Abstract
Purpose:
To describe the prevalence, ocular characteristics, and associated risk factors of moderate/high hyperopia in early childhood.
Design:
Pooled analysis of individual participant data from population-based studies.
Subjects:
6- to 72-month-old, multiethnic children who participated in four population-based studies of pediatric eye diseases.
Methods:
The pooled studies conducted comparable parental interviews and ocular examinations including cycloplegic autorefraction. Presence of hyperopia was defined based on cycloplegic refractive error in the worse eye. Multivariable analyses were performed to evaluate the association of potential risk factors with hyperopia risk.
Main Outcome Measures:
Prevalence and odds ratios of moderate/high hyperopia (≥4.0 D).
Results
Cycloplegic refraction was completed in 15,051 children 6 to 72 months of age. Among these children, the overall prevalence of moderate/high hyperopia (≥4.0 D) in the worse eye was 3.2% (95% confidence interval, 2.9%-3.5%), accounting for 15.6% of all hyperopia (≥2.0 D). Among children with moderate/high hyperopia, 64.4% had both eyes affected, 28.9% had spherical anisometropia of 1.0 D or greater, and 19.5% had astigmatism of 1.5 D or greater. Among 36- to 72-month-old children with moderate/high hyperopia, 17.6% wore glasses. Prevalence of moderate/high hyperopia was slightly less in 12- to 23-month-old children and was relatively stable in children aged 24 months and older. Non-Hispanic and Hispanic white race/ethnicity, family history of strabismus, maternal smoking during pregnancy, and being a participant in the U.S. studies were associated with a higher risk of moderate/high hyperopia (P<0.05).
Conclusions:
By assembling similarly-designed studies, our consortium provides robust estimates of the prevalence of moderate/high hyperopia in the general population and shows that in 6- to 72-month-old children moderate/high hyperopia is not uncommon and its prevalence does not decrease with age. Risk factors for moderate/high hyperopia differ from those for low/moderate hyperopia (2.0 D to <4.0 D) in preschool children, with family history of strabismus and maternal smoking during pregnancy more strongly associated with moderate/high hyperopia than low/moderate hyperopia.
INTRODUCTION
Hyperopia is a common refractive error in young children and its prevalence varies by race/ethnicity1 and geographic region.2 Hyperopia, especially moderate/high hyperopia, has been adversely associated with children’s vision development, various visual functions, and academic performance.3–6 Children 30 to 72 months of age with ≥4.0 D of hyperopia were found to have 10.8 times the risk of decreased bilateral visual acuity compared to emmetropic children.3 The presence of esotropia was associated with hyperopia in a severity-dependent manner, with relative risks increasing from 6.4 for 2.0 D to <3.0 D of hyperopia up to 59.8 for ≥4.0 D of hyperopia in children 6 to 72 months of age.4 The Vision in Preschoolers-Hyperopia in Preschoolers (VIP-HIP) Study5 found that the proportion of children with reduced visual functions (visual acuity, near stereoacuity, and/or accommodative response) increased from 17% in emmetropic children to 82% in children with uncorrected hyperopia of ≥4.0 D to ≤6.0 D. They also found that ≥4.0 D of uncorrected hyperopia was associated with worse performance on early literacy tests in 4- and 5-year-old children.6
Early detection and treatment of hyperopia may help reduce these adverse impacts of hyperopia on young children. Visual-motor function delays7 and poor literacy8, 9 in children with uncorrected hyperopia have been shown to improve with spectacle correction in some studies. The risk of developing strabismus and amblyopia in 9-month-old children with hyperopia was also reduced by wearing a partial hyperopic spectacle correction without affecting emmetropization in one study.10 Although there is no consensus regarding the threshold for prescribing spectacles11, 12 and the effect of spectacle correction on the risk of developing amblyopia and strabismus13–16 for young children with hyperopia, it is generally agreed that hyperopia of 4.0 D or greater warrants consideration of spectacle correction or at least close monitoring.17 With automated methods of refraction increasingly available for screening preschool children, it is important to identify populations at risk for moderate/high hyperopia that may benefit from early detection and intervention. Major risk factors for preschool hyperopia of 2.0 D or greater have been examined previously18 and factors such as white race, maternal smoking during pregnancy, and having health insurance were found to be associated with a higher risk of hyperopia. However, it remains unclear whether the same risk factors are also important for moderate/high hyperopia, and whether hereditary factors play a significant role in moderate/high hyperopia,19 in part because individual studies generally do not have adequate numbers of moderate/high hyperopia cases for statistical analyses.
To fill these gaps, we pooled individual participant data from four population-based studies,20–23 to generate a broader and more precise estimate of the prevalence and characteristics of moderate/high hyperopia (defined as 4.0D or greater) in 6- to 72-month-old children and identify demographic, behavioral, and clinical risk factors that may be associated with a higher risk of moderate/high hyperopia. We also conducted a comparable analysis for low/moderate hyperopia (2.0 D to <4.0 D).
MATERIALS AND METHODS
Study Cohort
Systematic review of the literature identified four studies that met the following inclusion criteria: (1) population-based study, (2) conducted among pre-school age children, (3) performed standardized, comprehensive ocular examinations on all eligible participants with outcomes assessed in both eyes, and (4) assessed potential risk factors in sufficient detail. These four studies formed the Population-based Pediatric Eye Disease Study Consortium (POPEYE consortium): the Multi-Ethnic Pediatric Eye Disease Study (MEPEDS) conducted in Los Angeles, California, United States;20 the Baltimore Pediatric Eye Disease Study (BPEDS) conducted in Baltimore, Maryland, United States;21 the Strabismus, Amblyopia and Refractive Error in Singaporean Children Study (STARS) conducted in Singapore;22 and the Sydney Pediatric Eye Disease Study (SPEDS) conducted in Sydney, Australia.23 The details of participant recruitment in each study have been reported elsewhere.20–23 Written informed consent was obtained from a parent or guardian (hereafter referred to as parent) of each participating child in the original studies.
De-identified individual participant data were pooled from these four studies and compiled into a central database. Study protocols, questionnaires, and data dictionaries were obtained from each study to facilitate data harmonization. These studies shared core eye examination and clinical interview protocols for all primary outcome measures at the design stage, and similar questionnaires with identical core questions were administered. Data from each study were checked for consistency across studies before pooling. Meetings were organized between individual study investigators to adjudicate differences in phenotype classifications and variable definitions. The protocols for this investigation were reviewed and approved by the Institutional Review Board /Ethics Committee at the Keck School of Medicine of the University of Southern California and adhered to the tenets of the Declaration of Helsinki.
Ocular Examination and Interview
Details of ocular examination and interviews conducted in the individual studies have been reported previously.18, 20–23 Briefly, for all studies, a comprehensive eye examination was performed by optometrists or ophthalmologists, trained and certified using standardized protocols. The eye examination included assessments of visual acuity, stereoacuity, ocular alignment and ocular motility testing, ocular health, ocular biometry measurement, and cycloplegic refraction. Details of cycloplegic refraction are summarized in the Supplemental Method. For children whose parents refused cycloplegic eye drops, noncycloplegic refraction measurements were obtained. However, for the current analyses, only participants that received at least two drops of cyclopentolate were included due to concerns of variable refraction measurements resulting from incomplete cycloplegia.24 After cycloplegia, axial length was measured using a noncontact partial coherence interferometer (IOLMaster) in all four studies and the average of 3 to 5 repeated readings was used for analyses. Keratometry was conducted using a handheld Retinomax to measure corneal power and corneal curvature.
