Growth Trajectories During Early Childhood and Ocular Development in High-Risk Infants

Abstract

Purpose

High-risk infants often show atypical postnatal growth and may be vulnerable to visual impairment, yet evidence linking early-life systemic growth with ocular development remains limited. This study investigated the associations between growth trajectories during early childhood (birth to 3 years) and ocular development in high-risk infants.

Methods

In this population-based prospective study, 10,030 high-risk infants underwent repeated measurements of weight, length, body mass index (BMI), and head circumference from birth to 36 months. Growth velocities were calculated as changes in World Health Organization (WHO)-standardized z scores across developmental intervals. Uncorrected visual acuity (UCVA; logMAR) and cycloplegic spherical equivalent refraction (SER; diopters [D]) were assessed at 3 to 4 years. Associations between growth increments and ocular outcomes were examined using multivariable linear regression adjusting for demographic and perinatal factors.

Results

Of the total, 36.6% were preterm, 6.9% were term low-birth-weight, and 31.7% were macrosomic. UCVA exhibited a graded pattern (P < 0.001): poorest in preterm infants (0.229 [0.224–0.234]), followed by term low-birth-weight (0.225 [0.215–0.235]) and term normal-birth-weight infants (0.219 [0.213–0.225]), whereas macrosomic infants showed comparable or slightly better UCVA (0.213 [0.207–0.219]). SER was most hyperopic among preterm infants (0.433 [0.410–0.456]) and least hyperopic among macrosomic infants (0.389 [0.360–0.419]). Accelerated growth during 0 to 6 and 6 to 12 months was consistently associated with poorer visual acuity and a more hyperopic refractive profile, with no significant associations observed beyond 12 months.

Conclusions

Accelerated early-life growth was linked to less favorable visual acuity and refractive outcomes, highlighting the first postnatal year as a critical period for coordinated systemic and ocular development.

Child development is a multifaceted process in which vision and ocular health play pivotal roles in shaping overall growth and functional outcomes.^1,2 It is well known that children with adverse birth outcomes frequently exhibit atypical growth trajectories; for instance, preterm and low birth weight infants often experience catch-up growth during early childhood.^3,4 Notably, emerging evidence demonstrates that such adverse neonatal outcomes, particularly preterm birth, are also associated with a heightened risk of developing visual impairment and ocular morbid conditions.^5,6 This convergence has led to the hypothesis that ocular development and physical growth may be coordinated by shared biological mechanisms. Specifically, ocular maturation may be modulated by physical growth processes, as critical windows of eye development coincide with periods of rapid physical growth and depend on adequate nutritional, hormonal, and neural support.^7,8

Epidemiological evidence linking physical growth and ocular development remains limited. Existing studies have predominantly relied on cross-sectional anthropometric assessments, and have failed to yield conclusive results.^9–12 Furthermore, most investigations have focused on children older than 5 years or on adults, developmental stages during which visual function is already influenced by education and other environmental confounders.^9–12 By contrast, early childhood (birth to 3 years) constitutes a critical window of rapid and dynamic maturation for both body stature and the visual system, particularly in children with adverse birth outcomes.¹³ Longitudinal studies within this early developmental period are urgently needed to elucidate how early growth trajectories shape ocular biometry, to clarify underlying biological mechanisms, and to inform preventive strategies and timely interventions.

To fill this gap, the present study aimed to evaluate visual function and ocular development in a large, well-characterized cohort of children aged 3 to 4 years who were born with adverse birth outcomes, and to investigate their associations with growth trajectories from birth onward.

Methods

Study Population

This study was embedded within the High-Risk Infant Neurodevelopment Outcomes Cohort, an ongoing prospective birth cohort initiated in January 2009 by the Children's Hospital of Fudan University in collaboration with 16 community health service centers in Minhang District, Shanghai, People's Republic of China. In clinical practice, high-risk infants are those who, due to maternal, perinatal, or neonatal factors, have an increased risk of mortality, serious complications, or adverse long-term developmental outcomes and therefore require enhanced follow-up, assessment, and early intervention. The cohort was designed to investigate growth, neurodevelopment, cognition, and behavioral outcomes in high-risk infants up to 6 years of age, with the broader goal of delineating their disease spectrum. The study protocol was approved by the Institutional Review Board of Fudan University (Approval No. [2025]42) and was acknowledged by all participating institutions. All child health records were stored and managed within the Minhang District Child Health Information Management System, from which data for the present analysis were extracted. Because only de-identified data were analyzed, informed consent was waived at all study sites in accordance with institutional and national regulations.

