Subfoveal Choroidal Thickness is Independently Associated with Axial Length and Myopia in Preschool Children: A Swept-Source Optical Coherence Tomography Study

Abstract

Introduction

The current study assessed subfoveal choroidal thickness (SFCT) in Chinese preschool children using swept-source optical coherence tomography (OCT) and evaluated its associations with age, sex, axial length, refractive status, family history of myopia, and corneal curvature.

Methods

In this cross-sectional observational study, 303 eyes of 303 preschool children with no ocular diseases underwent comprehensive ophthalmic examinations, including cycloplegic refraction. SFCT was measured using swept-source OCT. Factors affecting changes in SFCT were analyzed using a general linear model.

Results

The mean SFCT was 326.33 ± 43.01 μm (range 198.67–438.33 μm) in children aged 3–6 years. The myopic group had significantly thinner SFCT than the emmetropic group (305.66 ± 41.18 μm vs. 347.41 ± 33.68 μm; p < 0.05). After adjusting for covariates, SFCT remained significantly lower in the myopic group (317.23 μm vs. 338.42 μm; mean difference: 21.19 μm, 95% confidence interval 7.09–35.28 μm). Myopia significantly affected SFCT (F = 8.75, p = 0.003, η² = 0.029). SFCT was positively correlated with spherical equivalent refraction (β = 6.47, p = 0.036) and negatively associated with axial length, particularly in the myopic group (β = − 10.29, p = 0.023). There were no significant differences in SFCT associated with age group, sex, or family history of myopia (all p > 0.05).

Conclusion

In preschool children, myopia is associated with significantly thinner SFCT, which is independently influenced by spherical equivalent refraction and axial length, but not by age, sex, or family history of myopia. These findings suggest that SFCT may serve as a responsive indicator of early myopic changes in preschoolers, though longitudinal studies are needed to determine its temporal relationship with axial elongation.

Plain Language Summary

Myopia (nearsightedness) is becoming more common in young children, but the early structural changes in the eye associated with it are not fully understood. Myopia occurs when the eyeball grows longer than normal, causing light to focus incorrectly in front of the retina instead of directly on it. This elongation of the eye affects structures such as the choroid, a thin layer of tissue that supplies blood to the retina. This study examined the choroid—a blood-rich layer in the posterior ocular wall that nourishes the retina—in 303 healthy Chinese children aged 3–6 years. Using a non-invasive imaging method called optical coherence tomography, researchers measured the thickness of the choroid under the center of the retina. Children with myopia had a thinner choroid (average 306 μm) than those with normal vision (average 347 μm). This difference remained after accounting for factors like the child’s age or sex. A thinner choroid was also linked to longer eyeballs and greater nearsightedness, particularly in myopic children. The association between a thin choroid and myopia was not affected by the child’s age, sex, or family history of myopia. These findings suggest that choroidal thinning may be an early sign of myopia development in preschool children. However, long-term studies are needed to clarify whether choroid changes happen before or after the eye starts to elongate.

Key Summary Points

Why carry out this study?

Myopia is increasingly observed in preschool children, yet early structural ocular changes, particularly in subfoveal choroidal thickness (SFCT), remain insufficiently characterized in this age group.

Understanding whether SFCT is associated with refractive status and ocular biometric parameters in early childhood may help identify responsive markers of early myopic development.

This study asked whether myopia, axial length, and spherical equivalent refraction are independently associated with SFCT in Chinese preschool children with normal best-corrected visual acuity.

What was learned from the study?

Children with myopia had significantly thinner SFCT than emmetropic children, and SFCT was independently associated with spherical equivalent refraction (β = 6.47, p = 0.036) and axial length (β = − 10.29, p = 0.023), but not with age, sex, or family history of myopia.

These findings suggest that choroidal thinning is present at the early stage of myopia and that SFCT may serve as a responsive indicator of early myopic changes, although longitudinal studies are needed to clarify its temporal relationship with axial elongation.