On the day of the clinical examination, participating child’s parent(s) were interviewed by trained interviewers using standardized questionnaires to collect data on demographic characteristics, family history of eye diseases, and ocular and medical history of the participants. Supplemental questions that were not included for all four studies are not considered in this report.
Statistical Analysis
The analysis cohort consisted of all children in whom reliable cycloplegic refraction could be obtained. Children with reported diagnosis of Down syndrome or cerebral palsy were excluded from analyses (Supplemental Figure 1). Race/ethnicity as identified by parental report was grouped as Hispanic white (racially white Americans with Hispanic descent, i.e., Cuban, Mexican, Puerto Rican, South or Central American, or other Spanish culture or origin), non-Hispanic white (European Caucasians and Australians having origins in Europe or Middle East, white Americans that are not of Hispanic descent), African Americans, Asians (Americans having origins from the Far East, Southeast Asia, or the Indian subcontinent, Singapore Chinese, Australian with East and South Asian ethnicity). In terms of Asian participant’s ethnicity, 60% of Asians in SPEDS (based on parental report), 59% of Asians in MEPEDS (based on the reported language spoken at home), and all STARS participants (Chinese only) were East Asian. Prevalence of any hyperopia or moderate/high hyperopia was not significantly different between East Asians and other Asians in MEPEDS and SPEDS (Ps>0.05). Children from other racial/ethnic backgrounds were excluded due to their small sample size. Characteristics of participants included in the analysis cohort were compared to those of excluded children using chi-square test, t-test, and analysis of variance where appropriate.
Spherical equivalent (SE) refractive error was calculated as the sphere power plus ½ of the cylinder power. Any hyperopia was defined as SE hyperopia ≥2.0 D, low/moderate hyperopia as SE hyperopia of 2.0D to <4.0D, and moderate/high hyperopia as SE hyperopia ≥4.0 D in the worse eye for the primary analysis. SE hyperopia <2.0D was considered as non-significant hyperopia. Emmetropia was defined as SE refractive error between −0.5 D and +0.5 D, non-inclusive, in both eyes. The worse eye was defined as the eye with the greater hyperopic SE refractive error. If only one eye had refractive error data, that eye was considered to be the worse eye. The 95% confidence interval (CI) for prevalence was estimated using the exact method. The contribution to hyperopic refractive error of axial length, corneal curvature radius (CCR), and the ratio of axial length-to-CCR was evaluated by assessing the magnitude of the coefficient of determination (R2) in multivariable linear regression of refractive error among 36- to 72-month-old children without myopia (SE >−0.5 D). R2 measures the proportion of the variance in the refractive error that can be explained by a given biometric variable. Logistic regression was used to assess whether glasses wear was associated with the following parameters: study, child’s age, sex, race/ethnicity, education level of primary caregiver, low income, gestational age, and birth weight in children with hyperopia.
Sociodemographic, behavioral, and clinical risk factors that have been previously suspected or associated with refractive error in young children18, 25–28 and were collected in the four participating studies were evaluated for their association with hyperopia. Sociodemographic factors evaluated were age group; sex; racial/ethnic group (Hispanic white, non-Hispanic white, African American, and Asian); low income (yes, no); and education level (< high school graduate, high school graduate, college/university graduate or more) of the primary caregiver (biological mothers accounted for 83% and 97% of primary caregivers in MEPEDS and BPEDS, respectively) or the biological mother (SPEDS and STARS). Behavioral and clinical risk factors evaluated were maternal smoking during pregnancy (yes or no); maternal alcohol consumption during pregnancy (yes or no); gestational diabetes (yes or no); preeclampsia (yes or no); maternal anemia during pregnancy (yes or no); maternal hypertension (yes or no) during pregnancy; other pregnancy complications (yes or no); maternal age at childbirth (<35 years; 35-<40 years; and 40+ years); gestational age (≤36 weeks; 37-<42 weeks; ≥42 weeks); birth weight (<2.5kg; 2.5-4.2 kg; and >4.2 kg); breastfeeding (yes or no); family history of strabismus (yes or no) or amblyopia (yes or no) in first-degree relatives.
To identify independent factors associated with hyperopia risk, multivariable logistic regression analyses with forward stepwise selection were performed with a P ≤0.20 criterion for entry into the model and P ≤0.05 for retention in the model. Model selection was performed with any hyperopia (≥2.0 D) as the outcome. Model selection with moderate/high hyperopia as the outcome was also performed and no additional risk factors were identified. For risk factors with missing data, a missing-value indicator29 was created for each risk factor to keep the most participants’ data in the analyses. Risk factors identified through this method were further validated through complete-case analyses and analyses with imputed data. For the variables selected in the final model, there were no material differences in the analysis results of the full data using missing indicator method, those of the complete data, and those of the imputed data; therefore, results from analyses with complete data were reported. Odds ratios with 95% CI’s were reported for the risk factors included in the final model. Between-study heterogeneity in the association between independent risk factors and hyperopia was tested by including proper cross-product terms in the regression models. When significant heterogeneity was present, subgroup analysis stratified by study was performed for generating study-specific estimates and influence of data from an individual study was assessed through leave-one-out-analyses by iteratively removing one study at a time.
Quantile regression was used to explore the impact of potential risk factors at different levels of hyperopia using cutoff points determined by internal percentile ranks, rather than prespecified thresholds (such as the 2.0 D or 4.0 D thresholds used for logistic regression) or the mean value of refractive error (as in ordinary linear regression). Multivariable quantile regression, which estimates the conditional median or other quantiles of an outcome variable with respect to covariates, was used to evaluate the effect of identified risk factors on 0.05 to 0.95 percentile (by 0.05 interval) of refractive error after controlling for other covariates. The 95% CIs were estimated using the sparsity function.
All statistical analyses were conducted using SAS 9.4 (SAS Institute Inc., Cary, NC) with a significance level set at P ≤0.05. All reported P values are two-sided.
RESULTS
In total, the consortium included 17,214 children with ocular examination data from four studies, representing 74% of the eligible participants (Supplemental Figure 1). There were 456 children from racial/ethnic groups (e.g., Pacific Islander) with few participants, 43 children with Down’s syndrome or cerebral palsy, and 371 children who were younger than 6 months or older than 72 months at the time of examination; these children were excluded from subsequent analyses. Among the remaining 16,344 participants, 15,051 (92%) had undergone cycloplegic refraction of at least one eye measured after two or more drops of cyclopentolate. Refraction was measured using the Retinomax autorefractor, retinoscopy, or the Canon autorefractor in 74%, 16%, and 10% of children, respectively. Cycloplegic refraction was more commonly completed in older children (95.1% in 60- to 72-month-olds vs. 89.5% in 6- to 11-month-olds), Hispanic whites (98.4%) and African Americans (95.2%) versus non-Hispanic whites (89.1%) and Asians (87.8%), and BPEDS (94.1%) and MEPEDS (93.7%) versus SPEDS (89.2%) and STARS (87.7%) (P < 0.05 for all). Completion of cycloplegic refraction did not vary by sex or ocular history (e.g., amblyopia, strabismus, myopia, wearing glasses or contact lenses). The demographic characteristics of the participants with cycloplegic refraction data are shown in Table 1. Each of the four racial/ethnic groups had more than 3,000 participants; 48% were female and 8% were less than 1 year of age with nearly equal proportions of participants in the remaining age groups.
Table 1.
