To assess correlations between associated factors and treatment outcome of isoametropic amblyopia attributable to high hyperopia in children and to evaluate changes in retinal thickness during amblyopic treatment.
Retrospective (visual outcome) and prospective (retinal thickness) study.
Isoametropic (spherical equivalent ≥5.00 diopters in both eyes) amblyopic subjects (n = 217) with corrected visual acuity (VA) ≤0.5 (20/40) for children <5 years old and VA ≤0.6 (20/32) for children 6-10 years old were included. Sixty-nine of these children had refractive accommodative esotropia. All subjects were treated by full-time wearing of optical correction. The magnitude of spherical equivalent hyperopia, age at first treatment, VA, treatment duration, and binocular alignment were analyzed. Thicknesses of the retinal macula and nerve fiber layer were measured using optical coherence tomography in selected amblyopic children (n = 27) during treatment and in control subjects (n = 31).
The mean follow-up was 28.8 months. The age at first optical correction ranged from 3 to 10 years. The mean VA improved from 0.38 to 0.82, with 74.7% having acuity ≥0.8 and 28.6% having acuity ≥1.0. VA improvement was positively correlated with duration of treatment and negatively correlated with age at first correction. Foveal thickness slightly decreased after treatment; however, it was not correlated with visual improvement.
Visual acuity of isoametropic amblyopia improved satisfactorily with spectacle correction and vision therapy. Treatment duration had the greatest impact on VA improvement. Age at first correction also influenced VA improvement but was not a good clinical predictor. Foveal thinning occurring with treatment was not correlated with visual improvement.
Amblyopia is the most common cause of reduced vision in children, affecting approximately 1.6% to 3.6% of the population. The causes for amblyopia are generally thought to be attributable to rearing in an abnormal visual environment associated with strabismus, refractive error, and visual form deprivation. When high but equal or nearly equal uncorrected refractive error is present in both eyes, isoametropic amblyopia occurs. Extracellular recordings from striate neurons in monkeys with bilateral amblyopia showed that the lack of binocular form stimulation early in life leaves the primate visual cortex with very few neurons excited by binocular stimulation. Additionally, the cortical neurons associated with the amblyopic eyes had blunted resolution. Besides the cortical deficit, cells of lateral geniculate nucleus of amblyopic eyes were smaller than normal. These alterations are sufficient reason to account for the low vision and poor stereopsis in patients with isoametropic amblyopia without esotropia.
Binocular high hyperopia is the most common risk factor for the development of isoametropic amblyopia, although the incidence reported varies greatly. There is evidence in the literature for a lower accommodation in bilateral hypermetropic amblyopic children than predicted measurements for their age. The foveal images are blurred to an equal degree and capacity to bring images into focus is limited because of under-accommodation. Emmetropization is a rapid phenomenon. Cross-sectional and longitudinal human studies suggest that most emmetropization takes place between 3 and 12 months of age. In human studies, the most highly hyperopic infants tend to not emmetropize effectively, especially those in excess of +5 diopters (D). These eyes chronically experience a high degree of hyperopic defocus, which initiates other visual system alterations that interfere with emmetropization, and end up with bilateral isoametropic amblyopia. While correction of hypermetropic errors is needed to obtain a clear image, visual acuity does not initially normalize after the refractive error is corrected. Further, there is often a real need to correct hyperopia to avoid accommodative esotropia.
Little information is available concerning visual outcome in high hyperopic isoametropic amblyopia. Fern evaluated 45 binocular high hyperopic (≥5.00 D) patients whose best-corrected visual acuity (BCVA) was poorer than 1.0 (20/20). She considered the magnitude of refraction to be the greatest influence on the visual acuity outcome. However, Ziylan and associates assessed 31 hyperopic (≥5.00 D) isoametropic amblyopic patients (visual acuity [VA] ≤0.4) and reported that the duration of optical correction was the only factor that was positively correlated with the VA outcome measurement. The discrepancy between these 2 studies may be attributable to the small number of subjects and different criteria of amblyopia each used. One goal of the present study was to use a relatively large sample of subjects to evaluate the factors affecting visual outcome in hyperopic isoametropic amblyopia.
