To compare visual performance between emmetropic and uncorrected moderately hyperopic preschool-age children without strabismus or amblyopia.
setting : Multicenter, institutional. patient or study population : Children aged 4 or 5 years. intervention or observation procedures : Visual functions were classified as normal or reduced for each child based on the 95% confidence interval for emmetropic individuals. Hyperopic (≥3.0 diopters [D] to ≤6.0 D in the most hyperopic meridian; astigmatism ≤1.50 D; anisometropia ≤1.0 D) and emmetropic status were determined by cycloplegic autorefraction. main outcome measures : Uncorrected monocular distance and binocular near visual acuity (VA); accommodative response; and near random dot stereoacuity.
Mean (± standard deviation) logMAR distance visual acuity (VA) among 248 emmetropic children was better than among 244 hyperopic children for the better (0.05 ± 0.10 vs 0.14 ± 0.11, P < .001) and worse eyes (0.10 ± 0.11 vs 0.19 ± 0.10, P < .001). Mean binocular logMAR near VA was better in emmetropic than in hyperopic children (0.13 ± 0.11 vs 0.21 ± 0.11, P < .001). Mean accommodative response for emmetropic children was lower than for hyperopic subjects for both Monocular Estimation Method (1.03 ± 0.51 D vs 2.03 ± 1.03 D, P < .001) and Grand Seiko (0.46 ± 0.45 D vs 0.99 ± 1.0 D, P < .001). Median near stereoacuity was better in emmetropic than in than hyperopic children (40 sec arc vs 120 sec arc, P < .001). The average number of reduced visual functions was lower in emmetropic than in hyperopic children (0.19 vs 1.0, P < .001).
VA, accommodative response, and stereoacuity were significantly reduced in moderate uncorrected hyperopic preschool children compared to emmetropic subjects. Those with higher hyperopia (≥4 D to ≤6 D) were at greatest risk, although more than half of children with lower magnitudes (≥3 D to <4 D) demonstrated 1 or more reductions in function.
Hyperopia is the most frequently occurring refractive error in young children, with a prevalence in preschool children of up to 12% for ≥+3.00 diopters (D), depending on the population and definition of hyperopia used. Based on population estimates from the U.S. Census Bureau and the prevalence of moderate hyperopia, an estimated 550 000 to 1.7 million preschool children each year are expected to have hyperopia of at least +3.00 D in the U.S.
The association between moderate hyperopia and the presence of ocular conditions such as amblyopia or strabismus has been documented previously for young children. The additional accommodative effort associated with uncorrected hyperopia and the resultant accommodative convergence places these children at an increased risk for esotropia, with the risk increasing with the magnitude of hyperopia. Hyperopic children are also at an increased risk for development of unilateral or bilateral amblyopia or reduced visual acuity (VA) that is dependent on the magnitude of the hyperopia. Even in the absence of esotropia or amblyopia, preschool children with greater than 3.25 D of uncorrected hyperopia have been shown to have an increased risk for other vision disorders, including anisometropia, astigmatism, and reduced stereoacuity, the latter worsening with higher magnitudes of hyperopia. Children with moderate to high amounts of uncorrected hyperopia may show differences in accommodative convergence–to-accommodation (AC/A) ratios. This represents less accurate accommodative responses (increased lag of accommodation) and an increase in variability of accommodative effort to maintain clarity for distance and near viewing compared to children with lower magnitudes of hyperopia or emmetropia.
The Vision in Preschoolers – Hyperopia in Preschoolers (VIP-HIP) Study Group compared early literacy in nonstrabismic, nonamblyopic 4- and 5-year-old children with emmetropia (hyperopia ≤1.0 D, anisometropia, astigmatism, and myopia all <1.0 D) vs moderate hyperopia (≥3.0 D to ≤6.0 D in the most hyperopic meridian of at least 1 eye, with astigmatism ≤1.5 D and anisometropia ≤1.0 D) and found that children with hyperopia ≥4.0 D alone or hyperopia ≥3.0 D to ≤6.0 D associated with reduced near VA (20/40 or worse) or reduced stereoacuity (240 sec arc or worse) performed worse on a test of early literacy as compared with emmetropic children who demonstrated average performance. On the other hand, children with moderate hyperopia and good near VA and stereoacuity performed similarly to emmetropic children. The present analysis provides important new insights into associations between moderate uncorrected hyperopia and near visual function in the absence of amblyopia or strabismus. It also shows the effect of increasing hyperopia by comparing these visual functions (distance VA, binocular near VA, accommodative response, and near stereoacuity) in this large population of moderately hyperopic and emmetropic, nonstrabismic, and nonamblyopic 4- and 5-year-old children enrolled in the VIP-HIP Study.
