Quantifying refractive error in companion dogs with and without nuclear sclerosis: 229 eyes from 118 dogs

Abstract

Objective:

To evaluate the relationship between nuclear sclerosis (NS) and refractive error in companion dogs.

Animals studied:

118 companion dogs.

Procedures:

Dogs were examined and found to be free of significant ocular abnormalities aside from NS. NS was graded from 0 (absent) to 3 (severe) using a scale developed by the investigators. Manual refraction was performed. The effect of NS grade on refractive error was measured using a linear mixed effects analysis adjusted for age. The proportion of eyes with >1.5D myopia in each NS grade was evaluated using a chi-square test. Visual impairment score (VIS) was obtained for a subset of dogs and compared against age, refractive error, and NS grade.

Results:

Age was strongly correlated with NS grade (P <0.0001). Age-adjusted analysis of NS grade relative to refraction showed a mild but not statistically significant increase in myopia with increasing NS grade, with eyes with grade 3 NS averaging 0.58–0.88D greater myopia than eyes without NS. However, myopia of >1.5D was documented in 4/58 (6.9%) eyes with grade 0 NS, 12/91 (13.2%) eyes with grade 1 NS, 13/57 (22.8%) eyes with grade 2 NS, and 7/23 (30.4%) eyes with grade 3 NS. Risk of myopia >1.5D was significantly associated with increasing NS grade (P=0.02). VIS was associated weakly with refractive error, moderately with age, and significantly with NS grade.

Conclusions:

NS is associated with visual deficits in some dogs but is only weakly associated with myopia. More work is needed to characterize vision in aging dogs.

The average lifespan of pet dogs in North America has increased significantly in recent decades, from 8.5 years in 1997 to reportedly greater than 15 years in 2020.^1–5 As dogs live longer, age-related conditions are more commonly encountered in veterinary practice, including ocular changes such as senile cataracts and corneal endothelial decompensation.^6,7 Loss of sensory functions such as sight has been cited by pet owners as a significant cause of distress regarding their relationship with their aging pet.⁸ Little is known regarding age-related vision changes in dogs, however, and it has been difficult to separate true vision derangements related to ocular changes from behavioral alterations caused by cognitive dysfunction or other neurologic consequences of aging.^9–12

Perhaps the most common age-related ocular change in dogs is nuclear sclerosis (NS; sometimes also called lenticular sclerosis), which manifests clinically as an opalescent or cloudy appearance to the lens.^6,13 Over the course of a normal canine lifespan, increasing numbers of lens fibers occupy the relatively finite space within the lens capsule, subsequently compressing the lens nucleus and increasing its density.^13,14 This age-related increase in nuclear density is seen clinically as NS.¹³ By 10 years of age, over 50% of dogs show signs of NS, and nearly all dogs over 13 years of age are affected.¹⁵

Nuclear sclerosis, in which lens fibers are normal in their structure and arrangement but increased in their number and density, is often differentiated from cataract, in which lens fibers are abnormal in their structure and/or arrangement.¹³ Cataracts attenuate and scatter incoming light and produce unequivocal visual impairment if their extent and opacity is significant enough. Veterinary ophthalmologists have traditionally believed that NS causes minimal visual impairment in dogs,^6,13,16 but in people, significant NS is acknowledged to cause visual impairment and is considered an indication for phacoemulsification.^17–21 People with NS report multiple types of visual disturbance, including presbyopia, decreased color perception and contrast sensitivity, and increased glare.^19,22,23 While there are differences between dogs and people, such as the relative lack of yellowing or brunescence in the canine lens compared to the human lens,¹³ it is likely that at least some of these visual disturbances are experienced by dogs with NS.