Infants were eligible if they were younger than 3 months at enrollment and presented with one or more high-risk factors, including (1) adverse birth outcomes (preterm birth, low birth weight, small or large for gestational age, or macrosomia); (2) multiple gestation; (3) complicated delivery or perinatal events (dystocia, breech presentation, forceps- or vacuum-assisted delivery, or cesarean delivery recorded as high risk); (4) neonatal diseases or complications (birth asphyxia, hyperbilirubinemia, infection or pneumonia, seizures/convulsions, or congenital heart disease); and (5) maternal high-risk conditions (advanced maternal age, gestational hypertension, gestational diabetes, anemia, infections, or thyroid disorders). Families were required to have resided in Shanghai for at least 1 year and to plan ongoing child healthcare within the participating centers. Exclusion criteria included intention to relocate outside the district within 3 years and the presence of major congenital anomalies, such as central nervous system malformations, or severe metabolic and genetic disorders. By the end of December 2023, a total of 51,937 infants has been recruited into the cohort. For the present analysis, 10,030 infants who completed serial follow-up assessments of physical growth and development during early childhood (birth to 3 years), had ophthalmic biometry measurements available at ages 3 to 4 years, and had no clinically documented congenital infections or other severe systemic acute or chronic diseases with potential ocular involvement were included.

Physical Growth Assessment

Birth parameters, including gestational age, birth weight, birth length, and head circumference (HC), were obtained from medical records. Adverse birth outcomes were defined according to established criteria: preterm birth as delivery before 37 completed weeks of gestation,¹⁴ term low birth weight as <2500g at term,¹⁵ and macrosomia as >4000g.¹⁶ Postnatal growth was assessed at 6, 12, 18, 24, and 36 months during routine visits to community health service centers, following standardized measurement protocols.

For both birth and follow-up assessments, z scores for anthropometric indicators were derived from the World Health Organization Child Growth Standards, with corrected age (chronological age adjusted for gestational age) applied for preterm infants.^17,18 To characterize growth dynamics, we calculated differences in z scores between consecutive time points from birth to 36 months, with positive values indicating accelerated growth relative to the reference population and negative values indicating growth deceleration.

Ocular Development Assessment

Ophthalmic evaluation was performed as part of the comprehensive clinical and psychological assessment at 3 to 4 years of age, following a standardized protocol. This assessment included measurement of uncorrected visual acuity (UCVA) using a crowded logMAR chart and evaluation of refractive status using a handheld vision screener. UCVA was measured monocularly (the right eye followed by the left eye) with Lea Symbols 3-m Set charts (250300; Goodlite, Elgin, IL, USA) under standardized room lighting conditions. Refractive status, including spherical and cylindrical errors, was determined using the Welch Allyn Spot Vision Screener (Welch Allyn, Skaneateles Falls, NY, USA) 30 minutes after instillation of a cycloplegic solution containing cyclopentolate (0.85%) and phenylephrine (1.5%). For each participant, at least two measurements were obtained, and the final value was the average of valid readings; measurements were repeated if the device indicated inadequate quality. All examinations were conducted by uniformly trained ophthalmologists and optometrists under standardized procedures, with periodic supervision for quality control. The spherical equivalent refraction (SER) was calculated as the spherical power plus half of the cylindrical power.

Statistical Analysis

Descriptive statistics were computed to summarize general characteristics of the study population. Continuous variables are presented as means with standard deviations (SDs), and categorical variables as counts with percentages (N, %). Because visual acuity and refraction between the two eyes were highly correlated (Pearson correlation coefficients >0.95), analyses were restricted to the right eye to avoid redundancy.

