Macular vessel density in highly myopic eyes with a history of uncomplicated cataract surgery: a cross-sectional OCTA study

Abstract

Background

High myopia is associated with structural and microvascular retinal alterations. However, macular microvascular characteristics in highly myopic eyes with a history of cataract surgery remain insufficiently characterized, particularly in comparison with non-operated highly myopic eyes. This study aimed to compare macular optical coherence tomography angiography (OCTA) parameters between highly myopic eyes with and without a history of uncomplicated cataract surgery.

Methods

This cross-sectional, observational study included 61 eyes of 37 highly myopic patients who underwent uncomplicated cataract surgery as the study group and 71 eyes of 42 non-operated highly myopic patients as the control group. Retinal vascular structures in the macular regions were examined in detail and quantitatively assessed using optical coherence tomography and OCTA. Data from these two groups were compared.

Results

There were no statistically significant differences between the groups in terms of age, sex, best corrected visual acuity, or axial length. The mean surgery-to-imaging interval in the study group was 18.2 months. Significantly lower superficial and deep capillary plexus vessel density was observed in most evaluated macular regions in the study group compared with the control group. However, no significant difference was found in the foveal avascular zone area between the groups (p = 0.808).

Conclusions

Highly myopic eyes with a history of uncomplicated cataract surgery had lower macular vessel density than non-operated highly myopic eyes at a later postoperative assessment. Because of the cross-sectional design, these findings represent between-group differences and do not establish postoperative change or a causal effect of cataract surgery. Further longitudinal studies are needed to determine the clinical significance of these vascular differences and their potential relationship with visual function.

Introduction

Myopia is one of the leading causes of visual impairment worldwide and has reached epidemic levels in some countries [1]. High myopia is a significant global public health concern, and as of 2023, it has been shown to affect approximately 4% of the population (13 million people) in the United States [2]. It is estimated that the number of people experiencing vision loss due to high myopia may increase sevenfold from 2000 to 2050 [3]. High myopia is characterised by severe anatomical changes in the retina that limit the best corrected visual acuity (BCVA), which can lead to complications resulting in blindness [4].

In high myopia, refractive error is ≤ -6.00 diopters (D) and/or axial length of the eye is ≥ 26 mm. Findings of myopic degeneration include vitreous liquefaction, posterior vitreous detachment, retinal pigment epithelium thinning, prominence of choroidal vessels, peripheral retinal degenerations, retinal detachment, lacquer cracks in Bruch’s membrane, choroidal neovascularization, Fuchs spot, macular hole, optic disc changes, posterior staphyloma, crescent formation, chorioretinal atrophy, and macular atrophy [5].

Optical coherence tomography angiography (OCTA) is an emerging diagnostic tool in clinical practice that utilises infrared wavelengths to provide non-invasive, high-resolution imaging of the retinal microvascular structure [6]. Recent advancements in OCTA technology have enabled the mapping of ocular circulation at the capillary level. OCTA holds great potential for measuring microvascular changes in response to physiological and pathological alterations.

With the rapid increase in the myopic population, cataract surgery has also become more common in individuals with high myopia. However, cataract surgery in highly myopic eyes presents unique challenges because of excessive axial elongation, deep anterior chamber, zonular weakness, posterior staphyloma, and vitreoretinal interface abnormalities. These anatomical characteristics may increase the risk of both surgical and postoperative complications, including retinal detachment, macular hole formation, and progression of myopic maculopathy. Furthermore, pre-existing retinal and choroidal alterations in highly myopic eyes may limit postoperative visual recovery despite technically successful surgery [7, 8]. Therefore, both the surgeon and the patient should thoroughly evaluate the benefits and risks of the surgery and determine the most appropriate timing for the procedure. With proper patient selection, phacoemulsification surgery is frequently preferred for cataracts in highly myopic patients due to its good visual acuity outcomes, low complication rates, high predictability, and stability [9].