Introduction

Myopia is a rapidly escalating global public health concern that affects hundreds of millions of children and adolescents worldwide [1, 2]. It has been projected that approximately 50% of the global population will be affected by myopia by 2050 [2]. Myopia prevalence is increasing, especially in younger populations, and presents risks that could lead to vision-threatening complications and reduced quality of life [3]. Environmental factors, particularly insufficient time outdoors and excessive near work, play a key modifiable role in childhood myopia development; however, their complex interplay necessitates an objective assessment to clarify causal pathways and inform prevention strategies [4, 5]. Understanding the multifactorial etiology of myopia, encompassing both ocular structural changes and environmental exposures, is crucial for effective intervention. As a common refractive error, axial myopia is often associated with alterations in ocular structures, with the choroid being a notable example. The choroid, a vascular layer located between the retina and sclera at the back of the eyeball, plays a critical role in maintaining photoreceptor health by supplying vascular support and essential nutrients to the outer retina, as well as facilitating the removal of metabolic waste from the region [6–8]. In addition to this basic function, it is increasingly recognized for its potential role in the onset and progression of myopia [9]. Changes in choroidal thickness and blood flow may be closely associated with the development of myopia [9, 10]. Specifically, choroidal thinning has been observed in myopic eyes, indicating a possible link between choroidal structure and refractive error [11–15]. Choroidal thinning may compromise the delivery of oxygen and metabolic substrates to the retinal pigment epithelium and photoreceptors, potentially leading to localized hypoxia and altered biochemical signaling [16, 17]. Such metabolic stress may trigger the release of growth-regulating factors [18, 19]—such as vascular endothelial growth factor, transforming growth factor-beta, or matrix metalloproteinases—that may diffuse posteriorly to influence scleral fibroblast activity, promote extracellular matrix remodeling, and ultimately contribute to scleral thinning and excessive axial elongation. Thus, the choroid may reflect myopic structural change and actively participate in the signaling cascade that drives eye growth.

The advent of optical coherence tomography (OCT) has significantly advanced our understanding of ocular morphology. OCT enables non-invasive visualization and assessment of the choroid, facilitating quantitative studies in humans [20–22]. Despite growing interest in the choroid’s role in myopia, existing research lacks age-specific data derived from pediatric populations. In particular, there is a dearth of evidence for children of preschool age, a critical developmental window marked by rapid ocular maturation and heightened susceptibility to myopiagenic stimuli. Most studies to date have focused on adults or combined pediatric, adolescent, and adult participants into heterogeneous cohorts. While valuable, these studies fail to capture the unique structural and functional dynamics of the early childhood eye. Age-aggregated analyses may obscure whether reduced subfoveal choroidal thickness (SFCT) is part of the incipient myopic phenotype or simply a secondary consequence of prolonged axial elongation. Robust evidence on the independent effects of axial length (AL) and refractive error on SFCT in this age group remains scarce, largely because preschool children are rarely included in choroidal imaging studies.

The present study investigated SFCT and its associations with demographic factors (age, sex, and family history of myopia) and key ocular biometric parameters, including AL, corneal curvature, and spherical equivalent refraction (SER), in a well-characterized cohort of Chinese preschool children. We aimed to determine whether AL and refractive error are independently associated with SFCT in early myopia. We hypothesized that longer AL and more myopic SER would be independently associated with thinner SFCT, even at the onset of myopia in preschool-aged children.

Methods

Study Design and Participants

This cross-sectional, observational study was conducted at the ophthalmology outpatient department of Tongzhou Maternal and Child Health Hospital of Beijing, from March 5, 2024, to May 30, 2025. Ethical approval was obtained from the Ethics Review Committee of Tongzhou Maternal and Child Health Hospital of Beijing (No. 2024-TZFY-001–01), in accordance with the tenets of the Declaration of Helsinki. The study was conducted after the protocol was explained to the participants and their parents or legal guardians, and informed consent was obtained. Children aged 3 to 6 years with best-corrected visual acuity (BCVA) appropriate for their age, subnormal uncorrected visual acuity, normal intraocular pressure (IOP) (10–21 mmHg), and a SER of ≤ − 0.50 D after cycloplegic refraction, as well as phoria within ± 15 prism diopters (△) were enrolled in the study. Children were considered to have normal visual development if their presenting visual acuity met the age-appropriate norms: ≥ 20/40 (0.5) for children aged 3 to 4 years, and ≥ 20/32 (0.6) for children aged 5 years and older, consistent with AAPOS guidelines. Subnormal visual acuity was defined as presenting visual acuity below these thresholds. Age-matched healthy children with BCVA meeting the same age-appropriate standards and SER within the normal range (− 0.50 D to + 1.75 D after cycloplegia) were enrolled as controls (the emmetropic group). Participants were excluded if they had congenital ocular structural abnormalities, active ocular lesions, a history of myopia control treatment (e.g., orthokeratology lenses or low-dose atropine eye drops), other ocular diseases (such as retinal pathology), neurodevelopmental conditions such as autism that could impair cooperation with examinations, or any significant systemic disease. The initial visual acuity screening served two purposes for all potential participants: (1) to exclude children with underlying visual impairment unrelated to refractive error (e.g., amblyopia, retinal diseases); and (2) to assess visual potential (i.e., confirming that BCVA met age-appropriate norms as defined above). For the emmetropic group, this confirmed they were healthy controls with normal visual function. For the myopic group, it ensured that subnormal uncorrected visual acuity was solely attributable to uncorrected myopia rather than to other pathologies.