Multivariable models were adjusted for potential confounders identified a priori from medical records and standardized questionnaires administered at enrollment and follow-up. Covariates included age at ocular measurements, infant sex, age at breastfeeding cessation, paternal age, maternal age, fetal plurality, and mode of delivery. These covariates were selected based on their established relevance in previous studies.^7,9,11

Geometric means of ocular parameters were estimated from least-squares means using general linear regression models, stratified by birth outcome groups and adjusted for the identified covariates. To assess longitudinal patterns, change in body stature parameters were modeled across consecutive developmental intervals (0–6, 6–12, 12–18, 18–24, and 24–36 months), enabling assessment of time-dependent associations between physical growth and ocular biometry and the identification of sensitive developmental windows. All statistical analyses were conducted using R software (version 4.0.5), with statistical significance defined as 2-sided P < 0.05.

Results

A total of 10,030 infants were included in the analysis, of whom 4376 (43.6%) were girls. The majority were delivered by cesarean section (62.9%). The mean age at breastfeeding cessation was 12.3 ± 7.8 months. Based on birth outcomes, 2492 (24.9%) were term with normal-birth-weight, 3674 (36.6%) were preterm, 690 (6.9%) were term with low-birth-weight, and 3174 (31.7%) were macrosomic (Supplementary Table S1).

The Table presents geometric means with 95% confidence intervals (CIs) for ocular parameters at 3 to 4 years of age across birth outcome groups, reported as unadjusted, minimally adjusted, and multivariable-adjusted estimates. After fully adjusting for potential confounders, UCVA displayed a graded pattern (P < 0.001): compared with term normal-birth-weight infants (0.219 [0.213–0.225] logMAR), UCVA was poorest in preterm infants (0.229 [0.224–0.234] logMAR), followed by term low-birth-weight infants (0.225 [0.215–0.235] logMAR). By contrast, macrosomic infants had UCVA comparable to, or marginally better than, the reference group (0.213 [0.207–0.219] logMAR). Refractive status also varied significantly by birth outcome. For spherical equivalent, preterm infants exhibited the greatest hyperopic shift (0.433 [0.410–0.456] diopter [D]), term low-birth-weight infants were intermediate (0.407 [0.358–0.456] D), and macrosomic infants had the lowest values (0.389 [0.360–0.419] D). Consistent patterns were found for both spherical and cylindrical components, with preterm infants showing the most pronounced alterations, followed by term low-birth-weight infants.

Table.

Distribution of Ocular Parameters at Ages 3 to 4 Years Across Birth Outcome Groups

	Mean (95% CI)
Ocular Parameters	Term Normal Birth Weight (N = 2492)	Preterm Birth (N = 3674)	Term Low Birth Weight (N = 690)	Macrosomia (N = 3174)	P Value*
Visual acuity (logMAR units)
Unadjusted	0.219 (0.214 to 0.224)	0.229 (0.225 to 0.233)	0.229 (0.219 to 0.238)	0.214 (0.210 to 0.219)	<0.001
Age and sex adjusted†	0.219 (0.214 to 0.224)	0.229 (0.225 to 0.233)	0.228 (0.219 to 0.237)	0.215 (0.211 to 0.219)	<0.001
Multivariable adjusted‡	0.219 (0.213 to 0.225)	0.229 (0.224 to 0.234)	0.225 (0.215 to 0.235)	0.213 (0.207 to 0.219)	<0.001
Spherical equivalent
Unadjusted	0.403 (0.379 to 0.428)	0.433 (0.413 to 0.453)	0.409 (0.363 to 0.456)	0.391 (0.370 to 0.413)	0.04
Age and sex adjusted†	0.404 (0.379 to 0.428)	0.433 (0.413 to 0.453)	0.409 (0.362 to 0.455)	0.393 (0.371 to 0.414)	0.05
Multivariable adjusted‡	0.395 (0.366 to 0.423)	0.433 (0.410 to 0.456)	0.407 (0.358 to 0.456)	0.389 (0.360 to 0.419)	0.03
Spherical power
Unadjusted	0.753 (0.726 to 0.779)	0.804 (0.782 to 0.827)	0.779 (0.728 to 0.830)	0.725 (0.701 to 0.749)	<0.001
Age and sex adjusted†	0.753 (0.726 to 0.780)	0.805 (0.783 to 0.827)	0.778 (0.727 to 0.829)	0.728 (0.704 to 0.752)	<0.001
Multivariable adjusted‡	0.746 (0.714 to 0.777)	0.804 (0.779 to 0.829)	0.773 (0.719 to 0.827)	0.727 (0.694 to 0.759)	<0.001
Cylindrical power
Unadjusted	−0.698 (−0.726 to −0.670)	−0.744 (−0.767 to −0.721)	−0.740 (−0.793 to −0.686)	−0.668 (−0.693 to −0.643)	<0.001
Age and sex adjusted†	−0.699 (−0.727 to −0.671)	−0.745 (−0.768 to −0.721)	−0.738 (−0.792 to −0.685)	−0.670 (−0.696 to −0.645)	<0.001
Multivariable adjusted‡	−0.702 (−0.735 to −0.669)	−0.743 (−0.770 to −0.717)	−0.732 (−0.788 to −0.675)	−0.676 (−0.710 to −0.642)	0.003