Although short-term OCTA findings after cataract surgery have previously been investigated in highly myopic eyes, macular microvascular characteristics at later postoperative assessments remain insufficiently characterized, particularly in comparison with non-operated highly myopic eyes. Therefore, the present study aimed to compare macular OCTA parameters between highly myopic eyes with a history of uncomplicated cataract surgery and highly myopic eyes without previous cataract surgery. The use of a high-myopia comparison group was intended to reduce the confounding influence of vascular alterations related to axial elongation and high myopia itself.

Methods

Study design and participants

This cross-sectional, prospective observational study was reviewed and approved by the Clinical Research Ethics Committee of Şişli Hamidiye Etfal Training and Research Hospital, University of Health Sciences, with number 2640. The study was conducted between 2020 and 2024 following ethics committee approval. All patients had provided informed written consent. This study was conducted in accordance with the tenets of the Declaration of Helsinki.

Patients aged 40 years or older with high myopia, defined as a preoperative spherical equivalent refractive error of ≤ − 6.00 D and/or an axial length of ≥ 26 mm, were eligible for inclusion. In pseudophakic eyes, high-myopia eligibility was determined using preoperative refractive records and axial length measurements rather than postoperative refraction. All pseudophakic eyes included in the study had a preoperative spherical equivalent of ≤ − 6.00 D and an axial length of ≥ 26 mm. The study group consisted of highly myopic patients who had undergone uncomplicated phacoemulsification cataract surgery, defined as phacoemulsification with intraocular lens implantation without intraoperative complications such as posterior capsule rupture, vitreous loss, zonular dialysis, or dropped nuclear fragments. Surgical history and the intraoperative course were verified from operative reports and electronic surgical records. All eyes in the study group underwent standard phacoemulsification with posterior chamber intraocular lens implantation. The control group consisted of phakic highly myopic eyes without cataract or a history of previous cataract surgery. Each participant was assigned to only one study group. When both eyes met the eligibility criteria, both eyes were included; when only one eye was included, the fellow eye was excluded because of inadequate OCTA image quality.

Patients younger than 40 years; those unable to maintain adequate fixation for OCTA imaging due to low visual acuity; those with anterior or posterior segment pathologies affecting image quality (including corneal opacity, vitreous opacity, or vitreous hemorrhage); patients with systemic diseases known to affect macular microcirculation (such as diabetes mellitus, chronic obstructive pulmonary disease, or congestive heart failure); those using systemic medications; patients with ocular diseases other than myopia, including diabetic retinopathy, hypertensive retinopathy, glaucoma (including primary open-angle and normal-tension glaucoma), or uveitis; eyes with myopic choroidal neovascularization, macular hole, retinal detachment, Meta-Analysis for Pathologic Myopia (META-PM) category 4 macular atrophy [10], myopic traction maculopathy, or posterior staphyloma; poor-quality OCTA images with a signal strength index below 6/10; patients who had undergone ocular surgery within the previous six months; and those with complicated cataract surgery were excluded from the study.

Clinical examination

The gender, age, laterality, and lens status of all included patients were recorded. Refractive measurements were performed using an autorefractometer (Canon RK-F1 Auto Ref-Keratometer, Tokyo, Japan). In the study group, refractive and spherical equivalent measurements were obtained postoperatively. These postoperative refractive values were reported descriptively and were not used to determine eligibility for high myopia. The spherical equivalent of the refractive error was calculated using the formula: Spherical equivalent = Spherical value + (Cylindrical value/2). The BCVA was determined using the Snellen chart and converted to Logarithm of the Minimum Angle of Resolution (logMAR) equivalents. Axial length measurements were obtained using optical biometry (AL-Scan, Nidek Co, Ltd., Gamagori, Japan). For eyes in the study group, intraoperative surgical parameters, including cumulative dissipated energy (CDE) and surgery duration, were retrieved from operative reports and electronic surgical records. Surgery duration was defined as the time from the initiation of phacoemulsification to completion of the surgical procedure. These parameters were included in additional analyses to evaluate their potential associations with visual, structural, and OCTA outcomes.