Huynh and associates compared retinal thicknesses in 33 treated unilateral and binocular amblyopic subjects and 12 untreated similar subjects. They found that the fovea and macula were thicker in amblyopic eyes than in normal fellow eyes. Further, the differences between the 2 eyes were larger in untreated than in treated subjects. They proposed that foveal center thickening occurs after the onset of amblyopia and regresses with treatment. If so, we wondered if the macular thickness could affect the visual outcome of amblyopia treatment. Thus the other goal of the current study was to determine if the thickness of retinal macula and nerve fiber layer changes during amblyopia treatment and if these values affect the visual outcome of isoametropic amblyopia.
The present study included visual outcome and retinal thickness study. Visual outcome study is retrospective whereas retinal thickness study is prospective. The prospective study was approved by the Review Board of Wenzhou Medical College and was performed according to the tenets of the Declaration of Helsinki for research involving human subjects. Prior to enrolling their children in the study, the parents were informed about the purpose and methods of the study. Each parent then provided a signed informed consent agreement.
The charts of all patients seen at the Amblyopia and Strabismus Department of Wenzhou Medical College Affiliated Eye Hospital from January 2003 to December 2010 were retrospectively reviewed. The criteria for selection included the following: (1) spherical equivalent (SE) of hyperopia ≥5.00 D in both eyes; (2) interocular SE differences ≤1.00 D; (3) cylinder <1.50 D in both eyes; and (4) amblyopia in both eyes. For children younger than 5 years old the corrected VA was ≤0.5 and for children between 6 and 10 years of age the corrected VA was ≤0.6. Subjects who had undergone previous treatment for amblyopia or with evidence of ocular pathology were not included in this study. A total of 217 children met the inclusion criteria, including 69 children with refractive accommodative esotropia. All refractions were performed using retinoscopy after pupillary dilation with 1% atropine, twice per day for 3 days. Full dilated optical corrections were made in the strabismic subjects, and the sphere was reduced by 1.00-1.50 D in nonstrabismic children. The astigmatism for each subject was fully corrected.
All subjects could perform VA assessments at either the first or a subsequent visit. The patients wore the spectacles full time and were re-refracted every 3 months during therapy. Data were collected on the magnitude of hyperopia SE, age at correction, VA with Tumbling E charts, treatment duration, and binocular alignment.
Among these subjects, 27 amblyopic children who could cooperate enough to participate in examination by optical coherence tomography (OCT), along with 31 normal-vision control subjects, were selected at their initial visit for retinal thickness measurements. These amblyopic subjects were followed prospectively, and the data were included in the chart review. Clinical examinations included VA, retinoscopy after papillary dilation, slit-lamp examination, visuoscopy and an orthoptic evaluation including Hirschberg test, cover test, and extraocular movements. The IOL-Master (version 5.0; Carl Zeiss, Jena, Germany) was used to measure axial length. The thickness of the retinal macula and retinal nerve fiber layer (RNFL) were measured using OCT. All of these examinations were performed at first optical correction and every 3 months during treatment.
OCT (version 3.0; Zeiss Humphrey, Dublin, California, USA) was used to measure the thicknesses of the fovea and RNFL. The resolution for the OCT-3 was less than 10 μm in an axial scan and 20 μm in the transverse direction. Retinal thickness was defined as the distance between the reflection at the vitreoretinal interface and the anterior boundary of the reflective layer corresponding to the retinal pigment epithelium and choriocapillaris. RNFL thickness was defined as the number of pixels between the anterior and posterior edges of the RNFL. All measurements were performed by the same examiner.
A fast macular map scan protocol was used to scan the macular thickness. It consisted of 6 consecutive 6-mm radial scans that focused on the fovea. Each scan was rotated by 30 degrees. Retinal thickness mapping software was used to calculate the macular thickness. Data for macular retinal thickness were displayed in 3 concentric circles, with radii of 1 mm (fovea), 3 mm (inner ring), and 6 mm (outer ring). A “fast RNFL thickness (3.4)” scan protocol was used to calculate the RNFL thickness. It consisted of 3 consecutive 360-degree circular scans with a diameter of 3.4 mm centered on the optic disc. The RNFL thickness analysis program was used to evaluate the superior, inferior, temporal, nasal, and average RNFL thickness.