Demographic Characteristics of Children with Cycloplegic Refraction Included in the Analysis of Hyperopia Prevalence
| N (column percent) | BPEDS (Baltimore, MD, USA 2003-2007) | MEPEDS (Los Angeles, CA, USA 2003-2010) | STARS (Singapore 2006-2008) | SPEDS (Sydney, Australia 2007-2009) | Total |
|---|---|---|---|---|---|
| Overall | 2231 (100%) | 8448(100%) | 2639(100%) | 1733(100%) | 15051 (100%) |
| Racial/ethnic groups | |||||
| Asian | 27 (1%) | 1309 (15%) | 2639 (100%) | 667 (38%) | 4642 (31%) |
| African American | 1194 (54%) | 2861 (34%) | - | - | 4055 (27%) |
| Hispanic white | 140 (6%) | 3003 (36%) | - | - | 3143 (21%) |
| Non-Hispanic white | 870 (39%) | 1275 (15%) | - | 1066 (62%) | 3211 (21%) |
| Sex | |||||
| Male | 1153 (52%) | 4333 (51%) | 1375 (52%) | 934 (54%) | 7795 (52%) |
| Female | 1078 (48%) | 4115 (49%) | 1264 (48%) | 799 (46%) | 7256 (48%) |
| Age (months) | |||||
| 6-11 | 164 (7%) | 660 (8%) | 163 (6%) | 209 (12%) | 1196 (8%) |
| 12-23 | 351 (16%) | 1467 (17%) | 451 (17%) | 259 (15%) | 2528 (17%) |
| 24-36 | 426 (19%) | 1532 (18%) | 442 (17%) | 325 (19%) | 2725 (18%) |
| 36-47 | 430 (19%) | 1559 (18%) | 513 (19%) | 313 (18%) | 2815 (19%) |
| 48-59 | 451 (20%) | 1570 (19%) | 537 (20%) | 304 (18%) | 2862 (19%) |
| 60-72 | 409 (18%) | 1660 (20%) | 533 (20%) | 323 (19%) | 2925 (19%) |
Abbreviations: BPEDS = the Baltimore Pediatric Eye Disease Study; MEPEDS = the Multiethnic Pediatric Eye Disease Study; STARS = the Strabismus, Amblyopia and Refractive Error in Singaporean Children Study; SPEDS = the Sydney Pediatric Eye Disease Study.
Prevalence of low/moderate and moderate/high hyperopia by demographic factors
Table 2 presents the prevalence of any (≥2.0 D), low/moderate (2.0 D to <4.0 D), and moderate/high (≥4.0 D) hyperopia for the worse eye, stratified by race/ethnicity, sex, and age group. Among these 6- to 72-month-old children, the overall prevalence of moderate/high hyperopia in the worse eye was 3.2%, accounting for 15.6% of all hyperopic children. Similar to low/moderate hyperopia, there was a significant race/ethnicity difference in the prevalence of moderate/high hyperopia (P<0.001), with the highest prevalence observed in Hispanic whites and non-Hispanic whites and the lowest prevalence in Asians. This race/ethnicity difference was consistently observed in all four studies (Supplemental Figure 2) and remained after adjusting for age (data not shown). There was a higher prevalence of low/moderate hyperopia in girls versus in boys (P<0.001), but similar sex-difference was not observed for moderate/high hyperopia (P=0.96). This was found in all four studies for all race/ethnicity groups except Hispanic whites (P for se-difference in low/moderate hyperopia=0.83). While the prevalence of low/moderate hyperopia was greatest in 6- to 11-month-old children and then stabilized after 24 months, the prevalence of moderate/high hyperopia was the same or greater for all age groups older than 6-11 months, with the exception of 12- to 23-month-old children. The associations with race/ethnicity, sex, and age group showed the same patterns in our multivariate analyses (see below and Table 4).
Table 2.
Prevalence of Hyperopia Severity in the Worse Eye by Race/ethnicity, Sex, and Age of Child
| Any hyperopia (≥2.0 D) | Low/moderate hyperopia (2.0 D to < 4.0 D) | Moderate/high hyperopia (≥4.0 D) | ||||
|---|---|---|---|---|---|---|
| N | % (95% CI) | N | % (95% CI) | N | % (95% CI) | |
| Overall | 3086 | 20.5% (19.9%-21.2%) | 2604 | 17.3% (16.7%-17.9%) | 482 | 3.2% (2.9%-3.5%) |
| Race/Ethnicity | ||||||
| Non-Hispanic white | 884 | 27.5% (26.0%-29.1%) | 748 | 23.3% (21.8%-24.8%) | 136 | 4.2% (3.6%-5.0%) |
| Hispanic white | 838 | 26.7% (25.1%-28.3%) | 679 | 21.6% (20.2%-23.1%) | 159 | 5.1% (4.3%-5.9%) |
| African American | 824 | 20.3% (19.1%-21.6%) | 695 | 17.1% (16.0%-18.3%) | 129 | 3.2% (2.7%-3.8%) |
| Asian | 540 | 11.6% (10.7%-12.6%) | 482 | 10.4% (9.5%-11.3%) | 58 | 1.3% (1.0%-1.6%) |
| P for race/ethnic difference1 | <0.001 | <0.001 | <0.001 | |||
| Sex | ||||||
| Male | 1476 | 18.9% (18.1%-19.8%) | 1227 | 15.7% (14.9%-16.6%) | 249 | 3.2% (2.8%-3.6%) |
| Female | 1610 | 22.2% (21.2%-23.2%) | 1377 | 19.0% (18.1%-19.9%) | 233 | 3.2% (2.8%-3.6%) |
| P for sex difference1 | <0.001 | <0.001 | <0.96 | |||
| Age | ||||||
| 6-11 | 323 | 27.0% (24.5%-29.6%) | 286 | 23.9% (21.5%-26.4%) | 37 | 3.1% (2.2%-4.2%) |
| 12-23 | 502 | 19.9% (18.3%-21.5%) | 447 | 17.7% (16.2%-19.2%) | 55 | 2.2% (1.6%-2.8%) |
| 24-35 | 512 | 18.8% (17.3%-20.3%) | 434 | 15.9% (14.6%-17.4%) | 78 | 2.9% (2.3%-3.6%) |
| 36-47 | 581 | 20.6% (19.2%-22.2%) | 484 | 17.2% (15.8%-18.6%) | 97 | 3.5% (2.8%-4.2%) |
| 48-59 | 593 | 20.7% (19.3%-22.3%) | 470 | 16.4% (15.1%-17.8%) | 123 | 4.3% (3.6%-5.1%) |
| 60-72 | 575 | 19.7% (18.2%-21.2%) | 483 | 16.5% (15.2%-17.9%) | 92 | 3.2% (2.5%-3.8%) |
| P for age trend1 | 0.012 | <0.001 | 0.007 | |||
Abbreviation: CI=confidence interval.
P values for race/ethnic and sex differences were estimated from chi-square tests and the P values for age trends were estimated from Mantel-Haenszel chi-square tests. 95% confidence interval values were estimated based on the exact method.
Table 4.