Optical coherence tomography and OCTA imaging

Spectral-domain optical coherence tomography (OCT) imaging (AngioVue, Optovue, Fremont, CA) was performed after pupil dilatation (using tropicamide 0.5% and phenylephrine 2.5% eye drops). All OCT scans were performed by the same retina specialist (M.E.K), during the same time interval of the day (between 10:00 A.M. and 2:00 P.M.). The foveal centre was defined as the location at which the foveal pit is deepest. Central macular thickness (CMT) was measured using the Retina Map mode of OCTA, and subfoveal choroidal thickness (SFCT) was manually measured in the High Definition Line mode. While CMT was automatically provided by the device, SFCT was manually measured using the caliper tool, defined as the perpendicular distance between the outer boundary of the hyperreflective retinal pigment epithelium - Bruch’s membrane complex and the outer boundary of the large choroidal vessels extending to the scleral hyperreflectivity. All SFCT measurements were performed manually by the same retina specialist (M.E.K.). A single measurement was obtained for each eye and used for analysis. To exclude diurnal variations, all examinations were performed at the same time of day.

Macular imaging with OCTA was performed using a scan of 3 × 3 mm centred on the fovea (AngioVue Avanti RTVue-XR, OptoVue, Fremont, CA; software version 2018.1.1.85). In the central macula, the superficial foveal avascular zone (FAZ) area, as well as the vessel density (VD) of the superficial capillary plexus (SCP) and deep capillary plexus (DCP) were evaluated. The VD of SCP and DCP were analysed in subregions, including the whole scan area, fovea, parafoveal region, temporal quadrant, superior quadrant, nasal quadrant, and inferior quadrant. The superficial FAZ area (in mm²) was automatically measured from en face OCTA images using the nonflow analysis mode of the device software. All OCTA images were reviewed by a single retina specialist who was masked to group assignment. Automatic segmentation was visually verified for all scans. Segmentation boundaries were not manually corrected; when segmentation errors were identified, image acquisition was repeated, and scans with persistent segmentation errors or inadequate image quality despite repeated acquisition were excluded from the analysis. Vessel density measurements were automatically generated by the device software. The device included the projection artifact removal algorithm.

The SCP was automatically measured by the device between 3 microns below the internal limiting membrane and 15 microns below the inner plexiform layer. The DCP was automatically measured by the device from 15 microns below the inner plexiform layer to 70 microns beneath. Using the algorithm in the device software, VD (%) in SCP and DCP was measured within areas divided by 1 mm and 3 mm diameter circles centred on the FAZ (foveal and parafoveal regions).

Two patients (3 eyes) in the control group and three patients (4 eyes) in the study group were excluded because of inadequate image quality or persistent segmentation errors.

Statistical analysis

In descriptive statistics, categorical variables were presented as counts and percentages, while continuous variables were expressed as mean ± standard deviation, median (minimum–maximum values), ratio, and frequency. The normality of continuous variables was assessed using the Kolmogorov-Smirnov or Shapiro-Wilk test. The chi-square test was used to compare categorical variables between groups, and Fisher’s exact test was applied when chi-square test assumptions were not met. The Mann-Whitney U test was used for non-normally distributed continuous variables, while the independent samples t-test was used for normally distributed continuous variables. To account for the correlation between eyes from the same patient, statistical analyses were performed using Generalized Estimating Equations (GEE) models.

Participant-level variables, including age and sex, were compared at the participant level, whereas eye-level outcomes were analyzed using GEE models accounting for within-patient correlation. Between-group comparisons of continuous eye-level outcomes were performed using GEE models with a Gaussian distribution and identity link function. Patient identification was specified as the clustering variable to account for within-patient correlation between fellow eyes. An exchangeable working correlation structure and a robust (sandwich) covariance estimator were used. The between-group models were unadjusted, with study group as the sole explanatory variable. GEE group coefficients (β) with 95% confidence intervals (CI) were reported for the principal OCTA outcomes. Associations of age, axial length, CDE, and surgery duration with visual, structural, and OCTA outcomes were evaluated using separate univariable GEE models. Regression coefficients (β) with 95% CI were reported for these analyses. For the 15 primary between-group OCTA comparisons presented in Table 2, Bonferroni correction was applied, resulting in an adjusted significance threshold of p < 0.0033. The univariable association analyses presented in Table 3 were considered exploratory and were not adjusted for multiple comparisons; for these analyses, p < 0.05 was considered statistically significant. Sample size calculation was performed using G*Power software version 3.1 (Heinrich Heine University, Düsseldorf, Germany). Assuming a two-tailed α level of 0.05, 80% statistical power, and a medium effect size (Cohen’s d = 0.50), the initial sample size calculation yielded a minimum of 128 eyes under an independent-observation assumption. Because both eyes of some participants were included, this calculation did not account for within-participant inter-eye correlation and should therefore be regarded as an approximate planning estimate rather than a cluster-adjusted calculation. Within-patient correlation was accounted for in all final eye-level analyses using GEE models with patient-level clustering. A sensitivity assessment based on the observed effect sizes of the primary OCTA outcomes was additionally performed using the actual sample sizes and the Bonferroni-adjusted significance threshold. For statistical analysis, SPSS 27.0 (IBM Corp., Armonk, NY, USA) statistical software package was used. For analyses not otherwise specified, p < 0.05 was considered statistically significant.