Evaluation of visual parameters like glare, contrast sensitivity, or color perception requires patient participation and verbal feedback (i.e. viewing and responding to or describing an image on a screen) and is not currently possible for veterinary patients undergoing routine clinical evaluation. However, myopia is also a reported consequence of NS in people, and is positively correlated with severity of NS in people.^{17–19,24–28} Studies of people with severe NS have documented a mean NS-associated myopia of 7D or greater:^17,19,25 to put this in context, “legal blindness,” or the point at which individuals qualify for government assistance due to poor vision, is defined as corrected visual acuity of less than 20/200, or about 2.5–3.5D of myopia.²⁹ Myopia is believed to develop in people with NS due both to density-induced changes in the refractive index of the lens and to age-related failure of the normal homeostatic mechanisms of the lens to maintain homogeneity of refractive indices between lens fiber membranes and cytoplasm in the nucleus, which in turn disrupts the refractive gradient within the lens.^30,31

Evaluation of myopia and other forms of refractive error is possible in pet dogs. Refractive error has been investigated as a cause of performance deficiencies in sporting and agility dogs.^32,33 Inducing 1.5D or more of myopia in field trial dogs was shown to lead to marked decreases in retrieving performance and obvious confusion and inability to navigate.³² It is reasonable to assume that significant myopia could affect behavior and quality of life in companion dogs as well.

Prior studies evaluating refractive error have shown that most pet dogs are emmetropic, but that myopia increases with age in some dogs, particularly within specific breeds.^34–37 Nuclear sclerosis was suggested as the cause for this phenomenon by some investigators, but the relationship was not examined further in published studies.^34,36 In the one study to date that did not find a positive correlation between age and myopia in dogs, the authors theorized that an unusual lack of clinically evident NS in the older dogs evaluated in their work could account for their findings.³⁸

The goal of the current study was to evaluate the relationship between age, NS, and refractive error in a cross-sectional study of adult companion dogs without other ocular abnormalities.

Materials and methods

This study was conducted in accordance with the Guidelines for Ethical Research in Veterinary Ophthalmology.

Client-owned pet dogs were evaluated at two sites for this study. Approval for work done at the Cummings School of Veterinary Medicine at Tufts University was granted by the Institutional Animal Care and Use Committee (G2021–30) and the Clinical Sciences Research Committee. Approval for work done at the University of Wisconsin School of Veterinary Medicine was granted by the Institutional Animal Care and Use Committee (V006521). Informed consent was obtained from owners prior to enrollment of dogs. Reporting adheres to ARRIVE guidelines.³⁹

At site 1 (Tufts), dogs over 5 years of age without a previous diagnosis of ongoing ocular disease were recruited. Because minimal information exists in dogs regarding refractive error and NS, human data suggesting NS produces up to 7–10D of myopia in severely affected individuals was used to calculate sample size for this site.^17,25 Based on the work of Ofri et al., 1.5D or greater of myopia was assumed to cause visual disability in dogs.³² If the relationship between NS and myopia were to hold across species, a sample size as low as 3 dogs (and 6 eyes) would yield over 90% power to detect a difference in refractive error of 1.5D or greater between dogs without NS and dogs with grade 2 or higher NS. However, to allow for the possibility of undocumented differences between canine and human eyes, a desired sample size of at least 30 dogs with grade 2 or higher NS and 30 dogs with grade 0 (i.e. without) NS was elected for this study site. At site 2 (Wisconsin-Madison), dogs were enrolled from a larger population enlisted for a study on canine aging, and were selected from the larger group if they were over 2 years of age with at least one eye with clear ocular media. Dogs from both sites whose behavior would not permit examination or refraction without use of sedation were excluded as were eyes with keratoconjunctivitis sicca, significant corneal opacity, cataract (aside from isolated punctate lesions that were judged unlikely to affect refraction), or other ocular conditions (aside from NS) that could potentially affect results of refraction. Dogs that were currently receiving topical ocular medications or that had undergone sedation or general anesthesia in the previous 72 hours were also excluded, as were dogs that had undergone phacoemulsification or other lens removal surgery or corneal surgery.

Patient breed, sex, neuter status, and age were recorded. An ocular examination including Schirmer tear testing (Merck Animal Health), fluorescein staining (I-Glo, Jorvet), slit lamp biomicrosopy (Kowa SL-17), indirect fundoscopy (Keeler All-Pupil binocular indirect headset, Volk 20D or 28D indirect lens), and rebound tonometry (TonoVet, iCare) was performed on all dogs prior to refraction by a diplomate of the American College of Veterinary Ophthalmologists or an ophthalmology resident. Pharmacologic mydriasis was not utilized as the examiners felt that posterior segment visualization through the nondilated pupil in a darkened room was sufficient to rule out significant disease likely to affect refraction results or visual function. The investigators had previously determined that grading of NS was readily performed without mydriasis using retroillumination with a dim light source and biomicroscopy with a low intensity slit beam in a darkened room, and that the retinoscopic reflex was easily seen through the nondilated pupil even with higher grades of NS. Furthermore, use of topical cycloplegic/mydriatic drugs has been found to have minimal impact on results of refraction.^34,40

