Determinants of axial length growth in infantile persistent fetal vasculature following early lensectomy–vitrectomy: a retrospective cohort

Abstract

Purpose

To identify clinical determinants of axial length (AL) before and after early surgery in infantile persistent fetal vasculature (PFV).

Methods

Single-center retrospective cohort of 58 unilateral PFV eyes (2019–2022). All infants underwent lensectomy, posterior capsulectomy, vitrectomy, and targeted cauterization under general anesthesia before 12 months. Baseline AL, phenotype, and structural findings (enlarged ciliary processes, retrolental membrane, posterior capsular defect, corneal opacity, posterior synechiae, retinal detachment, Morning Glory disc anomaly) were recorded. Mean follow-up was 21.9 ± 11.4 months (12–48). Primary endpoint: poor growth, defined a priori as an absolute postoperative inter-eye AL difference ≥ 2.0 mm at last visit. Secondary outcomes: preoperative inter-eye asymmetry and postoperative AL change (ΔAL). Multivariable models evaluated predictors.

Results

All surgeries occurred < 12 months. Preoperatively, anterior PFV, corneal opacity, posterior synechiae, retinal detachment, and Morning Glory anomaly were associated with shorter AL versus the fellow eye (all P < 0.05), whereas posterior capsular defect correlated with longer AL (P < 0.05). Postoperatively, greater ΔAL occurred when enlarged ciliary processes, retrolental membrane, and posterior capsular defect were absent (all P < 0.05). Poor growth occurred in 12.1%; presentation ≤ 6 months independently predicted poor growth (OR 11.62; 95% CI 1.20–112.46; P = 0.034).

Conclusions

In infantile PFV, age at onset, phenotype, and specific structural abnormalities determine AL trajectories. Early standardized surgery may favorably modify growth, whereas complicated cataract in early infancy signals risk for subnormal postoperative elongation and warrants closer surveillance.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12886-026-05093-0.

Introduction

Persistent fetal vasculature (PFV)—also known as persistent hyperplastic primary vitreous (PHPV)—is a congenital anomaly caused by failure of the primary fetal vasculature to regress [1]. PFV is classically categorized as anterior, posterior, or combined according to the location of persistent tissue [2], and may affect the lens, ciliary processes, anterior chamber, vitreous, retina, and optic nerve [3]. These structural changes can influence ocular growth, leading to variability in axial length (AL) between PFV and fellow eyes [4, 5].

Indeed, several studies have shown shorter AL and smaller corneal diameter in PFV, particularly when anterior segment collapse or retinal detachment is present [6, 7], whereas others reported minimal AL differences [8], —implying that growth trajectories are phenotype- and structure-dependent rather than uniform across PFV.

Surgical management is undertaken to relieve traction, clear the visual axis, and prevent/treat complications such as retinal detachment [9, 10]. Although postoperative increases in AL have been described [9, 11, 12], the determinants of AL development after a standardized lensectomy–vitrectomy protocol in infants have not been fully elucidated. Moreover, few reports have combined paired inter-eye analyses with clinically interpretable endpoints for growth failure.

In this retrospective study of unilateral PFV operated before 12 months of age, we used fellow eyes as internal controls to: (1) describe pre- and postoperative inter-eye AL differences; (2) quantify axial growth (ΔAL) after a uniform surgical approach; and (3) identify structural and clinical predictors of poor postoperative growth—defined a priori as postoperative inter-eye AL difference ≥ 2.0 mm at last follow-up—with an age-adjusted sensitivity analysis reported in the Supplement. Therefore, this study aimed to identify clinical and structural determinants of axial length growth after early lensectomy-vitrectomy in infants with unilateral PFV using a fellow-eye control design.

Materials and methods

Ethics approval and consent to participate

This study adhered to the tenets of the Declaration of Helsinki and was approved by the Ethics Committee of Beijing Tongren Hospital, Capital Medical University (TREC2022-XJS05). The requirement for informed consent for participation in this retrospective study was waived by the Ethics Committee because only de-identified clinical data and images were used.