In emmetropic individuals, data from the right eye were analyzed. In individuals with myopia, the eye with the higher degree of myopia was selected, or if refractive error was symmetric, the right eye was chosen. After excluding 17 children due to non-cooperation or poor-quality examination results, a total of 303 children were included in this study.

To ensure the accuracy of our data, all visual acuity measurements (uncorrected and presenting VA) were conducted prior to the administration of cycloplegic agents (1% Atropine). The cycloplegic refraction was performed subsequently to determine the accurate spherical equivalent (SER) for grouping. Therefore, the pharmacological agents did not affect the vision measurements.

Ocular Examinations and Child-Centered Imaging Protocol

Ocular examinations included visual acuity, AL, IOP, slit-lamp examination, ocular alignment assessment, fundus examination with OCT imaging, and refraction measurement. IOP was measured by a non-contact tonometer (model NT-510, Nidek, Japan). Refraction and corneal radius of curvature (CR) were obtained using an auto kerato-refractometer (KR-1 Topcon, Tokyo, Japan). All children enrolled in this study received 1% Atropine Eye Gel (Shenyang Xingqi Pharmaceutical Co., Ltd., National Drug Approval Number H20052295, China) twice daily for 3 consecutive days to achieve maximal cycloplegia, followed by refractive measurement. Atropine was selected because it provides more reliable cycloplegia in young children than cyclopentolate, thereby minimizing the risk of underestimating hyperopia or overestimating myopia during refractive assessment [23]. SER was used to classify refractive status.

Given the challenges of obtaining reliable data from children aged 3–6 years, all imaging procedures, including AL measurement, fundus photography, and OCT, were conducted using a child-centered protocol designed to maximize cooperation, comfort, and image quality. The core strategy was to transform examinations into engaging, game-like experiences. For example, fixation targets were described as “chasing the light bug,” and procedures were demonstrated first on a parent or doll to reduce anxiety. Children were encouraged through verbal cues (e.g., “Let’s play the statue game: hold still on three!”), with parents present for reassurance and, when necessary, gentle physical support. Immediate praise and small, non-food rewards (e.g., stickers) reinforced participation.

Technical adaptations further optimized success. The most cooperation-sensitive tests (e.g., OCT, high-resolution fundus imaging) were performed first during the child’s peak alertness; imaging protocols prioritized speed (e.g., the fastest available OCT volume scan); and flash intensity in fundus photography was minimized to avoid startling the child. Multiple brief attempts, interspersed with short breaks as needed, were permitted, with the best-quality image selected for analysis. AL was measured using the Lenstar LS 900 (Haag-Streit AG, Koeniz, Switzerland). To sustain a steady gaze, the standard fixation target was replaced with a brightly colored, animated toy positioned at the instrument’s focal point. Five valid readings per eye (quality score ≥ 7/10) were acquired, and the median value was used for analysis.

Choroidal Thickness Measurements

Choroidal thickness was measured using swept-source OCT (NIDEK RS-3000 Lite, Japan), a non-invasive imaging modality based on low-coherence interferometry. This system uses a rapidly tunable near-infrared laser (~ 1050 nm) to generate depth-resolved reflectivity profiles via low-coherence interferometry, enabling enhanced choroidal visualization owing to deeper tissue penetration. All participants underwent imaging after pharmacological mydriasis (pupil diameter ≥ 6 mm) under standardized ambient lighting conditions. To control for diurnal variation in choroidal thickness, all scans were performed within a standardized time window of 9:00 AM to 2:00 PM. Each eye underwent three consecutive high-definition horizontal B-scans centered on the fovea (scan type: single 6-mm line scan, 1024 A-scans per B-scan). Scans were acquired with the participant fixating on an internal target, and real-time eye tracking was used to minimize motion artifacts.

Image selection and measurement followed a standardized protocol. The best-quality scan among the three was selected based on the following criteria: (1) Clear visualization of both Bruch’s membrane and the choroid–scleral interface across the central 1 mm; (2) signal strength index ≥ 7/10; and (3) absence of significant motion artifacts, shadowing, and segmentation errors.

