Longitudinal development of ocular biometric components and refractive error in hyperopic children with infantile versus late-onset accommodative esotropia

Abstract

Purpose:

To examine the developmental patterns of refractive error and optical components in hyperopic children with infantile (onset ≤12 months of age) accommodative or late-onset (18–48 months of age) accommodative esotropia.

Methods:

This prospective longitudinal study included children with infantile (n = 34) or late-onset (n = 63) accommodative esotropia. Axial length (AL), anterior chamber depth (ACD), lens thickness (LT) and keratometry (K1, K2) were obtained with a Lenstar LS 900. Lenstar measures were recorded <6 months after cycloplegic spherical equivalent refraction (SER) was derived. An initial examination was conducted at 5.8 ± 1.5 years of age, with a follow-up duration of 4.8 ± 0.8 years. A linear mixed-effects model was used to estimate the rate of individual development for each ocular component and SER, and to compare the two groups.

Results:

All biometric components changed with age. The rates of change with age for SER and AL were significantly different between the infantile and late-onset groups (SER: −0.18 vs. −0.12D/year, p < 0.001; AL: 0.16 vs. 0.14 mm/year, p < 0.01). The rate of change with age of the AL/CR ratio was significantly different between the infantile and late-onset groups (0.019 vs. 0.016, p < 0.001). No significant differences in the rates of change in ACD, LT, K1 or K2 were identified.

Conclusions:

Major ocular biometric components in children continue to mature in both infantile and late-onset accommodative esotropia. Annual change in axial length is smaller in late-onset accommodative esotropia than for infantile accommodative esotropia, consistent with less change in the SER with age.

INTRODUCTION

Emmetropisation often fails in children with strabismus, particularly those with hyperopia and accommodative esotropia.^1,2 Hyperopic refractive errors persist in children with accommodative esotropia (ET), with only a slight decrease after 7 years of age.^3,4 Furthermore, the pattern of change in refractive error with age differs between children with infantile (onset ≤12 months of age) versus late-onset (onset 18–48 months of age) accommodative ET. Children with infantile-onset maintain a stable level of hyperopia until 7 years of age, at which point it begins to decrease, whereas children with the late-onset condition exhibit little change in hyperopia up to 12 years of age.³

To date, little is known about how the failure of emmetropisation relates to changes in optical components with age in these hyperopic children. Prior work has predominantly focused on axial length (AL). Uretmen et al. investigated 112 children with accommodative esotropia aged 3–14 years, revealing a weak correlation between AL and age.⁵ Similarly, Ahn et al. tracked 27 Korean children with accommodative esotropia and 40 hyperopic children without esotropia for 3 years, observing that all children had larger AL at the final visit, indicating an increase with age.⁶ Our previous cross-sectional study observed that ocular biometric parameters changed with age in hyperopic children (N = 212), whereby axial growth was offset by a reduction in corneal power.⁷ A recent retrospective longitudinal study involving Chinese hyperopic children (≥+5.00D) observed an increase in AL alongside a decrease in spherical equivalent refraction (SER).⁸ However, their cohort included both children with and without accommodative ET, and it remains uncertain whether these two types of accommodative ET exhibit different patterns of growth.

Based on this literature review, it is likely that the ocular biometric components continue to grow with age and may compensate for one another, but one cannot determine whether the growth pattern differs across different types of accommodative ET. Without understanding how ocular biometric factors change over time, it is difficult to optimise clinical management and to design personalised treatment strategies, such as prescribing corrective lenses or implementing interventions to alter eye growth in these children.

Building upon our previous observations on refractive error development in children with infantile (onset ≤12 months of age) and late-onset (18–48 months of age) accommodative ET and the literature, we hypothesise that major ocular biometric components develop differently between infantile and late-onset accommodative ET. To test this hypothesis, this investigation will compare developmental rates of these components in each group. This prospective study aimed to: (1) investigate the pattern of refractive errors and optical component development in these two groups of children with ET utilising a longitudinal mixed-effects linear model and (2) understand the relationship between refractive errors and optical components throughout the development process.

METHODS

This research protocol observed the tenets of the Declaration of Helsinki, was approved by the Institutional Review Board of the University of Texas Southwestern Medical Center and conformed to the requirements of the United States Health Insurance Portability and Privacy Act. Participants were referred to the Retina Foundation of the Southwest by 16 Dallas-Fort Worth paediatric ophthalmologists. Informed consent was obtained from each participant’s parent or guardian after explanation of the nature and possible consequences of the study. Assent was obtained if the child was 10 years of age or older.