Multivariable Analyses of Risk Factors Associated with Different Severities of Hyperopia in the Worse Eye
| Total N of participants | Low/moderate Hyperopia (2.0 D to <4.0 D) |
Moderate / High Hyperopia (≥4.0 D) |
|||||
|---|---|---|---|---|---|---|---|
| % | OR (95% CI)1 | P1 | % | OR (95% CI) 1 | P1 | ||
| Race | |||||||
| Asian | 3584 | 10.3% | 1.00 (ref) | - | 1.3% | 1.00 (ref) | - |
| African American | 3351 | 16.9% | 1.19 (0.98-1.44) | 0.082 | 3.3% | 1.27 (0.83-1.93) | 0.28 |
| Hispanic white | 2958 | 21.5% | 1.68 (1.37-2.05) | <0.001 | 5.1% | 1.97 (1.28-3.03) | 0.002 |
| Non-Hispanic white | 2291 | 24.4% | 2.16 (1.78-2.61) | <0.001 | 4.4% | 2.32 (1.52-3.52) | <0.001 |
| Sex | |||||||
| Male | 6316 | 16.0% | 1.00 (ref) | 3.3% | 1.00 (ref) | ||
| Female | 5868 | 19.1% | 1.26 (1.14-1.39) | <0.001 | 3.4% | 1.09 (0.89-1.33) | 0.42 |
| Age | |||||||
| 6-11 | 1018 | 24.0% | 1.51 (1.25-1.81) | <0.001 | 2.9% | 0.87 (0.56-1.34) | 0.52 |
| 12-23 | 2108 | 17.4% | 0.99 (0.84-1.15) | 0.85 | 2.2% | 0.59 (0.41-0.86) | 0.006 |
| 24-35 | 2230 | 16.0% | 0.89 (0.76-1.05) | 0.16 | 2.9% | 0.77 (0.55-1.08) | 0.13 |
| 36-47 | 2229 | 17.4% | 1.01 (0.86-1.18) | 0.94 | 3.7% | 1.05 (0.77-1.44) | 0.76 |
| 48-59 | 2267 | 16.6% | 0.97 (0.83-1.14) | 0.70 | 4.5% | 1.30 (0.96-1.75) | 0.091 |
| 60-72 | 2332 | 17.2% | 1.00 (ref) | - | 3.5% | 1.00 (ref) | - |
| Study | |||||||
| STARS (Singapore) | 2112 | 8.7% | 1.00 (ref) | - | 0.7% | 1.00 (ref) | - |
| SPEDS (Sydney, Australia) | 1026 | 16.5% | 1.24 (0.95-1.62) | 0.12 | 2.4% | 2.28 (1.09-4.78) | 0.028 |
| BPEDS (Maryland, US) | 1772 | 21.8% | 1.79 (1.38-2.32) | <0.001 | 3.4% | 3.18 (1.58-6.42) | 0.001 |
| MEPEDS (California, US) | 7274 | 19.2% | 1.75 (1.40-2.19) | <0.001 | 4.2% | 4.48 (2.36-8.49) | <0.001 |
| Family history of Strabismus | |||||||
| No | 11763 | 17.3% | 1.00 (ref) | - | 3.3% | 1.00 (ref) | - |
| Yes | 421 | 22.6% | 1.20 (0.94-1.52) | 0.14 | 5.7% | 1.54 (1.00-2.38) | 0.050 |
| Educational level of primary caregiver | |||||||
| ≥College graduate | 2565 | 14.7% | 1.00 (ref) | - | 2.4% | 1.00 (ref) | - |
| High school graduate | 6636 | 17.8% | 1.16 (1.01-1.33) | 0.032 | 3.3% | 1.20 (0.88-1.63) | 0.26 |
| < High school | 2983 | 19.1% | 1.16 (0.98-1.37) | 0.082 | 4.3% | 1.32 (0.92-1.90) | 0.13 |
| Maternal smoking during pregnancy | |||||||
| No | 11360 | 17.0% | 1.00 (ref) | - | 3.2% | 1.00 (ref) | - |
| Yes | 824 | 23.7% | 1.28 (1.07-1.54) | 0.007 | 5.2% | 1.64 (1.16-2.33) | 0.005 |
| Number of cigarettes per day2 | |||||||
| 0 | 11360 | 17.0% | 1.00 (ref) | - | 3.2% | 1.00 (ref) | - |
| 1 to <5 | 336 | 22.3% | 1.26 (0.96-1.65) | 0.091 | 5.4% | 1.72 (1.04-2.83) | 0.035 |
| ≥5 | 437 | 25.2% | 1.34 (1.06-1.70) | 0.015 | 5.3% | 1.62 (1.02-2.58) | 0.040 |
| Period of smoking2 | |||||||
| Nonsmokers | 11360 | 17.0% | 1.00 (ref) | - | 3.2% | 1.00 (ref) | - |
| 1st or 2nd trimester only | 349 | 25.5% | 1.44 (1.12-1.86) | 0.005 | 3.7% | 1.17 (0.66-2.08) | 0.59 |
| 3rd trimester | 463 | 22.9% | 1.21 (0.95-1.53) | 0.13 | 6.3% | 2.03 (1.32-3.11) | 0.001 |
| Duration of smoking, months2 | |||||||
| 0 | 11360 | 17.0% | 1.00 (ref) | - | 3.2% | 1.00 (ref) | - |
| >0 to ≤3 | 317 | 25.9% | 1.45 (1.11-1.89) | 0.006 | 3.5% | 1.10 (0.59-2.05) | 0.77 |
| 4 to 6 | 88 | 22.7% | 1.24 (0.74-2.07) | 0.42 | 5.7% | 1.79 (0.71-4.52) | 0.22 |
| 7 to 9 | 413 | 22.5% | 1.19 (0.93-1.53) | 0.17 | 6.5% | 2.11 (1.36-3.28) | <0.001 |
| Cigarette-Months of smoking2 | |||||||
| 0 | 11360 | 17.0% | 1.00 (ref) | - | 3.2% | 1.00 (ref) | - |
| >0 to <9 | 180 | 27.2% | 1.61 (1.15-2.27) | 0.006 | 4.4% | 1.49 (0.72-3.09) | 0.28 |
| 9 to <45 | 304 | 21.7% | 1.17 (0.88-1.56) | 0.28 | 4.6% | 1.41 (0.80-2.48) | 0.24 |
| ≥45 | 287 | 24.7% | 1.30 (0.97-1.73) | 0.079 | 6.6% | 2.11 (1.26-3.52) | 0.005 |
Abbreviations: BPEDS = the Baltimore Pediatric Eye Disease Study; CI=confidence interval; D=diopter; MEPEDS = the Multiethnic Pediatric Eye Disease Study; OR=odds ratio; STARS = the Strabismus, Amblyopia and Refractive Error in Singaporean Children Study; SPEDS = the Sydney Pediatric Eye Disease Study.
Odds ratios (ORs), 95% confidence intervals (CIs) and p values were estimated from multivariable polytomous logistic regression with severity of hyperopia (categorized as emmetropia or nonsignificant hyperopia, low/moderate hyperopia, and moderate/high hyperopia) as the outcome variable. Estimates for each of the maternal smoking parameters were generated from multivariable regression models replacing the binary smoking status variable with the corresponding categorical variable and adjusting for all other non-smoking covariates.
Cigarette-months of smoking was calculated as the product of number of cigarettes smoked per day and the number of months of smoking during pregnancy. Data on number of cigarettes smoked per day, period of smoking, duration of smoking, and cigarette-months were missing for ≤53 participants.
Prevalence data for hyperopia in the right eye and in the better eye are available in Supplemental Table 1.
We examined the proportion of 36- to 72-month-old children who were reported to wear glasses (Table 3). Glasses wear was reported for 17.6% of children with moderate/high hyperopia in the worse eye and 20.5% (data not shown in tables) of children with moderate/high hyperopia in both eyes. Glasses wear was less common in African American children with moderate/high hyperopia (6.6%) than children from other racial/ethnic groups with moderate/high hyperopia (28.4%, 23.4%, and 15.2% in non-Hispanic whites, Asians, and Hispanic whites, respectively; P for race=0.039) and less common in 36- to 47-month-olds (11.7%) than in 60- to 72-month-olds (24.3%) (P for age trend=0.048), but was not significantly different by study, household income level, education level of the primary caregiver, and other factors evaluated (data not shown in tables).
