- Email: [email protected]
Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County
Journal Pre-proof Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County J Ben Margines, Connie Huang, Andrew Young, Shiva Mehravaran, Fei Yu, Bartley J. Mondino, Anne L. Coleman PII:
S0002-9394(19)30506-9
DOI:
https://doi.org/10.1016/j.ajo.2019.10.013
Reference:
AJOPHT 11105
To appear in:
American Journal of Ophthalmology
Received Date: 15 July 2019 Revised Date:
7 October 2019
Accepted Date: 9 October 2019
Please cite this article as: Margines JB, Huang C, Young A, Mehravaran S, Yu F, Mondino BJ, Coleman AL, Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County, American Journal of Ophthalmology (2019), doi: https://doi.org/10.1016/ j.ajo.2019.10.013. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Abstract: Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County (Margines, et. al.) Purpose: To summarize the results of five years of vision screening with the University of California, Los Angeles (UCLA) Preschool Vision Program (UPVP). Design: Retrospective evaluation of a screening program Methods: The UPVP performed visual acuity and undilated non-cycloplegic refractive screening using a Retinomax 3 on 93,097 children between 2012 and 2017. Of these, 79,451 children who were between 3- and 5-years-old and were screened for the first time, and 14,259 were referred for full cycloplegic examination if they met specific refractive criteria for myopia, hyperopia, astigmatism or anisometropia. UPVP performed 6,779 cycloplegic examinations on this population. Data from the right eye only were included in this analysis. Results: Of the examined population, hyperopia was found in 61% (4,018), myopia in 20% (1,336), and astigmatism in 93% (6,122) children. Latino children had higher rates of astigmatism and worse visual acuity compared to all other races/ethnicities. An astigmatism cutoff of ≥ 1.50 D in either eye correctly predicted the need for glasses 93% of the time; increasing this cutoff to ≥ 1.50 D in both eyes increased the positive predictive value to 96%. Refractive amblyopia was found in 780 children (1.0% of the screened population and 11.5% of the examined population), and of these, 211 (27%) were bilateral amblyopes. Conclusions: These data represent the largest published sample of vision screening results on preschool-aged children, provide additional insight on the proportion of common refractive errors and their association with race/ethnicity, and can inform screening criteria to more accurately identify children who need intervention to prevent permanent vision loss.
1
Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County
J Ben Margines1, Connie Huang2, Andrew Young3∆, Shiva Mehravaran3, Fei Yu3, 4, Bartley J. Mondino3, Anne L Coleman3, 4
1
David Geffen School of Medicine at the University of California, Los Angeles 10833 Le Conte Ave, Los Angeles, CA 90095
2
University of California, Los Angeles 410 Charles E Young Dr E, Ste 2131, Los Angeles, CA 90024
3
Department of Ophthalmology, Stein Eye Institute, David Geffen School of Medicine at UCLA, University of California, Los Angeles, Los Angeles, California 100 Stein Plaza Driveway, Los Angeles, CA 90095
4
Fielding School of Public Health at the University of California, Los Angeles 650 Charles E Young Dr S, Los Angeles, CA 90095
Corresponding Author: Anne L. Coleman, M.D., Ph.D. Stein Eye Institute 100 Stein Plaza, 2-124 Los Angeles, California 90095 Phone: 310-825-5298 Fax: 310-206-7773 Email: [email protected]
Short Title: Refractive Errors and Amblyopia in Los Angeles County Word Count: 3,811 ∆
Andrew Young, M.D. has since transitioned to the Warren Alpert Medical School of Brown University
2
Introduction It is estimated that in 2015, the prevalence of visual impairment in children 3- to 5- years old was 1.5%, or 174,600 individuals, and that the incidence is projected to increase substantially in the near future.1 Unlike strabismus, refractive errors, especially unilateral cases, can often go unnoticed in children as this patient population may be unaware or unable to adequately describe their condition. Vision screening in this population allows for early diagnosis and treatment to prevent the long-term effects of visual impairment. For these reasons, the United States Preventive Services Task Force (USPSTF) as well as the American Academy for Pediatric Ophthalmology and Strabismus (AAPOS) recommend that all children between the ages of 3 and 5 years undergo vision screening.2 Amblyopia is a reduction of visual acuity that results from a disruption of the normal development of the visual pathway. Refractive amblyopia is the most common, the most difficult to detect, and the most straightforward to correct. The benefit to prompt correction of refractive errors has been reported on previously: children have improved academic performance and quality of life.3-5 Between 2012-2017, the University of California, Los Angeles (UCLA) Preschool Vision Program (UPVP) screened over 90,000 preschoolers in the Los Angeles area using a Retinomax Autorefractor (Righton, Japan), which has been shown to be sufficiently sensitive and specific for this purpose.6 The details of the UPVP program have been reported on previously, as have preliminary results from the first year.6,7 This comprehensive five-year summary represents the largest sample size of vision screening in a population of children aged 3-5. The data enable a more precise analysis of the distribution of refractive errors and proportion of amblyopia in a vulnerable population. Further, it allows for comparisons to prevalence levels previously published in other large studies such as the Multiethnic Pediatric Eye Disease Study (MEPEDS) and the Baltimore Pediatric Eye Disease Study (BPEDS). Importantly, this represents the largest sample of Latino children, who have been previously reported to have higher rates of astigmatism; the data can expand on this disproportionate proportion.8 Methods This retrospective cross-sectional evaluation of a screening program was approved by the UCLA Institutional Review Board and conformed to the requirements of the United States Health Insurance Portability and Accountability Act. Between 2012 and 2017, UPVP performed 93,097 screening procedures among preschool-aged children at subsidized preschools in Los Angeles County. Of these records, 3,305 were excluded for being either younger than 3 or older than 5 years old, and 10,341 were excluded for being repeat visits of the same individual child; for each of these children, only their first visit with UPVP was included in the analysis. The remaining 79,451 records met inclusion criteria.
3
The UPVP has been described in detail elsewhere.7 In brief, the visit protocol at each school comprised an initial screening day for all children and a subsequent examination day for any child who failed the screening day. On the screening day, the Retinomax 3 autorefractor (Rmax) (Righton, Japan) was used to screen children for refractive errors. The criteria for failing the initial screening were as follows: sphere ≤ 3.25 diopters (D) or ≥ +1.75 D, cylinder ≥ 1.50 D, or an interocular difference of ≥ 1.50 D of sphere or ≥ 1.50 D of cylinder error* (Figure 1). For those who failed the initial screening, a secondary screening was performed on the examination day using Rmax with the same criteria. If the child failed the secondary screening and consent had been obtained, a comprehensive ophthalmologic exam was performed. For some schools, consent for vision screening was obtained at the time of school enrollment. For others, a separate consent form for vision screening was provided to the parent or legal guardian to return signed. All children for whom an ophthalmic examination was recommended were provided another consent form for their parent or legal guardian to sign. The comprehensive examination consisted of testing with Allen pictures with occlusion glasses at a ten-foot distance, color vision, stereoacuity with the Titmus fly test, ocular motility, ocular alignment performed with cover test, slit-lamp examination, cycloplegic retinoscopy, and indirect ophthalmoscopy. Cycloplegic refraction was performed at least 30 minutes after instillation of cycloplegic eye drops, and the bestcorrected visual acuity (BCVA) was assessed. Children who would benefit from refractive correction selected a frame style and were fitted with frames. UPVP then arranged to have the glasses made and delivered to the child within a few weeks. Children with more complex ocular pathology were referred to an ophthalmologist, and all children were recommended to follow up with an eye care professional at a time interval dependent on the child’s diagnosis. An initial analysis of the data from the first year of the program was published previously; the analyses performed here are similar.6 Snellen visual acuity (VA) format was converted to logMAR in order to facilitate the calculations. Hyperopia and myopia were calculated based on cycloplegic refraction spherical equivalent (SE). SE was calculated as the sum of the spherical plus half the cylindrical error. On cycloplegic exam, hyperopia was defined as ≥ 0.50 D, myopia as ≤ -0.50 D and emmetropia in between -0.50 D and +0.50 D. Refractive error in minus cylinder, as reported by Rmax, was converted to plus cylinder; astigmatism was defined as having a cylinder ≥ 1.50 D. Anisometropia was defined as ≥ 1.50 D and ≥ 1.00 D in the difference in spherical or cylindrical error, respectively. Amblyopia was defined as unilateral if there was a ≥ 2 line interocular difference in the BCVA, with at least 1 refractive risk factor (anisohyperopia ≥ 1.00 D, anisoastigmatism ≥ 0.50 D, or anisomyopia ≥ 2.00 D), and bilateral if BCVA in each eye *
In the first year of the program, the screening failure criteria included spherical anisometropia of ≥ 2.00 D (≥ 1.50 if antimetropic) and cylindrical anisometropia of ≥ 1.0 D; the subsequent change occurred to align our screening criteria with the UCSD Shiley Eye Institute EyeMobile Program.