Participants

Eligible children ≥3 years of age, with hyperopia ≥+1.00D in at least one eye were identified as having infantile or late-onset accommodative ET by their referring paediatric ophthalmologist, based on the American Academy of Ophthalmology Preferred Practice Patterns for Pediatric Eye Evaluations⁹ and for Esotropia and Exotropia.¹⁰

Onset of accommodative ET

Based on the onset of accommodative ET documented in their medical records, children were classified as having infantile accommodative ET if onset occurred at ≤12 months of age and late-onset accommodative ET if onset occurred at 18–48 months of age.³ Note, diagnoses made between 12 and 18 months of age were excluded to avoid confusion. Additionally, children were excluded if they had a concurrent ophthalmic or systemic disease, developmental delay or were born preterm (≤36 weeks).

Patient care

Patient care was managed at the discretion of the referring ophthalmologist, following the guidelines of the American Academy of Ophthalmology Preferred Practice Pattern.¹⁰ Hyperopic refractive errors were corrected with spectacles, and amblyopia treatment (including patching and/or atropine penalisation) was administered if deemed necessary.

Ocular biometric components

Ocular biometric components were assessed under natural pupil conditions by trained examiners with the Lenstar LS 900 (Haag Streit Diagnostics, haag-streit.com/). The following parameters were measured: axial length (AL, mm), anterior chamber depth (ACD, mm), lens thickness (LT, mm), central corneal thickness (CCT, microns) and anterior corneal power for the flattest and steepestmeridians, denoted as $\mathrm{K}1$ and $\mathrm{K}2$, respectively (D). In addition, mean corneal power (K, D) was calculated as $\mathrm{K}=(\mathrm{K}1+\mathrm{K}2)/2$, and the mean anterior corneal radius (CR, mm) was calculated using the formula: $\mathrm{C}\mathrm{R}=1000\mathrm{*}(1.3375-1)/\mathrm{K}$. Further, the axial length-corneal radius ratio (AL/CR), which has been reported to have a stronger correlation with SER than AL alone,¹¹ was also calculated. Ocular biometric data from children who had a minimum follow-up of 2 years and up to <13 years, three or more visits and a cycloplegic refraction performed by the referring ophthalmologist within 6 months of the optical component measurements were included for analysis.

Cycloplegic refraction

Medical records were obtained from the referring paediatric ophthalmologist to extract the cycloplegic refraction, where it was recorded as sphere (S), plus cylinder (C) and axis (α). Using a custom spreadsheet (Microsoft Excel, microsoft.com), the refractive findings were converted into power vector components, that is, M (SER), J₀ (positive J₀ indicates with-the-rule [WTR] astigmatism), negative J₀ (indicates against-the-rule [ATR] astigmatism) and J₄₅ (oblique astigmatism: positive J₄₅ indicates 135° astigmatism while negative J₄₅ indicates 45° astigmatism).¹²

Sensory function

Visual acuity was measured monocularly at each visit using an age appropriate logMAR letter chart; generally the Amblyopia Treatment Study (ATS) HOTV chart for children <7 years old and the ATS-Early Treatment of Diabetic Retinopathy Study (ETDRS) for children ≥7 years of age.^13,14 Stereoacuity was measured with the Stereo Optical Randot Preschool Stereotest (stereooptical.com/products/stereotests-colour-tests/randot-preschool/) with polarised 3D glasses and habitual spectacles.

Data analysis was performed and plotted using R 4.2 Statistics (R Core Team. R: R-project.org). Descriptive statistics are presented as mean ± standard deviation (SD). Comparison of prevalences was conducted with the Chi-squared test. Comparison among groups, the initial and final visit and sex for each parameter (including both eyes) were conducted using generalised estimating equation (GEE) analysis, which is an extension of generalised linear models (GLM).¹⁵