^*P values indicate the statistical significance of group differences and were derived from 1-way ANOVA based on linear models.

^†Adjusted for age at ocular measurements and infant sex.

^‡Multivariable adjusted for age at ocular measurements, infant sex, age at breastfeeding cessation, paternal age, maternal age, fetal plurality, and mode of delivery.

To further contextualize these findings, longitudinal growth trajectories for weight, length, body mass index (BMI), and HC were plotted by birth outcome group (Supplementary Fig. S1). Preterm infants demonstrated marked catch-up growth during the early postnatal period, term low-birth-weight infants showed a more modest catch-up pattern, and macrosomic infants displayed mild growth deceleration. These developmental patterns aligned with the group-wise ocular findings: groups exhibiting greater catch-up growth in early life also showed poorer visual outcomes at 3 to 4 years, whereas groups with minimal catch-up growth showed smaller ocular differences.

After additional adjustment for birth outcome categories, we evaluated direct associations between postnatal growth trajectories and ocular parameters (Figs. 1²³–4). Increases in weight, length, BMI, and HC were positively associated with visual acuity at ages 3 to 4 years, but only within the first year of life (see Figs. 1A–4A). Significant associations were only evident during 0 to 6 months and 6 to 12 months, with regression coefficients of approximately 0.004 per z-score increment. Specifically, during 0 to 6 months, greater gains in weight (β = 0.005, 95% CI = 0.003 to 0.007, P < 0.001), length (β = 0.005, 95% CI = 0.002 to 0.007, P = 0.001), and BMI (β = 0.004, 95% CI = 0.002 to 0.006, P < 0.001) were associated with visual acuity. During 6 to 12 months, greater gains in length (β = 0.004, 95% CI = 0.001 to 0.008, P = 0.008) and HC (β = 0.004, 95% CI = 0.001 to 0.008, P = 0.02) were associated with visual acuity. These sensitive windows overlapped with the periods of altered growth observed among infants with adverse birth outcomes (see Supplementary Fig. S1). Comparable time-dependent associations were identified for refractive status (see Figs. 1B–4B). Spherical power showed a positive association with growth (see Figs. 1C–4C), whereas cylindrical power demonstrated an inverse association (see Figs. 1D–4D).

Figure 1.

Association between weight growth and ocular parameters. (A) Weight growth and visual acuity. (B) Weight growth and spherical equivalent. (C) Weight growth and spherical error. (D) Weight growth and cylindric error. Ocular parameters were measured at 3 to 4 years of age. Weight growth during different developmental periods was evaluated using changes in weight-for-age z scores, reflecting relative weight change over time. Estimates represent the change in ocular parameters per 1-unit increase in weight-for-age Z-score change, expressed as diopters (D) for refraction and logMAR for visual acuity. Models were adjusted for age at breastfeeding cessation, paternal age, maternal age, fetal plurality, mode of delivery, and birth outcome.

Figure 2.