NS was graded in each eye as absent (0), mild (1), moderate (2), or severe (3) using a scale developed by the investigators (Figure 1). Manual refraction was performed in both horizontal and vertical medians using a streak retinoscope (Welch Allyn Elite) and Luneau skiascopy bars (Wilson Ophthalmic) as previously described, with a working distance of 67 cm.^34,35

Figure 1:

Representative images of NS grades. The lens is evaluated in a darkened room using retroillumination with a dim light source followed by a narrow slit beam at low light intensity. Grade 0 (absent): clear lens with no visible corticonuclear distinction grossly on retroillumination or on slit lamp biomicroscopy. Grade 1 (mild): faint refractile changes to nucleus visible on retroillumination, defined junction at corticonuclear boundary on slit lamp biomicroscopy. Grade 2 (moderate): grossly visible nuclear opalescence not interfering with fundoscopy (fundic detail readily visible), narrow beam faintly visible throughout nucleus on slit lamp biomicroscopy. Grade 3 (severe): prominent nuclear opacity visible grossly, not interfering with retroillumination, fundus visible but difficult to discern fine detail due to light scatter, narrow beam readily visible throughout nucleus on slit lamp biomicroscopy.

At site 1, refraction was performed by a diplomate of the American College of Veterinary Ophthalmologists and an ophthalmology resident who were blinded to one another’s findings. A two-way mixed-effects model was used with refraction as the outcome and evaluators were modeled as fixed effects and dog was used as a random effect to calculate the intraclass correlation coefficient (ICC) between the two evaluators at site 1 for each of the 4 refraction measurements (OS and OD horizontal and vertical meridians) (calculated using Wolak, Package ‘ICC’ (2015) in RStudio).⁴¹ Because ICCs suggested good agreement (average ICC for 4 refraction measures was 0.91), values obtained by the two evaluators were averaged to provide a single value per eye for site 1 for the overall analysis of refractive error. A single examiner at site 1 performed NS grading for all dogs. At site 2, refraction and NS grading was performed by a single, different diplomate of the American College of Veterinary Ophthalmologists. T-tests were used to compare refraction results between sites, both overall and within each NS grade.

Refraction data from sites 1 and 2 was then pooled for further analysis. Astigmatism was defined as >0.5D difference between horizontal and vertical meridians within an eye and anisometropia was defined as >0.5D difference between eyes. Because astigmatism was rare in the population, results for vertical and horizontal medians were averaged to yield a single value for each eye. Right and left eyes from each dog were then considered individually. Descriptive statistics were generated regarding age, sex, refractive error, and number of eyes with >1.5D myopia for each grade of NS.

At site 2, a previously published visual function assessment survey was completed by owners of all participating dogs at the time of examination (Supplemental File 1).⁴² Visual Impairment Score (VIS) responses were scored from 0–3 or 0–4 based on published criteria.⁴² For the VIS, higher scores denote poorer vision. Some owners did not provide an answer to all questions or responded that their dog did not perform a specific behavior for reasons unrelated to vision. Scores were therefore adjusted to account for incomplete responses or non-relevant behaviors by dividing the aggregate score for the VIS by the number of questions answered. To widen the scale of the VIS due to this adjustment method, generated VIS were multiplied by 100 to create the final adjusted scores.

All statistical analyses were performed using RStudio ((R Core Team (2021). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. https://www.R-project.org/). P values less than 0.05 were considered statistically significant. An analysis of variance (ANOVA) with age as the outcome was used to determine if there was a significant association between age and NS grade. Because NS grade did not vary between eyes in any of the dogs evaluated, this analysis was performed at dog level. The effect of NS grade on refractive error was then evaluated for left and right eyes individually using a linear mixed effects analysis with refractive error as the outcome and with the individual eye incorporated as a random effect after adjusting for age and evaluator as fixed effects, and also for pooled left and right eyes using a mixed effects linear regression model with refractive error as the outcome and with age and evaluator included as fixed effects but individual dog included as a random effect to account for any nonindependence between eyes.⁴³ The proportion of eyes with significant (>1.5D) myopia in each group was also compared using a chi-square test, and the difference in age between dogs with and without >1.5D myopia was compared using a t-test.