To estimate the rate of individual change for each variable (SER, J₀, J₄₅ and the ocular components) with age, a linear mixed-effects model was applied, using the iterative weighted least square (IWLS) method. Not all of the children provided data in every age group, and not every child was followed up to 12 years of age. Under the assumption that data were missed at random, the mixed-effects model was selected, because it is one of the preferred methods when data is missing at random. The model provides comparisons within each group as well as between the infantile and late-onset groups. The mixed-effects model uses longitudinal information from both eyes of each individual.¹⁶ Onset, age, sex and group were treated as a fixed effect and individual and interocular correlation as a random effect.^15,16 Thus, the mixed-effects model explicitly accounts for the interocular correlations by adding a random effect (such as a random intercept, assuming that the intercept is the same for both eyes of a subject, but different across different subjects) in addition to the fixed effects of interest that may be either person- or eye-specific. Similar analyses were conducted for astigmatism (J₀ and J₄₅) and the ocular biometric components. Further, stereoacuity data labelled as ‘nil’ (representing missing or zero values) were replaced with a value of 10,000. The transformed data were then subjected to a log transformation (base 10) to normalise the distribution. Therefore, GEE analysis could be applied.

RESULTS

Demographics and general information

In this prospective longitudinal study, a total of 113 children with accommodative esotropia were enrolled (35 infantile onset and 78 late onset). A total of 97 children who met the criteria (minimum follow-up of 2 years, three or more visits) were included for analysis in either the infantile group (n = 34) or the late-onset group (n = 63). Table 1 summarises the demographic distribution of the cohort studied. The majority of the cohort was non-Hispanic white (76.3%), with no significant difference between the two groups in terms of racial distribution (χ² = 2.96, p = 0.08). Both groups had a higher proportion of female participants than males, but no significant difference in sex distribution was observed between the two groups (χ² = 0, p > 0.99). On average, participants were 5.8 ± 1.5 years old at their initial ocular biometry visit, with the age at the initial visit being similar in both groups (t = −1.5, p = 0.10). Each participant completed 7 ± 3 ocular biometry visits with an average follow-up duration of 4.8 ± 0.8 years (range: 2.3–5.7 years).

TABLE 1.

Demographic characteristics of the two esotropic groups based on age of onset.

	Total, n = 97	Infantile group, n = 34	Late-onset group, n = 63	Comparison of the two groups
Sex, N (%)
Female	57 (58.8%)	20 (58.8%)	37 (58.7%)	χ2 = 0, p > 0.99
Male	40 (41.2%)	14 (41.2%)	26 (41.3%)
Race/Ethnicity N (%)
African American	4 (4.1%)	2 (5.75%)	2 (3.2%)
Asian	4 (4.1%)	2 (5.75%)	2 (3.2%)
American Indian, Alaskan Native	0	0	0
Non-Hispanic White	74 (76.3%)	22 (65.0%)	52 (82.5%)	χ2 = 2.96, p = 0.08
Hispanic White	8 (8.2%)	5 (14.7%)	3 (4.8%)
Other	7 (5.2%)	3 (8.8%)	4 (6.3%)
Age at first biometric measurement (year)
Mean (SD) (min, max)	5.8 (1.5) (3.4 to 10.6)	5.4 (1.2) (3.4, 7.9)	5.9 (1.6) (3.9, 10.6)	t = −1.88, p = 0.06
Age at final biometric measurement (year)
Mean (SD) (min, max)	10.5 (1.7) (6.2 to 12.9)	10.8 (1.5) (7.1, 12.8)	10.4 (1.8) (6.2, 12.9)	t = 1.10, p = 0.27

Initial visit versus final visit

To characterise the refractive development of the two groups, data were compared between the initial and final visits to discern major trends (Figure 1). Table 2 summarises the group mean and standard deviations for refractive errors and biometric parameters in each eye separately, p-values comparing the initial and final visits, as well as comparing the two groups using GEE analysis. Overall (combining both groups), AL and ACD were significantly larger and SER and LT were significantly smaller at the final visit compared with the initial visit. In addition, the GEE analysis also identified significant differences between the infantile and late-onset groups in the amount of change that occurred between the initial and final visits for SER, AL, ACD, LT, K1 and K. Overall, the AL/CR ratio was significantly higher at the final visit, and the GEE analysis identified a significantly larger increase in AL/CR in the infantile compared with the late-onset group. Taken together, these results indicate that the infantile group showed more eye development than the late-onset group. There were no changes in J₀, J₄₅, or CCT either between the initial and final visits or between the two groups.

FIGURE 1.

SER (a) and ocular biometry components (b–g) at the initial and final visits in two groups (late: Late-onset; inf: Infantile). The cyan colour indicates the infantile group; the red colour indicates the late-onset group. ACD, Anterior chamber depth; AL, Axial length; AL_CR ratio, Axial length-corneal radius ratio or AL/CR; K1 and K2, Anterior corneal power for the flattest and steepest meridians; LT, Lens thickness; SER, Spherical equivalent refraction.