Study design and participants

We retrospectively reviewed consecutive children with unilateral persistent fetal vasculature (PFV) managed at our tertiary care facility from January 1, 2019 to January 1, 2022. Inclusion criteria: clinical PFV diagnosis; age at surgery < 12 months; follow-up ≥ 12 months. Exclusion: PFV mimickers (e.g., retinoblastoma, retinal dysplasia), severe systemic illness affecting treatment/surgery, or incomplete records.

Clinical assessment and phenotyping

PFV was categorized as anterior or posterior/mixed using standard clinical criteria. Predefined structural features recorded at baseline: corneal opacity, shallow anterior chamber, posterior synechiae, enlarged ciliary processes, retrolental vascularized membrane, posterior capsular defect, retinal detachment, and Morning Glory disc anomaly. Demographic and ocular variables (age, sex, laterality, history) were extracted.

Surgical protocol

All procedures were performed by a single surgeon (Hai Lu) under general anesthesia and consisted of lensectomy, posterior capsulectomy, and vitrectomy via pars plicata/plana (25-gauge) or limbal (25-gauge) approach, with targeted cauterization/resection of the PFV stalk. Adjunctive procedures (pupilloplasty, anterior chamber reconstruction, synechiolysis, retinal laser) were performed as indicated.

Axial length measurement

All axial length measurements (preoperative and postoperative) were performed under general anesthesia using standardized A-scan ultrasonography by experienced examiners following a uniform protocol. When multiple postoperative visits were available, the last visit was used for postoperative AL.

Outcomes

Primary endpoint: poor growth, defined a priori as absolute postoperative inter-eye AL difference ≥ 2.0 mm at last follow-up (|AL_PFV − AL_fellow| ≥2.0 mm).

Secondary outcomes (main text): (1) inter-eye AL difference at pre- and postoperative timepoints; (2) axial growth ΔAL for each eye (postoperative − preoperative, mm).

Sensitivity analysis (Supplement): age-adjusted criterion using annualized PFV-eye AL gain; poor growth defined as below the age-stratum 10th percentile (P10) for ≤ 6 months vs. > 6 months at presentation (reported in Supplementary Table S1).

Statistical analysis

Analyses were performed in SPSS v17. Normality was assessed using Shapiro–Wilk. Within-patient inter-eye comparisons (PFV vs. fellow) at pre/post used paired t-tests (or Wilcoxon signed-rank when non-Gaussian). ΔAL comparisons between fellow and PFV eyes used paired tests. Between-group (feature present vs. absent) comparisons used independent-samples t-tests (or Mann–Whitney U). Multivariable logistic regression (backward selection) identified predictors of poor growth (primary endpoint), entering age at presentation (≤ 6 vs. > 6 months), clinical classification, surgical approach, follow-up duration, and prespecified structural features. Results are reported as odds ratios (ORs) with 95% CIs and exact P values. Two-sided α = 0.05.

Results

Demographics and baseline characteris

We included 58 children with unilateral PFV and congenital cataract (36 boys, 22 girls). The mean age at presentation was 7.88 ± 3.31months (range, 3–12), with a mean follow-up of 21.95 ± 11.37 months (range, 12–48). Phenotypes comprised anterior PFV (38/58, 65.5%)and posterior/mixed PFV (20/58, 34.5%). Preoperative structural findings included corneal opacity (2/58, 3.4%), shallow anterior chamber(12/58, 20.7%), posterior synechiae (18/58, 31.0%), enlarged ciliary processes (36/58, 62.1%), retrolental vascularized membrane (35/58,60.3%), posterior capsular defect (12/58, 20.7%), retinal detachment (20/58, 34.5%), and Morning Glory disc anomaly (4/58, 6.9%). Allpatients underwent lensectomy, posterior capsulectomy, and vitrectomy; the surgical route was limbal in 22/58 (37.9%) and parsplicata/plana in 36/58 (62.1%). Table 1 summarizes baseline characteristics.

