Meditec, Inc., Dublin, CA, USA). Each OCT image was reviewed and automated segmentations were reviewed. For patients with OCT data, optic nerve atrophy was determined based on the normative database provided by the OCT machine software. The average thicknesses for RNFL and GCL-IPL in the normative databases are 92.9 μm (standard deviation 9.4 μm) and 84.7 μm (standard deviation 7.1 μm), respectively.

Visual evoked potentials

Pattern-reversal visual evoked potentials were performed as per published ISCEV protocols ¹⁶. Briefly, electrodes were placed on the participants’ head in predetermined locations dictated by the ISCEV protocol; signals were then measured as the patient focused on a pattern-reversal grid of 1° checkered boxes, and 0.25° checkered boxes.

Vineland Adaptive Behavior Scales 3^rd Edition and associated statistical analysis

The use of the Vineland Adaptive Behavior Scales (VABS) is described in more detail in Sudnawa et al. ¹⁵ Briefly, individuals’ adaptive function using the Vineland Adaptive Behavior Scales 3rd Edition comprehensive interview form (VABS) were administered by phone interview. Adaptive Behavior Composite (ABC) standard score, communication standard score, daily living skills standard score, and socialization standard score were used for analysis. Correlation of these scores with quantitative ophthalmic measurements were analyzed using IBM SPSS (version 28). Pearson’s correlation coefficient was used to assess the relationship between continuous variables. A p-value ≤ 0.05 was considered as significant.

Results

KAND ophthalmic cohort

The cohort for this ophthalmic assessment study included 24 participants who were able to tolerate an ophthalmic examination and/or were able to report on subjective visual acuity and follow instructions for ophthalmic imaging. The average overall age of the cohort was 14.0 years. Of these participants, twenty were participants under 18 years of age with an average age of 9.1 years, and four were adults with an average age of 38.3 years. The participant characteristics are provided in Table 1. Age, neurocognitive ability and/or attention prevented some participants from participating in some aspects of the examination. This limitation is noted where relevant in this text, and in detail in Supplemental Table 2. Examination findings are summarized in Table 2.

TABLE 1.

Demographics of KOALA Participants Undergoing Ophthalmic Examination (N = 24)

Variable	Value
Age
Overall, y, mean (range)	14.0 (2-59)
Age group, n (%); mean age
<18 y	20 (83); 9.1 y
≥18 y	4 (17); 38.3 y
Gender, n (%)
Male	15 (62.5)
Female	9 (37.5)
Race/ethnicity, n (%)
European	18 (75)
African American	1 (4)
Asian	3 (13)
Hispanic/Latino	0 (0)
Mixed	2 (8)

KOALA = KIF1A Outcome measures, Assessments, Longitudinal And endpoints Study.

TABLE 2.

Ophthalmic Clinical Findings in KIF1A Participants

Clinical Component	Fraction of Patients Tested, n/n (%)	Major Clinical Findings
Average subjective visual aculty, logMAR (Snellen)
All participants	18/24 (75)	0.428 (20/54)
<18 y	14/20 (70)	0.329 (20/43)
>18 y	4/4 (100)	0.773 (20/119)
Refractive error, SphEq, D		Astigmatism >3 D cyl 4/24 patients total (16%)
<18 y	14/20 (70)	+1.03 D
<18 y	4/4 (100)	−5.32 D
Color vision (HRR color plate testing), all participants	12/24 (50)	11/12 (92%) with color deficits: average HRR score 2.5/6
Stereo vision (Titmus viewer), all participants	10/24 (42)	Average stereopsis 918″
History of strabismus, all participants	24/24 (100)	9/24 (38%) with strabismus or history of strabismus
Cataract, all participants	3/24 (13)	1 posterior polar, bilateral 1 traumatic, unilateral 1 age-related, bilateral
RNFL thickness on OCT, μm, mean (range)
All participants (31 eyes total)	16/24 (67)	61.9 (44-102)
<18 y (26 eyes)	13/20 (65)	63.7 (44-102)
>18 y (5 eyes)	3/4 (75)	54.5 (46-63)
GCL-IPL thickness on OCT, μm, mean (range)
All participants (33 eyes total)	17/24 (71)	58.2 (46-76)
<18 y (26 eyes)	13/20 (65)	60.7 (48-76)
>18 y (6 eyes)	4/4 (100)	50.3 (46-56)

cyl = cylinder, D = diopters, GCL-IPL = ganglion cell layer–inner plexiform layer, HRR = Hardy-Rand-Rittler, OCT = optical coherence tomography, RNFL = retinal nerve fiber layer, SphEq = spherical equivalent.