SFCT was defined as the vertical distance from the outer boundary of Bruch’s membrane to the inner boundary of the choroid–scleral interface at the foveal center (shown in Fig. 1). Measurements were performed manually using the built-in caliper tool in the device software (RS-3000 Viewer v.1.5). Two trained graders masked to participant characteristics independently measured SFCT on the selected B-scan. The final value for each eye was calculated as the mean of the two graders’ measurements. In cases where the inter-grader difference exceeded 15 μm (approximately 5% of typical SFCT in this age group), a third grader was consulted, and a consensus value was determined via adjudication.

Fig. 1.

Measurement of subfoveal choroidal thickness (SFCT). The definition of SFCT was the vertical distance between the outer layer of the retinal pigment epithelium (RPE) and the choroscleral interface at the center of the fovea

The participant recruitment and examination process was illustrated in Fig. 2. All ocular images underwent real-time quality assessment by the technician. Predefined exclusion criteria, including poor centration, motion blur, and significant artifacts, prompted immediate, child-friendly reattempts. Final inclusion in the analysis required agreement between two masked graders on image quality, using modality-specific thresholds (e.g., clear visualization of optic disc and macula for fundus photographs). The number of exclusions because of inadequate image quality or poor cooperation is reported in the Results section.

Fig. 2.

Flow diagram of participant recruitment and examination sequence

Statistical Analysis

All statistical analyses were performed using SPSS v.27.0 software (IBM Corp., Armonk, NY, USA). Data from only one eye of each participant were included in statistical analyses. The normality of measurement data was assessed using the Shapiro–Wilk test. Normally distributed data are presented as means ± the standard deviation, and group differences were compared using the t-test for independent samples. Non-normally distributed data (specifically SER) are reported as medians and interquartile ranges (IQR, 25th–75th percentiles), and differences were investigated using the Wilcoxon–Mann–Whitney U test. Between-group categorical variable comparisons were conducted using the chi-squared (χ²) test. To investigate associations between SFCT and refractive and biometric factors while adjusting for potential confounders, a general linear model was used with SFCT as the dependent variable. The model included the following predictors: myopia status (coded as 0 = emmetropia, 1 = myopia), family history of myopia (coded as 0 = no family history, 1 = positive family history), age (years), sex (male vs. female), SER (diopters), and centered AL (AL_C). An interaction term between myopia status and AL_C was also included to evaluate whether the association between AL and SFCT varied by myopia status.

AL_C was calculated as follows, where AL represents the individual participant’s AL (in mm), and $\overline{\text{AL}}$ represents the arithmetic mean of AL across all study participants:

$${\text{AL}}_{\text{C}}=\text{AL}-\overline{\text{AL}}$$

This centering procedure was applied to reduce multicollinearity between AL and its interaction term, and to ensure that the main effect of myopia reflected the difference in SFCT at the average AL in the cohort. All statistical tests were two-sided, with a significance level set at α = 0.05 (2-tailed). The values of effect were reported using partial eta-squared (η²), and 95% confidence intervals (CIs) were calculated where appropriate.

Results

Demographic and Baseline Characteristics

Altogether, 303 eyes of 303 children (151 male children, 152 female children) aged 3–6 years (mean 5.14 ± 0.86 years) were evaluated in this cross-sectional study. Of these, 153 were assigned to the myopia group, and 150 were assigned to the emmetropia group. Their general characteristics and the differences between the two groups are presented in Table 1. Mean parameters included SFCT 326.33 ± 43.01 μm (range 198.67–438.33 μm), AL 22.96 ± 0.91 mm (range 20.84–26.73 mm), and CR 7.72 ± 0.25 mm (range 6.94–8.35 mm). Mean SFCT was significantly thinner in the myopia group (305.66 ± 41.18 μm) than in the emmetropia group (347.41 ± 33.68 μm) (t = 9.65, p < 0.001). Statistically significant differences were observed in SER, AL, K, and CR between the two groups (p < 0.05; Table 1). When analyzing the entire cohort regardless of refractive status, AL was significantly longer for male children than that for female children (Male: 23.18 ± 0.90 mm vs. Female: 22.74 ± 0.86 mm, p < 0.01). No significant between-group differences were observed in subfoveal retinal thickness (SFRT), sex, or age distribution between the two groups.

Table 1.