This is a secondary data analysis of the multicenter, prospective, cross-sectional VIP-HIP study data collected over a 3-year period (2011–2014) on uncorrected moderately hyperopic and emmetropic children without strabismus or amblyopia. The VIP-HIP Study adhered to the tenets of the Declaration of Helsinki and was approved prospectively by the appropriate local Institutional Review Boards associated with each study site. Prior to participation, parents or legal guardians of participating children provided written informed consent. This study was conducted in accordance with HIPAA regulations.
Children attending preschool or kindergarten identified as potentially having emmetropia or moderate hyperopia through a school vision screening were invited to participate in the VIP-HIP Study. All children were 4 or 5 years of age at the time of enrollment and eligibility eye examination. Children with an Individualized Education Program for developmental, educational, or behavioral issues; those with a known history of strabismus or amblyopia; or children who previously wore correction for refractive error were excluded.
Children were included in the analysis if they were found to have either hyperopia (defined as ≥3.0 D to ≤6.0 D in the most hyperopic meridian of at least 1 eye, with astigmatism ≤1.5 D and anisometropia ≤1.0 D) or emmetropia (defined as hyperopia ≤1.0 D, and astigmatism, anisometropia, and myopia each <1.0 D) on cycloplegic autorefraction. Children with strabismus or suspected amblyopia (VA worse than 20/40 or 2 or more lines worse than the contralateral eye) were excluded from this analysis and did not participate in the early literacy assessments that were the primary aim of the VIP-HIP Study.
Eligibility eye examinations were performed at 3 participating clinical centers (Pennsylvania College of Optometry at Salus University, Philadelphia, Pennsylvania, USA; The Ohio State University College of Optometry, Columbus, Ohio, USA; or New England College of Optometry, Boston, Massachusetts, USA) or in specially equipped Vision in Preschooler vans parked at the child’s school. The vans provided an environment similar to the examination rooms at the clinical centers and allowed greater access to participation for families who were unable to bring their child to a clinical center. The eye examination procedures were delineated in a Manual of Procedures and were performed by study-certified licensed eye care professionals experienced in working with young children. The order of testing was distance VA, binocular near VA, accommodative response, near stereoacuity, ocular alignment, cycloplegic refraction, and ocular health. Refraction was performed at the end of the eye examination so the child’s refractive status was unknown at the time of VA, accommodative, and stereoacuity testing.
Distance Visual Acuity Assessment
Distance VA thresholds were determined monocularly, with computerized crowded HOTV testing using the Amblyopia Treatment Study (ATS) protocol. Children were asked to point to each letter on a matching lap card. Children with reduced VA on initial testing were retested with full correction after cycloplegia to rule out amblyopia. All other testing was performed without correction.
Binocular Near Visual Acuity Assessment
Near VA was tested binocularly at 40 cm with crowded HOTV letters, using a modified ATS4 protocol, in order to obtain a measure of the child’s clarity of near vision under habitual conditions when performing near work. Children pointed to the letters with a double-edged Q-tip “pointer” that helped maintain their working distance, increased their interest in the task, and kept the cards free of fingerprints and smudges.
Accommodative response (target at 33 cm) was measured while the child viewed high-interest, colorful targets (a detailed sticker of popular cartoon characters) using Monocular Estimation Method (MEM) dynamic retinoscopy (2 readings of the horizontal meridian of the right eye) and Grand Seiko open field autorefraction (5 readings of the right eye). The examiner asked the child to describe what was on the sticker and asked questions about the characters to keep the child engaged during each task.
Assessment of sensory stereopsis at near was performed with the Preschool Assessment of Stereopsis with a Smile (PASS), a 2-alternative forced-choice random dot stereopsis test (blank, demonstration, and 480, 240, 120, 60, 40, and 30 sec arc test cards). While wearing colorful pediatric-size polarized glasses, children viewed a demonstration card with no disparity, “Card A,” that was initially presented paired next to a blank card (random dot pattern only) to ascertain testability. This was followed by paired presentations of the blank and a test card with a random-dot smile-face target beginning with Card B (480 sec arc) and continuing to Card G (30 sec arc). Children were required to pass 4 of 4 or 4 of 5 trials at each level. The order of testing was streamlined, to reduce fatigue and the number of trials required, by skipping the 240 and 40 sec arc cards (Cards C and F) unless the subsequent card was missed.