Ophthalmic clinical assessment

Visual acuity and refraction

Participant-reported best corrected visual acuity (BCVA) with Snellen or Allen optotypes was obtainable in 18 participants, all over the age of four. The overall average visual acuity measured was logMAR 0.428 (Snellen 20/54). The average visual acuity in the pediatric sub-cohort (28 eyes of 14 participants) was logMAR 0.329 (range 0.0-1.0), corresponding to an average Snellen acuity of 20/43, whereas the average visual acuity in adult participants was logMAR 0.773 (range 0.471-1.351), corresponding to an average Snellen acuity of 20/119. The difference in visual acuity between the pediatric and adult sub-cohorts was statistically significant using analysis of variance, with a p-value of 0.02. Despite this statistical significance between the two groups, there was only poor correlation between age and best corrected visual acuity with R² = 0.0664. Of the remaining six participants for whom a patient-reported BCVA was not possible, three had reliable fix-and-follow visual behavior indicative of good vision, all preverbal children. Of the remaining participants, one 4 year-old child was able to fix but unable to follow; one 8-year-old was unable to cooperate with the examination due to attention deficit; and one 16 year-old child was neither able to fix nor follow a penlight due to both visual and cortical issues, and it was not possible to gauge her visual acuity. Visual acuity is reported in detail for each participant in Supplemental Table 2.

The average spherical equivalent refractive error was obtainable in 14/20 pediatric participants and was slightly hyperopic at +1.03 D (range −3.00 D to +6.75 D). In the adult participants, the average spherical equivalent in 4/4 participants was −5.32 D (range −7.00 D to −3.25 D). Three of the four adults had significant astigmatism (one with −4.00 D cyl OU, and two with −2.75 D cyl in at least one eye), presumably from some component of corneal or lenticular astigmatism, skewing the spherical equivalent refractive error in the adult sub-cohort toward a more negative value. Fourteen percent of participants overall had high astigmatism (> 3 D cyl) (Table 2).

Color vision, stereopsis and visual fields

Color vision was tested using Hardy-Rand-Rittler (HRR) pseudoisochromatic screening color plates and was obtainable on 12/24 participants. Participants who were unable to be tested were too young, could not cooperate with this testing or had central visual acuity that was too poor for testing. Of the 12 participants tested, 11 (91.6% of those tested) had color deficits, with one able to identify all color plates correctly. The average score for participants who were able to participate in testing was 2.5/6 (as per scoring system for the HRR screening plates; range 0-6). Visual fields to confrontation were also tested; twelve participants were able to participate in this testing. Of these participants, 11 (91.6% of those tested) had visual fields that were constricted. Only two participants were able to undergo formal visual field testing with the Humphrey automated perimeter; both participants had generalized decrease in sensitivity throughout their field. Ten participants were able to undergo stereo testing using the Stereo Optical Stereo Fly Test booklet. The average stereopsis within the group was 918 seconds of arc (range 40-3540). Of these 10 participants, four had a history of strabismus, and two had a history of bilateral refractive amblyopia. Further information regarding the nature of the strabismus is provided in Supplemental Table 2.