The general characteristics of the study participants

Characteristics	Total	Range	Groups		χ2/t/Z	p-value
			Myopia(n = 153)	Emmetropia(n = 150)
Demographic
Sex, n (%)
Male	151 (49.83)	–	72 (47.68)	79 (52.32)	0.95	0.329
Female	152 (50.17)	–	81 (53.29)	71 (46.71)
Age (years)	5.14 ± 0.86	3–6	5.22 ± 0.82	5.06 ± 0.90	− 1.64	0.102
Ocular biometric
SER (D)	− 0.50 (− 1.38, 0.38)	− 5.37 to 1.50	− 1.38 (− 2.13, − 1.00)	0.38 (0.13, 0.50)	− 15.07	< 0.001
AL (mm)	22.96 ± 0.91	20.84–26.73	23.47 ± 0.78	22.44 ± 0.70	− 12.12	< 0.001
K (D)	43.75 ± 1.45	40.43–48.66	43.97 ± 1.47	43.52 ± 1.40	− 2.73	0.007
CR (mm)	7.72 ± 0.25	6.94–8.35	7.68 ± 0.25	7.76 ± 0.25	− 2.73	0.007
SFRT (μm)	212.51 ± 13.80	178.33–270.67	212.27 ± 14.09	212.75 ± 13.54	0.30	0.763
SFCT (μm)	326.33 ± 43.01	198.67–438.33	305.66 ± 41.18	347.41 ± 33.68	9.65	< 0.001

Categorical variables are presented as n (%). Continuous variables are presented as mean ± SD, except for SER which is presented as median (P25, P75)

SER spherical equivalent refraction, AL axial length, K corneal curvature, CR corneal radius, SFRT subfoveal retinal thickness, SFCT subfoveal choroidal thickness

Among all study participants, 183 had at least one parent with a family history of myopia, while 120 had no such family history. No significant between-group differences in SFCT were observed between the groups in terms of sex classification or whether parents had a family history of myopia (p > 0.05; Table 2).

Table 2.

Effects of gender and family history of myopia on SFCT

Parameters	Groups	N	SFCT (μm)	t	p-value
Sex	Male	152	323.40 ± 41.53	− 1.19	0.234
	Female	151	329.28 ± 44.40
Family history of myopia	Yes	120	327.78 ± 42.30	0.47	0.636
	No	183	325.38 ± 43.57

Factors Associated with SFCT in General Linear Model Analysis

Correlations between SFCT and ocular parameters are shown in Table 3. SFCT was positively correlated with SER (r = 0.47, p < 0.001) and negatively correlated with AL (r = − 0.40, p < 0.001; Table 3). No significant correlations were observed between SFCT and age, sex, or CR. Scatterplots for SFCT by SER, AL, CR, and age are shown in Fig. 3. Levene’s test confirmed homogeneity of error variance (F = 1.66, p = 0.118), supporting the assumptions of the general linear model.

Table 3.

Correlations between subfoveal choroidal thickness and ocular parameters

Parameters	SFCT
Age
r	− 0.08
P	0.165
Sex
rs	0.10
P	0.096
SER
rs	0.47
P	< 0.001
AL
r	− 0.40
P	< 0.001
CR
r	0.01
P	0.828

SFCT subfoveal choroidal thickness, SER spherical equivalent refraction, AL axial length, CR corneal radius

Fig. 3.

Relationships between subfoveal choroidal thickness and spherical equivalent, axial length, corneal radius of curvature, and age

The overall model was statistically significant (F = 16.80, p < 0.001), with an adjusted R² of 0.268, indicating that approximately 26.8% of the variance in SFCT was explained by the variables included (Table 4). Myopia status had a significant main effect on SFCT (F = 8.75, p = 0.003, η² = 0.029). After adjusting for covariates, the mean SFCT in the myopic group (317.23 ± 4.14 μm, 95% CI 309.08–325.37) was significantly lower than that in the emmetropia group (338.42 ± 4.75 μm, 95% 329.07–347.76), with a mean difference of 21.19 μm (95% CI 7.09–35.28) (shown in Fig. 4).

Table 4.

The results of the inter-subject effect test

Parameters	F	df	p	partial η2	95% CI (adjusted mean difference)
Myopia status	8.75	1	0.003	0.029	[7.09, 35.28]
${\text{AL}}_{\text{C}}$	4.02	1	0.046	0.013	–
SER	4.44	1	0.036	0.015	–
Sex	2.38	1	0.124	0.008	–
Family history of myopia	0.41	1	0.521	0.001	–
Myopia status × ${\text{AL}}_{\text{C}}$	1.16	1	0.282	0.004	–
Age	0.01	1	0.927	0.000	–
Modified model	16.80	7	< 0.001	0.285	–

Subfoveal choroidal thickness (SFCT) was the dependent variable, Covariates: ${\text{AL}}_{\text{C}}$, age, SER

R² = 0.285(adjusted R² = 0.268)

95% CI for the adjusted mean difference in SFCT between groups (emmetropia vs. myopia)

AL_C centered axial length, SER spherical equivalent refraction

Fig. 4.