Table 1.

Basic information, various manifestations, clinical classification and surgical approach of patients with PFV

Variable (N = 58)	Patients, n (%)
Gender
Male patients	36 (62.07%)
Female patients	22 (37.03%)
Age at presentation	7.88 ± 3.31 months (3 to 12 months)
≤ 6 months	22 (37.93%)
> 6 months	36 (62.07%)
Eye
Right eyes	32 (55.17%)
Left eyes	26 (44.83%)
Various manifestations
Corneal opacity	2 (3.45%)
Shallow anterior chamber	12 (20.69%)
Posterior synechia	18 (31.03%)
Enlarged ciliary processes	36 (62.07%)
Lens opacity	58 (100%)
Retrolental vascularized membrane	35 (60.34%)
Posterior capsular defect	12 (20.69%)
Retinal detachment	20 (34.48%)
Morning glory syndrome	4 (6.90%)
Clinical classification
Anterior PFV	38 (65.52%)
Posterior PFV	0
Combined PFV	20 (34.48%)
Surgical approach
Translimbal lensectomy and vitrectomy	22 (37.93%)
Pars plicata lensectomy and vitrectomy	36 (62.07%)

Preoperative inter-eye AL comparisons

Overall, baseline AL did not differ significantly between fellow and PFV eyes (19.95 ± 1.41 mm vs. 19.73 ± 2.02 mm; t = 1.534, P = 0.131). PFV eyes were shorter than fellow eyes in the presence of corneal opacity (t = 15.000, P = 0.042), posterior synechiae (t = 2.163, P = 0.045), retinal detachment (t = 2.929, P = 0.006), Morning Glory disc anomaly (t = 3.498, P = 0.040), and anterior PFV (t = 2.929, P = 0.006). In contrast, posterior capsular defect was associated with longer AL in PFV eyes than fellow eyes (t = − 3.134, P = 0.010). Stratified analyses indicated that the absence of enlarged ciliary processes or retrolental membrane was associated with greater inter-eye asymmetry favoring the fellow eye (P ≈ 0.02–0.03). Detailed pre-/postoperative inter-eye comparisons by feature are provided in Table 2 and visualized in Fig. 1.

Table 2.