4
was < 20/50 for children 3 years old and < 20/40 for children ≥ 4 years old with at least 1 refractive risk factor (hyperopia > 4.00 D for 3-year-olds and > 3.50 D for 4- to 5-yearolds, astigmatism of > 2.00 D for 3-year-olds and > 1.50 D for 4- to 5-year-olds, myopia > 3.00 D for 3-year-olds and > 1.50 D for 4- to 5-year-olds).9 Amblyopia classification is detailed in Figure 2. Based on BCVA of the amblyopic eye, unilateral amblyopia was categorized as mild (≥ 20/30), moderate (20/40 - 20/80) and severe (≤ 20/100). Results are reported as the mean (standard deviation) with a 95% confidence interval (CI) where applicable. Statistical analyses include the student’s t-test, ANOVA with Bonferroni correction, and Kruskal-Wallis (KW) testing for variables which are not normally distributed, performed in Microsoft Excel (Redmond, WA) and R Studio (Boston, MA). Because the results between the left eye and right eye were comparable, we report results for the right eye only. Results During this five-year period, 79,451 children were screened and met eligibility criteria for the first time by the UPVP using the autorefractor with the failure criteria previously mentioned and met inclusion criteria for the study. A summary of the population demographics can be found in Table 1. The mean age was 4.3 (0.57) years, and 49% were female. Of the screened population, 79% (63,129 children) self-reported their ethnicity as Latino, and 31% (24,937 children) spoke Spanish as their primary language. Of the 79,451 children who were screened and met eligibility criteria, 18% (14,549 children) failed the initial screening and were offered to return on a subsequent day for a repeat screening and, if indicated, cycloplegic examination. Of the children who failed screening and were offered examination, 56% (8,101) attended their examination appointment after parental consent was obtained: 84% (6,779) of whom actually received a cycloplegic examination, and 16% (1,322) of whom did not receive a cycloplegic exam nor receive glasses after determining that there was no significant visual pathology based on re-screening of VA, refraction, and per the ophthalmologist’s clinical judgement. In total, 86% (5,883) of children who received a cycloplegic exam needed glasses due to myopia, hyperopia, and/or astigmatism, and an additional 498 children (0.6% of the screened population) received glasses by virtue of already being under the care of an optometrist or ophthalmologist. In sum, 7.9% of the screened population received glasses. Visual Acuity Uncorrected VA (UCVA) in logMAR among examined children was 0.38 (0.19) (20/48) (95% CI: 0.01 – 0.75) and BCVA was 0.19 (0.14) (20/31) (95% CI: -0.09 – 0.46). Visual acuity did not vary by gender: the average male BCVA was 0.18 (0.14) (20/30)
5
(95% CI: -0.09 – 0.45), and the average female BCVA was 0.19 (0.14) (20/31) (95% CI: 0.08 – 0.46) (p = 0.23) (Table 2). The average VA appeared to improve with older age, such that average UCVA for 3-year-olds was 0.40 (0.19) (20/50) (95% CI: 0.03 – 0.77), for 4-year-olds was 0.38 (0.19) (20/48) (95% CI: 0.08 – 0.75), and for 5-year-olds was 0.37 (0.20) (20/47) (95% CI: -0.02 – 0.76). There was a statistically significant improvement between 3 and 4 years of age (p < 0.001), but not between 4 and 5 years of age (p = 0.34). BCVA also improved as follows: for 3-year-olds was 0.21 (0.13) (20/32) (95% CI: -0.04 – 0.46), for 4-year-olds it was 0.18 (0.14) (20/30) (95% CI: -0.01 – 0.45), and for 5-year-olds it was 0.16 (0.13) (20/29) (95% CI: -0.09 – 0.41). This represented a statistically significant improvement between 3 years and 4 years, and between 4 years and 5 years of age (p < 0.001 for both comparisons). In our sample, 71% of examined children corrected to 20/30 or better. Race/ethnicity also correlated with VA: the UCVA was 0.37 (0.20) (20/47) (95% CI: -0.02 – 0.76) for Asians, 0.36 (0.20) (20/46) (95% CI: -0.03 – 0.75) for Blacks, 0.39 (0.19) (20/49) (95% CI: 0.02 – 0.76) for Latinos, 0.32 (0.19) (20/41) (95% CI: -0.05 – 0.69) for non-Latino Whites (p = 0.001). A Bonferroni post-hoc analysis yielded that the statistically significant difference arose only between Latinos and Whites (p < 0.001). The average BCVA was 0.14 (0.12) (20/29) (95% CI: -0.11 – 0.38) for Asians, 0.15 (0.14) (20/30) (95% CI: -0.13 – 0.44) for Blacks, 0.17 ± 0.14 (20/31) (95% CI: -0.10 – 0.45) for Latinos, and 0.17 (0.16) (20/31) (95% CI: -0.14 – 0.47) for non-Latino Whites. The statistically significant difference arose only between Asians and Latinos (p = 0.007). Refraction With non-cycloplegic screening, myopia, emmetropia, and hyperopia were found in 5,168 (7%), 73,245 (92%), and 932 (1%), respectively, and the average spherical refraction was -1.46 D (1.32 D) (95% CI: -4.05 – 1.13). Astigmatism (evaluated at ≥ 1.50 D) was found in 8,957 (11%), and the average cylindrical refraction was 0.65 D (0.78 D) (95% CI: -0.88 – 2.18). Spherical and cylindrical anisometropia were found in 4,843 (6%) and 1,666 (2%), respectively. Of the children who received a full cycloplegic exam (n = 6,779), 20% (1,336) were myopic (SE ≤ -0.50 D, minimum -10.75 D), 19% (1,261) were emmetropic, 59% (4,018) were hyperopic (SE ≥ 0.50 D, maximum 9.50 D), and 164 (2.4%) were missing spherical equivalent data (Table 3). Of cyclopleged emmetropes, 1,051 children were still prescribed glasses due to astigmatism. Astigmatism was the most common refractive error among those given a full cycloplegic exam. When considering a threshold of ≥ 0.50 D, it appeared in 53% (42,152) of the screened population, while it appeared in 90% (6,122) of the examined population. A total of 162 (2.4%) were missing cylinder data. Among examined children, the mean cylinder was 1.89 D (1.10
6
D) (95% CI: -0.27 – 4.05, maximum 8.0 D) and astigmatism ≥ 1.50 D was found in 66% (4,506). Furthermore, spherical anisometropia ≥ 1.0 D was present in 20% (1,383), while cylindrical anisometropia ≥ 1.0 D was present in 24% (1,632). The mean SE was not significantly different between boys at 1.12 D (2.0 D) (95% CI: -2.80 – 5.04) and girls at 1.16 D (2.0 D) (95% CI: -2.76 – 5.08 D). (KW p = 0.18). SE varied significantly by age, such that the mean SE for 3-year-olds, 0.94 D (2.0 D) (95% CI: -2.98 – 4.86 D) was significantly lower than that of 4-year-olds, 1.2 D (2.0 D) (95% CI: -2.72 – 5.12 D) (KW p < 0.001), and for 5-year-olds, 1.2 D (2.1 D) (95% CI: -2.92 – 5.32 D) (KW p < 0.001). There was no statistically significant difference in the SE between 4-year-olds and 5-year-olds (KW p = 0.90). Refractive errors also varied significantly by ethnicity, as follows: the mean SE for Asian children was 0.83 D (2.0 D) (95% CI: -3.09 – 4.75 D), for Black children was 1.0 D (2.4 D) (95% CI: -3.70 – 5.70 D), for Latino children was 1.1 D (2.0 D) (95% CI: -2.92 – 5.02 D), and for White children was 2.1 D (2.1 D) (95% CI: -2.02 – 6.22) (KW p < 0.01). Individual KW testing further yielded that the significant differences occur between White children and children of other racial and ethnic groups (maximum KW p-value was < 0.01), but not between those other races/ethnicities (minimum KW p-value was 0.08). Astigmatism has been reported to be higher in Hispanic children than in AfricanAmerican or non-Hispanic white children.10,11 Within the Latino population in our screened sample, the mean cylinder was 0.68 D (0.82 D) (95% CI: -0.93 – 2.29 D). Cylinder of 1.50 D or greater in either eye on non-cyclopleged Retinomax screening, independent of other refractive errors, triggered 12% of Latino children to be referred for cycloplegic examination. Of those examined, 68% of those examined had astigmatism ≥ 1.50 D. This is a significantly higher percentage compared to all other ethnic/racial groups (e.g., the next highest level was in Blacks, at 61%) (KW p < 0.001). Amblyopia Among those who received a cycloplegic examination, amblyopia was found in 780 children (11.5% of examined population, 1.0% of the screened population) (Table 4). Of these, 568 (8.4% of examined population) were unilateral and 212 (3.1% of examined population) bilateral. Twelve children met the definition of unilateral and bilateral amblyopia and were thus categorized as bilateral. Of children with unilateral amblyopia, 92 (16%) were classified as mild, 421 (74%) were classified as moderate, and 55 (10%) were classified as severe. The proportion of amblyopia increased significantly from 3-year-olds (8.5%) to 4-year-olds (12.2%, p < 0.001) but not between 4-year-olds and 5-year-olds (13.4%, p = 0.28). As a proportion of screened children, unilateral or bilateral amblyopia was found most frequently in children of Latino ethnicity, at 1.1%, while it was found in children of other races less frequently: 0.6% of Asian children, 0.5% of White children, and 0.4% of
7