Cover testing at near was performed using fixation sticks with small, colorful, age-appropriate, high-interest accommodative targets. The targets were periodically changed to maintain the child’s interest in the task. Cover testing at a 10-foot distance was performed while the child looked at a small colorful character on a monitor. Magnitude of any phoria was neutralized with a prism bar or loose prism.
Following completion of binocular near VA, accommodative response, near stereoacuity, and ocular alignment testing, 1% cyclopentolate was administered and cycloplegic refraction was measured with the Retinomax after 30–45 minutes.
Assessments of pupils and anterior and posterior ocular health were performed to rule out the presence of ocular abnormalities or disease.
Distributions of demographic characteristics and the results of visual function testing were compared between refractive error groups using Fisher exact test for categorical variables and analysis of variance for continuous variables. Calculations for VA scores were performed using the logMAR (logarithm of the minimum angle of resolution) transformation; similarly, calculations of stereoacuity were performed using a logarithmic transformation of the seconds of arc. Near stereoacuity results of “Unable” for children who could not complete the demonstration card were considered missing. A score of 1600 seconds of arc was assigned when children could complete only the demonstration card. The closest meridian to the target was used in the analysis of accommodative response by Grand Seiko, which was selected for assessment of accommodative response because Grand Seiko has been shown to be an effective and objective measure of accommodative response in previous clinical trials, such as The Collaborative Longitudinal Evaluation of Ethnicity and Refractive Error (CLEERE) and Correction of Myopia Evaluation Trial 2 (COMET2). Visual functions were considered reduced if they were worse than the limit of the 95% confidence interval in the emmetropic subjects; specifically, accommodative lag >1.35 D, binocular near VA 20/40 or worse, and near stereoacuity 240 sec arc or worse. A two-sided P value of < .05 is considered to be statistically significant.
Eight hundred and fifty-eight children were examined and a total of 492 children (mean age = 58 months) were confirmed to have either emmetropia (n = 248) or hyperopia (n = 244) in the absence of amblyopia or strabismus. The majority of children were enrolled in Head Start (n = 439; 89%). There were no significant differences between the hyperopic and emmetropic children in mean age, sex, race, ethnicity, parent or guardian educational level, or percent enrolled in Head Start ( Table 1 ). The mean (± SD) cycloplegic refraction of the most hyperopic meridian was +3.78 (± 0.81) D for the hyperopic children and +0.51 (± 0.48) D for the emmetropic children.
N = 248
N = 244
|Age in months, n (%)|
|48–53||58 (23%)||73 (30%)|
|54–59||86 (35%)||79 (32%)|
|60–65||80 (32%)||72 (30%)|
|66 to <72||24 (10%)||20 (8%)|
|Mean (SD)||58.7 (5.5)||58.0 (5.8)||.43|
|Sex, n (%)|
|Male||128 (52%)||114 (47%)|
|Female||120 (48%)||130 (53%)||.28|
|Ethnicity and race, n (%)|
|Non-Hispanic black||150 (60%)||140 (57%)|
|Non-Hispanic white||20 (8%)||28 (11%)|
|Hispanic||61 (25%)||63 (26%)|
|Other or unknown||17 (7%)||13 (5%)||.53|
|Education level of parent or caregiver, n (%)|
|Less than high school||18 (7%)||27 (11%)||.35|
|High school||92 (37%)||102 (42%)|
|Some college||55 (22%)||39 (16%)|
|2-year college||21 (8%)||22 (9%)|
|4-year college||23 (9%)||23 (9%)|
|Graduate degree||19 (8%)||12 (5%)|
|Unknown||20 (8%)||19 (8%)|
|Preschool/Kindergarten, n (%)|
|Head Start||224 (90%)||215 (88%)|
|Other preschool/kindergarten||24 (10%)||29 (12%)||.43|
|Most hyperopic meridian, more hyperopic eye, diopters, mean (SD)||0.51 (0.48)||3.78 (0.81)||—|
|Spherical equivalent, more hyperopic eye, diopters, mean (SD)||0.37 (0.50)||3.47 (0.81)||—|
|Anisometropia, diopters, mean (SD)||0.17 (0.19)||0.40 (0.30)||<.001|
|Astigmatism, more astigmatic eye, diopters, Mean (SD)||0.40 (0.24)||0.72 (0.41)||<.001|
a P values for sphere and equivalent are not provided because they define the 2 groups. The eligibility criterion for astigmatism was different for emmetropic (<1 diopter) and hyperopic subjects (≤1.5 diopters); thus they were unlikely to be similar.