Other elements of the clinical examination

Nine of 24 participants (38%) had a history of strabismus treated with patching, prisms or surgery. Of these nine participants, one patient had a fourth nerve palsy, one had esotropia and seven had intermittent exotropia. Three participants had cataract on slit lamp examination; of these participants, one demonstrated bilateral posterior polar cataract, one had a traumatic cataract owing to falls secondary to spastic paraplegia, and one had bilateral age-related cataract. All participants had intraocular pressure within normal range using iCare tonometry or by palpation of the globes. Dilated examination revealed no notable clinical retinal changes in any of the participants. None of the adult patients with myopia exhibited findings consistent with pathologic myopia, such as lacquer cracks, Fuchs spots or staphyloma.

Optical coherence tomography (OCT) studies

The status of the optic nerve was further defined using optical coherence tomography (OCT), given the superior sensitivity of this modality of imaging in detecting subtle deficits in the retinal nerve fiber layer (RNFL) and ganglion cell layer (GCL). For RNFL thickness, 31 eyes of 16 participants were imaged. For participants who had both eyes imaged (15/16 participants), the average of the two eyes was used for graphing and statistical analysis, given that the values between the two eyes were largely symmetric. The average RNFL thickness in this cohort overall was 61.9 μm. For the pediatric participants within this subgroup, the average RNFL thickness was 63.7 μm, whereas the average thickness for adults was 54.5 μm (Table 2; Figure 2). RNFL thinning appeared to be largely confined to the superior, inferior and temporal quadrants with sparing of the nasal RNFL (Figure 2). The ganglion cell layer-inner plexiform layer (GCL-IPL) thickness was also imaged in 33 eyes of 17 participants. Fifteen of 17 participants were able to have both eyes imaged, and the average of the two eyes was used for graphing and statistical analysis. The overall average thickness of the GCL-IPL was 58.2 μm, with a pediatric average of 60.7 μm and an adult average of 50.3 μm (Table 2; Figure 3). The pattern of thinning for both RNFL and GCL-IPL appeared to be generalized (Figure 3). Individual RNFL and GCL-IPL thickness values along with patient genotype are presented in Table 3. There were otherwise no structural abnormalities noted within any of the other retinal layers on macular OCT.

FIGURE 2.

A. Retinal nerve fiber layer (RNFL) thickness (average of 2 eyes in each participant) plotted against the age of the participant. B. Output of parameters of RNFL scan for one of the adult participants showing areas of thinning using the Cirrus 6000 OCT device (Carl Zeiss Meditec, Inc).

FIGURE 3.

A. Ganglion cell layer–inner plexiform layer (GCL-IPL) thickness (average of 2 eyes in each participant) plotted against the age of the participant. B. Output of parameters of macular GCL-IPL scan for one of the adult participants showing areas of thinning using the Cirrus 6000 OCT device (Carl Zeiss Meditec. Inc).

TABLE 3.

Participant Genotypes and Corresponding RNFL and GCL-IPL Thicknesses on OCT.

Genotype	Age, y	RNFL Thickness, μM		Average RNFL, μM	GCL-IPL, μM		Average GCL, μM
		OD	OS		OD	OS
p.Gly102Arg	6	85	80	82.5	67	67	67
p.Glu267Gln	7	77	66	71.5	65	64	64.5
p.Thr258Met	7	108	102	105	76	76	76
p.Pro305Leu	8	60	58	59	57	60	58.5
p.Arg254Pro	9	54	52	53	56	55	55.5
p.Arg11Gln	10	57	55	56	59	60	59.5
p.Asn211His	11	66	63	64.5	53	54	53.5
p.Trp313Ser	11	45	47	46	55	55	55
p.Arg316Trp	13	42	46	44	48	48	48
p.His171Pro	13	72	63	67.5	69	68	68.5
p.Gly117Val	14	60	54	57	57	56	56.5
p.Arg254Gln	14	57	58	57.5	64	65	64.5
p.Arg13His	17	66	62	64	61	62	61.5
p.Lys103Thr	23	Unable	Unable	Unable	51	Unable	51
p.Arg316Trp	30	55	Unable	55	49	Unable	49
p.Arg216His	41	45	47	46	44	47	45.5
p.Arg254Gln	59	63	62	62.5	57	54	55.5

GCL-IPL = ganglion cell layer–inner plexiform layer, RNFL = retinal nerve fiber layer. Three-letter code for amino acid abbreviations were used.