Black children. When considering this same analysis only among examined children, 11.6% of examined Latino children were diagnosed with amblyopia, compared to 7.3% of Blacks, 12.3% of Asians, and 13.3% of Whites. Of those with unilateral amblyopia, 423 (75%) were hyperopic, 62 (11%) were emmetropic, 83 (15%) were myopic, and 520 (92%) were astigmatic in the amblyopic eye (Table 5). In these children, 74% had visual acuities in the amblyopic eye which were correctable to 20/40 - 20/80; the overall mean BCVA was 0.40 ± 0.18 (20/50). The mean cylinder of the amblyopic eye was 2.7 D (1.3 D) (95% CI: 0.15 – 5.25 D) (range 0 - 8 D). Of those who were emmetropic, 44 (71%) were astigmatic ≥ 1.50 D. The results show that the proportion of children with unilateral and bilateral amblyopia increases with increasing refractive errors, and that amblyopia was more strongly associated with a hyperopic SE than a myopic SE (p < 0.05). Of the children with anisohyperopia ≥ 1.00 D, anisomyopia ≥ 2.00 D, and anisoastigmatism ≥ 0.5 D, 37%, 51%, and 15% either had unilateral or bilateral amblyopia, respectively. Astigmatism was found most frequently as an initial cause of screening failure in the amblyopic children: it appeared in 617 children (79%). Other causes of screening failure in this group included hyperopia (n = 188, 24%), myopia (n= 216, 28%), interocular difference in sphere (n = 298, 38%) or cylinder (n = 244, 31%). Among the 780 amblyopes, 69% (n = 540) failed multiple criteria, and 3% (n = 23) were referred for reasons other than refractive error. Discussion Over 5 years, the UPVP identified 780 preschoolers in Los Angeles County with undiagnosed, untreated amblyopia; this represented 1.0% of screened, eligible children (including 1.1% of Latinos) and 11.5% of examined children. The majority of these children had moderate severity unilateral amblyopia, meaning that their VA in the worse eye was, at best, correctable to 20/40. Another 27% of these children were diagnosed with bilateral amblyopia, meaning that the BCVA in either eye was 20/40. Within these children, astigmatism was the most common cause of referral, although the majority failed multiple screening criteria. Excluding cases of obvious visual pathology, these are children who were likely not complaining about their vision and who would have suffered permanent visual loss without intervention. This work is consistent with previously published work on the first year of this program, which noted a lower proportion of amblyopia than in other populations. MEPEDS researchers found a prevalence rate of 2.6% for amblyopia among Latino preschool-aged children, compared with 1.5% among African American children.12 This difference may be due to variations in study protocol. For UPVP, over half of the children referred for cycloplegic examination did not undergo that examination. This likely represents an enriched sample of visual pathology, however if it is assumed that this sample has equal proportion of visual pathology, then of the 6,448 children who
8
failed their screening but did not return for subsequent rescreening and examination, an estimated 5,396 would have qualified for cycloplegic examination, and 621 would have been diagnosed with amblyopia. In contrast, MEPEDS, which also operated in Los Angeles County, offered cycloplegic examination to all participants, and while they found that non-examined participants were more likely to be male, wealthier, and have a higher self-reported health status, the overall rate of examination was 77%. Astigmatism, which was extremely prevalent in this population, has previously been shown to be associated with ethnicity.8,10 Specifically, this refractive error is more prevalent in Latino children than in African-American or non-Latino white children. 6,8,13 The UPVP data support this finding: although statistical differences were found between the proportion of astigmatism between each ethnicity/race, Latino children who underwent cycloplegic examination were astigmatic ≥ 1.50 D most often at 68%, and they have the highest cylindrical refraction at 1.93 D (1.1 D) (95% CI: -0.23 – 4.09 D). In addition, astigmatism was identified in 61% of Black children, with a mean cylinder of 1.7 D (1.0 D) (95% CI: -0.26 D – 3.66 D), whereas it was identified in 43% of White children, with a mean cylinder 1.3 D (1.2 D) (95% CI: -1.05 – 3.65). Because these proportions pull from the examined population, which have a higher level of refractive error, the proportion of astigmatism ≥1.50 D among the screened population was more comparable to other studies: 11.5% overall, 12.3% in Latinos, and 8.9% in Blacks. These rates were similar, but slightly lower than previously published in MEPEDS (16.8% in Latinos, 12.7% in Blacks) and in Vision In Preschoolers (VIP) (16.4% in Latinos, 12.7% in Blacks).8,10 Further, there was a lower proportion of cylindrical anisometropia in our population (2.3% in Latinos, 1.4% in Blacks) compared to data from MEPEDS (5.6% in Latinos, 4.5% in Blacks). The differences in these rates may have arisen from age-related changes in astigmatism, for which there is little consensus among previously published research. MEPEDS found an age-dependent decrease in rates of astigmatism, while VIP found the opposite, and BPEDS found no decrease in astigmatism among Black children, but a minor decrease among White children.12-14 With respect to the previously published report on the first year of the UPVP program, the data are similar. For example, Latino children represented 89% of examined preschoolers in the first year of the program, and 88% over all five years. Hyperopic children represented 58% of the preschoolers in the first year versus 59% over all five years. Astigmatism ≥ 0.5 D was present in 93% of the population over the first year and 90% over all five years. As the screening failure cutoff for anisoastigmatism decreased from the first to subsequent years of the program, the data reflect a minor increase in the proportion of anisoastigmatism. Two primary limitations of this study which pertain to its external validity were previously published: first, children who were asked to return for a re-screening and cycloplegic examination did not return, and second, UPVP only performed examinations on children who failed screening, creating a biased sample.6 Implicit in this is the fact
9
that non-cycloplegic refraction with Retinomax is less accurate than cycloplegic retinoscopy. Children are known to have a strong accommodative ability, leading to a falsely low proportion of hyperopia in the screened population. Additionally, previous research suggests that Allen symbols overestimate VA, however they were used in this research because the protocol was closely developed based off the UCSD Shiley Eye Mobile Protocol and the tools which the UCLA Mobile Eye Clinic had already been using; they remained the same for internal consistency.15,16 Despite there being methodologic differences between this research and the UPVP, the point remains that estimates of the proportion of amblyopia in this population may be underestimated. An additional limitation of this research is that the purpose of the UPVP was to identify visual pathology and prescribe glasses or refer to an ophthalmologist for further workup; as such, it’s methodology was not designed at its outset to identify the prevalence of amblyopia or refractive errors in this population. For example, the UPVP only performed cycloplegic examinations on those children who failed screening. Further, because most children with more obvious causes of amblyopia (e.g., strabismic) were already under the care of an ophthalmologist, they were also not examined. As a consequence, the comparisons between data from UPVP and other previously published studies are limited in their accuracy because of the difference in methodology; this likely explains why the proportion of amblyopia identified in this study is lower than in previously published research. In conclusion, the data from five years of screening over 90,000 children ages 35 reflect the following: astigmatism was the most common refractive error, occurring in 53% of all children and 90% of children who were referred for cycloplegic examination. Latinos had a higher proportion of astigmatism compared with other groups in our sample, and they had worse average UCVA and BCVA. Amblyopia is found in 1.0% of screened and 11.5% of examined preschoolers, and unilateral amblyopes, hyperopia and astigmatism were the most commonly found refractive errors. These data suggest that vision screening programs using non-cycloplegic autorefractor technology should evaluate closely their screening criteria for hyperopia, astigmatism and anisomyopia, especially in those populations which are known to have higher rates of amblyopia. Further, the results implore vision screening programs to identify factors which depress follow-up rates and to enact protocols to counteract these factors.
10
Acknowledgements: Funding / Support: The authors would like to thank and disclose funding from a grant from First5LA. The research was also supported by an unrestricted grant from Research to Prevent Blindness to the UCLA Stein Eye Institute. Financial Disclosures by Author: Margines – No financial disclosures Huang – No financial disclosures Young – No financial disclosures Mehravaran – No financial disclosures Yu – No financial disclosures Mondino – No financial disclosures Coleman – No financial disclosures Other Acknowledgements: The authors would like to acknowledge the help of the UPVP staff, including Kara Mondino and Pamela Duarte, for their help in administering the program.
11
References: 1.
2.
3. 4.
5.
6.
7. 8.
9.
10. 11.
12.
13.
14. 15.
16.