Pre- and postoperative inter-eye AL by clinical features

Correlation analysis		preoperative				postoperative
		AL of the fellow eye (mm)	AL of the affected eye (mm)	t	P value	AL of the fellow eye (mm)	AL of the affected eye (mm)	t	P value
		19.95 ± 1.41	19.73 ± 2.02	1.534	0.131	21.69 ± 1.36	21.68 ± 2.09	0.037	0.970
Clinical manifestations
Corneal opacity	+	20.85 ± 1.63	19.35 ± 1.48	15.000	0.042*	22.85 ± 2.05	22.00 ± 0.57	0.810	0.567
	-	19.92 ± 1.41	19.74 ± 2.05	1.212	0.231	21.65 ± 1.33	21.67 ± 2.13	-0.109	0.914
Shallow anterior chamber	+	20.05 ± 1.30	19.38 ± 2.01	1.823	0.096	21.92 ± 1.58	21.29 ± 2.31	1.994	0.072
	-	19.93 ± 1.45	19.82 ± 2.04	0.698	0.489	21.63 ± 1.30	21.79 ± 2.05	-0.655	0.516
Posterior synechia	+	19.72 ± 1.34	19.42 ± 1.77	2.163	0.045*	21.55 ± 1.67	20.79 ± 2.03	3.328	0.004*
	-	19.94 ± 1.46	19.86 ± 2.14	0.453	0.653	21.75 ± 1.21	22.08 ± 2.02	-1.277	0.209
Enlarged ciliary processes	+	20.08 ± 1.51	19.95 ± 2.29	0.602	0.551	21.76 ± 1.35	21.60 ± 2.07	0.693	0.493
	-	19.74 ± 1.24	19.36 ± 1.46	2.386	0.027*	21.58 ± 1.39	21.83 ± 2.17	-0.631	0.535
Lens opacity	+	19.95 ± 1.41	19.73 ± 2.02	1.534	0.131	21.69 ± 1.36	21.68 ± 2.09	0.037	0.970
	-	/	/	/	/	/	/	/	/
Rentrolental vascularized membrane	+	20.11 ± 1.54	19.98 ± 2.33	0.563	0.577	21.83 ± 1.30	1.67 ± 2.05	0.662	0.512
	-	19.71 ± 1.17	19.33 ± 1.41	2.481	0.021*	21.48 ± 1.44	21.70 ± 2.20	-0.601	0.554
Rupture of posterior lens capsule	+	21.23 ± 1.12	22.26 ± 1.35	-3.134	0.010*	22.74 ± 1.02	23.40 ± 1.41	-1.52	0.157
	-	19.62 ± 1.29	19.06 ± 1.61	4.352	< 0.001*	21.42 ± 1.31	21.24 ± 2.02	0.799	0.429
Retinal detachment	+	19.75 ± 1.35	19.37 ± 1.79	2.929	0.006*	22.30 ± 1.21	22.47 ± 1.75	-0.498	0.624
	-	20.34 ± 1.48	20.40 ± 2.30	-0.172	0.865	21.37 ± 1.34	21.27 ± 2.16	0.413	0.682
Morning glory syndrome	+	20.10 ± 2.76	18.45 ± 3.63	3.498	0.040*	22.89 ± 1.10	22.06 ± 1.26	1.661	0.195
	-	19.94 ± 1.31	19.82 ± 1.88	0.831	0.410	21.60 ± 1.34	21.66 ± 2.15	-0.248	0.805
Clinical classification
Anterior PFV		19.75 ± 1.35	19.37 ± 1.79	2.929	0.006*	21.37 ± 1.34	21.27 ± 2.16	0.413	0.682
Combined PFV		20.34 ± 1.48	20.40 ± 2.30	-0.172	0.865	22.30 ± 1.21	22.47 ± 1.75	-0.498	0.624

Fig. 1.

Postoperative inter-eye axial-length (AL) difference (PFV − fellow) by anatomic features. Circles show the mean and horizontal bars the 95% CI for “Present” and “Absent” strata; the dashed vertical line at 0 mm denotes no inter-eye difference. Positive values indicate the PFV eye is longer. Paired, within-patient data. Abbreviations: PFV, Persistent fetal vasculature; AL, Axial length; CI, Confidence interval

Postoperative AL and axial growth (ΔAL)

At last follow-up, both eyes elongated, and the overall inter-eye AL difference was no longer significant (fellow 21.69 ± 1.36 mm vs. PFV 21.68 ± 2.09 mm; t = 0.037, P = 0.970). Posterior synechiae remained associated with shorter postoperative AL in PFV eyes (t = 3.328, P = 0.004). For axial growth, ΔAL (postoperative − preoperative) was 1.74 ± 1.30 mm in fellow eyes versus 1.96 ± 1.92 mm in PFV eyes (t = − 1.361, P = 0.179). PFV eyes showed greater ΔAL than fellow eyes when enlarged ciliary processes were absent (t = − 2.320, P = 0.031), retrolental membrane was absent (t = − 2.321, P = 0.030), or posterior capsular defect was absent (t = − 2.073, P = 0.044). ΔAL did not differ by clinical classification or surgical route. These comparisons are summarized in Table 3. Figure 2 shows a clean mean-trajectory plot of PFV-eye AL (mean ± SD at pre-op and last visit) stratified by enlarged ciliary processes.

Table 3.