For patients who were unable to cooperate for an OCT scan (7 participants total), clinical examination of the optic nerve was performed, and optic nerve atrophy was assessed based on optic nerve pallor. No cupping was observed in any patient. An example of the nerve appearance in a KAND participant is shown in Supplemental Figure 1.

Visual evoked potentials

Visual evoked potentials (VEPs) were performed on three participants with varying degrees of disease severity (mild, moderate and severe). The first participant was a seven-year-old boy with no discernible cognitive impairment on psychometric testing, but with clinically significant hereditary spastic paraplegia. His optic nerves appeared healthy on clinical examination. OCT revealed RNFL thickness in the right eye of 108 μm, and in the left eye of 102 μm, whereas GCL-IPL thickness was 76 μm in both eyes, consistent with mild to no structural damage. Pattern reversal VEP showed a normal waveform with a P100 peak at approximately 100 μm with 1° grid pattern; however, this peak shifted to approximately 130 ms with the 0.25° grid pattern, revealing subtle deficits consistent with optic neuropathy (Figure 4A). His genetic testing revealed a p.Thr258Met missense variant.

FIGURE 4.

Visual evoked potentials (VEPs) for patients with (A) mild, (B) moderate, and (C) severe systemic and ocular manifestations of KAND. Major tick marks on the x axis represent 100-millisecond (ms) increments, starting at zero; tick marks on the y axis represent 10 microvolts increments. The dashed line represents the 110-ms mark, beyond which a VEP P100 peak is considered delayed. Asterisks mark the P100 peak of a tracing; some tracings are too diffuse to accurately pinpoint a peak. A. Pattern-reversal VEPs for participant 1, who has little to no thinning of the RNFL or GCL-IPL complex on OCT; the upper graphs show a 1° grid pattern, and the lower graphs show a 0.25° grid pattern. B. Pattern-reversal VEPs for participant 2, who has thinning of the RNFL and GCL-IPL complex effectively equal to the KAND cohort average; the upper graphs show a 1° grid pattern, and the lower graphs show a 0.25° grid pattern. C. Flash VEPs for participant 3, who has severe systemic and ocular disease.

The second participant was a 10-year-old boy with BCVA 20/20 in both eyes, and with mild cognitive impairment. His optic nerves on clinical examination revealed no obvious pallor. Despite this finding, OCT revealed RNFL thickness in the right eye of 57 μm and in the left eye of 55 μm, slightly lower than the average pediatric KAND cohort value of 63.7 μm. GCL-IPL thickness was 59 μm in the right eye and 60 μm in the left, similar to the average pediatric KAND cohort value of 60.7 μm. Pattern-reversal VEP studies with a 1° grid showed a significantly slowed and diffuse waveform, with the P100 peak occurring at approximately 141 ms with both eyes fixating (Figure 4B), consistent with significant optic neuropathy. Similar results were seen using pattern-reversal VEP studies with a 0.25° grid. Genetic testing revealed an p.Arg11Gln missense variant.

The last participant, the most severely affected, was a 16-year-old girl with indeterminable vision, due to her inability to report subjective vision as well as lack of blink response to light. The patient was nonverbal and non-ambulatory, with severe cognitive impairment on psychometric testing. Clinical examination revealed severely atrophic optic nerves, and OCT imaging was not possible due to her severe impairment. Flash VEP was performed on this participant due to limitations in her ability to focus on a target. Results showed minimal response to flashes bilaterally (Figure 4C). Her genetic testing revealed a p.Ala351Pro missense variant.

Systemic extraocular associations of KAND and correlation with ophthalmic findings

The systemic extraocular associations of KAND can be found in the recent study by Sudnawa et al ¹⁵. Systemic neurological findings, as well as results of magnetic resonance imaging (MRI) of the brain, are reported for each participant in this study in Supplemental Table 2.