Varma R, Tarczy-Hornoch K, Jiang X. Visual Impairment in Preschool Children in the United States: Demographic and Geographic Variations From 2015 to 2060. JAMA Ophthalmol. 2017;135(6):610-616. U. S. Preventive Services Task Force. Vision screening for children 1 to 5 years of age: US Preventive Services Task Force Recommendation statement. Pediatrics. 2011;127(2):340-346. Williams WR, Latif AH, Hannington L, Watkins DR. Hyperopia and educational attainment in a primary school cohort. Arch Dis Child. 2005;90(2):150-153. Roch-Levecq AC, Brody BL, Thomas RG, Brown SI. Ametropia, preschoolers' cognitive abilities, and effects of spectacle correction. Arch Ophthalmol. 2008;126(2):252-258; quiz 161. Cotter SA, Pediatric Eye Disease Investigator G, Edwards AR, et al. Treatment of anisometropic amblyopia in children with refractive correction. Ophthalmology. 2006;113(6):895-903. Hendler K, Mehravaran S, Lu X, Brown SI, Mondino BJ, Coleman AL. Refractive Errors and Amblyopia in the UCLA Preschool Vision Program; First Year Results. Am J Ophthalmol. 2016;172:80-86. Mehravaran S, Duarte PB, Brown SI, Mondino BJ, Hendler K, Coleman AL. The UCLA preschool vision program, 2012-2013. J AAPOS. 2016;20(1):63-67. Fozailoff A, Tarczy-Hornoch K, Cotter S, et al. Prevalence of astigmatism in 6- to 72month-old African American and Hispanic children: the Multi-ethnic Pediatric Eye Disease Study. Ophthalmology. 2011;118(2):284-293. Donahue SP, Arthur B, Neely DE, Arnold RW, Silbert D, Ruben JB. Guidelines for automated preschool vision screening: A 10-year, evidence-based update. J AAPOS. 2013;17(1):4-8. Huang J, Maguire MG, Ciner E, et al. Risk factors for astigmatism in the Vision in Preschoolers Study. Optom Vis Sci. 2014;91(5):514-521. McKean-Cowdin R, Varma R, Cotter SA, et al. Risk factors for astigmatism in preschool children: the multi-ethnic pediatric eye disease and Baltimore pediatric eye disease studies. Ophthalmology. 2011;118(10):1974-1981. Multi-ethnic Pediatric Eye Disease Study G. Prevalence of amblyopia and strabismus in African American and Hispanic children ages 6 to 72 months the multi-ethnic pediatric eye disease study. Ophthalmology. 2008;115(7):1229-1236 e1221. Giordano L, Friedman DS, Repka MX, et al. Prevalence of refractive error among preschool children in an urban population: the Baltimore Pediatric Eye Disease Study. Ophthalmology. 2009;116(4):739-746, 746 e731-734. Pascual M, Huang J, Maguire MG, et al. Risk factors for amblyopia in the vision in preschoolers study. Ophthalmology. 2014;121(3):622-629 e621. Mocan MC, Najera-Covarrubias M, Wright KW. Comparison of visual acuity levels in pediatric patients with amblyopia using Wright figures, Allen optotypes, and Snellen letters. J AAPOS. 2005;9(1):48-52. Anstice NS, Jacobs RJ, Simkin SK, Thomson M, Thompson B, Collins AV. Do picturebased charts overestimate visual acuity? Comparison of Kay Pictures, Lea Symbols, HOTV and Keeler logMAR charts with Sloan letters in adults and children. PLoS One. 2017;12(2):e0170839.
12
Figure Captions: Figure 1: Criteria during initial non-cycloplegic screening with Retinomax which would prompt referral for re-screening and possibly cycloplegic examination. Figure 2: Amblyopia Classification: Patients were classified as amblyopic according to this flow diagram. To qualify for bilateral amblyopia, while a maximum BCVA was specified for both eyes, only one eye had to meet refraction requirements.
Table 1: Population Demographics: A summary of population data for the children who were screened, sub-grouped by self-reported ethnicity, age and gender. Ethnicity Asian (5%) Black / African American (5%) Latino (79%) Other / Unknown (6%) White (4%) Total
3 (37%) 1,587 1,624 23,791 1,492 1,222 29,716
Age in Years 4 (54%) 1,987 2,203 34,425 2,361 1,777 42,753
5 (9%) 466 519 4,913 618 466 6,982
Female (49%) 1,924 2,151 30,705 2,124 1,683 38,587
Gender Male (51%) 2,080 2,177 31,823 2,295 1,762 40,137
Unk. (1%) 36 18 601 52 20 727
Total 4,040 4,346 63,129 4,471 3,465 79,451
Table 2: Visual Acuity Results (N, %) for the Right Eyes of the Examined Population.
Age (Years) UCVA (logMAR) ≤ 0.2 (20/32) 0.3 to 0.4 (20/40 to 20/50) 0.5 to 0.6 (20/63 to 20/80) ≥ 0.7 (20/100)
3 n = 1,598 312 (20%) 618 (39%) 436 (27%) 232 (15%)
4 n = 4,039 1,012 (25%) 1,607 (40%) 1,032 (26%) 388 (10%)
5 n = 1,098 316 (29%) 446 (41%) 238 (22%) 98 (9%)
Total n = 6,735 1,640 (24%) 2,671 (40%) 1,706 (25%) 718 (11%)
BCVA (logMAR) ≤ 0.2 (20/32) 0.3 to 0.4 (20/40 to 20/50) 0.5 to 0.6 (20/63 to 20/80) ≥ 0.7 (20/100)
n = 1,445 895 (62%) 460 (32%) 53 (4%) 37 (3%)
n = 3,756 2,716 (72%) 878 (23%) 108 (3%) 54 (1%)
n = 1,041 799 (77%) 203 (20%) 30 (3%) 9 (1%)
n = 6,242 4,410 (71%) 1,541 (25%) 191 (3%) 100 (2%)
Table 3: Refractive Error (OD) of the Examined Population (N, %) Categorized by Age. The data shown do not include 164 children for whom SE data were missing and 162 for whom cylinder data were missing. Age: Hyperopia SE (D) 0.5 to 1.99 2.00 to 3.99 4.00 to 5.99 ≥ 6.00 Total (≥ +0.50) Emmetropia SE (D) -0.49 to 0.49 Myopia SE (D) -0.5 to -1.99 -2.00 to -3.99 -4.00 to -5.99 ≤ -6.00 Total (≤ -0.50) Astigmatism (D) 0.50 to 1.25 1.50 to 3.75 4.00 to 5.75 ≥ 6.00 Total (≥ 0.50) Anisometropia (D) Spherical (≥ 1.5) Spherical (≥ 1.0) Cylindrical (≥ 1.5) Cylindrical (≥ 1.0)
3 (n = 1,616)
4 (n = 4,065)
5 (n = 1,098)
Total (n = 6,779)
476 (29%) 282 (17%) 96 (6%) 32 (2%) 886 (55%)
1,236 (30%) 856 (21%) 272 (7%) 105 (3%) 2,469 (61%)
316 (29%) 245 (22%) 82 (7%) 20 (2%) 663 (60%)
2,028 (30%) 1,383 (20%) 450 (7%) 157 (2%) 4,018 (59%)
300 (19%)
754 (19%)
207 (19%)
1,261 (19%)
334 (21%) 40 (2%) 8 (0%) 3 (0%) 385 (24%)
629 (15%) 92 (2%) 13 (0%) 8 (0%) 742 (18%)
174 (16%) 25 (2%) 5 (0%) 5 (0%) 209 (19%)
1,137 (17%) 157 (2%) 26 (0%) 16 (0%) 1,336 (20%)
390 (24%) 991 (61%) 72 (4%) 1 (0%) 1,454 (90%)
994 (24%) 2,491 (61%) 179 (4%) 4 (0%) 3,668 (90%)
232 (21%) 712 (65%) 55 (5%) 1 (0%) 1,000 (91%)
1,616 (24%) 4,194 (62%) 306 (5%) 6 (0%) 6,122 (90%)
125 (8%) 285 (18%) 131 (8%) 348 (22%)
345 (8%) 843 (21%) 392 (10%) 1,003 (25%)
116 (11%) 255 (23%) 139 (13%) 281 (26%)
586 (9%) 1,383 (20%) 662 (10%) 1,632 (24%)
Table 4: Amblyopia Proportion and Severity Among Examined Children
Unilateral Amblyopia Age 3 (n = 1,616) 4 (n = 4,065) 5 (n = 1,098) Total (n = 6,779)
Mild 12 57 23 92
0.7% 1.4% 2.1% 1.4%
Moderate 92 5.7% 254 6.2% 75 6.8% 421 6.2%
Severe 14 0.9% 30 0.7% 11 1.0% 55 0.8%
Bilateral Amblyopia
Unilateral or Bilateral Amblyopia
20 154 38 212
138 495 147 780
1.2% 3.8% 3.5% 3.1%
8.5% 12.2% 13.4% 11.5%
Table 5: Refractive Error in Children with Unilateral Amblyopia (N, %) Categorized by the Amblyopic Eye Refractive Error Hyperopia SE 0.5 to 1.99 2.00 to 3.99 4.00 to 5.99 ≥ 6.00 Total (≥ +0.50) Emmetropia SE -0.49 to 0.49 Myopia SE -0.5 to -1.99 -2.00 to -3.99 -4.00 to -5.99 ≤ -6.00 Total (≤ -0.50) Astigmatism 0.50 to 1.25 1.50 to 3.75 4.00 to 5.75 ≥ 6.00 Total (≥ 0.50)
Unilateral Amblyopes (n = 568) 127 143 103 50 423
22.4% 25.2% 18.1% 8.8% 74.5%
62
10.9%
57 9 7 10 83
10.0% 1.6% 1.2% 1.8% 14.6%
112 332 74 2 520
19.7% 58.5% 13.0% 0.4% 91.5%
Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County (Margines, et. al.)
Table of Contents Statement: This study analyzes the proportion of refractive errors and amblyopia in the largest sample of preschoolaged children published to date. Subjects in this study failed auto-refraction screening and subsequently received cycloplegic exams. Refractive errors are characterized by age, gender, and race/ethnicity in order to further identify the most at-risk groups, allowing clinicians to more effectively screen for and treat these children.