Statistical analysis was performed using SPSS software (version 17.0; SPSS Inc, Chicago, Illinois, USA). The data for the less ametropic, dominant eye were used for analysis to avoid the possibility of strabismus or anisometropia affecting the final visual acuity. Descriptive statistics for continuous variables were calculated as means ± standard deviations. Visual acuity data were converted to logarithms of the minimal angle of resolution (logMAR) for statistical calculation and analysis. The results were displayed as Snellen visual acuity, which was calculated using logMAR visual acuity equivalents. Associations between age at first optical correction, magnitude of hyperopia, treatment duration, retinal thickness, and VA improvement or visual outcome were examined by multivariate regression analysis. Mann-Whitney U test, independent t test, analysis of covariance (ANCOVA), and χ 2 test were used to evaluate differences between groups when appropriate. Differences of retinal thickness before and after treatment were compared by paired t tests. P values < .05 were considered to be significant.
A total of 217 subjects (136 male and 81 female) were enrolled, and the mean age at initial visit was 6.0 ± 1.7 years (range: 3-10 years). The mean refractive error was +7.71 ± 1.59 D (range: +5.00 to +13.50 D), and the VA at first correction ranged from 0.1 to 0.6 (mean 0.38). The largest number, 61 of 217 eyes (27.1%), had hyperopic refractive errors of 8-9 D.
Sixty-nine of these children had refractive accommodative esotropia. This gives an estimated prevalence of strabismus of 32% in children who presented at the eye clinic with at least 5.00 D of hyperopia in both eyes. The mean age at presentation of these children, 5.4 ± 1.6 years, was significantly lower than for the nonstrabismic amblyopic group, 6.3 ± 1.6 years ( P < .001). The mean magnitude of hyperopia was +7.52 ± 1.52 D, which did not differ significantly from that of the nonstrabismic group ( P = .257). The VA of the better eyes of strabismic and nonstrabismic groups was similar ( P = .855). Strabismus was slightly, but not significantly, more prevalent among children with SEs of refraction less than +8.00 D compared with children with refraction equal to or more than +8.00 D (34.5% vs 28.7%, P = .363 by χ 2 ).
The mean treatment duration of all subjects was 28.8 ± 21.7 months (range: 3-89 months). The mean VA at the last visit was 0.82 (range: 0.3-1.2), which was an improvement of more than 4 lines on Tumbling E charts from the initial visit ( P < .001). Of 217 children, 192 (88.5%) achieved a final VA of better than 0.6. A total of 162 (74.7%) had final VAs better than or equal to 0.8, and 62 (28.6%) had final VAs better than or equal to 1.0.
By multivariate regression analysis, we evaluated the influence of the magnitude of hyperopia and age at first correction on initial VA. The VA was negatively correlated with magnitude of hyperopia (r = −0.318, P < .001) and positively correlated with age at first correction (r = 0.317, P < .001). In addition, the mean VA of children with hyperopia less than 8.00 D was significantly better than that of children with hyperopia of 8.00 D or more (0.40 vs 0.35, P = .012). The mean VA in children of 6 years or younger was significantly worse than in children 7 to 12 years old (0.35 vs 0.43, P < .001).
Associations With Visual Improvement and Outcome
The presence of strabismus did not influence either the VA improvement or final VA of the better eye ( P ≥ .840, Table 1 ). We evaluated the influence of the magnitude of hyperopia, age at first correction, and treatment duration on the VA improvement by multivariate regression analysis. The VA improvement was positively correlated with treatment duration (r = 0.301, P < .001) but negatively correlated with age at first correction (r = −0.293, P < .001, Table 1 ). However, the magnitude of hyperopia was excluded from the equation ( P = .05). Consistent with multivariate regression analysis, the VA improvement in children with refraction equal to or more than +8.00 D was similar to children with less than +8.00 D ( P = .424). The improvement of VA in children 6 years old or younger, 4 lines, was greater than in children older than 6 years old, 3 lines ( P = .001).