Association between height growth and ocular parameters. (A) Height growth and visual acuity. (B) Height growth and spherical equivalent. (C) Height growth and spherical error. (D) Height growth and cylindric error. Ocular parameters were measured at 3 to 4 years of age. Height growth during different developmental periods was evaluated using changes in height-for-age z scores, reflecting relative height change over time. Estimates represent the change in ocular parameters per 1-unit increase in height-for-age Z-score change, expressed as diopters (D) for refraction and logMAR for visual acuity. Models were adjusted for age at breastfeeding cessation, paternal age, maternal age, fetal plurality, mode of delivery, and birth outcome.

Figure 3.

Association between BMI growth and ocular parameters. (A) BMI growth and visual acuity. (B) BMI growth and spherical equivalent. (C) BMI growth and spherical error. (D) BMI growth and cylindric error. Ocular parameters were measured at 3 to 4 years of age. BMI growth during different developmental periods was evaluated using changes in BMI-for-age z scores, reflecting relative BMI change over time. Estimates represent the change in ocular parameters per 1-unit increase in BMI-for-age Z-score change, expressed as diopters (D) for refraction and logMAR for visual acuity. Models were adjusted for age at breastfeeding cessation, paternal age, maternal age, fetal plurality, mode of delivery, and birth outcome.

Figure 4.

Association between head circumference growth and ocular parameters. (A) Head circumference growth and visual acuity. (B) Head circumference growth and spherical equivalent. (C) Head circumference growth and spherical error. (D) Head circumference growth and cylindric error. Ocular parameters were measured at 3 to 4 years of age. Head circumference growth during different developmental periods was evaluated using changes in head circumference-for-age z scores, reflecting relative head circumference change over time. Estimates represent the change in ocular parameters per 1-unit increase in head circumference-for-age Z-score change, expressed as diopters (D) for refraction and logMAR for visual acuity. Models were adjusted for age at breastfeeding cessation, paternal age, maternal age, fetal plurality, mode of delivery, and birth outcome.

Discussion

In this large, population-based birth cohort, we identified robust and biologically coherent associations between early-life growth trajectories and subsequent ocular development, supporting a coordinated regulation of systemic and visual maturation. Children born with adverse birth outcomes, particularly those who were preterm or of low birth weight, exhibited measurable deficits in visual acuity and a more hyperopic refractive status at 3 to 4 years of age. Notably, the magnitude and timing of postnatal growth acceleration, especially within the first year of life, were strongly linked to later visual performance, highlighting a sensitive developmental window during which systemic growth exerts a formative influence on ocular development.

The Costa and Ventura group conducted longitudinal comparisons of visual development between term and preterm infants and examined the effects of early stimulation on visual-motor pathways, thereby establishing longitudinal links among birth status, early experience, and visual development.^19,20 Our study provides further prospective longitudinal evidence linking postnatal growth trajectories with early visual development in high-risk infants. By integrating detailed anthropometric data during early childhood with standardized ophthalmic assessments, we identified that rapid catch-up growth was associated with suboptimal visual acuity and refractive alterations. These results extend previous observations of impaired visual outcomes in preterm and low-birth-weight infants, suggesting that accelerated or dysregulated somatic growth may disturb the synchronized maturation of ocular and systemic development. For example, Jain et al.⁶ reported persistent visual deficits in extremely preterm individuals up to 19 years of age, independent of retinopathy of prematurity, whereas Raffa et al.¹² showed that preterm children exhibited altered macular morphology correlated with HC and body size, highlighting the interdependence of somatic and ocular growth. Similarly, O'Connor et al.²¹ and Spierer et al.²² found that low birth weight and preterm infants displayed long-term visual impairments even in the absence of severe neonatal complications. Collectively, these findings identify early postnatal growth as a critical determinant of visual development and highlight the need for integrated monitoring of physical growth and visual function in high-risk infants to guide timely, growth-sensitive interventions.