TABLE 2.

Comparison of refractive and ocular biometric components (mean ± standard deviation) at the initial and final visits between the two groups.

	Infantile group, N = 34, OD		Infantile group, N = 34, OS		Late-onset group, N = 63, OD		Late-onset group, N = 63, OS		p-Value (initial vs. final)	p-Value (groups)
	Initial	Final	Initial	Final	Initial	Final	Initial	Final
Refractive errors
SER(D)	3.76 ± 2.07	2.73 ± 2.56	3.95 ± 2.07	3.06 ± 2.38	4.89 ± 1.85	4.37 ± 1.85	4.84 ± 2.17	4.30 ± 2.01	p = 0.001*	p < 0.001*
J0(D)	0.28 ± 0.36	0.25 ± 0.48	0.29 ± 0.44	0.37 ± 0.64	0.34 ± 0.43	0.41 ± 0.58	0.32 ± 0.43	0.41 ± 0.50	p = 0.22	p = 0.16
J45(D)	−0.04 ± 0.22	−0.03 ± 0.18	0.03 ± 0.28	0.01 ± 0.35	0.05 ± 0.33	0.06 ± 0.40	0.01 ± 0.29	−0.02 ± 0.34	p = 0.82	p = 0.25
Measured biometric parameters
Axial length (mm)	21.50 ± 0.98	22.44 ± 0.97	21.36 ± 0.94	22.26 ± 0.97	21.20 ± 0.96	21.86 ± 1.05	21.19 ± 1.06	21.91 ± 1.09	p < 0.001*	p = 0.001*
Anterior chamber depth (mm)	3.31 ± 0.32	3.61 ± 0.26	3.40 ± 0.28	3.57 ± 0.28	3.23 ± 0.28	3.48 ± 0.26	3.19 ± 0.40	3.50 ± 0.28	p < 0.001*	p = 0.003*
Lens thickness (mm)	3.72 ± 0.29	3.47 ± 0.22	3.72 ± 0.28	3.48 ± 0.19	3.72 ± 0.34	3.52 ± 0.25	3.81 ± 0.48	3.59 ± 0.33	p < 0.001*	p = 0.02*
Central cornea thickness (μm)	554.00 ± 37.52	559.91 ± 32.89	553.55 ± 36.13	560.81 ± 31.55	559.81 ± 35.82	562.56 ± 37.38	558.66 ± 36.57	564.84 ± 36.48	p = 0.16	p = 0.21
Least corneal Power (D)	42.82 ± 1.49	42.59 ± 1.52	42.88 ± 1.67	42.55 ± 1.53	42.25 ± 1.63	42.05 ± 1.60	42.35 ± 1.68	42.19 ± 1.62	p = 0.20	p = 0.005*
Greatest corneal power (D)	44.08 ± 1.60	44.03 ± 1.42	44.37 ± 1.64	44.24 ± 1.50	44.00 ± 1.68	43.84 ± 1.64	43.95 ± 1.60	43.88 ± 1.61	p = 0.54	p = 0.13
Calculated biometric parameters
Mean corneal power (D)	43.45 ± 1.47	43.31 ± 1.39	43.62 ± 1.53	43.39 ± 1.41	43.13 ± 1.59	42.94 ± 1.52	43. 51 ± 1.57	43.04 ± 1.52	p = 0.31	p = 0.02*
AL/CR ratio	2.76 ± 0.11	2.88 ± 0.12	2.75 ± 0.11	2.86 ± 0.11	2.70 ± 0.10	2.78 ± 0.12	2.70 ± 0.11	2.79 ± 0.12	p < 0.001*	p < 0.001*

Note: Generalised Estimating Equations (GEE) analysis was used to compare the two groups and initial and final visits after adjusting with sex; data from both eyes were included.

Abbreviations: AL/CR, axial length to corneal radius ratio; OD, right eye; OS, left eye; SER, spherical equivalent refraction.

^*indicates statistically significant difference.

Longitudinal models of the change of SER and optical components

Having demonstrated that several ocular biometric parameters change with age between the initial and final visits, the rate of the change of each parameter was compared between the infantile and late-onset groups. To illustrate the trend of change, the follow-up data were normalised against the initial visit. Figure 2 shows individual growth curves along with the mean change of SER and AL with age for the infantile and late-onset groups. Table 3 presents the mixed-effects model for the change of the ocular biometric components in both groups. The ‘Age’ column shows whether the parameter changed with age; the ‘Group Late-onset’ column shows if the late-onset group differed from the infantile group; the ‘Sex Male’ column shows the effect of sex while the ‘Age × Group Late-onset’ column shows if the rate of the change differed between two groups.