Table 2.

OCTA superficial capillary plexus vessel density (SCP-VD), deep capillary plexus vessel density (DCP-VD), and foveal avascular zone (FAZ) area values in both groups

Parameters	Control Group (mean ± SD)	Study Group (mean ± SD)	GEE β (95% CI)	p value
Whole image SCP-VD (%)	40.98 ± 4.98	37.27 ± 5	−3.47 (− 5.31 to − 1.62)	< 0.001*
Foveal SCP-VD (%)	19.52 ± 5.63	18.25 ± 6.25	−1.23 (− 3.40 to 0.93)	0.265
Parafoveal SCP-VD (%)	43.67 ± 5.14	39.98 ± 5.31	−3.66 (− 5.59 to − 1.74)	< 0.001*
Parafoveal temporal SCP-VD (%)	42.92 ± 4.46	39.44 ± 6.49	−3.45 (− 5.60 to − 1.29)	0.001*
Parafoveal superior SCP-VD (%)	45.66 ± 5.65	40.97 ± 6.54	−4.68 (− 6.94 to − 2.42)	< 0.001*
Parafoveal nasal SCP-VD (%)	42.85 ± 6.60	38.57 ± 5.51	−4.28 (− 6.34 to − 2.22)	< 0.001*
Parafoveal inferior SCP-VD (%)	42.96 ± 6.64	39.58 ± 6.52	−3.36 (− 5.70 to − 1.02)	0.003*
Whole image DCP-VD (%)	48.35 ± 4.96	42.60 ± 5.05	−5.70 (− 7.66 to − 3.75)	< 0.001*
Foveal DCP-VD (%)	34.66 ± 6.65	30.03 ± 8.09	−3.98 (− 7.13 to − 0.83)	0.003*
Parafoveal DCP-VD (%)	50.77 ± 5.25	45.38 ± 5.50	−5.41 (− 7.54 to − 3.28)	< 0.001*
Parafoveal temporal DCP-VD (%)	52.15 ± 5.44	45.30 ± 5.70	−6.98 (− 9.35 to − 4.62)	< 0.001*
Parafoveal superior DCP-VD (%)	50.17 ± 5.85	45.13 ± 5.54	−5.13 (− 7.22 to − 3.04)	< 0.001*
Parafoveal nasal DCP-VD (%)	50 ± 5.91	45.16 ± 6.19	−4.79 (− 7.09 to − 2.48)	< 0.001*
Parafoveal inferior DCP-VD (%)	50.34 ± 5.98	44.23 ± 6.42	−6.09 (− 8.30 to − 3.87)	< 0.001*
FAZ (mm2)	0.25 ± 0.098	0.26 ± 0.093	+ 0.005 (− 0.035 to 0.045)	0.808

β: Group coefficient, CI: Confidence interval, DCP: Deep capillary plexus, FAZ: Foveal avascular zone, SCP: Superficial capillary plexus, VD: Vessel density. β (95% CI) represents the GEE-estimated between-group difference. Mean ± SD values are descriptive eye-level statistics

All p-values were derived from GEE analyses. Bonferroni correction was applied to the 15 primary OCTA comparisons

(adjusted significance threshold: p < 0.0033)

* Statistically significant after Bonferroni correction

Table 3.