All children were able to complete testing on each of the visual functions assessed except for 1 child with hyperopia who was unable to complete testing on the PASS.
Differences in Visual Function Between Hyperopic and Emmetropic Children
Differences in visual function (VA, accommodative response, and stereoacuity) are shown in Table 2 and described as follows:
|Visual Function||Emmetropic (N = 248)||Hyperopic (N = 244)||P Value|
|Distance visual acuity, better eye|
|Better than 20/20||35 (14%)||10 (4%)|
|20/20||98 (40%)||43 (18%)|
|20/25||75 (30%)||72 (30%)|
|20/32||34 (14%)||84 (34%)|
|20/40||6 (2%)||35 (14%)|
|Median (25%, 75%), Snellen||20/20 (20/20, 20/25)||20/25 (20/25, 20/32)|
|Mean (SD), logMAR||0.05 (0.10)||0.14 (0.11)||<.001|
|Distance visual acuity, worse eye|
|Better than 20/20||18 (7%)||1 (0%)|
|20/20||67 (27%)||20 (8%)|
|20/25||87 (35%)||64 (26%)|
|20/32||57 (23%)||83 (34%)|
|20/40||19 (8%)||74 (30%)|
|Worse than 20/40||0 (0%)||2 (1%)|
|Median (25%, 75%), Snellen||20/25 (20/20, 20/32)||20/32 (20/25, 20/40)|
|Mean (SD), logMAR||0.10 (0.11)||0.19 (0.10)||<.001|
|Near visual acuity, binocular|
|20/20||59 (24%)||13 (5%)|
|20/25||95 (38%)||69 (28%)|
|20/32||67 (27%)||77 (32%)|
|20/40||22 (9%)||64 (26%)|
|Worse than 20/40||5 (2%)||21 (9%)|
|Median (25%, 75%), Snellen||20/25 (20/25, 20/32)||20/32 (20/25, 20/40)|
|Mean (SD), logMAR||0.13 (0.11)||0.21 (0.11)||<.001|
|Accommodative response, by Monocular Estimation Method, diopters|
|0 to 0.5||58 (23%)||19 (8%)|
|>0.5 to 1||84 (34%)||26 (11%)|
|>1 to 1.5||75 (30%)||37 (15%)|
|>1.5 to 2.01||22 (9%)||53 (22%)|
|>2.01||9 (4%)||109 (45%)|
|Median (25%, 75%)||1.00 (0.63, 1.38)||2.00 (1.38, 2.50)|
|Mean (SD)||1.03 (0.51)||2.03 (1.03)||<.001|
|Accommodative response, by Grand Seiko, diopters|
|0 to 0.5||161 (65%)||93 (38%)|
|>0.5 to 1||62 (25%)||66 (27%)|
|>1 to 1.35||18 (7%)||26 (11%)|
|>1.35 to 2||3 (1%)||30 (12%)|
|>2||4 (2%)||29 (12%)|
|Median (25%, 75%)||0.32 (0.16, 0.63)||0.67 (0.34, 1.31)|
|Mean (SD)||0.46 (0.45)||0.99 (1.00)||<.001|
|Stereoacuity by Preschool Assessment of Stereopsis with a Smile, seconds of arc|
|Unable||0 (0%)||1 (0%)|
|Demonstration card only||3 (1%)||30 (12%)|
|480||2 (1%)||27 (11%)|
|240||8 (3%)||41 (17%)|
|120||50 (20%)||81 (33%)|
|60||38 (15%)||24 (10%)|
|40||87 (35%)||24 (10%)|
|30||60 (24%)||16 (7%)|
|Median (25%, 75%)||40 (40, 120)||120 (60, 240)||<.001 a|
a For stereoacuity the median and interquartile range are provided because the data are skewed, and the P value calculation for the comparison of means was calculated on log-transformed seconds of arc.