Table 3.
Ocular Characteristics of Children with Different Levels of Hyperopia in the Worse Eye
| Emmetropia or nonsignificant hyperopia (0.5D to <2.0D) | Low/moderate hyperopia (2.0 D to <4.0 D) | Moderate/high hyperopia (≥4.0 D) | P4 | |
|---|---|---|---|---|
| N1 | 10443 | 2589 | 481 | |
| N (%) with hyperopia in both eyes1 | - | 1759 (67.9%) | 310 (64.4%) | 0.14 |
| RE of the worse eye, mean ± SD | 0.96±0.54 D | 2.59±0.54 D | 5.19±1.33 D | - |
| RE of the better eye, mean ± SD | 0.74±0.55 D | 2.23±0.59 D | 4.36±1.51 D | - |
| Interocular difference in RE, mean ± SD | 0.22±0.21 D | 0.36±0.38 D | 0.82±0.90 D | <0.001 |
| N (%) with spherical anisometropia (≥1.0 D) | 90 (0.9%) | 165 (6.4%) | 139 (28.9%) | <0.001 |
| N (%) with astigmatism (≥1.5 D) | 542 (5.2%) | 264 (10.2%) | 94 (19.5%) | <0.001 |
| N (%) with either spherical anisometropia or astigmatism2 | 619 (5.9%) | 395 (15.3%) | 213 (44.3%) | <0.001 |
| Among children aged 3-5 years | ||||
| N (%) with either spherical anisometropia or astigmatism2 | 327 (5.3%) | 210 (14.6%) | 134 (43.0%) | <0.001 |
| N (%) of children wearing glasses3 | 50 (1.0%) | 30 (2.6%) | 45 (17.6%) | <0.001 |
| Axial length (mm) | 22.2±0.7 | 21.6±0.7 | 20.8±0.7 | <0.001 |
| Corneal power (D) | 44.0±1.5 | 43.9±1.6 | 43.7±1.5 | <0.001 |
| Corneal curvature radius (mm) | 7.68±0.27 | 7.70±0.28 | 7.74±0.26 | <0.001 |
| Axial length/corneal curvature radius | 2.89±0.07 | 2.81±0.07 | 2.69±0.08 | <0.001 |
Abbreviation: SD = standard deviation.
Data on bilateral involvement were missing for 101 children due to missing refraction data from one eye.
Astigmatism of 1.5 D or greater
Data on glasses wearing were not collected for 1483 children aged 36-72 months.
P values for dichotomous and continuous variables were estimated using Mantel-Haenszel chi-square and analysis of variance tests.
Ocular characteristics of low/moderate and moderate/high hyperopia Of the children with moderate/high hyperopia, 64% (310/481) had moderate/high hyperopia in both eyes, and 36% had moderate/high hyperopia in one eye only (Table 3). Both spherical anisometropia ≥1.0 D and astigmatism ≥1.5 D were much more common in children with moderate/high hyperopia (28.9% and 19.5%, respectively) than children with low/moderate hyperopia (6.4% and 10.2%) and emmetropia/no significant hyperopia (0.9% and 5.2%). Similar proportions were observed in the 36- to 72-month-old subset of older children. There were no age differences in the prevalence of astigmatism and anisometropia among children with moderate/high hyperopia (P for age trend= 0.18 and 0.87, respectively).
We compared axial length, corneal power, and CCR across different levels of hyperopia severity in 36- to 72-month-old children, from whom reliable biometry measurements were obtained (Table 3). Shorter axial length, lower corneal power, greater CCR, and smaller axial length-to-CCR ratio were observed for hyperopia in a severity-dependent manner (P<0.001 for all). These differences remained even after adjustment for age, sex, race/ethnicity, and study (data not shown). The proportion of variance in hyperopic refractive error that can be attributed (R2) to axial length alone, CCR alone, or axial length/CCR ratio was estimated to be 0.251, 0.002, and 0.419, respectively.
Risk factors associated with low/moderate and moderate/high hyperopia
Race/ethnicity, sex, age, study, level of education of primary caregiver, maternal smoking during pregnancy, and family history of strabismus were identified to be associated with the prevalence of any hyperopia. Family history of amblyopia was positively associated with any hyperopia and maternal hypertension during pregnancy was negatively associated with any hyperopia in the analysis of full data using missing-indicator method. However, these two factors did not remain associated with any hyperopia in the analyses with imputed data for missing values (P= 0.078 and 0.055, respectively) and were not associated with moderate/high hyperopia, and therefore were removed from the final model. Estimates for other identified risk factors were not materially changed with and without these two factors.
The results of the multivariable analysis of the associations between the identified risk factors and having either low/moderate or moderate/high levels of hyperopia are shown in Table 4. Results were similar with additional adjustment for astigmatism (data not shown). The patterns of association of low/moderate and moderate/high hyperopia with race/ethnicity, sex, and age were discussed above, and remained similar after adjustment for other covariates. Family history of strabismus conferred a 54% higher risk for moderate/high hyperopia but was not associated with low/moderate hyperopia. Having a less educated primary caregiver was also associated with high prevalence of any hyperopia; however, the association was significant only for low/moderate hyperopia. Maternal smoking during pregnancy was associated with a 28% higher risk of low/moderate hyperopia and 64% higher risk of moderate/high hyperopia. The impact of intensity, period, duration, and cigarette-months of smoking were also evaluated (Table 4). Even a modest amount (<5 cigarettes/day) of smoking was associated with a higher risk for moderate/high hyperopia. While a short duration of smoking conferred a higher risk only for low/moderate hyperopia, longer duration of smoking, especially smoking in the third trimester, was associated with a much higher risk of moderate/high hyperopia.
Significant differences in low/moderate and moderate/high hyperopia were also observed between studies, after adjusting for race/ethnicity and other risk factors that might vary by study. The highest risk of hyperopia was associated with the MEPEDS and the BPEDS in the U.S., and the lowest with the STARS in Singapore (Table 4). We also evaluated between-study heterogeneity in the association between independent risk factors and hyperopia. No significant heterogeneity was found except for the associations of age and education level of primary caregiver with low/moderate hyperopia (P’s for heterogeneity=0.002 and 0.003, respectively; Supplemental Table 2).
To further explore the effect of the identified risk factors at different levels of hyperopia using cutoff points determined by internal percentile ranks, multivariable quantile regression was performed for refractive error in the worse eye. Over the entire cohort, refractive error was distributed such that −0.625 D, −0.125 D, +1.125 D, +2.625 D and +3.45 D represented the 5th, 10th, 50th, 90th, and 95th percentile values, respectively. However, the percentile values for levels of refractive error varied with age and race/ethnicity (Supplemental Figure 3). After controlling for other factors associated with hyperopia, age was associated with refractive error in a complex manner (Figure 1). Overall, there was a hyperopic shift with age in the refractive error levels defining the lower quantiles (more myopic/less hyperopic end) (Figure 1F). Refractive errors associated with non-Asian ethnicity, U.S. studies, maternal smoking during pregnancy, and family history of strabismus were consistently more hyperopic at the upper quantiles (hyperopic end) of the distribution, and their effect sizes were generally greater at more extreme quantiles (Supplemental Figure 4).