FIGURE 2.

Change of spherical equivalent refraction (SER) (a) and change of axial length (AL) (b) from the baseline and their best-fit models plotted against age for individuals of two groups (late: Late-onset; inf: Infantile). The left column is for the right eye (OD), and the right column is for the left eye (OS). The cyan colour indicates the infantile group; the red colour indicates the late-onset group. The fine dashed lines indicates individual longitudinal data; the solid line indicates the best-fit linear model for each group. The shaded areas indicate 95% confidence intervals (CI).

TABLE 3.

Summary of mixed-effects model results for the change of refractive and ocular biometric components between the two groups.

	Infantile group, N = 34	Late-onset group, N = 63	Linear mixed-effects model fit by maximum likelihood
			Age	Group, late-onset	Sex, male	Age × group late-onset
Refractive errors
Change of SER (D)	SER = 1.00 −0.18*Age +0.08 (Male)	SER = 0.65 −0.12*Age +0.08 (Male)	−0.18 (p < 0.001*)	−0.35 (p = 0.045)	0.08 (p = 0.48)	0.06 (p < 0.001*)
Change of J0	J0 = 0.003 −0.0009*Age +0.07 (Male)	J0 = −0.17 +0.026*Age +0.07 (Male)	−0.0009 (p = 0.84)	−0.17 (p = 0.007*)	0.07 (p = 0.08)	0.026 (p < 0.001*)
Change of J45	J45 = 0.048 −0.004*Age +0.02 (Male)	J45 = −0.018 +0.0002*Age +0.07 (Male)	−0.0047 (p = 0.30)	−0.066 (p = 0.25)	0.02 (p = 0.50)	0.0049 (p = 0.39)
Measured biometric parameters
Change of axial length (AL, mm)	AL = −0.84 +0.16*Age −0.02 (Male)	AL = −0.82 +0.14*Age −0.02 (Male)	0.16 (p < 0.001*)	0.02 (p = 0.73)	−0.02 (p = 0.71)	−0.02 (p = 0.002*)
Change of anterior chamber depth (ACD, mm)	ACD = −0.30 +0.06*Age −0.04 (Male)	ACD = −0.25 +0.05*Age −0.04 (Male)	0.06 (p < 0.001*)	0.05 (p = 0.54)	−0.04 (p = 0.51)	−0.009 (p = 0.25)
Change of lens thickness (LT, mm)	LT = 0.27 −0.05*Age −0.04 (Male)	LT = 0.31 −0.06*Age −0.04 (Male)	−0.05 (p < 0.001*)	0.04 (p = 0.64)	−0.04 (p = 0.46)	−0.007 (p = 0.48)
Change of central cornea thickness (CCT, pm)	CCT = −9.1 +1.7*Age −2.0 (Male)	CCT = −8.4 +1.4*Age −2.0 (Male)	1.7 (p < 0.001*)	0.7 (p = 0.80)	−2.0 (p = 0.89)	−0.27 (p = 0.37)
Change of least corneal Power (K1, D)	K1 = 0.21 −0.03*Age −0.04 (Male)	K1 = 0.23 −0.04*Age −0.04 (Male)	−0.03 (p < 0.001*)	0.02 (p = 0.82)	−0.04 (p = 0.44)	−0.01 (p = 0.07)
Change of greatest corneal power (K2, D)	K2 = 0.19 −0.02*Age +0.08 (Male)	K2 = 0.07 −0.016*Age +0.08 (Male)	−0.02 (p = 0.003*)	−0.12 (p = 0.26)	0.08 (p = 0.18)	0.004 (p = 0.69)
Calculated biometric parameters
Change of mean corneal power (K, D)	K = 0.19 −0.02*Age +0.02 (Male)	K = 0.15 −0.026*Age +0.02 (Male)	−0.02 (p < 0.001*)	−0.04 (p = 0.60)	0.02 (p = 0.64)	−0.006 (p = 0.43)
Change of AL/CR ratio	AL/CR = −0.097 +0.019*Age −0.002 (Male)	AL/CR = 0.086 +0.015*Age −0.002 (Male)	0.019 (p < 0.001*)	0.01 (p = 0.30)	−0.002 (p = 0.75)	−0.004 (p < 0.001 *)