Previous studies have reported correlations between ocular biometrics and various systemic growth indicators, yet the findings have been inconsistent. Several have reported negative correlations between refraction and stature, suggesting that taller and heavier individuals tend to have less hyperopic refractive status.^23,24 Secular trends showing concurrent increases in average height and weight alongside the global rise in myopia point toward potential shared biological or environmental determinants.^25,26 In contrast, several investigations have failed to replicate these associations, reporting no significant relationship between height and refraction.^9,27 Similarly, Wong et al. observed that body weight and BMI were not significantly associated with ocular parameters among Singaporean Chinese adults.²⁸ These inconsistencies likely reflect differences in age composition and study design. Most prior studies were conducted among school-aged children or adults, in whom visual outcomes are already shaped by environmental and educational factors. Evidence from the critical early-life period, during early childhood (birth to 3 years), remains limited, primarily due to challenges in obtaining reliable ophthalmic measurements in infants. More importantly, previous research has largely relied on cross-sectional designs, overlooking the dynamic nature of growth. It is increasingly evident that longitudinal trajectories of somatic development, rather than static anthropometric indicators, may exert a formative influence on ocular growth and subsequent visual function.⁷

We also found that the associations between physical growth and visual outcomes were confined to the first year of life, corresponding to the period of most rapid postnatal development. This stage also overlapped with the phase during which growth trajectory differences between infants with adverse birth outcomes and their term counterparts were most pronounced. Consistent with previous studies, catch-up growth in preterm and small-for-gestational-age infants peaks within the first postnatal year and is closely linked to subsequent developmental and metabolic outcomes.^29–32 For instance, Burstein et al. reported that preterm birth confers sustained disruptions in visual attention across the first 2 years of life,³² whereas Tang et al. demonstrated that rapid, accelerated catch-up growth is associated with elevated inflammatory and metabolic markers in adolescence among previously institutionalized children, underscoring long-term cardiometabolic risk.³¹ Biologically, the first postnatal year likely represents a critical window when ocular structures and neural circuits mature in concert with systemic growth under active hormonal and metabolic regulation. Thereafter, visual development is increasingly shaped by environmental factors, such as nutrition, caregiving, and light exposure.^33–35 Hence, the attenuation of associations at older ages suggests that systemic growth predominates in early ocular development, whereas environmental visual input subsequently becomes the key driver of functional refinement.

Several strengths are evident in this study, including its prospective design, large sample size, and the standardized measurement of refractive error under cycloplegia together with uniform protocols for assessing visual acuity. By longitudinally and continuously tracking multiple growth indicators throughout early childhood in high-risk infants, the study identified critical developmental windows when systemic growth most strongly influences ocular maturation, offering new evidence for the integrated regulation of early physical and visual development.

Our study also has some limitations. First, despite comprehensive adjustment for a wide range of potential confounders, residual confounding from inaccurately measured or unmeasured factors cannot be fully excluded. Although infants with clinically documented congenital infections and other severe systemic diseases with potential ocular involvement were excluded, detailed information on all infections and later-acquired non-congenital conditions that may influence visual development was not uniformly available across the entire cohort. Second, the assessment of ocular development was limited by the scope of available measures. We focused on visual acuity and refractive status because they are clinically informative and biologically sensitive indicators that reflect both optical and neural aspects of early visual maturation. These parameters provide a meaningful window into overall visual system development during infancy, when more advanced imaging techniques are often impractical. However, the absence of complementary biometric data, such as axial length, corneal curvature, and retinal morphology, limited our ability to clarify the structural mechanisms linking systemic growth to ocular outcomes. Third, the cohort consisted predominantly of Han Chinese participants, which restricts the examination of potential ethnic variations in growth and ocular development. Future investigations across more diverse populations, combined with mechanistic and imaging-based studies, are warranted to validate these findings and further clarify the biological basis of growth-vision interactions.

Conclusions

This large prospective cohort provides robust longitudinal evidence that early-life systemic growth trajectories are closely associated with subsequent ocular development in high-risk infants. Accelerated catch-up growth during the first year of life was linked to poorer visual acuity and a more hyperopic refractive profile at 3 to 4 years of age, suggesting that rapid somatic growth may disrupt the coordinated maturation of ocular and systemic systems. These findings highlight the early childhood (birth to 3 years) as a critical window for the integrated regulation of physical and visual development. From a clinical perspective, incorporating visual assessments into routine early growth surveillance, particularly for infants born preterm or with low birth weight, may facilitate early identification of children at risk for suboptimal visual outcomes and inform timely, individualized interventions to promote healthy development.