Univariable associations of age, axial length, cumulative dissipated energy (CDE), and surgery duration with best corrected visual acuity (BCVA), central macular thickness (CMT), subfoveal choroidal thickness (SFCT), and OCTA parameters

Parameters	Age		Axial Length		CDE		Surgery Duration
	β (95% CI)	p value	β (95% CI)	p value	β (95% CI)	p value	β (95% CI)	p value
BCVA (logMAR)	0.012 (0.002 to 0.022)	0.019*	0.076 (0.027 to 0.124)	0.002*	0.004 (− 0.005 to 0.013)	0.361	0.003 (− 0.003 to 0.009)	0.287
Central macular thickness	−0.992 (− 1.676 to − 0.308)	0.004*	−3.458 (− 5.790 to − 1.126)	0.004*	−0.417 (− 1.436 to 0.602)	0.422	−0.308 (− 1.227 to 0.611)	0.511
Subfoveal choroidal thickness	−0.950 (− 1.820 to − 0.080)	0.034*	−7.183 (− 12.570 to − 1.796)	0.009*	−0.703 (− 2.016 to 0.610)	0.294	−0.624 (− 1.682 to 0.434)	0.247
Whole image SCP-VD	−0.101 (− 0.185 to − 0.018)	0.017*	−0.611 (− 1.101 to − 0.121)	0.014*	−0.121 (− 0.303 to 0.061)	0.193	−0.083 (− 0.246 to 0.080)	0.319
Foveal SCP-VD	−0.116 (− 0.197 to − 0.035)	0.005*	−0.261 (− 0.733 to 0.212)	0.280	−0.079 (− 0.296 to 0.138)	0.476	−0.047 (− 0.245 to 0.151)	0.642
Parafoveal SCP-VD	−0.097 (− 0.182 to − 0.012)	0.025*	−0.693 (− 1.220 to − 0.167)	0.010*	−0.112 (− 0.291 to 0.067)	0.219	−0.091 (− 0.255 to 0.073)	0.276
Parafoveal temporal SCP-VD	−0.055 (− 0.147 to 0.037)	0.238	−0.667 (− 1.198 to − 0.136)	0.014*	−0.132 (− 0.311 to 0.047)	0.148	−0.102 (− 0.273 to 0.069)	0.243
Parafoveal superior SCP-VD	−0.123 (− 0.235 to − 0.010)	0.033*	−0.914 (− 1.524 to − 0.304)	0.003*	−0.119 (− 0.294 to 0.056)	0.183	−0.088 (− 0.251 to 0.075)	0.291
Parafoveal nasal SCP-VD	−0.140 (− 0.221 to − 0.059)	< 0.001*	−0.645 (− 1.242 to − 0.048)	0.034*	−0.096 (− 0.270 to 0.078)	0.279	−0.072 (− 0.239 to 0.095)	0.397
Parafoveal inferior SCP-VD	−0.086 (− 0.187 to 0.014)	0.092	−0.718 (− 1.298 to − 0.139)	0.015*	−0.108 (− 0.279 to 0.063)	0.216	−0.081 (− 0.249 to 0.087)	0.344
Whole image DCP-VD	−0.147 (− 0.262 to − 0.031)	0.013*	−0.823 (− 1.283 to − 0.363)	< 0.001*	−0.151 (− 0.345 to 0.043)	0.127	−0.118 (− 0.297 to 0.061)	0.196
Foveal DCP-VD	−0.166 (− 0.330 to − 0.002)	0.048*	−0.126 (− 0.641 to 0.388)	0.630	0.058 (− 0.153 to 0.269)	0.589	0.041 (− 0.158 to 0.240)	0.687
Parafoveal DCP-VD	−0.132 (− 0.258 to − 0.006)	0.040*	−0.871 (− 1.363 to − 0.378)	< 0.001*	−0.143 (− 0.345 to 0.059)	0.165	−0.129 (− 0.315 to 0.057)	0.174
Parafoveal temporal DCP-VD	−0.145 (− 0.287 to − 0.003)	0.046*	−0.849 (− 1.303 to − 0.394)	< 0.001*	−0.169 (− 0.376 to 0.038)	0.109	−0.141 (− 0.337 to 0.055)	0.158
Parafoveal superior DCP-VD	−0.150 (− 0.272 to − 0.027)	0.017*	−0.930 (− 1.453 to − 0.407)	< 0.001*	−0.139 (− 0.342 to 0.064)	0.179	−0.113 (− 0.301 to 0.075)	0.239
Parafoveal nasal DCP-VD	−0.130 (− 0.256 to − 0.005)	0.042*	−0.750 (− 1.297 to − 0.203)	0.007*	−0.154 (− 0.359 to 0.051)	0.141	−0.121 (− 0.307 to 0.065)	0.202
Parafoveal inferior DCP-VD	−0.125 (− 0.247 to − 0.002)	0.046*	−1.057 (− 1.590 to − 0.524)	< 0.001*	−0.181 (− 0.393 to 0.031)	0.094	−0.148 (− 0.348 to 0.052)	0.147
Foveal avascular zone	0.001 (− 0.001 to 0.003)	0.474	−0.006 (− 0.012 to 0.001)	0.104	0.001 (− 0.002 to 0.004)	0.536	0.001 (− 0.002 to 0.004)	0.419