Comparison of subfoveal choroidal thickness (SFCT) between emmetropia and myopia groups. The emmetropia group had a mean SFCT of 338.42 μm (95% confidence interval [CI] 329.07–347.76 μm), whereas the myopia group had a mean SFCT of 317.23 μm (95% CI 309.08–325.37 μm)

In the myopic subgroup, AL_C was significantly independently negatively associated with SFCT (β =  − 10.29, p = 0.023), and each 1-mm increase in AL was associated with a 10.29-μm reduction in SFCT. SER was positively correlated with SFCT (β = 6.47, P = 0.036), consistent with a thinner choroid in eyes with greater myopic refractive error. In contrast, sex, family history of myopia, age, and the aforementioned interaction term (myopia status × AL_C) were not significantly associated with SFCT (all p > 0.05; Table 5).

Table 5.

Estimated values of each parameter

Parameters	B	Standard error	t	p-value
Constant	327.83	15.03	21.81	0.000
[Myopia status = 0] × ${\text{AL}}_{\text{C}}$	− 3.87	4.73	− 0.82	0.414
[Myopia status = 1] × ${\text{AL}}_{\text{C}}$	− 10.29	4.51	− 2.28	0.023
Sex	− 7.24	4.70	− 1.54	0.124
Myopia status	21.18	7.16	2.96	0.003
Family history of myopia	− 2.80	4.36	− 0.64	0.521
${\text{AL}}_{\text{C}}$	0a
Age	− 0.25	2.67	− 0.09	0.927
SER	6.47	3.07	2.11	0.036

AL_C centered axial length, SER spherical equivalent refraction

^aThis parameter was redundant, so it was set to zero

Discussion

The current study investigated the determinants of SFCT in a cohort of preschool-aged children (3–6 years), a critical yet underrepresented period of ocular development. We found that, unlike in older populations, chronological age did not significantly influence SFCT within this narrow age range. Instead, our analysis identified axial elongation and refractive status as the primary drivers of choroidal thickness in this demographic.

Consistent with our hypotheses, no differences in SFCT were observed among different age groups, between sexes, or based on family history of myopia. Crucially, myopia status emerged as an independent predictor of SFCT. After adjusting for covariates, SFCT was significantly lower in the myopia group than in the emmetropia group (mean difference 21.19 μm). Furthermore, in the myopia subgroup, AL was significantly negatively associated with SFCT (β = − 10.29), suggesting that axial elongation leads to an accelerated decline in choroidal thickness.

These findings provide novel insights that extend beyond previous studies of older children and adults. Unlike studies of school-aged children, in which age and axial elongation are often conflated, our data in this preschool cohort reveal a critical distinction: axial elongation, rather than chronological age, is the primary determinant of choroidal thinning. While previous research has established the link between SFCT and AL in older populations, our study uniquely demonstrates that this relationship is already fully established at the very onset of myopia in preschool children. This suggests that the mechanical stretching of the posterior segment occurs immediately with the initiation of axial growth, rather than being a consequence of long-term myopia progression. Therefore, our data provide new evidence that SFCT serves as a responsive, real-time structural indicator of the early emmetropization disruption in young children.

Age-Related Changes in SFCT

In some studies, changes in SFCT have varied by age. Xiong et al. [24] investigated changes in the characteristics of SFCT over time in different age groups between 6 to 18 years, and reported that among children aged 6–9 years, age was significantly negatively correlated with SFCT. The specific nature of the relationship between age and choroidal thickness in preschool children remains unclear, and the results of the few studies that have investigated choroidal thickness in preschool children are inconsistent. A study conducted in Taiwan by Huang et al. [25] included participants with a similar mean age to that of those in the present study. That study investigated 374 eyes in 187 children with a mean age of 5.6 ± 1.9 years. The range of ages (2 to 16 years) was larger than that in the current study. They reported a significant negative association between age and SFCT in univariate regression analyses, but no significant association in multiple linear regression analyses. However, a study on Chinese preschool children aged 3–6 years found no evident change in SFCT with age [26]. Similarly, in a study of 109 Malay children aged 7–17 years, there was no significant correlation between age and SFCT [27]. We also found no significant correlation between age and SFCT within the 3- to 6-year-old range investigated (F = 0.01, p = 0.927). However, this observation is limited to this specific developmental window, and age-SFCT relationships may differ across broader age ranges or with longer-term follow-up owing to varying stages of ocular development and growth [28].