Note: A significant age effect indicates that the parameter changed with age; a significant group effect indicates the parameter differed between the two groups, where the infantile group was the reference group. If the sex effect : (Sex column) was significant, it indicates that the parameter differed between boys and girls. If the interaction effect between age and late-onset group (the last column) was significant, it indicates that the rate of change differed between the two groups; this represents how the effect of age on the parameter differed between the late-onset group and the infantile group. Additionally, since the data were normalised to the initial visit, we did not report the intercept, as it would indicate extrapolation beyond the available data.

Abbreviations: AL/CR, axial length to corneal radius ratio; SER, spherical equivalent refraction.

The change of SER as well as all measured and calculated ocular biometric components changed significantly with age (all p < 0.01). A significant effect of group (infantile vs. late-onset) was identified for AL and the AL/CR ratio (p < 0.001). Additionally, a significant interaction of age and group, indicative of different rates of change, was found for SER, J₀, AL and AL/CR (p < 0.01). Compared with the infantile group, the slope of the SER change in the late-onset group, as well as the AL growth rate, was slower.

The relationship between stereoacuity and change of SER and AL

At the initial visit, the mean (±SD) stereoacuity was 3.81 ± 0.52 and 3.54 ± 0.76 log arcsecs in the infantile group and the late-onset group, respectively. At the final visit, the respective values were 3.56 ± 0.93 and 3.21 ± 0.87 log arcsecs. ANOVA showed the late-onset group had significantly better stereoacuity than the infantile group (F = 6.34, p = 0.01), and that stereoacuity at the final visit was significantly better than at the initial visit (F = 6.75, p = 0.01), i.e., it improved with age. No significant correlation between the change of SER and stereoacuity at the final visit was observed (r =−0.09, p = 0.40), nor was a significant correlation found between the change in AL and the final visit stereoacuity (r = 0.05, p = 0.70).

DISCUSSION

To our knowledge, this is the first longitudinal prospective study investigating ocular biometric components in children with two types of accommodative ET during their critical period. Notably, this study stands out for its extended prospective follow-up period, with an average duration of 4.8 ± 0.8 years, tracking refractive error and ocular biometric components up to 13 years of age. During childhood, children with accommodative ET showed developmental changes in major ocular components, including increased AL and ACD, lens thinning and flattening of the cornea.

The late-onset group exhibited a slower rate of change in SER than the infantile group, confirming our previous findings.³ Additionally, the slower rate of increase in AL in the late-onset group is consistent with the smaller change in SER. Hu et al. reported a mean rate of change of −0.35D/year in non-strabismic Chinese school-age children with moderate to high hyperopia.¹⁷ In our prior report on the longitudinal development of SER in children with accommodative ET, the rate of change in the infantile group was similar to that reported by Hu et al. for non-strabismic hyperopic children,¹⁷ i.e., −0.43D/year and −0.35D/year, respectively.³ However, in the present investigation, the infantile group changed at a slower rate than previously reported, i.e., −0.18D/year. The inclusion criteria in the present investigation had a relatively lower SER compared with our previous study. To determine whether the inclusion criteria influenced the rate of SER change, children with SER < +4D in both eyes were excluded (Table S1) and the rate of SER change was recalculated for the infantile and late-onset groups. The results were similar and did not alter the overall conclusion (−0.21D/year in the infantile group, compared to −0.13D/year in the late-onset group). As can be appreciated in Figure 2A, it is important to note that the pattern of eye growth in children with accommodative ET had high intrasubject variability. We previously reported three distinct patterns of refractive development (increase, decrease or no change) in individuals with accommodative ET.³ This has also been noted by others.^18,19 Improved image quality with early spectacle correction in the infantile onset group may be linked with ocular growth and emmetropisation. To understand the mechanisms underlying the different trends in children with infantile and accommodative ET, we are currently evaluating ocular shape (relative peripheral hyperopia) and accommodative lag as signals that may guide different patterns of refractive development in the two groups.