S0002-9394(19)30506-9
DOI:
https://doi.org/10.1016/j.ajo.2019.10.013
Reference:
AJOPHT 11105
To appear in:
American Journal of Ophthalmology
Received Date: 15 July 2019 Revised Date:
7 October 2019
Accepted Date: 9 October 2019
Please cite this article as: Margines JB, Huang C, Young A, Mehravaran S, Yu F, Mondino BJ, Coleman AL, Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County, American Journal of Ophthalmology (2019), doi: https://doi.org/10.1016/ j.ajo.2019.10.013. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.
Abstract: Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County (Margines, et. al.) Purpose: To summarize the results of five years of vision screening with the University of California, Los Angeles (UCLA) Preschool Vision Program (UPVP). Design: Retrospective evaluation of a screening program Methods: The UPVP performed visual acuity and undilated non-cycloplegic refractive screening using a Retinomax 3 on 93,097 children between 2012 and 2017. Of these, 79,451 children who were between 3- and 5-years-old and were screened for the first time, and 14,259 were referred for full cycloplegic examination if they met specific refractive criteria for myopia, hyperopia, astigmatism or anisometropia. UPVP performed 6,779 cycloplegic examinations on this population. Data from the right eye only were included in this analysis. Results: Of the examined population, hyperopia was found in 61% (4,018), myopia in 20% (1,336), and astigmatism in 93% (6,122) children. Latino children had higher rates of astigmatism and worse visual acuity compared to all other races/ethnicities. An astigmatism cutoff of ≥ 1.50 D in either eye correctly predicted the need for glasses 93% of the time; increasing this cutoff to ≥ 1.50 D in both eyes increased the positive predictive value to 96%. Refractive amblyopia was found in 780 children (1.0% of the screened population and 11.5% of the examined population), and of these, 211 (27%) were bilateral amblyopes. Conclusions: These data represent the largest published sample of vision screening results on preschool-aged children, provide additional insight on the proportion of common refractive errors and their association with race/ethnicity, and can inform screening criteria to more accurately identify children who need intervention to prevent permanent vision loss.
1
Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County
J Ben Margines1, Connie Huang2, Andrew Young3∆, Shiva Mehravaran3, Fei Yu3, 4, Bartley J. Mondino3, Anne L Coleman3, 4
1
David Geffen School of Medicine at the University of California, Los Angeles 10833 Le Conte Ave, Los Angeles, CA 90095
2
University of California, Los Angeles 410 Charles E Young Dr E, Ste 2131, Los Angeles, CA 90024
3
Department of Ophthalmology, Stein Eye Institute, David Geffen School of Medicine at UCLA, University of California, Los Angeles, Los Angeles, California 100 Stein Plaza Driveway, Los Angeles, CA 90095
4
Fielding School of Public Health at the University of California, Los Angeles 650 Charles E Young Dr S, Los Angeles, CA 90095
Corresponding Author: Anne L. Coleman, M.D., Ph.D. Stein Eye Institute 100 Stein Plaza, 2-124 Los Angeles, California 90095 Phone: 310-825-5298 Fax: 310-206-7773 Email: [email protected]
Short Title: Refractive Errors and Amblyopia in Los Angeles County Word Count: 3,811 ∆
Andrew Young, M.D. has since transitioned to the Warren Alpert Medical School of Brown University
2
Introduction It is estimated that in 2015, the prevalence of visual impairment in children 3- to 5- years old was 1.5%, or 174,600 individuals, and that the incidence is projected to increase substantially in the near future.1 Unlike strabismus, refractive errors, especially unilateral cases, can often go unnoticed in children as this patient population may be unaware or unable to adequately describe their condition. Vision screening in this population allows for early diagnosis and treatment to prevent the long-term effects of visual impairment. For these reasons, the United States Preventive Services Task Force (USPSTF) as well as the American Academy for Pediatric Ophthalmology and Strabismus (AAPOS) recommend that all children between the ages of 3 and 5 years undergo vision screening.2 Amblyopia is a reduction of visual acuity that results from a disruption of the normal development of the visual pathway. Refractive amblyopia is the most common, the most difficult to detect, and the most straightforward to correct. The benefit to prompt correction of refractive errors has been reported on previously: children have improved academic performance and quality of life.3-5 Between 2012-2017, the University of California, Los Angeles (UCLA) Preschool Vision Program (UPVP) screened over 90,000 preschoolers in the Los Angeles area using a Retinomax Autorefractor (Righton, Japan), which has been shown to be sufficiently sensitive and specific for this purpose.6 The details of the UPVP program have been reported on previously, as have preliminary results from the first year.6,7 This comprehensive five-year summary represents the largest sample size of vision screening in a population of children aged 3-5. The data enable a more precise analysis of the distribution of refractive errors and proportion of amblyopia in a vulnerable population. Further, it allows for comparisons to prevalence levels previously published in other large studies such as the Multiethnic Pediatric Eye Disease Study (MEPEDS) and the Baltimore Pediatric Eye Disease Study (BPEDS). Importantly, this represents the largest sample of Latino children, who have been previously reported to have higher rates of astigmatism; the data can expand on this disproportionate proportion.8 Methods This retrospective cross-sectional evaluation of a screening program was approved by the UCLA Institutional Review Board and conformed to the requirements of the United States Health Insurance Portability and Accountability Act. Between 2012 and 2017, UPVP performed 93,097 screening procedures among preschool-aged children at subsidized preschools in Los Angeles County. Of these records, 3,305 were excluded for being either younger than 3 or older than 5 years old, and 10,341 were excluded for being repeat visits of the same individual child; for each of these children, only their first visit with UPVP was included in the analysis. The remaining 79,451 records met inclusion criteria.
3
The UPVP has been described in detail elsewhere.7 In brief, the visit protocol at each school comprised an initial screening day for all children and a subsequent examination day for any child who failed the screening day. On the screening day, the Retinomax 3 autorefractor (Rmax) (Righton, Japan) was used to screen children for refractive errors. The criteria for failing the initial screening were as follows: sphere ≤ 3.25 diopters (D) or ≥ +1.75 D, cylinder ≥ 1.50 D, or an interocular difference of ≥ 1.50 D of sphere or ≥ 1.50 D of cylinder error* (Figure 1). For those who failed the initial screening, a secondary screening was performed on the examination day using Rmax with the same criteria. If the child failed the secondary screening and consent had been obtained, a comprehensive ophthalmologic exam was performed. For some schools, consent for vision screening was obtained at the time of school enrollment. For others, a separate consent form for vision screening was provided to the parent or legal guardian to return signed. All children for whom an ophthalmic examination was recommended were provided another consent form for their parent or legal guardian to sign. The comprehensive examination consisted of testing with Allen pictures with occlusion glasses at a ten-foot distance, color vision, stereoacuity with the Titmus fly test, ocular motility, ocular alignment performed with cover test, slit-lamp examination, cycloplegic retinoscopy, and indirect ophthalmoscopy. Cycloplegic refraction was performed at least 30 minutes after instillation of cycloplegic eye drops, and the bestcorrected visual acuity (BCVA) was assessed. Children who would benefit from refractive correction selected a frame style and were fitted with frames. UPVP then arranged to have the glasses made and delivered to the child within a few weeks. Children with more complex ocular pathology were referred to an ophthalmologist, and all children were recommended to follow up with an eye care professional at a time interval dependent on the child’s diagnosis. An initial analysis of the data from the first year of the program was published previously; the analyses performed here are similar.6 Snellen visual acuity (VA) format was converted to logMAR in order to facilitate the calculations. Hyperopia and myopia were calculated based on cycloplegic refraction spherical equivalent (SE). SE was calculated as the sum of the spherical plus half the cylindrical error. On cycloplegic exam, hyperopia was defined as ≥ 0.50 D, myopia as ≤ -0.50 D and emmetropia in between -0.50 D and +0.50 D. Refractive error in minus cylinder, as reported by Rmax, was converted to plus cylinder; astigmatism was defined as having a cylinder ≥ 1.50 D. Anisometropia was defined as ≥ 1.50 D and ≥ 1.00 D in the difference in spherical or cylindrical error, respectively. Amblyopia was defined as unilateral if there was a ≥ 2 line interocular difference in the BCVA, with at least 1 refractive risk factor (anisohyperopia ≥ 1.00 D, anisoastigmatism ≥ 0.50 D, or anisomyopia ≥ 2.00 D), and bilateral if BCVA in each eye *
In the first year of the program, the screening failure criteria included spherical anisometropia of ≥ 2.00 D (≥ 1.50 if antimetropic) and cylindrical anisometropia of ≥ 1.0 D; the subsequent change occurred to align our screening criteria with the UCSD Shiley Eye Institute EyeMobile Program.