Comparative analysis of axial length changes before and after operation

Correlation analysis		D-value in AL between pre-operation and post-operation in the fellow eye (mm)	D-value in AL between pre-operation and post-operation in the affected eye (mm)	t	P value
		1.74 ± 1.30	1.96 ± 1.92	-1.361	0.179
Age at presentation
≤ 6 months		2.46 ± 1.48	2.84 ± 2.59	-1.166	0.256
> 6 months		1.27 ± 0.91	1.38 ± 1.00	-0.705	0.486
Clinical manifestations
Corneal opacity	+	2.00 ± 0.42	2.65 ± 0.92	-0.684	0.618
	-	1.73 ± 1.31	1.93 ± 1.95	-1.238	0.221
Shallow anterior chamber	+	1.87 ± 0.81	1.91 ± 0.96	-0.158	0.878
	-	1.71 ± 1.40	1.97 ± 2.11	-1.388	0.172
Posterior synechia	+	1.58 ± 0.85	1.37 ± 1.14	1.029	0.318
	-	1.81 ± 1.46	2.22 ± 2.15	-1.977	0.055
Enlarged ciliary processes	+	1.68 ± 1.40	1.65 ± 2.02	0.149	0.882
	-	1.84 ± 1.13	2.46 ± 1.68	-2.320	0.031*
Lens opacity	+	1.74 ± 1.30	1.96 ± 1.92	-1.361	0.179
	-	/	/	/	/
Rentrolental vascularized membrane	+	1.72 ± 1.42	1.69 ± 2.04	0.157	0.877
	-	1.77 ± 1.11	2.37 ± 1.69	-2.321	0.030*
Rupture of posterior lens capsule	+	1.52 ± 1.29	1.14 ± 0.79	1.256	0.235
	-	1.80 ± 1.30	2.17 ± 2.08	-2.073	0.044*
Retinal detachment	+	1.96 ± 1.73	2.08 ± 2.45	-0.396	0.696
	-	1.62 ± 1.00	1.90 ± 1.62	-1.405	0.168
Morning glory syndrome	+	2.79 ± 3.02	3.61 ± 4.75	-0.925	0.423
	-	1.66 ± 1.10	1.84 ± 1.57	-1.079	0.286
Clinical classification
Anterior PFV		1.62 ± 1.00	1.90 ± 1.62	-1.405	0.168
Combined PFV		1.96 ± 1.73	2.08 ± 2.45	-0.396	0.696
Surgical approach
Translimbal lensectomy and vitrectomy		1.96 ± 1.69	2.00 ± 2.66	-0.148	0.884
Pars plicata lensectomy and vitrectomy		1.60 ± 0.98	1.93 ± 1.33	-1.665	0.105

Fig. 2.

PFV-eye axial-length growth stratified by enlarged ciliary processes. Points indicate group means ± SD at preoperative and final postoperative visits; the line connects the two means within each stratum. The x-axis shows follow-up (years); individual trajectories are omitted for clarity. Abbreviations: PFV, persistent fetal vasculature; AL, axial length; SD, standard deviation

Poor growth (primary endpoint)

Poor growth—defined a priori as absolute postoperative inter-eye AL difference ≥ 2.0 mm at last follow-up—occurred in 7/58 (12.1%). In multivariable logistic regression (Table 4), age ≤ 6 months independently increased the odds of poor growth (OR 11.62; 95% CI 1.20–112.46; P = 0.034), whereas clinical classification, surgical route, and follow-up duration were not significant.

Table 4.

Multivariable logistic analysis results for factors associated with poor axial growth in patients with PFV of the Eye

Factors	Poor axial growth (AL1 - AL2 ≥ 2 mm; n = 7)	OR (95% CI)	P value
Age at presentation			0.034*
≤ 6 months	6 (85.71%)	11.62 (1.20–112.46)
> 6 months	1 (14.29%)	1.000
Clinical classification			0.224
Anterior PFV	6 (85.71%)	4.58 (0.39–53.31)
Combined PFV	1 (14.29%)	1.000
Surgical approach			0.226
Translimbal lensectomy and vitrectomy	4 (57.14%)	3.25 (0.48–21.82)
Pars plicata lensectomy and vitrectomy	3 (42.86%)	1.000
The follow-up period	-	1.03 (0.95–1.13)	0.465

Sensitivity analysis

As a robustness check, we also evaluated age-adjusted annualized AL-gain percentiles (≤ 6 vs. > 6 months at presentation). Using the 10th percentile within each stratum as a threshold yielded similar poor-growth proportions; full results are reported in Supplementary Table S1.