To determine whether there was a relationship between the severity of ophthalmic findings and the severity of systemic findings, we correlated the RNFL and GCL-IPL thickness measurements with a quantitative measure of systemic severity described in Sudnawa et al., the Vineland Adaptive Behavior Scales (VABS) (3^rd edition) score ¹⁵. Briefly, this score is a measure of neurocognitive and adaptive function, with higher VABS scores corresponding to better adaptive functioning. The score is a composite of 4 variables: one overarching adaptive behavior composite (ABC), which reflects total adaptive functioning, and three subdomains that include communication, daily living skills (DLS) and socialization. We found that there is statistically significant positive correlation between GCL-IPL thickness and ABC (r=0.54, p<0.05), GCL-IPL thickness and communication (r=0.51, p<0.05) and GCL-IPL thickness and DLS (r=0.70, p<0.01). Additionally, there is a statistically significant positive correlation between RNFL thickness and DLS (r=0.62, p<0.05). The correlation coefficients and associated p-values are listed in Supplemental Table 3.

Discussion

The ophthalmic phenotype in KAND has been previously poorly characterized. Of approximately 44 publications describing the clinical phenotype in KAND (Supplemental Table 1), only 21 address the ocular phenotype. The aggregate prevalence of optic nerve atrophy is approximately 50% in the studies that address this finding and is largely based on clinical chart review. Of these 21 studies, only six address ophthalmic manifestations separate from optic nerve atrophy, and only one publication was performed by ophthalmologists in one patient ¹⁴. Almost no information is available regarding the level of visual disability experienced by patients, nor are there data to aid in prognostication of visual outcome. Our study addresses these gaps through a detailed, multidimensional, quantitative ophthalmic examination in KAND study participants.

In this study, we report the precise ophthalmic phenotype in participants with KAND, including both subjective and objective components of the examination. This study presents the largest cohort of KAND participants with an ophthalmic assessment to date. Using a combination of clinical examination and optical coherence tomography, we find that the prevalence of optic nerve atrophy to be much higher than previously reported, with 95% having evidence of atrophy on clinical exam and/or with OCT imaging of the RNFL and of the GCL-IPL complex. Although normative databases for pediatric RNFL and GCL-IPL thicknesses are not as well established as in the adult population, there is a substantial decrease between the average thicknesses obtained in the pediatric participants of this cohort compared to studies measuring RNFL and GCL-IPL in healthy pediatric cohorts as well as to the normative databases used by the OCT software, indicating that thinning of these layers is present in childhood and possibly progresses with age ^17–19. There appeared to be a general trend of RNFL and GCL-IPL thinning with age; however, as our study was cross-sectional and participants have heterogeneous genetic mutations, it is difficult to ascertain whether this represents a progressive loss of RNFL and GCL-IPL with time. Given that progression of other neurological symptoms has been observed in patients in KAND, we hypothesize that with longitudinal data we will observe a loss of RNFL and GCL. Statistical analysis is further complicated as RNFL and GCL-IPL thicknesses approach the floor of measurement of OCT imaging, as this precludes quantification of RNFL and GCL-IPL beyond the limits of measurement ^20,21. However, since individuals with KAND are often not able to undergo formal visual field testing, OCT of the nerve and ganglion cell layer may be the most sensitive and quantitative method to monitor patients, so long as they remain above the “floor effect” threshold of the OCT machine. Given that there is a statistically significant correlation between elements of the VABS score, which is a measure of neurocognitive and adaptive function, and RNFL and GCL-IPL thicknesses from OCT, OCT may also offer the advantage of providing a quantitative proxy for systemic severity of the disease that is relatively easy to obtain. The utility of this use of OCT will need to be investigated further with larger cohorts.