Due to the skewness of the VIS, a Spearman correlation was used to determine if there was an association between VIS and age or refractive error. A Kruskal-Wallis test was used to determine if there was an association between VIS as the outcome and NS grade.

Results

One hundred twenty-one dogs were initially enrolled. Three dogs were excluded from evaluation (two with bilateral cataracts, one that would not tolerate ocular examination). Seven dogs had only one eye available for refraction (five dogs had previously undergone unilateral enucleation for intraocular neoplasia or corneal disease, two dogs had unilateral cataract).

Refraction was performed on 229 eyes from 118 dogs (Figure 2). Forty-two breeds were included. Mixed breed dogs were most commonly represented (n=45), followed by Golden Retrievers (n=7) and German Shepherds (n=6). The population included 68 (58%) female dogs (63 spayed, 5 intact) and 50 (42%) male dogs (44 neutered, 6 intact). NS grade did not vary between eyes in any of the dogs evaluated. Age was strongly correlated with NS grade (ANOVA P <0.0001) (Table 1).

Figure 2:

Distribution of refractive error within the study population. Numbers in boxes represent number of eyes. One eye, from a Toy Poodle, had a refractive error of −4.5D; the other eye of this dog had a mature cataract and was not refracted.

Table 1.

Number of dogs, number of eyes, and mean age for each grade of NS.

NS grade	Site 1 (eyes (dogs))	Site 2 (eyes (dogs))	Total population (eyes (dogs))	Mean age (+/− standard deviation)
All	128 (67)	101 (51)	229 (118)	106 +/− 42 mos
0	32 (16)	26 (13)	58 (29)	61 +/− 30 mos
1	51 (26)	40 (20)	91 (46)	99 +/− 27 mos
2	30 (16)	27 (14)	57 (30)	134 +/− 26 mos
3	15 (9)	8 (4)	23 (13)	162 +/− 22 mos

Mean refractive error across all eyes was −0.21 +/− 1.16D. Astigmatism greater than 0.5D was noted in 2/229 eyes (0.9%), while anisometropia greater than 0.5D was noted in 2/111 dogs (1.8%). Hyperopia of greater than 1.5D was uncommon, occurring in 5 eyes from 3 dogs, including a Golden Retriever, a Dachshund, and a mixed breed dog (Table 2, Figure 3). No significant differences were found between sites for results of refraction overall (P=0.76) or within each NS grade (NS grade 0, P=0.78; NS grade 1, P=0.14; NS grade 2, P=0.53; NS grade 3, P=0.94).

Table 2.

Results of refraction by NS grade (mean +/− standard deviation) for dogs from sites 1 and 2 combined.

NS grade (dogs (eyes))	OD vertical	OD horiz.	OD avg	OS vertical	OS horiz.	OS avg	OU avg
All (229 (118))	−0.17 +/− 0.95 D	−0.21 +/− 1.18D	−0.19 +/− 1.14D	−0.18 +/− 1.22 D	−0.25 +/− 1.18D	−0.22 +/− 1.18D	−0.21 +/− 1.16D
0 (58 (29))	0.06 +/− 0.64D	0.08 +/− 0.61D	0.06 +/− 0.61D	0.04 +/− 0.60D	0.03 +/− 0.60D	0.03 +/− 0.59D	0.04 +/− 0.59D
1 (91 (46))	−0.07 +/− 0.97D	−0.11 +/− 1.03D	−0.09 +/− 0.98D	−0.07 +/− 1.05D	−0.19 +/− 0.96D	−0.14 +/− 0.99D	−0.11 +/− 0.98D
2 (57 (30))	−0.39 +/− 1.35D	−0.44 +/− 1.49D	−0.42 +/− 1.41D	−0.39 +/− 1.51D	−0.49 +/− 1.47D	−0.45 +/− 1.48D	−0.44 +/− 1.42D
3 (23 (13))	−0.56 +/− 1.83D	−0.79 +/− 1.73D	−0.68 +/− 1.78D	−0.61 +/− 1.93D	−0.52 +/− 2.00D	−0.57 +/− 1.96D	−0.63 +/− 1.79D

Figure 3:

Box and whisker plot showing results of refraction by NS grade. The horizontal line within the box shows the median, the X indicates the mean, and the box demarcates the second and third quartiles of values. Number of eyes per group is indicated in parentheses below the box and whisker. The set of boxes labeled as OU average shows the results of refraction for right and left eyes combined within each NS grade.