4
was < 20/50 for children 3 years old and < 20/40 for children ≥ 4 years old with at least 1 refractive risk factor (hyperopia > 4.00 D for 3-year-olds and > 3.50 D for 4- to 5-yearolds, astigmatism of > 2.00 D for 3-year-olds and > 1.50 D for 4- to 5-year-olds, myopia > 3.00 D for 3-year-olds and > 1.50 D for 4- to 5-year-olds).9 Amblyopia classification is detailed in Figure 2. Based on BCVA of the amblyopic eye, unilateral amblyopia was categorized as mild (≥ 20/30), moderate (20/40 - 20/80) and severe (≤ 20/100). Results are reported as the mean (standard deviation) with a 95% confidence interval (CI) where applicable. Statistical analyses include the student’s t-test, ANOVA with Bonferroni correction, and Kruskal-Wallis (KW) testing for variables which are not normally distributed, performed in Microsoft Excel (Redmond, WA) and R Studio (Boston, MA). Because the results between the left eye and right eye were comparable, we report results for the right eye only. Results During this five-year period, 79,451 children were screened and met eligibility criteria for the first time by the UPVP using the autorefractor with the failure criteria previously mentioned and met inclusion criteria for the study. A summary of the population demographics can be found in Table 1. The mean age was 4.3 (0.57) years, and 49% were female. Of the screened population, 79% (63,129 children) self-reported their ethnicity as Latino, and 31% (24,937 children) spoke Spanish as their primary language. Of the 79,451 children who were screened and met eligibility criteria, 18% (14,549 children) failed the initial screening and were offered to return on a subsequent day for a repeat screening and, if indicated, cycloplegic examination. Of the children who failed screening and were offered examination, 56% (8,101) attended their examination appointment after parental consent was obtained: 84% (6,779) of whom actually received a cycloplegic examination, and 16% (1,322) of whom did not receive a cycloplegic exam nor receive glasses after determining that there was no significant visual pathology based on re-screening of VA, refraction, and per the ophthalmologist’s clinical judgement. In total, 86% (5,883) of children who received a cycloplegic exam needed glasses due to myopia, hyperopia, and/or astigmatism, and an additional 498 children (0.6% of the screened population) received glasses by virtue of already being under the care of an optometrist or ophthalmologist. In sum, 7.9% of the screened population received glasses. Visual Acuity Uncorrected VA (UCVA) in logMAR among examined children was 0.38 (0.19) (20/48) (95% CI: 0.01 – 0.75) and BCVA was 0.19 (0.14) (20/31) (95% CI: -0.09 – 0.46). Visual acuity did not vary by gender: the average male BCVA was 0.18 (0.14) (20/30)
5
(95% CI: -0.09 – 0.45), and the average female BCVA was 0.19 (0.14) (20/31) (95% CI: 0.08 – 0.46) (p = 0.23) (Table 2). The average VA appeared to improve with older age, such that average UCVA for 3-year-olds was 0.40 (0.19) (20/50) (95% CI: 0.03 – 0.77), for 4-year-olds was 0.38 (0.19) (20/48) (95% CI: 0.08 – 0.75), and for 5-year-olds was 0.37 (0.20) (20/47) (95% CI: -0.02 – 0.76). There was a statistically significant improvement between 3 and 4 years of age (p < 0.001), but not between 4 and 5 years of age (p = 0.34). BCVA also improved as follows: for 3-year-olds was 0.21 (0.13) (20/32) (95% CI: -0.04 – 0.46), for 4-year-olds it was 0.18 (0.14) (20/30) (95% CI: -0.01 – 0.45), and for 5-year-olds it was 0.16 (0.13) (20/29) (95% CI: -0.09 – 0.41). This represented a statistically significant improvement between 3 years and 4 years, and between 4 years and 5 years of age (p < 0.001 for both comparisons). In our sample, 71% of examined children corrected to 20/30 or better. Race/ethnicity also correlated with VA: the UCVA was 0.37 (0.20) (20/47) (95% CI: -0.02 – 0.76) for Asians, 0.36 (0.20) (20/46) (95% CI: -0.03 – 0.75) for Blacks, 0.39 (0.19) (20/49) (95% CI: 0.02 – 0.76) for Latinos, 0.32 (0.19) (20/41) (95% CI: -0.05 – 0.69) for non-Latino Whites (p = 0.001). A Bonferroni post-hoc analysis yielded that the statistically significant difference arose only between Latinos and Whites (p < 0.001). The average BCVA was 0.14 (0.12) (20/29) (95% CI: -0.11 – 0.38) for Asians, 0.15 (0.14) (20/30) (95% CI: -0.13 – 0.44) for Blacks, 0.17 ± 0.14 (20/31) (95% CI: -0.10 – 0.45) for Latinos, and 0.17 (0.16) (20/31) (95% CI: -0.14 – 0.47) for non-Latino Whites. The statistically significant difference arose only between Asians and Latinos (p = 0.007). Refraction With non-cycloplegic screening, myopia, emmetropia, and hyperopia were found in 5,168 (7%), 73,245 (92%), and 932 (1%), respectively, and the average spherical refraction was -1.46 D (1.32 D) (95% CI: -4.05 – 1.13). Astigmatism (evaluated at ≥ 1.50 D) was found in 8,957 (11%), and the average cylindrical refraction was 0.65 D (0.78 D) (95% CI: -0.88 – 2.18). Spherical and cylindrical anisometropia were found in 4,843 (6%) and 1,666 (2%), respectively. Of the children who received a full cycloplegic exam (n = 6,779), 20% (1,336) were myopic (SE ≤ -0.50 D, minimum -10.75 D), 19% (1,261) were emmetropic, 59% (4,018) were hyperopic (SE ≥ 0.50 D, maximum 9.50 D), and 164 (2.4%) were missing spherical equivalent data (Table 3). Of cyclopleged emmetropes, 1,051 children were still prescribed glasses due to astigmatism. Astigmatism was the most common refractive error among those given a full cycloplegic exam. When considering a threshold of ≥ 0.50 D, it appeared in 53% (42,152) of the screened population, while it appeared in 90% (6,122) of the examined population. A total of 162 (2.4%) were missing cylinder data. Among examined children, the mean cylinder was 1.89 D (1.10
6
D) (95% CI: -0.27 – 4.05, maximum 8.0 D) and astigmatism ≥ 1.50 D was found in 66% (4,506). Furthermore, spherical anisometropia ≥ 1.0 D was present in 20% (1,383), while cylindrical anisometropia ≥ 1.0 D was present in 24% (1,632). The mean SE was not significantly different between boys at 1.12 D (2.0 D) (95% CI: -2.80 – 5.04) and girls at 1.16 D (2.0 D) (95% CI: -2.76 – 5.08 D). (KW p = 0.18). SE varied significantly by age, such that the mean SE for 3-year-olds, 0.94 D (2.0 D) (95% CI: -2.98 – 4.86 D) was significantly lower than that of 4-year-olds, 1.2 D (2.0 D) (95% CI: -2.72 – 5.12 D) (KW p < 0.001), and for 5-year-olds, 1.2 D (2.1 D) (95% CI: -2.92 – 5.32 D) (KW p < 0.001). There was no statistically significant difference in the SE between 4-year-olds and 5-year-olds (KW p = 0.90). Refractive errors also varied significantly by ethnicity, as follows: the mean SE for Asian children was 0.83 D (2.0 D) (95% CI: -3.09 – 4.75 D), for Black children was 1.0 D (2.4 D) (95% CI: -3.70 – 5.70 D), for Latino children was 1.1 D (2.0 D) (95% CI: -2.92 – 5.02 D), and for White children was 2.1 D (2.1 D) (95% CI: -2.02 – 6.22) (KW p < 0.01). Individual KW testing further yielded that the significant differences occur between White children and children of other racial and ethnic groups (maximum KW p-value was < 0.01), but not between those other races/ethnicities (minimum KW p-value was 0.08). Astigmatism has been reported to be higher in Hispanic children than in AfricanAmerican or non-Hispanic white children.10,11 Within the Latino population in our screened sample, the mean cylinder was 0.68 D (0.82 D) (95% CI: -0.93 – 2.29 D). Cylinder of 1.50 D or greater in either eye on non-cyclopleged Retinomax screening, independent of other refractive errors, triggered 12% of Latino children to be referred for cycloplegic examination. Of those examined, 68% of those examined had astigmatism ≥ 1.50 D. This is a significantly higher percentage compared to all other ethnic/racial groups (e.g., the next highest level was in Blacks, at 61%) (KW p < 0.001). Amblyopia Among those who received a cycloplegic examination, amblyopia was found in 780 children (11.5% of examined population, 1.0% of the screened population) (Table 4). Of these, 568 (8.4% of examined population) were unilateral and 212 (3.1% of examined population) bilateral. Twelve children met the definition of unilateral and bilateral amblyopia and were thus categorized as bilateral. Of children with unilateral amblyopia, 92 (16%) were classified as mild, 421 (74%) were classified as moderate, and 55 (10%) were classified as severe. The proportion of amblyopia increased significantly from 3-year-olds (8.5%) to 4-year-olds (12.2%, p < 0.001) but not between 4-year-olds and 5-year-olds (13.4%, p = 0.28). As a proportion of screened children, unilateral or bilateral amblyopia was found most frequently in children of Latino ethnicity, at 1.1%, while it was found in children of other races less frequently: 0.6% of Asian children, 0.5% of White children, and 0.4% of
7