Sex-Related Changes in SFCT

A few studies have reported that male children have a greater SFCT than female children [12, 29, 30]. In one study involving a 2-year longitudinal follow-up, thinning of the choroid was significantly greater in female participants than in male participants [12]. The current study investigated the relationship between sex and SFCT, and there was no significant difference in male and female children. This is consistent with those reported in previous studies: one conducted in preschool children by He et al. [26], and another in children aged 2–16 years reported by Huang et al. [25]. it is worth noting that, in our cohort, male children had significantly longer AL than female children. Despite the known inverse correlation between AL and SFCT, we still observed comparable SFCT between the sexes. We speculate that in this pediatric population, the choroidal thinning effect associated with longer AL in male children may have offset the potential sex-based anatomical differences. We further speculate that the differences between these studies may be related to the age range of the participants, with a smaller effect of hormone levels in preschool children and a larger effect in adolescents and adults.

Refraction-Related Changes in SFCT

Multiple studies have reported an inverse correlation between refractive error and choroidal thickness, particularly SFCT [11, 13, 25, 31, 32], with increasing myopia (more negative spherical equivalent) associated with choroidal thinning. Variations in choroidal thickness are also implicated in the development of various pediatric ocular disorders [15]. In the present study, SER was significantly positively correlated with SFCT (β = 6.47, p = 0.036), indicating that greater myopia was associated with thinner SFCT, consistent with most previous studies [27, 33]. Additionally, myopia status was significantly associated with SFCT (F = 8.75, p = 0.003, η² = 0.029), and after adjusting for covariates, SFCT in the myopia group was significantly lower than that in the emmetropia group. The mean difference between the two groups was 21.19 μm (95% CI 7.09–35.28). These findings suggest that choroidal thinning is a consistent structural feature associated with myopia in preschool children, independent of age. This implies that SFCT may reflect the underlying ocular remodeling, such as scleral expansion and axial elongation, that accompanies the early development of myopia. While the cross-sectional data yielded by the present study cannot determine whether choroidal changes precede or follow refractive shifts, the strong association between SFCT and myopic status supports its potential use as a non-invasive indicator of current myopic phenotype in young children. Future longitudinal studies are needed to evaluate whether baseline or changing SFCT can predict the rate of myopia progression and to clarify temporal relationships with axial elongation.

Notably, some studies have reported thicker choroids in hyperopic eyes than in emmetropic or myopic eyes [9, 34, 35]. Because our cohort included only myopic and emmetropic children, we were unable to assess the full spectrum of refractive–choroidal relationships across the entire refractive range. Including hyperopic children in future investigations may yield a more comprehensive understanding of how choroidal thickness varies with refractive status during early ocular development.

AL-Related Changes in SFCT

AL has been consistently shown to be inversely associated with SFCT across multiple populations [11, 27, 31, 36], with a longer AL typically associated with a thinner SFCT, especially in myopic eyes [37]. Concordantly, in the current cohort of Chinese preschool children, there was a moderate negative correlation between AL and SFCT (r = − 0.40). Although interaction between myopia status and AL_C was not statistically significant, indicating that the slope of the AL-SFCT relationship did not differ markedly between myopic and non-myopic children, the association between AL and SFCT was significant within the myopic subgroup. After adjusting for SER, each 1-mm increase in AL was associated with a 10.29-μm reduction in SFCT (β = − 10.29, p = 0.023), suggesting that axial elongation exerts an independent influence on choroidal thinning, even in early myopia. Although transverse magnification effects related to AL can influence lateral scaling in OCT imaging, our primary outcome (SFCT) represents an axial (depth) measurement, which is theoretically independent of transverse magnification.

Age was not significantly associated with SFCT in the present study. This is consistent with a previous study conducted in preschool children, in which AL increased with age, but choroidal thickness remained relatively stable throughout this narrow age range [26]. This implies that axial elongation, not chronological age per se, is the primary ocular parameter linked to choroidal thinning during the initial stages of myopia. The choroid is known to be dynamically involved in ocular growth regulation and emmetropization, potentially through modulation of retinal position and scleral signaling [38]. However, in the context of excessive axial elongation, as seen in progressive myopia, this regulatory capacity may be overwhelmed. Indeed, a study of older children showed that the protective effect of age-related choroidal thickening diminishes in eyes with myopia exceeding –2.00 D, after which AL becomes the dominant determinant of SFCT [39].