Age-adjusted analysis of NS grade relative to refraction showed a mild but not statistically significant increase in myopia with increasing NS grade (Table 3). However, myopia >1.5D was documented in 36 eyes from 20 dogs from 15 breeds, including 4/58 (6.9%) eyes with grade 0 NS, 12/91 (13.2%) eyes with grade 1 NS, 13/57 (22.8%) eyes with grade 2 NS, and 7/23 (30.4%) eyes with grade 3 NS. Risk of myopia >1.5D was significantly associated with increasing NS grade (P=0.02). Dogs with >1.5D myopia included 3 mixed breeds, 2 Miniature Dachshunds, 2 Pembroke Welsh Corgis, 2 Beagles, and single representatives from 11 other breeds (Rough Collie, Golden Retriever, Miniature Poodle, Toy Poodle, Bassett Hound, Field Spaniel, Toy Rat Terrier, Shetland Sheepdog, Maltese, Yorkshire Terrier, and Labrador Retriever). Age distribution for dogs with and without >1.5D myopia is shown in Figure 4. There was a significant difference in age between dogs with and without >1.5D myopia (P=0.02).

Table 3.

Age-adjusted analysis of refractive error relative to NS grade, with NS grade 0 used as reference group, for dogs from sites 1 and 2 combined.

NS grade	OD average relative to NS grade 0		OS average relative to NS grade 0		Pooled OS and OD relative to NS grade 0
	Estimate (SE) in D	P value	Estimate (SE) in D	P value	Estimate (SE) in D	P value
1	−0.09 (0.31)	0.771	−0.16 (0.32)	0.625	−0.10 (0.33)	0.379
2	−0.36 (0.42)	0.381	−0.47 (0.43)	0.278	−0.41 (0.44)	0.173
3	−0.58 (0.55)	0.298	−0.58 (0.59)	0.321	−0.88 (0.58)	0.065

Figure 4:

Box and whisker plot showing age distribution of dogs with and without >1.5D myopia. The horizontal line within the box shows the median, the X indicates the mean, and the box demarcates the second and third quartiles of values. Number of dogs per group is indicated in parentheses below the box and whisker.

VIS were available for 51 dogs. There was a moderate correlation between VIS scores and age (Spearman correlation = 0.56, P = 1.61 × 10⁻⁵) and a weak correlation between VIS scores and refractive error (Spearman correlation = −0.08, P = 0.576). The association between NS grade and VIS score was statistically significant (χ² = 24.57, df=3, P=1.896 × 10⁻⁵) with the median VIS score increasing as NS grade increased (Table 4).

Table 4.

VIS score relative to NS grade for dogs from site 2. A higher VIS score indicates worse vision. Values reported are median (1st quartile, 3^rd quartile).

	Overall (51 dogs)	Grade 0 (13 dogs)	Grade 1 (20 dogs)	Grade 2 (14 dogs)	Grade 3 (4 dogs)
VIS score	12 (0,29)	0 (0,6)	6 (0.13.5)	35 (12,63.5)	70.5 (50,108.75)

Discussion

Results of this study suggest that NS in dogs is associated overall with mild, statistically insignificant increases in myopic refractive error, but that dogs with higher grades of NS are more likely to display significant myopia than those with less NS. Dogs with higher grades of NS also appear to have worse vision than dogs without NS based on owner-reported behaviors. Refractive error of less than 0.5D has not been considered to have a meaningful effect on visual function in past studies evaluating dogs.^34,35 It therefore seems unlikely that the mild myopia found in the majority of the dogs with NS in this study is visually debilitating. Because dogs were examined in an undilated state, undocumented mild and/or peripheral lenticular or retinal changes may be responsible for some of the reported visual decreases in the dogs with NS in this study. However, since visual impairment and other sensory or neurologic deficits can have mutually reinforcing impacts on affected individuals, it is possible that even mild refractive error may be behaviorally problematic in older dogs with cognitive dysfunction or other age-related conditions such as hearing loss.^6,10