Monocular distance and binocular near visual acuity
Mean (± SD) logMAR uncorrected distance VA was significantly better for emmetropic than for hyperopic children for both the better eye (0.05 ± 0.10 vs 0.14 ± 0.11, P < .001) and the worse eye (0.10 ± 0.11 vs 0.19 ± 0.10, P < .001). More emmetropic than hyperopic children were able to achieve a VA of at least 20/20 in both the better-seeing eye (133 [54%] vs 53 [22%], P < .001) and the worse-seeing eye (85 [34%] vs 21 [9%], P < .001). The mean (± SD) logMAR binocular near VA of emmetropic children was also significantly better than for hyperopic children (0.13 ± 0.11 vs 0.21 ± 0.11, P < .001). Binocular near VA of 20/25 or better was attained by 154 (62%) of the emmetropic children and 82 (33%) of the hyperopic children ( P < .001).
The mean (± SD) accommodative response for a 33 cm viewing distance and a naturalistic cartoon image was significantly more accurate (closer to the +3.00 D optical demand for the 33 cm viewing distance) in emmetropic than in hyperopic children. The difference or “lag” between the optical demand (+3.00 D) and the actual accommodative response was significantly different between emmetropic and hyperopic children for both MEM (1.03 D ± 0.51 D vs 2.03 D ± 1.03 D lag on average, P < .001) and for Grand Seiko (0.46 D ± 0.45 D vs 0.99 D ±1.00 D lag on average, P < .001). In addition to a larger lag, there was greater variability in lag across hyperopic children (as indicated by a twice larger SD in hyperopic children than in emmetropic children) on both MEM and Grand Seiko ( P < .001).
Median [interquartile range] stereoacuity was better in emmetropic than in hyperopic children (40 [40, 120] sec arc vs 120 [60, 240] sec arc, P < .001). More emmetropic than hyperopic children (147 [59%] vs 40 [17%], P < .001) attained 1 of the 2 best tested levels of stereoacuity (30 or 40 sec arc), whereas more hyperopic than emmetropic subjects (99 [40%] vs 13 [5%], P < .001) had 240 sec arc or worse stereoacuity.
Reduced near visual function
The mean number of reduced visual functions was higher in hyperopic than in emmetropic children (1.00 vs 0.19, P < .001), with more hyperopic individuals having 1 or more reduced near visual functions compared to emmetropic children (157 [64%] vs 42 [17%], P < .001, Table 3 ).
|Sphere||N||Number of Reduced Visual Functions||P Value a|
|Emmetropic||248||206 (83%)||37 (15%)||5 (2%)||0 (0%)||0.19 (0.03)||<.001|
|Hyperopic (all)||244||87 (36%)||87 (36%)||54 (22%)||16 (7%)||1.00 (0.06)|
|Hyperopic (by sphere):|
|3–3.25 D||102||52 (51%)||35 (34%)||12 (12%)||3 (3%)||0.67 (0.08)||<.001|
|3.5–3.75 D||57||20 (35%)||20 (35%)||16 (28%)||1 (2%)||0.96 (0.11)|
|≥4–6 D||85||15 (18%)||32 (38%)||26 (31%)||12 (14%)||1.41 (0.10)|
Effect of increasing magnitude of hyperopia
Within the hyperopic group (3.0 to ≤6.0 D), as the magnitude of the refractive error increased, the mean number of reduced visual functions more than doubled from 0.67 (for 3.0–3.25 D) to 1.41 (for 4.0–6.0 D) ( P < .001) and the number of children with 1 or more reduced visual functions also increased from 50 of 102 (49%) (for 3.0–3.25 D) to 37 of 57 (65%) (for 3.5–3.75 D) to 70 of 85 (82%) (for 4.0–6.0 D) ( P < .001) ( Table 3 ).
Decreasing visual function was also found on average with increasing hyperopia (3.0 to <4.0 D vs 4.0–6.0 D), as shown in Table 4 . Differences in the means for the 3 VA measures were approximately 0.05 logMAR (about a half line on Snellen) worse for children with 4.0–6.0 D hyperopia vs those with hyperopia of 3.0 to <4.0 D (all P < .01). Similarly, accommodative lag was on average 0.45 D greater by Grand Seiko and 0.86 D greater using MEM dynamic retinoscopy ( P < .001 for both comparisons). The median near stereoacuity of the children with ≥4.0 D hyperopia was 1 card worse than that of children with 3.0 to <4.0 D hyperopia (240 vs 120 arc sec; P < .001).