|VA Improvement (logMAR)||Final VA (Snellen)|
(n = 116)
(n = 108)
(n = 109)
|Duration of treatment (mo)|
|Mean||24.8 a||33.5||0.301 b||22.4 a||35.1||0.316 b|
|Age at first optical correction (y)|
|Mean||6.4 a||5.6||−0.293 b||6.0||6.0||0.073|
|Magnitude of hyperopia (D)|
|Prevalence of strabismus (%)||31.03||32.67||31.78||31.82|
By multivariate regression analysis, we determined the correlation between final VA outcome and the magnitude of hyperopia, age at first correction, and treatment duration. Final VA was negatively correlated with magnitude of hyperopia (r = −0.243, P < .001) and positively correlated with treatment duration (r = 0.316, P < .001, Table 1 ). Age at first correction was excluded from the equation ( P = .478). Consistent with regression analysis, children with refractive errors less than +8.00 D had better final VAs, 0.86, than children with refractive errors equal to or more than +8.00 D, 0.78 ( P = .001).
We divided the subjects into 4 groups according to treatment duration: ≤12 months, 12-24 months, 24-36 months, and more than 36 months. The final VAs for treatment durations 12-24, 24-36, and >36 months were similar (0.85, 0.85, and 0.89, respectively), and all were greatly improved over the initial VAs and VAs at ≤12 months treatment (0.38 and 0.72, respectively). The similarity of final VAs for the last 3 treatment periods reflects the reduced rates of improvements that occurred during those periods compared with the initial rapid improvements at ≤1 year of treatment. The initial VA tended to increase with age; however, the final VAs, 0.8-0.9, for all age groups except those 9 years or older were similar to one another. For the oldest children, the final VA was 0.78, which was similar to ages 3-4 (0.81) and 5-6 (0.81) but lower than ages 7-8 (0.86), although the difference was not significant.
Thickness of the Retinal Macula and Nerve Fiber Layer
The thicknesses of the retinal macula and RNFL were measured by OCT in 27 amblyopic children and 31 normal-vision control subjects. OCT showed normal sensory retina, which consists of a series of distinct layers with a typical architecture in both groups. The mean age at first correction of the amblyopic group was 6.3 ± 0.9 years old (range 5-8 years old), which was significantly younger than the normal control group with the same age range (mean 7.0 ± 0.9 years old, P = .003, Table 2 ). The mean refraction was +7.48 ± 1.51 D in the amblyopic group at first correction and +0.37 ± 0.37 D in the normal control group ( P < .001, Table 2 ). The mean axial length was 20.41 ± 0.85 mm in the amblyopic group at the first visit and 22.72 ± 0.76 mm in the normal control group ( P < .001, Table 2 ). The initial VA of the amblyopic group ranged from 0.16 to 0.60 (mean 0.36), which was significantly less than the normal control group, 1.02 (range: 1.0-1.2, Table 2 ).
|Age (y)||Sex||Refraction (D)||Axial Length (mm)||VA (logMAR)|
|Amblyopic group||6.3 ± 0.9||17||10||+7.48 ± 1.51||20.41 ± 0.85||+0.44 ± 0.21|
|Control group||7.0 ± 0.9||16||15||+0.37 ± 0.37||22.72 ± 0.76||−0.01 ± 0.03|
The thicknesses of fovea and inner ring macula in the amblyopic group were 173.11 ± 14.02 μm and 267.31 ± 14.61 μm, respectively ( Table 3 ). These thicknesses were not significantly different from the control group before or after adjustment for age, sex, axial length, or refraction. For the amblyopic group, the thickness of the outer ring macula, 261.77 ± 12.65 μm, was significantly greater than for the normal group ( P = .006 by independent t test, Table 3 ). Similarly, the average RNFL thickness of the amblyopic group, 126.28 ± 10.19 μm, was greater than the control group ( P = .017 by independent t test, Table 3 ). However, after adjustment for age, sex, axial length, and refraction, the thickness differences between amblyopic and control children for both the outer ring macula and the RNFL were no longer significant ( P ≥ .720 by ANCOVA).