Newly acquired myopia, even if mild, might also have a more significant impact on behavior than longstanding refractive error. The single-timepoint nature of this study is therefore a limitation. In people, longitudinal studies have shown that refractive error changes over time, and that myopia develops due to (or at least in conjunction with) NS.^18,24–26 Because dogs in this study were not sequentially evaluated, it is not possible to determine whether measured refractive error predated observed NS or was temporally correlated.

This study might also have failed to accurately capture the degree of refractive error associated with higher NS grades. In people, increasing NS is associated with an increase in the prevalence and severity of myopia.^18,27 Although overall relationship of NS grade to myopia was not found to be significant in the current study, dogs with NS grades of 2 or 3 were nevertheless more likely to exhibit myopia of >1.5D than dogs with lesser grades of NS. There were fewer dogs enrolled with NS grades of 2 or 3, likely because many of these dogs had concurrent age-related ocular conditions such as cataracts, endothelial degeneration, or corneal fibrosis that excluded them prior to recruitment into the study (recruitment materials specified dogs without known ocular disease). Because sample size was calculated based on the assumption of a large difference in refractive error across NS grades, which seems not to hold true in dogs, it is possible that the study population was too small to capture significant differences in refractive error.

Evaluation of larger numbers of dogs with high-grade NS may reveal a more significant relationship between NS and myopia, or may establish that a subset of dogs with NS have significant myopia warranting intervention. In people, the relationship between NS, axial globe length, and relative globe dimensions seems to dictate the degree of myopic shift.²⁵ Measurement of additional ocular parameters in aging dogs may identify specific individuals that experience a more significant change in their refractive state with onset of NS.

This study also did not seek to assess vision more globally in dogs with NS. Developing novel methods of evaluating other parameters such as glare or contrast sensitivity in veterinary patients, and establishing more consistent, reproducible means of assessing owner-reported behaviors, is critical to understanding vision changes in aging dogs with varying degrees of NS.¹² Phacoemulsification could be of benefit in ameliorating NS-associated phenomena other than refractive error and is commonly performed in people with NS, who generally report improvements in vision.^17–21,25 However, because both morbidity and mortality associated with general anesthesia and risk of phacoemulsification-associated ocular complications increases with age in dogs,^44–50 better characterization of the effects of NS on canine vision is needed.

Overall, refraction results in this study agreed with those of previous multi-breed studies in finding that most dogs had little refractive error.^34,35 The distribution of breeds in this study did not allow for breed-specific analysis, although a few suggestive findings were observed. An increased prevalence of myopia has been reported in companion German Shepherd Dogs in prior studies, but the 6 German Shepherd Dogs refracted for this study were emmetropic.^34,35 The group of dogs with myopia >1.5D in this study included representatives from breeds in which myopia has previously been documented, including the Toy Poodle, Collie, and Beagle.^34–36 However, significant myopia was also found in this study in 4 eyes from 2 Miniature Dachshunds and 3 eyes from 2 Pembroke Welsh Corgis (one of the Corgis had previously undergone unilateral enucleation for uveal melanoma). Myopia has not been reported previously in Pembroke Welsh Corgis, and information on refractive error in Miniature Dachshunds is scant.⁵¹ These breeds, in addition to others found to be myopic in previous studies (Rottweiler, English Springer Spaniel, Miniature Schnauzer) are also breeds in which behavioral complaints such as anxiety, reactivity, and aggression are more common.^52–58 Evaluation of refractive error and other visual parameters in dogs with specific behavioral traits may therefore be warranted.

This study found that myopia did not increase significantly with increasing NS grade in dogs, although dogs with higher grades of NS had poorer owner-reported vision and were more likely to exhibit >1.5D of myopia than dogs with minimal NS. Further work in aging dogs is needed to characterize effects of NS on vision and to assess changes in vision over time.