Interestingly, the axial growth rates observed in both groups (0.16 and 0.14 mm/year) are similar to those of emmetropic children >7 years of age in the Orinda Longitudinal Study of Myopia (range from 0.06 to 0.15).²⁰ However, in a retrospective study, Sun et al. reported a higher axial growth rate of approximately 0.28 mm/year (range: 0.08–0.81) compared with the present investigation.⁸ This difference is not surprising because: (1) Chinese children generally experience more eye growth than children in the USA and (2) their cohort included hyperopic children both with and without accommodative ET. When compared with children having accommodative ET, those without accommodative ET show significantly greater decreases in SER.²¹

Children with accommodative ET often have abnormal stereoacuity. Despite the relative maturity of stereoacuity at 18 months of age, 40% of children with intermittent accommodative ET exhibit reduced or no stereoacuity, consistent with a possible congenital deficit.²¹ Only 18% had normal stereoacuity at the final visit some 4–11 years later.²² Nevertheless, stereopsis improved with age in both groups. Interestingly, the late-onset group had better stereoacuity than the infantile group, which confirms our previous results.²³ Binocularity is often thought to be related to emmetropisation, as children with strabismus may not emmetropise normally. However, we did not find any previous studies that specifically evaluated the relationship between binocularity and eye growth. In the current cohort, no significant correlation was observed between stereopsis and eye growth.

The growth of ocular biometric components, particularly AL, is an important outcome measure when evaluating interventions to modify refractive error. One randomised clinical trial has been conducted to accelerate eye growth in hyperopic children 8–13 years of age using multifocal contact lenses, based on the hypothesis that relative peripheral defocus during childhood promotes a myopic shift (increased AL).²⁴ These authors reported growth rates of 0.072 mm/year in the intervention group and 0.032 mm/year in the control group. In the subset of the present cohort being 8 to 12 years of age, similar growth rates of 0.10 mm/year were observed in the infantile group and 0.11 mm/year in the late-onset group (p = 0.45), which was more rapid than for either group in the randomised clinical trial. While it is not clear which baseline variables may be associated with the slower rate of AL change in the intervention cohort versus the present cohort, future interventional study designs may need to consider the effect of age at enrolment and age of onset.

There are several limitations of the models derived in this study. While the eligibility criteria for the late-onset group should define a cohort of primarily late-onset accommodative ET, we cannot exclude the possibility that a few children had a delayed diagnosis of infantile accommodative ET. We attempted to limit this by purposefully excluding children diagnosed between 12 and 18 months of age. However, growth patterns for children who were prescribed full refractive error correction were not analysed separately from those who were prescribed partial correction. Indeed, Bonafede et al. compared SER development in children with partial and fully corrected accommodative ET, and reported no significant difference in growth pattern.⁴ Compliance with spectacle wear could potentially influence refractive error development. While formal assessment of compliance with spectacle wear was not conducted in this study, it is reasonable to assume that the cohort, referred by 16 Dallas area paediatric ophthalmologists, including both those in private and academic practice, was representative of typical paediatric patients presenting with accommodative ET. Additionally, a high accommodative convergence/accommodation (AC/A) ratio might be a factor affecting eye growth. This ratio was not quantified here because AC/A measurements are variable in children and there is no universally accepted method of measurement.²⁵

Clinical meaning of this study

These findings could provide reference values for clinicians to describe the longitudinal prognosis of spectacle correction for children with different types of accommodative ET according to the age of onset. For the infantile onset group, a greater change in SER suggests that a half-yearly or annual eye examination is necessary. Eye growth varies among individuals, and faster growth may require personalised intervention. However, even when refractive error is relatively stable, ocular biometric components are still changing with age. It is probably as important to measure ocular biometric parameters in these children with accommodative ET as is the case for myopic children.

CONCLUSIONS

Although relatively slower than the normal population, ocular biometric components continue to grow in hyperopic children with accommodative ET between 3 and 12 years of age. The development of these ocular components occurs at different rates for children with infantile versus late-onset accommodative ET, supporting our hypothesis. AL elongation progresses more slowly in children with late-onset accommodative ET than infantile accommodative ET, consistent with less change in SER with age.

Supplementary Material

SUPPORTING INFORMATION

Additional supporting information can be found online in the Supporting Information section at the end of this article.

Key points.

• Even when the hyperopic refractive error is relatively stable, biometric components in children with accommodative esotropia change with age.

• Developmental changes in the ocular components differ for children with infantile-onset versus late-onset accommodative esotropia.

• Interventions aimed at elongating hyperopic eyes with accommodative esotropia should take into account the development of ocular biometric components.

DATA AVAILABILITY STATEMENT

The data used in this study are available from the corresponding author upon request.