Discussion

In this single-surgeon cohort of unilateral PFV with surgery performed before 1 year of age, we found that: (1) preoperative axial-length (AL) asymmetry was tightly linked to structural anatomy—PFV eyes were shorter when corneal opacity, posterior synechiae, tractional retinal detachment, or Morning Glory optic disc anomaly were present, but longer when posterior capsular defect had occurred; (2) after standardized lensectomy–posterior capsulectomy–vitrectomy, the inter-eye AL difference was no longer significant at the final visit, and ΔAL (postoperative − preoperative) was significantly greater in PFV eyes without enlarged ciliary processes, retrolental vascularized membrane, or posterior capsular defect; and (3) 12.1% of patients met our clinically interpretable poor-growth endpoint (absolute postoperative inter-eye AL difference ≥ 2.0 mm), with age ≤ 6 months as an independent risk factor, while surgical approach showed no association with postoperative growth (consistent with our regression results).

Prior reviews and retrospective studies generally indicate that anterior PFV carries a more favorable prognosis than posterior/combined PFV; the goals of surgery are to release traction, clear the visual axis, and address complications, with lensectomy plus vitrectomy commonly used for posterior/combined disease [4, 5, 13, 14, 15]. Our findings are consistent with this view: eyes exhibiting tractional features—such as a retrolental fibrovascular membrane, posterior synechiae, or enlarged ciliary processes—tend to have a shorter baseline axial length and attenuated postoperative elongation, whereas posterior capsular defect suggests that lenticulo-vitreoretinal traction has been released, corresponding to a longer preoperative axial length. These observations support a pathophysiologic framework of traction → scleral remodeling → axial-length regulation [16].

Early classic series reported that anterior PFV alone can achieve relatively better visual outcomes when managed with surgery plus optical correction and amblyopia therapy, whereas posterior involvement often limits function because of macular/posterior pole abnormalities; this aligns with our observation that anatomy shapes growth trajectories [17, 18]. Furthermore, recent multicenter studies and systematic reviews emphasize that surgical timing and anatomical complexity jointly determine anatomic and functional outcomes, and that early release of traction may reduce subsequent complications and structural decompensation [16, 19]. Our data provide quantitative biometry to support this paradigm: in the absence of traction-related signs, the PFV eye exhibits greater ΔAL, suggesting that timely intervention may steer axial growth toward the trajectory of the fellow eye.

Huang et al. reported in a retrospective study that PFV eyes tend to be shorter and are associated with poorer visual function, although differences across subtypes are substantial; Yeh et al. showed that vitrectomy and/or lensectomy for posterior/combined PFV yields acceptable anatomic outcomes but limited functional results. Using the fellow eye as an internal control, our study further delineates the coupling between anatomic signatures and axial-length trajectories, and introduces a 2.0-mm inter-eye difference as a clinically interpretable threshold for growth failure, complementing prior literature that focused primarily on anatomic and visual outcomes [4, 5]. In addition, regarding the relationship between surgical approach and outcomes, recent studies have centered on retinal detachment risk and long-term complications [20]; we did not observe a significant association between approach and ΔAL, which supports the clinical view that approach can be individualized when safety is ensured, though prospective data are still needed for confirmation.

Clinical implications. First, an inter-eye axial-length difference of 2.0 mm at the final postoperative visit can serve as a quantitative screening threshold for children at high risk of amblyopia and refractive abnormalities; presentation at ≤ 6 months signals a higher probability of growth failure and warrants closer follow-up and earlier amblyopia management. Second, traction-related structural signs (e.g., retrolental membrane, enlarged ciliary processes, posterior synechiae) indicate a risk of growth suppression; by contrast, when these signs are minimal or absent, the PFV eye demonstrates more robust ΔAL, supporting strategies that relieve traction early and restore a clear visual axis. Third, given the phenotypic heterogeneity of PFV and variability in growth trajectories, using the fellow eye as an internal control provides a more robust real-world evaluation framework in infants and young children. These conclusions are directionally consistent with prior observations on axial-length and corneal-diameter asymmetry and help explain why a subset of PFV eyes shows limited growth despite anatomical success [21].