In highly elongated eyes, choroidal thinning is commonly attributed to mechanical stretching of the posterior segment, which may reduce choroidal volume, compromise vascular perfusion, and alter the local expression of growth factors such as vascular endothelial growth factor [9, 40, 41]. Reduced choroidal blood flow and thinning have been linked to an increased risk of myopia-related complications, including choroidal neovascularization [42]. Collectively, these findings indicate that SFCT is closely tied to structural remodeling of the myopic eye. While our cross-sectional data cannot determine whether choroidal changes precede or follow axial elongation, the strong and independent association observed in preschool children supports the contention that SFCT reflects key features of early myopic development, and may be a responsive, non-invasive indicator of ongoing ocular changes.

The primary strength of the present study is its focus on SFCT in preschool-aged children—a critical yet underrepresented population in myopia research, in which the eye undergoes rapid emmetropization and is particularly susceptible to environmental and genetic influences on refractive development. Myopia status was independently associated with reduced SFCT, even after adjusting for potential confounders such as age, sex, and family history of myopia. This independence from known risk factors suggests that SFCT captures aspects of ocular structure that are closely associated with the myopic phenotype itself, beyond demographic factors and familial background.

Given the choroid’s known responsiveness to visual stimuli and its involvement in ocular growth regulation, SFCT may serve as a non-invasive indicator of concurrent structural changes in the early stages of myopia. In clinical settings, measuring SFCT could complement traditional assessments—particularly in children with borderline refraction or strong familial predisposition—by providing additional insight into the presence of myopic eye remodeling. However, it should be emphasized that the cross-sectional study design negated determination of whether choroidal thinning precedes, accompanies, or follows axial elongation.

Several study limitations warrant acknowledgment. First, the cross-sectional nature of the study precludes causal inference. Longitudinal follow-up is essential to clarify the temporal relationships between SFCT, refractive error, and AL. Second, we acknowledge that diurnal variation in choroidal thickness may still act as a confounding factor. It is well established that choroidal thickness exhibits a rhythmic fluctuation over 24 h, typically thinning during the day (reaching its minimum at approximately noon) and thickening at night. This physiological change is inversely correlated with AL. Given that our measurement window spanned a period of significant choroidal thinning, this inherent biological variability may have introduced noise into our biometric data, potentially affecting the precision of the observed correlations. Third, the study cohort was recruited from a single center, which may limit the generalizability of the study findings.

Conclusion

This study demonstrates that myopic preschool children exhibit significantly thinner SFCT compared to their emmetropic peers, and that this thinning is independently associated with SER and AL, rather than age, sex, or family history of myopia. These findings support the potential utility of SFCT as a responsive indicator for early myopic changes in this population. Future prospective studies should evaluate whether baseline SFCT or its rate of change can predict the onset or progression of myopia. Such work would help determine whether SFCT has a role as part of a multimodal approach to monitoring the early development of myopia in at-risk preschool children.

Author Contributions

Xiaolan Xie, Hongyang Li, and Bidan Zhu designed the study. Xin Wang, Shuting Wen, and Lin Tian performed the study. Xin Wang and Shuting Wen validated the dataset. Xiaolan Xie analyzed and interpreted the data, and subsequently wrote the initial draft. Bidan Zhu supervised the study, validated the results, and acquired funding. Hongyang Li also supervised the study and revised the manuscript. All authors provided a final review and approved the final manuscript.

Funding

This study was supported by Beijing Tongzhou District Outstanding Young Scientists and Technologists Project (No. JCQN2024030). The funder had no role in the design, data collection, data analysis, and reporting of this study. The journal’s Rapid Service Fee was funded by the authors.

Data Availability

The datasets used and analyzed in the current study are available from the corresponding author upon reasonable request.

Declarations

Conflict of Interest

Xiaolan Xie has nothing to disclose, Xin Wang has nothing to disclose, Shuting Wen has nothing to disclose, Lin Tian has nothing to disclose, Hongyang Li has nothing to disclose, Bidan Zhu has nothing to disclose.

Ethical Approval

Ethical approval was obtained from the Ethics Review Committee of Tongzhou Maternal and Child Health Hospital of Beijing (No. 2024-TZFY-001-01), in accordance with the tenets of the Declaration of Helsinki. The study was conducted after the protocol was explained to the participants and their parents or legal guardians, and informed consent was obtained.