Figure 1. Quantile Regression Plots (a-f) Demonstrating the Estimated Effect of Age Across the Distribution of Refractive Errors (D) in Children 6 to 72 Months of Age.
The effect estimates were generated from multivariable quantile regression adjusted for other risk factors. The x-axis represents the quantile level (e.g., 0.1 = 10th percentile) of refractive error, and the y-axis represents the difference in corresponding refractive error between age groups (refractive error for older age group minus refractive error for younger age group). The shaded area represents 95 percent confidence intervals.
The impact of risk factors for hyperopia on axial length, CCR, and axial length/CCR ratio are presented in Supplemental Table 3. Maternal smoking during pregnancy was associated with shorter axial length and smaller CCR.
DISCUSSION
Using cycloplegic refraction data from more than 15,000 participants of four population-based studies, we present much needed data characterizing the prevalence, ocular characteristics, and associated risk factors for moderate/high hyperopia (≥4.0 D) in children 6 to 72 months of age. We found moderate/high hyperopia was not uncommon in young children and its prevalence did not decrease with age. A family history of strabismus and maternal smoking during pregnancy, especially maternal smoking that persisted into the third trimester, were more strongly associated with moderate/high hyperopia than low/moderate hyperopia in this combined cohort. Non-Hispanic and Hispanic white race/ethnicity were associated with a higher risk of moderate/high hyperopia even after accounting for other factors.
Young children with moderate/high hyperopia are at much higher risk for strabismus,30 reduced visual acuity,3 near stereoacuity, and accommodative response,5 and deficits in early literacy; 6 thus, better surveillance for early identification and close monitoring by eye care professionals are needed for these children. Additionally, having anisometropia or astigmatism in addition to hyperopia increases the risk of strabismus or amblyopia further;3, 23, 30 we found that 44.3% of children with moderate/high hyperopia had either anisometropia of 1.0 D or greater or astigmatism of 1.5 D or greater. Similar observations were also reported by the VIP Study.31 Despite the high risk for strabismus and amblyopia in these children, only a small proportion (17.6%) of 36- to 72-month-old children with moderate/high hyperopia wore glasses. Furthermore, African American children with moderate/high hyperopia were less likely to be wearing glasses than children from other racial/ethnic groups, even after adjustment for household income and education level of the primary caregiver. This low proportion of optical correction in children with moderate/high hyperopia, even in those aged 36- to 72-months, could be due to the ongoing debate on the need for optical correction for these children among eye care providers, underestimation of the importance of optical correction by parents, and/or underdiagnosis of hyperopia. Data from MEPEDS showed that few preschool children had ever previously had a dilated eye examination, ranging from 2.8% among 6- to 11-month-old children to 11.6% among 61- to 72-month-old children.32 A lack of surveillance and treatment of moderate/high hyperopia in preschool children can result in irreversible vision impairment, as preschool age remains a critical period for the effective treatment of strabismus and amblyopia.
Hyperopia can result from variations in multiple ocular components (short axial length, flat cornea, low lens power, or a combination of these factors) which may be influenced by different genetic/environmental factors. We found that in 36- to 72-month-old children, from whom biometry data can be more reliably measured, an abnormally low axial length-to-CCR ratio is a more important contributor to hyperopia (explaining ~42% of the variation in hyperopic refractive error) than a short axial length or a greater CCR alone. Maternal smoking during pregnancy and a family history of strabismus both seem to contribute to higher hyperopia risk through associated shorter axial length. Racial/ethnic differences in hyperopia prevalence, on the other hand, can result from variations in either CCR or axial length (see below). Future studies with data on other ocular components (e.g., vitreous chamber depth, lens thickness, anterior chamber depth) are needed to further explore ocular determinants of preschool hyperopia.
Prevalence of both moderate/high hyperopia (≥4.0 D) and low/moderate hyperopia (2.0 D to <4.0 D) vary by race/ethnicity. Overall, non-Hispanic and Hispanic white children have a higher risk of hyperopia than African American and Asian children. This racial/ethnic difference was consistently seen across studies and remained after controlling for other potential risk factors for hyperopia, suggesting that genetic differences or racial/ethnic differences in other uncontrolled environmental or behavioral factors may play a role. The higher risk of hyperopia in non-Hispanic whites may be attributable in part to shorter axial length, whereas the higher prevalence of hyperopia in Hispanic whites may be attributable more to a disproportionately flatter corneal curvature. The observation of Hispanic whites having greater CCR and less corneal power and Asians and Hispanic whites having longer axial lengths is consistent with previous studies.33, 34 The observation of Asian children having longer axial lengths than other racial/ethnic groups even in preschool age, suggests that Asians are predisposed to myopia even before a significant school-age myopic shift occurs.
Even though longitudinal data were not available for us to investigate how an individual child’s refractive error develops in preschool age, our cross-sectional comparison of refractive error distributions across different age groups suggested the following age-related changes. First, children with different initial levels of refractive error may undergo different developmental changes in refraction from 6 months to 72 months of age. While myopic eyes may go through a progressive hyperopic shift toward emmetropia over the first few years of life, hyperopic eyes may undergo a myopic shift mostly in the first year of life. Second, it is likely that refractive error does not change significantly between 3 and 5 years of age. Consequently, as our prevalence estimates indicate, the prevalence of moderate/high hyperopia remains relatively stable or even increases slightly in children aged 2 years and older. These findings suggest that screening for hyperopia may reasonably be initiated as early as 2 years of age. Nonetheless, prospective studies that monitor longitudinal changes in preschool refraction are needed to confirm these observations.
In addition to confirming previous reports of an association between maternal smoking during pregnancy and a higher risk of hyperopia1 and greater amounts of hyperopia,35 we found that even a modest level (<5 cigarettes/day) of maternal smoking during pregnancy conferred a higher risk for moderate/high hyperopia, and third trimester smoking may be a more important determinant of moderate/high hyperopia risk than early-pregnancy smoking. Other studies have reported that late pregnancy was more sensitive to the adverse effects of smoking on fetal growth36 and the development of oculomotor control.37 The effect of maternal smoking during pregnancy on hyperopia risk does not seem to be mediated by a higher risk of preterm birth and/or low birth weight, neither of which was associated with hyperopia in this study. Biometric analysis revealed that maternal smoking during pregnancy was associated with reduced axial length and CCR. This is consistent with findings that nicotine may inhibit eye growth in animal models38 and that maternal smoking may be associated with an increased risk of anophthalmia and microphthalmia.39 Unfortunately, despite many tobacco control efforts, smoking during pregnancy is still prevalent in many countries.40 With the rapid increase of e-cigarette use in young adults and the perception of relative safety, e-cigarette use during pregnancy and therefore fetal exposure to nicotine is expected to increase.41 It remains unclear how maternal exposure to e-cigarettes will impact the vision of the developing fetus.
The finding of an association between family history of strabismus and a higher risk of moderate/high hyperopia was not surprising. It is known that hyperopia, especially moderate/high hyperopia, is a strong risk factor for esotropia.30 Among MEPEDS and BPEDS participants with esotropia, 73.5% (75 out of 102) had hyperopia of 2.0 D or greater and 40.2% (41 out of 102) had hyperopia of 4.0 D or greater. In the present analysis, family history of strabismus is likely a surrogate marker for family history of hyperopia, which was not directly assessed. Unfortunately, our knowledge of genetic variations contributing to strabismus and hyperopia remain limited and genetic susceptibility loci shared by the two traits have not been identified to date.42 Our data also support a higher risk of hyperopia for females, possibly due to a smaller AL/CCR ratio. However, the average sex-difference in hyperopia is small, with females being approximately 0.1 D more hyperopic than males, and there is no significant difference in the risk of moderate/high hyperopia.