Why some infants still show limited growth. We speculate several contributors: (i) earlier onset leading to a longer period of visual deprivation; (ii) an intrinsic microphthalmic tendency in a subset of PFV phenotypes; and (iii) residual or microscopic traction despite anatomically successful surgery. Our age-percentile (P10) sensitivity analysis showed that the proportion of eyes with annualized AL gain below the age-specific threshold was similar across age strata, suggesting that age distribution is not the sole driver (see Supplementary Table S1). This aligns with growing attention to the long-term complications of PFV (e.g., glaucoma, retinal detachment), underscoring the need to monitor structural status and growth in parallel during follow-up [22].

Strengths and limitations. The strengths of this study include: (i) a uniform age window (surgery before 1 year), which reduces age-related variability in ocular growth; (ii) a single-surgeon, standardized surgical protocol; and (iii) paired-eye analyses, which minimize inter-individual variability. Limitations are the retrospective design, lack of standardized visual/refractive data in very young children, and potential A-scan measurement error under general anesthesia. A limitation of this study is the variability in follow-up duration, which may influence postoperative axial length change and the primary endpoint; although statistical adjustment was performed, residual bias cannot be excluded. In addition, although stratification by structural features (e.g., retrolental membrane, enlarged ciliary processes) was prespecified, residual confounding by severity gradients cannot be excluded. Generalizability to bilateral PFV or non-tertiary settings also remains to be determined.

Future directions. Prospective, multicenter cohorts using more standardized biometry and imaging (e.g., swept-source biometry, quantitative traction scores) are needed to validate the 2.0-mm inter-eye difference as a clinically practical growth endpoint. Future studies should incorporate refractive and amblyopia outcomes and evaluate whether earlier surgery or targeted membrane management can further improve growth trajectories and functional outcomes in the highest-risk phenotypes.

Conclusion

In infants with unilateral PFV, axial growth is heterogeneous and tightly coupled to structural anatomy. Standardized surgery generally normalizes inter-eye AL differences, but early age at presentation marks a subgroup at risk for persistent asymmetry. Identifying and addressing traction-related features may optimize postoperative ocular growth and downstream visual rehabilitation.

Electronic Supplementary Material

Below is the link to the electronic supplementary material.

Abbreviations

AL

Axial length

CI

Confidence interval

ΔAL

Change in axial length (postoperative minus preoperative value)

NSFC

National Natural Science Foundation of China

OR

Odds ratio

PFV

Persistent fetal vasculature

SD

Standard deviation

Author contributions

Guangda Deng and Mingzhen Yuan: Designed the experiment and wrote the manuscript. Tianyu Wang: Collected data and organized imaging materials, performed statistical analysis and created figures. Jing Ma: Reviewed the manuscript. Liang Li: Conducted the research and performed statistical analysis. Jinghua Liu: Supervised the research and provided funding support. Songfeng Li: Provided administrative and material support, guidance, and supportive contributions. Hai Lu: Provided administrative and material support, guidance, and supportive contributions.All authors listed have made a substantial, direct, and intellectual contribution to the work and approved it for publication.

Funding

This study was supported by the National Natural Science Foundation of China (NSFC) [Grant No. 82571254]. The authors express gratitude to all participants for their help.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request, subject to institutional approval and the protection of patient confidentiality.

Declarations

Ethics approval and consent to participate

This study adhered to the tenets of the Declaration of Helsinki and was approved by the Ethics Committee of Beijing Tongren Hospital, Capital Medical University (TREC2022-XJS05). The requirement for informed consent for participation in this retrospective study was waived by the Ethics Committee because only de-identified clinical data and images were used.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.