Black children. When considering this same analysis only among examined children, 11.6% of examined Latino children were diagnosed with amblyopia, compared to 7.3% of Blacks, 12.3% of Asians, and 13.3% of Whites. Of those with unilateral amblyopia, 423 (75%) were hyperopic, 62 (11%) were emmetropic, 83 (15%) were myopic, and 520 (92%) were astigmatic in the amblyopic eye (Table 5). In these children, 74% had visual acuities in the amblyopic eye which were correctable to 20/40 - 20/80; the overall mean BCVA was 0.40 ± 0.18 (20/50). The mean cylinder of the amblyopic eye was 2.7 D (1.3 D) (95% CI: 0.15 – 5.25 D) (range 0 - 8 D). Of those who were emmetropic, 44 (71%) were astigmatic ≥ 1.50 D. The results show that the proportion of children with unilateral and bilateral amblyopia increases with increasing refractive errors, and that amblyopia was more strongly associated with a hyperopic SE than a myopic SE (p < 0.05). Of the children with anisohyperopia ≥ 1.00 D, anisomyopia ≥ 2.00 D, and anisoastigmatism ≥ 0.5 D, 37%, 51%, and 15% either had unilateral or bilateral amblyopia, respectively. Astigmatism was found most frequently as an initial cause of screening failure in the amblyopic children: it appeared in 617 children (79%). Other causes of screening failure in this group included hyperopia (n = 188, 24%), myopia (n= 216, 28%), interocular difference in sphere (n = 298, 38%) or cylinder (n = 244, 31%). Among the 780 amblyopes, 69% (n = 540) failed multiple criteria, and 3% (n = 23) were referred for reasons other than refractive error. Discussion Over 5 years, the UPVP identified 780 preschoolers in Los Angeles County with undiagnosed, untreated amblyopia; this represented 1.0% of screened, eligible children (including 1.1% of Latinos) and 11.5% of examined children. The majority of these children had moderate severity unilateral amblyopia, meaning that their VA in the worse eye was, at best, correctable to 20/40. Another 27% of these children were diagnosed with bilateral amblyopia, meaning that the BCVA in either eye was 20/40. Within these children, astigmatism was the most common cause of referral, although the majority failed multiple screening criteria. Excluding cases of obvious visual pathology, these are children who were likely not complaining about their vision and who would have suffered permanent visual loss without intervention. This work is consistent with previously published work on the first year of this program, which noted a lower proportion of amblyopia than in other populations. MEPEDS researchers found a prevalence rate of 2.6% for amblyopia among Latino preschool-aged children, compared with 1.5% among African American children.12 This difference may be due to variations in study protocol. For UPVP, over half of the children referred for cycloplegic examination did not undergo that examination. This likely represents an enriched sample of visual pathology, however if it is assumed that this sample has equal proportion of visual pathology, then of the 6,448 children who
8
failed their screening but did not return for subsequent rescreening and examination, an estimated 5,396 would have qualified for cycloplegic examination, and 621 would have been diagnosed with amblyopia. In contrast, MEPEDS, which also operated in Los Angeles County, offered cycloplegic examination to all participants, and while they found that non-examined participants were more likely to be male, wealthier, and have a higher self-reported health status, the overall rate of examination was 77%. Astigmatism, which was extremely prevalent in this population, has previously been shown to be associated with ethnicity.8,10 Specifically, this refractive error is more prevalent in Latino children than in African-American or non-Latino white children. 6,8,13 The UPVP data support this finding: although statistical differences were found between the proportion of astigmatism between each ethnicity/race, Latino children who underwent cycloplegic examination were astigmatic ≥ 1.50 D most often at 68%, and they have the highest cylindrical refraction at 1.93 D (1.1 D) (95% CI: -0.23 – 4.09 D). In addition, astigmatism was identified in 61% of Black children, with a mean cylinder of 1.7 D (1.0 D) (95% CI: -0.26 D – 3.66 D), whereas it was identified in 43% of White children, with a mean cylinder 1.3 D (1.2 D) (95% CI: -1.05 – 3.65). Because these proportions pull from the examined population, which have a higher level of refractive error, the proportion of astigmatism ≥1.50 D among the screened population was more comparable to other studies: 11.5% overall, 12.3% in Latinos, and 8.9% in Blacks. These rates were similar, but slightly lower than previously published in MEPEDS (16.8% in Latinos, 12.7% in Blacks) and in Vision In Preschoolers (VIP) (16.4% in Latinos, 12.7% in Blacks).8,10 Further, there was a lower proportion of cylindrical anisometropia in our population (2.3% in Latinos, 1.4% in Blacks) compared to data from MEPEDS (5.6% in Latinos, 4.5% in Blacks). The differences in these rates may have arisen from age-related changes in astigmatism, for which there is little consensus among previously published research. MEPEDS found an age-dependent decrease in rates of astigmatism, while VIP found the opposite, and BPEDS found no decrease in astigmatism among Black children, but a minor decrease among White children.12-14 With respect to the previously published report on the first year of the UPVP program, the data are similar. For example, Latino children represented 89% of examined preschoolers in the first year of the program, and 88% over all five years. Hyperopic children represented 58% of the preschoolers in the first year versus 59% over all five years. Astigmatism ≥ 0.5 D was present in 93% of the population over the first year and 90% over all five years. As the screening failure cutoff for anisoastigmatism decreased from the first to subsequent years of the program, the data reflect a minor increase in the proportion of anisoastigmatism. Two primary limitations of this study which pertain to its external validity were previously published: first, children who were asked to return for a re-screening and cycloplegic examination did not return, and second, UPVP only performed examinations on children who failed screening, creating a biased sample.6 Implicit in this is the fact
9
that non-cycloplegic refraction with Retinomax is less accurate than cycloplegic retinoscopy. Children are known to have a strong accommodative ability, leading to a falsely low proportion of hyperopia in the screened population. Additionally, previous research suggests that Allen symbols overestimate VA, however they were used in this research because the protocol was closely developed based off the UCSD Shiley Eye Mobile Protocol and the tools which the UCLA Mobile Eye Clinic had already been using; they remained the same for internal consistency.15,16 Despite there being methodologic differences between this research and the UPVP, the point remains that estimates of the proportion of amblyopia in this population may be underestimated. An additional limitation of this research is that the purpose of the UPVP was to identify visual pathology and prescribe glasses or refer to an ophthalmologist for further workup; as such, it’s methodology was not designed at its outset to identify the prevalence of amblyopia or refractive errors in this population. For example, the UPVP only performed cycloplegic examinations on those children who failed screening. Further, because most children with more obvious causes of amblyopia (e.g., strabismic) were already under the care of an ophthalmologist, they were also not examined. As a consequence, the comparisons between data from UPVP and other previously published studies are limited in their accuracy because of the difference in methodology; this likely explains why the proportion of amblyopia identified in this study is lower than in previously published research. In conclusion, the data from five years of screening over 90,000 children ages 35 reflect the following: astigmatism was the most common refractive error, occurring in 53% of all children and 90% of children who were referred for cycloplegic examination. Latinos had a higher proportion of astigmatism compared with other groups in our sample, and they had worse average UCVA and BCVA. Amblyopia is found in 1.0% of screened and 11.5% of examined preschoolers, and unilateral amblyopes, hyperopia and astigmatism were the most commonly found refractive errors. These data suggest that vision screening programs using non-cycloplegic autorefractor technology should evaluate closely their screening criteria for hyperopia, astigmatism and anisomyopia, especially in those populations which are known to have higher rates of amblyopia. Further, the results implore vision screening programs to identify factors which depress follow-up rates and to enact protocols to counteract these factors.
10
Acknowledgements: Funding / Support: The authors would like to thank and disclose funding from a grant from First5LA. The research was also supported by an unrestricted grant from Research to Prevent Blindness to the UCLA Stein Eye Institute. Financial Disclosures by Author: Margines – No financial disclosures Huang – No financial disclosures Young – No financial disclosures Mehravaran – No financial disclosures Yu – No financial disclosures Mondino – No financial disclosures Coleman – No financial disclosures Other Acknowledgements: The authors would like to acknowledge the help of the UPVP staff, including Kara Mondino and Pamela Duarte, for their help in administering the program.
11
References: 1.
2.
3. 4.
5.
6.
7. 8.
9.
10. 11.
12.
13.
14. 15.
16.