This study has unique strengths, including a very large sample size comprised of four population-based studies from three different countries that performed similar standardized ocular examinations and parental interviews allowing for comparisons by race/ethnicity. Because of these advantages, prevalence estimates from this study are less likely to be impacted by selection bias than those from clinic-based studies and may be more generalizable to other similar populations. Despite methodological similarities, significant between-study differences were found, both with regard to the adjusted risk of hyperopia, and (in the case of low/moderate hyperopia) its associations with age and education of primary caregiver. These differences may result from demographic or environmental differences not captured in our analyses, or from methodological differences. The differences in the cycloplegia protocol and refraction method across studies may have contributed in part to the between-study differences in the magnitude of age effect. In addition, heterogeneity within a given racial/ethnic group may have contributed to between-study variation persisting after adjustment for race/ethnicity. For example, there may be additional genetic differences between non-Hispanic whites residing in different geographical locations. Better assessment of ethnicity or genetic ancestry are needed. Furthermore, having a less-educated primary caregiver might in fact be an indirect surrogate for other factors such as maternal substance misuse in pregnancy,43 and as such may more closely reflect increased risk in some countries more than in others. Given these considerations, there is a theoretical concern that because almost all Hispanic white participants were recruited through MEPEDS and all African Americans were recruited through MEPEDS and BPEDS, our observed racial/ethnic differences in hyperopia prevalence may be confounded by methodological, demographic, or environmental differences between studies. However, the confounding is not likely to be substantial, as the racial/ethnic differences in hyperopia risk remained after adjusting for study and are furthermore consistent with those reported by MEPEDS,44 in which four different ethnic groups were recruited from the same geographical location (Los Angeles, CA) and evaluated using identical methods.
This large pooled study has a number of other limitations as well. Even though refractive error was measured by cycloplegic autorefraction for most children, retinoscopy was used for children for whom an autorefraction measurement was unsuccessful in MEPEDS and BPEDS or for the youngest children in SPEDS and STARS. The difference in the choice of refraction method between participants is unlikely to affect our results substantially, as no clinically significant differences in refraction results were found among different cycloplegic autorefraction methods and cycloplegic retinoscopy.24 Our threshold definitions of low/moderate hyperopia as 2.0 D to < 4.0 D and moderate/high hyperopia as ≥ 4.0 D are of necessity somewhat arbitrary and could have led to the misclassification of some participants. However, the associations that we observed between various risk factors and hyperopia are robust, as shown by our quantile regression analyses, which revealed consistent patterns at the hyperopic end of the refractive error distribution. Also, participants in this combined cohort were all recruited from urban areas and therefore prevalence estimates from this study may be not applicable to children from rural areas, as rural residency has been associated with a higher prevalence of hyperopia in some studies.26
In conclusion, using a large set of pooled individual participant data, we have provided prevalence estimates and described ocular characteristics of moderate/high hyperopia, and explored various demographic, behavioral, and clinical associations with moderate/high hyperopia (≥4.0 D) and low/moderate hyperopia (2.0 D to <4.0 D). We found that moderate/high hyperopia was not uncommon in 6- to 72-month-old children and its prevalence does not decrease with age. and a large proportion of these young children with moderate/high hyperopia had either anisometropia or astigmatism. We also characterized the association of maternal smoking during pregnancy with hyperopia risk and found that maternal smoking persisting into the third trimester was associated with a much higher risk of moderate/high hyperopia. Further studies are needed to clarify the longitudinal patterns of refractive change in early childhood and biological mechanisms underlying the association between gestational exposure to tobacco smoke and hyperopia, and to characterize other social/environmental factors that may account for variations in adjusted risk for hyperopia at different studies.
Supplementary Material
Précis.
Pooled analyses of population-based studies found that hyperopia ≥4.0D is not uncommon and often accompanied by anisometropia or astigmatism in 6- to 72-month-old children and has distinct risk factors.
ACKNOWLEDGEMENTS
We would like to acknowledge all students, staffs, collaborators, and advisory board members involved in the design and data collection of MEPEDS, BPEDS, SPEDS, and STARS, as well as all the study participants.
Support: Supported by the National Eye Institute, National Institutes of Health, Bethesda, MD (grant no. EY025313), and unrestricted grants from the Research to Prevent Blindness (New York, NY) to the Department of Ophthalmology at the University of Southern California and the Department of Ophthalmology at the University of Washington. The STARS project was funded by the National Medical Research Council (NMRC/1009 /2005 and NMRC/1112/2007) in Singapore.
Footnotes
Meeting Presentation: ARVO 2017 Annual Meeting, Baltimore, MD, USA.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. 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.
Conflicts of Interest: No conflicting relationship exists for any author.
REFERENCES
- 1.Borchert M, Varma R, Cotter S, et al. Risk factors for hyperopia and myopia in preschool children: the Multi-Ethnic Pediatric Eye Disease and Baltimore Pediatric Eye Disease studies. Ophthalmology 2011;118(10): 1966–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hashemi H, Fotouhi A, Yekta A, et al. Global and regional estimates of prevalence of refractive errors: systematic review and meta-analysis. J Curr Ophthalmol 2018;30(1):3–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Tarczy-Hornoch K, Varma R, Cotter SA, et al. Risk factors for decreased visual acuity in preschool children: the Multi-Ethnic Pediatric Eye Disease and Baltimore Pediatric Eye Disease studies. Ophthalmology 2011;118(11):2262–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cotter SA, Varma R, Tarczy-Hornoch K, et al. Risk factors associated with childhood strabismus: the Multi-Ethnic Pediatric Eye Disease and Baltimore Pediatric Eye Disease Studies. Ophthalmology 2011. ;118(11):2251–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ciner EB, Kulp MT, Maguire MG, et al. Visual function of moderately hyperopic 4- and 5- year-old children in the Vision in Preschoolers - Hyperopia in Preschoolers Study. Am J Ophthalmol 2016;170:143–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Kulp MT, Ciner E, Maguire M, et al. Uncorrected hyperopia and preschool early literacy: results of the Vision in Preschoolers-Hyperopia in Preschoolers (VIP-HIP) Study. Ophthalmology 2016;123(4):681–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Roch-Levecq AC, Brody BL, Thomas RG, Brown SI. Ametropia, preschoolers’ cognitive abilities, and effects of spectacle correction. Arch Ophthalmol 2008;126(2):252–8; quiz 161. [DOI] [PubMed] [Google Scholar]