Although many participants appeared to have decreased visual acuity in our cohort, most maintained vision better than 20/200. However, our cohort consisted mostly of pediatric participants with neurocognitive deficits, and the accuracy of the visual acuity measured may have a wider confidence interval. The adults exhibited lower visual acuity on average, which may suggest progression of KAND-associated visual loss with age; however this needs to be further evaluated in individuals over time. Participant reported history points to loss of vision with time, as all of the adults in this study reported subjective visual loss with age. One participant (participant #22, Supplemental Table 2) reported her vision in childhood as 20/60, but her best corrected visual acuity during our study was 20/400 OD and 20/500 OS, indicating that her vision had declined with age; no media opacity or other ocular pathology besides optic nerve atrophy and RNFL and GCL-IPL thinning was noted on her examination. Other aspects of vision were also affected in the majority of participants, including color, stereo and field of vision. A large proportion (38%) of participants had strabismus or a history of strabismus, implicating KIF1A in the development of binocular vision. Although the percent of participants exhibiting cataract was similar to previous reports (14%), the etiologies of observed cataract were diverse and included trauma and age, suggesting that cataractous changes may be multifactorial in this patient population.

Visual evoked potential (VEP) testing is another sensitive method of assessing optic nerve dysfunction in KAND, and revealed that even in the presence of 20/20 vision and a seemingly normal-appearing nerve, optic neuropathy or pathology of the visual pathway may still be present. With subtle pathology, pattern-reversal VEP with 0.25° grid pattern may be a more sensitive method to detect optic nerve dysfunction. Given that KIF1A affects both the optic nerve and the brain, any defects detected on VEP may not be specific to the optic nerve itself, but rather to the neuronal pathways posterior to the lateral geniculate nucleus.

Limitations associated with our study include the difficulty in ascertaining whether the thinning of the retinal nerve fiber and ganglion cell layers is due purely to a degenerative process, or whether these layers are dually affected due to an additional neurodevelopmental cause. It is possible that the retinal nerve fiber layer and ganglion cell layer in KAND patients never achieve the thickness expected in unaffected individuals, and that there is a hypoplastic element to the decreased thicknesses that are measured. Another limitation involves the heterogeneity in the severity of each unique pathogenic KIF1A variant, and our inability to dissect how much of the thinning measured on OCT is due the severity of the variant versus possible degeneration with time. It is reassuring that some participants in this cohort with the same pathogenic variant show similar phenotypes. Further studies will need to be conducted to assess more accurately the changes in RNFL and GCL-IPL with time in patients with KAND.

Future directions for this study include more precise correlation of the genotype with phenotypic severity, which can guide the frequency of ophthalmic follow up, anticipating that more severe phenotypes will lose nerve fiber tissue at a faster rate than milder phenotypes. The severity of nerve fiber tissue thinning may also inform clinicians regarding the anticipated severity in other affected systems, such as degree of spastic paraplegia, neurocognitive ability or frequency and severity of seizures. Outcomes of future genetically based treatments of KAND may use VEP to monitor for improvement or stability in the signal amplitude and latency with treatment; treatment efficacy may possibly also be measured using OCT to monitor for stability of RNFL or GCL-IPL thicknesses. It will also be important to determine the best delivery method of gene therapy in these patients; most likely, direct introduction of a gene therapy product into the eye will be most effective, given that intrathecal treatments that are currently under investigation will not reach the optic nerve.

Our study allows clinicians to better counsel patients and their families regarding the extent of visual disability experienced by KAND patients. Although there is possible visual acuity decline with age, for many KAND participants, central visual acuity remains sufficient to be ambulatory. Refractive error was common, and thus patients, particularly patients with neurocognitive deficit, may benefit from annual cycloplegic refraction to allow for the best visual potential. Patients with KAND and their families should be counseled that other aspects of vision may be affected as well, including color, stereopsis and field of vision, and that strabismus is common, necessitating close monitoring for possible amblyopia. Optic nerve atrophy occurs in the majority of patients with KAND, and clinicians wishing to follow the visual trajectory of these patients should consider using optical coherence tomography of the RNFL and GCL given that these modalities are more sensitive and better able to quantify loss of nerve fiber tissue. VEPs may be useful in confirming optic neuropathy in patients whose clinical examination is largely unremarkable.

FIGURE 1.

A. LogMAR visual acuity per eye plotted against age of the participant. Larger logMAR scores indicate poorer visual acuity. B. LogMAR visual acuity per eye grouped according to whether the participant is < 18 or ≥ 18 years of age.