Varma R, Tarczy-Hornoch K, Jiang X. Visual Impairment in Preschool Children in the United States: Demographic and Geographic Variations From 2015 to 2060. JAMA Ophthalmol. 2017;135(6):610-616. U. S. Preventive Services Task Force. Vision screening for children 1 to 5 years of age: US Preventive Services Task Force Recommendation statement. Pediatrics. 2011;127(2):340-346. Williams WR, Latif AH, Hannington L, Watkins DR. Hyperopia and educational attainment in a primary school cohort. Arch Dis Child. 2005;90(2):150-153. Roch-Levecq AC, Brody BL, Thomas RG, Brown SI. Ametropia, preschoolers' cognitive abilities, and effects of spectacle correction. Arch Ophthalmol. 2008;126(2):252-258; quiz 161. Cotter SA, Pediatric Eye Disease Investigator G, Edwards AR, et al. Treatment of anisometropic amblyopia in children with refractive correction. Ophthalmology. 2006;113(6):895-903. Hendler K, Mehravaran S, Lu X, Brown SI, Mondino BJ, Coleman AL. Refractive Errors and Amblyopia in the UCLA Preschool Vision Program; First Year Results. Am J Ophthalmol. 2016;172:80-86. Mehravaran S, Duarte PB, Brown SI, Mondino BJ, Hendler K, Coleman AL. The UCLA preschool vision program, 2012-2013. J AAPOS. 2016;20(1):63-67. Fozailoff A, Tarczy-Hornoch K, Cotter S, et al. Prevalence of astigmatism in 6- to 72month-old African American and Hispanic children: the Multi-ethnic Pediatric Eye Disease Study. Ophthalmology. 2011;118(2):284-293. Donahue SP, Arthur B, Neely DE, Arnold RW, Silbert D, Ruben JB. Guidelines for automated preschool vision screening: A 10-year, evidence-based update. J AAPOS. 2013;17(1):4-8. Huang J, Maguire MG, Ciner E, et al. Risk factors for astigmatism in the Vision in Preschoolers Study. Optom Vis Sci. 2014;91(5):514-521. McKean-Cowdin R, Varma R, Cotter SA, et al. Risk factors for astigmatism in preschool children: the multi-ethnic pediatric eye disease and Baltimore pediatric eye disease studies. Ophthalmology. 2011;118(10):1974-1981. Multi-ethnic Pediatric Eye Disease Study G. Prevalence of amblyopia and strabismus in African American and Hispanic children ages 6 to 72 months the multi-ethnic pediatric eye disease study. Ophthalmology. 2008;115(7):1229-1236 e1221. Giordano L, Friedman DS, Repka MX, et al. Prevalence of refractive error among preschool children in an urban population: the Baltimore Pediatric Eye Disease Study. Ophthalmology. 2009;116(4):739-746, 746 e731-734. Pascual M, Huang J, Maguire MG, et al. Risk factors for amblyopia in the vision in preschoolers study. Ophthalmology. 2014;121(3):622-629 e621. Mocan MC, Najera-Covarrubias M, Wright KW. Comparison of visual acuity levels in pediatric patients with amblyopia using Wright figures, Allen optotypes, and Snellen letters. J AAPOS. 2005;9(1):48-52. Anstice NS, Jacobs RJ, Simkin SK, Thomson M, Thompson B, Collins AV. Do picturebased charts overestimate visual acuity? Comparison of Kay Pictures, Lea Symbols, HOTV and Keeler logMAR charts with Sloan letters in adults and children. PLoS One. 2017;12(2):e0170839.
12
Figure Captions: Figure 1: Criteria during initial non-cycloplegic screening with Retinomax which would prompt referral for re-screening and possibly cycloplegic examination. Figure 2: Amblyopia Classification: Patients were classified as amblyopic according to this flow diagram. To qualify for bilateral amblyopia, while a maximum BCVA was specified for both eyes, only one eye had to meet refraction requirements.
Table 1: Population Demographics: A summary of population data for the children who were screened, sub-grouped by self-reported ethnicity, age and gender. Ethnicity Asian (5%) Black / African American (5%) Latino (79%) Other / Unknown (6%) White (4%) Total
3 (37%) 1,587 1,624 23,791 1,492 1,222 29,716
Age in Years 4 (54%) 1,987 2,203 34,425 2,361 1,777 42,753
5 (9%) 466 519 4,913 618 466 6,982
Female (49%) 1,924 2,151 30,705 2,124 1,683 38,587
Gender Male (51%) 2,080 2,177 31,823 2,295 1,762 40,137
Unk. (1%) 36 18 601 52 20 727
Total 4,040 4,346 63,129 4,471 3,465 79,451
Table 2: Visual Acuity Results (N, %) for the Right Eyes of the Examined Population.
Age (Years) UCVA (logMAR) ≤ 0.2 (20/32) 0.3 to 0.4 (20/40 to 20/50) 0.5 to 0.6 (20/63 to 20/80) ≥ 0.7 (20/100)
3 n = 1,598 312 (20%) 618 (39%) 436 (27%) 232 (15%)
4 n = 4,039 1,012 (25%) 1,607 (40%) 1,032 (26%) 388 (10%)
5 n = 1,098 316 (29%) 446 (41%) 238 (22%) 98 (9%)
Total n = 6,735 1,640 (24%) 2,671 (40%) 1,706 (25%) 718 (11%)
BCVA (logMAR) ≤ 0.2 (20/32) 0.3 to 0.4 (20/40 to 20/50) 0.5 to 0.6 (20/63 to 20/80) ≥ 0.7 (20/100)
n = 1,445 895 (62%) 460 (32%) 53 (4%) 37 (3%)
n = 3,756 2,716 (72%) 878 (23%) 108 (3%) 54 (1%)
n = 1,041 799 (77%) 203 (20%) 30 (3%) 9 (1%)
n = 6,242 4,410 (71%) 1,541 (25%) 191 (3%) 100 (2%)
Table 3: Refractive Error (OD) of the Examined Population (N, %) Categorized by Age. The data shown do not include 164 children for whom SE data were missing and 162 for whom cylinder data were missing. Age: Hyperopia SE (D) 0.5 to 1.99 2.00 to 3.99 4.00 to 5.99 ≥ 6.00 Total (≥ +0.50) Emmetropia SE (D) -0.49 to 0.49 Myopia SE (D) -0.5 to -1.99 -2.00 to -3.99 -4.00 to -5.99 ≤ -6.00 Total (≤ -0.50) Astigmatism (D) 0.50 to 1.25 1.50 to 3.75 4.00 to 5.75 ≥ 6.00 Total (≥ 0.50) Anisometropia (D) Spherical (≥ 1.5) Spherical (≥ 1.0) Cylindrical (≥ 1.5) Cylindrical (≥ 1.0)
3 (n = 1,616)
4 (n = 4,065)
5 (n = 1,098)
Total (n = 6,779)
476 (29%) 282 (17%) 96 (6%) 32 (2%) 886 (55%)
1,236 (30%) 856 (21%) 272 (7%) 105 (3%) 2,469 (61%)
316 (29%) 245 (22%) 82 (7%) 20 (2%) 663 (60%)
2,028 (30%) 1,383 (20%) 450 (7%) 157 (2%) 4,018 (59%)
300 (19%)
754 (19%)
207 (19%)
1,261 (19%)
334 (21%) 40 (2%) 8 (0%) 3 (0%) 385 (24%)
629 (15%) 92 (2%) 13 (0%) 8 (0%) 742 (18%)
174 (16%) 25 (2%) 5 (0%) 5 (0%) 209 (19%)
1,137 (17%) 157 (2%) 26 (0%) 16 (0%) 1,336 (20%)
390 (24%) 991 (61%) 72 (4%) 1 (0%) 1,454 (90%)
994 (24%) 2,491 (61%) 179 (4%) 4 (0%) 3,668 (90%)
232 (21%) 712 (65%) 55 (5%) 1 (0%) 1,000 (91%)
1,616 (24%) 4,194 (62%) 306 (5%) 6 (0%) 6,122 (90%)
125 (8%) 285 (18%) 131 (8%) 348 (22%)
345 (8%) 843 (21%) 392 (10%) 1,003 (25%)
116 (11%) 255 (23%) 139 (13%) 281 (26%)
586 (9%) 1,383 (20%) 662 (10%) 1,632 (24%)
Table 4: Amblyopia Proportion and Severity Among Examined Children
Unilateral Amblyopia Age 3 (n = 1,616) 4 (n = 4,065) 5 (n = 1,098) Total (n = 6,779)
Mild 12 57 23 92
0.7% 1.4% 2.1% 1.4%
Moderate 92 5.7% 254 6.2% 75 6.8% 421 6.2%
Severe 14 0.9% 30 0.7% 11 1.0% 55 0.8%
Bilateral Amblyopia
Unilateral or Bilateral Amblyopia
20 154 38 212
138 495 147 780
1.2% 3.8% 3.5% 3.1%
8.5% 12.2% 13.4% 11.5%
Table 5: Refractive Error in Children with Unilateral Amblyopia (N, %) Categorized by the Amblyopic Eye Refractive Error Hyperopia SE 0.5 to 1.99 2.00 to 3.99 4.00 to 5.99 ≥ 6.00 Total (≥ +0.50) Emmetropia SE -0.49 to 0.49 Myopia SE -0.5 to -1.99 -2.00 to -3.99 -4.00 to -5.99 ≤ -6.00 Total (≤ -0.50) Astigmatism 0.50 to 1.25 1.50 to 3.75 4.00 to 5.75 ≥ 6.00 Total (≥ 0.50)
Unilateral Amblyopes (n = 568) 127 143 103 50 423
22.4% 25.2% 18.1% 8.8% 74.5%
62
10.9%
57 9 7 10 83
10.0% 1.6% 1.2% 1.8% 14.6%
112 332 74 2 520
19.7% 58.5% 13.0% 0.4% 91.5%
Refractive Errors and Amblyopia Among Children Screened by the UCLA Preschool Vision Program in Los Angeles County (Margines, et. al.)
Table of Contents Statement: This study analyzes the proportion of refractive errors and amblyopia in the largest sample of preschoolaged children published to date. Subjects in this study failed auto-refraction screening and subsequently received cycloplegic exams. Refractive errors are characterized by age, gender, and race/ethnicity in order to further identify the most at-risk groups, allowing clinicians to more effectively screen for and treat these children.