- 8.Bruce A, Kelly B, Chambers B, et al. The effect of adherence to spectacle wear on early developing literacy: a longitudinal study based in a large multiethnic city, Bradford, UK. BMJ Open 2018;8(6):e021277. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Slavin RE, Collins ME, Repka MX, et al. In plain sight: reading outcomes of providing eyeglasses to disadvantaged children. J Educ Stud Placed Risk 2018;23(3):250–8. [Google Scholar]
- 10.Atkinson J, Braddick O, Robier B, et al. Two infant vision screening programmes: prediction and prevention of strabismus and amblyopia from photo- and videorefractive screening. Eye (Lond) 1996;10 (Pt 2):189–98. [DOI] [PubMed] [Google Scholar]
- 11.Jones-Jordan L, Wang X, Scherer RW, Mutti DO. Spectacle correction versus no spectacles for prevention of strabismus in hyperopic children. Cochrane Database Syst Rev 2014(8):CD007738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lambert SR. Should glasses be prescribed for all children with moderate hyperopia? Ophthalmology 2016;123(4):676–8. [DOI] [PubMed] [Google Scholar]
- 13.Anker S, Atkinson J, Braddick O, et al. Non-cycloplegic refractive screening can identify infants whose visual outcome at 4 years is improved by spectacle correction. Strabismus 2004;12(4):227–45. [DOI] [PubMed] [Google Scholar]
- 14.Ingram RM, Arnold PE, Dally S, Lucas J. Results of a randomised trial of treating abnormal hypermetropia from the age of 6 months. Br J Ophthalmol 1990;74(3):158–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Ingram RM, Walker C, Wilson JM, et al. A first attempt to prevent amblyopia and squint by spectacle correction of abnormal refractions from age 1 year. Br J Ophthalmol 1985;69(11):851–3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kulp MT, Holmes JM, Dean TW, et al. A randomized clinical trial of immediate versus delayed spectacles for moderate hyperopia in 1- and 2-year-olds. Ophthalmology 2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Donahue SP. Prescribing spectacles in children: a pediatric ophthalmologist’s approach. Optom Vis Sci 2007;84(2):110–4. [DOI] [PubMed] [Google Scholar]
- 18.Borchert MS, Varma R, Cotter SA, et al. Risk factors for hyperopia and myopia in preschool children: the Multi-Ethnic Pediatric Eye Disease and Baltimore Pediatric Eye Disease Studies. Ophthalmology 2011;118(10):1966–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Wang P, Yang Z, Li S, et al. Evaluation of MFRP as a candidate gene for high hyperopia. Mol Vis 2009;15:181–6. [PMC free article] [PubMed] [Google Scholar]
- 20.Varma R, Deneen J, Cotter S, et al. The multi-ethnic pediatric eye disease study: design and methods. Ophthalmic Epidemiol 2006;13(4):253–62. [DOI] [PubMed] [Google Scholar]
- 21.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):739–46, 46 e1-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Dirani M, Chan YH, Gazzard G, et al. Prevalence of refractive error in Singaporean Chinese children: the Strabismus, Amblyopia, and Refractive Error in Young Singaporean Children (STARS) study. Invest Ophthalmol Vis Sci 2010;51(3):1348–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Pai AS-I, Wang JJ, Samarawickrama C, et al. Prevalence and risk factors forvisual impairment in preschool children: the Sydney Paediatric Eye Disease Study. Ophthalmology 2011;118(8):1495–500. [DOI] [PubMed] [Google Scholar]
- 24.Prabakaran S, Dirani M, Chia A, et al. Cycloplegic refraction in preschool children: comparisons between the hand-held autorefractor, table-mounted autorefractor and retinoscopy. Ophthalmic Physiol Opt 2009;29(4):422–6. [DOI] [PubMed] [Google Scholar]
- 25.Sham WK, Dirani M, Chong YS, et al. Breastfeeding and association with refractive error in young Singapore Chinese children. Eye 2010;24(5):875–80. [DOI] [PubMed] [Google Scholar]
- 26.Castagno VD, Fassa AG, Carret ML, et al. Hyperopia: a meta-analysis of prevalence and a review of associated factors among school-aged children. BMC Ophthalmol 2014;14:163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Shulman JP, Weng C, Wilkes J, et al. Association of maternal preeclampsia with infant risk of premature birth and retinopathy of prematurity. JAMA Ophthalmol 2017;135(9):947–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Dai AI, Demiryurek S, Aksoy SN, et al. Maternal iron deficiency anemia as a rsk factor for the development of retinopathy of prematurity. Pediatr Neurol 2015;53(2):146–50. [DOI] [PubMed] [Google Scholar]
- 29.Greenland S, Finkle WD. A critical look at methods for handling missing covariates in epidemiologic regression analyses. Am J Epidemiol 1995;142(12):1255–64. [DOI] [PubMed] [Google Scholar]
- 30.Cotter S, Varma R, Tarczy-Hornoch K, et al. Risk factors associated with childhood strabismus: the Multi-Ethnic Pediatric Eye Disease and Baltimore Pediatric Eye Disease studies. Ophthalmology 2011. ;118(11):2251–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Kulp MT, Ying GS, Huang J, et al. Associations between hyperopia and other vision and refractive error characteristics. Optom Vis Sci 2014;91(4):383–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Jiang X, Tarczy-Hornoch K, Cotter SA, et al. Dilated eye examination among multi-ethnic preschool children. Invest Ophthalmol Vis Sci 2016;57(12):3095 [Google Scholar]
- 33.Scheiman M, Gwiazda J, Zhang Q, et al. Longitudinal changes in corneal curvature and its relationship to axial length in the Correction of Myopia Evaluation Trial (COMET) cohort. J Optom 2016;9(1):13–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Twelker JD, Mitchell GL, Messer DH, et al. Children’s ocular components and age, gender, and ethnicity. Optom Vis Sci 2009;86(8):918–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Stone RA, Wilson LB, Ying G-s, et al. Associations between childhood refraction and parental Smoking. Invest Ophthalmol Vis Sci 2006;47(10):4277–87. [DOI] [PubMed] [Google Scholar]
- 36.England LJ, Kendrick JS, Wilson HG, et al. Effects of smoking reduction during pregnancy on the birth weight of term infants. Am J Epidemiol 2001;154(8):694–701. [DOI] [PubMed] [Google Scholar]
- 37.Pathai S, Cumberland PM, Rahi JS. Prevalence of and early-life influences on childhood strabismus: Findings from the millennium cohort study. Arch Pediatr Adolesc Med 2010;164(3):250–7. [DOI] [PubMed] [Google Scholar]
- 38.Parker B, Connaughton VP. Effects of nicotine on growth and development in larval zebrafish. Zebrafish 2007;4(1):59–68. [DOI] [PubMed] [Google Scholar]
- 39.Källén B, Tornqvist K. The epidemiology of anophthalmia and microphthalmia in Sweden. Eur J Epidemiol 2005;20(4):345–50. [DOI] [PubMed] [Google Scholar]
- 40.Lange S, Probst C, Rehm J, Popova S. National, regional, and global prevalence of smoking during pregnancy in the general population: a systematic review and meta-analysis. Lancet Glob Health 2018;6(7):e769–e76. [DOI] [PubMed] [Google Scholar]
- 41.Collaco JM, Drummond MB, McGrath-Morrow SA. Electronic cigarette use and exposure in the pediatric population. JAMA pediatr 2015;169(2):177–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Babinsky E, Candy TR. Why do only some hyperopes become strabismic? Invest Ophthalmol Vis Sci 2013;54(7):4941–55. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Spiteri Cornish K, Hrabovsky M, Scott NW, et al. The short- and long-term effects on the visual system of children following exposure to maternal substance misuse in pregnancy. Am J Ophthalmol 2013;156(1):190–4. [DOI] [PubMed] [Google Scholar]
- 44.Wen G, Tarczy-Hornoch K, McKean-Cowdin R, et al. Prevalence of myopia, hyperopia, and astigmatism in non-Hispanic white and Asian children: Multi-Ethnic Pediatric Eye Disease Study. Ophthalmology 2013;120(10):2109–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Australian Bureau of Statistics. 2006 Census QuickStats. http://quickstats.censusdata.abs.gov.au/census_services/getproduct/census/2006/quickstat/0. 2007; Accessed Jun 18, 2018.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.