β: Regression coefficient, BCVA: Best corrected visual acuity, CDE: Cumulative dissipated energy, CI: Confidence interval, DCP: Deep capillary plexus, SCP: Superficial capillary plexus, VD: Vessel density. β (95% CI) represents the estimated change in the outcome per one-unit increase in the corresponding predictor. Age, axial length, CDE and surgery duration associations were evaluated using univariable GEE models accounting for within-patient correlation. These exploratory analyses were not adjusted for multiple comparisons; p < 0.05 was considered statistically significant

*Statistically significant

Results

The control group consisted of 14 (33.3%) male and 28 (66.6%) female patients, while the study group included 12 (32.4%) male and 25 (67.6%) female patients. A total of 71 eyes from 42 patients in the control group and 61 eyes from 37 patients in the study group were included in the analysis. In the control group, 29 participants contributed both eyes and 13 contributed one eye, whereas in the study group, 24 participants contributed both eyes and 13 contributed one eye. The mean age was 64.6 ± 9.9 years in the control group and 65.8 ± 10.4 years in the study group. The mean surgery-to-imaging interval in the study group was 18.2 ± 4.3 months (range: 12–30 months). The mean signal strength index of the included OCTA scans was 8.1 ± 0.4. Demographic, clinical, refractive, and ocular characteristics of the two groups, including preoperative spherical equivalent, axial length, CMT, SFCT, and META-PM categories, are presented in Table 1. Postoperative spherical equivalent was also reported descriptively for the study group.

Table 1.

Distribution of age, gender, best corrected visual acuity (BCVA), spherical equivalent, axial length, central macular thickness (CMT), subfoveal choroidal thickness (SFCT), and myopic maculopathy categories in both groups

Parameters	Control Group	Study Group	p value
Age (years), mean ± SD	64.6 ± 9.9	65.8 ± 10.4	0.602
Gender, Female/Male, (n)	28/14	25/12	0.932
BCVA (logMAR), mean ± SD	0.33 ± 0.67	0.38 ± 0.49	0.08
Preoperative Spherical Equivalent (D), mean ± SD	-8.52 ± 4.82	−9.21 ± 2.17	0.419
Postoperative Spherical Equivalent (D), mean ± SD	-	-1.91 ± 3.42	-
AL (mm), mean ± SD	28.97 ± 1.7	29.27 ± 1.9	0.126
CMT (µm), mean ± SD	164.12 ± 33.68	162.36 ± 52.85	0.224
SFCT (µm), mean ± SD	147.57 ± 53.07	137.40 ± 36.93	0.108
META-PM category, n (%)			0.694
Category 0	28 (39.4%)	25 (41.0%)
Category 1	24 (33.8%)	17 (27.9%)
Category 2	13 (18.3%)	10 (16.4%)
Category 3	6 (8.5%)	9 (14.8%)

AL: Axial length, BCVA: Best corrected visual acuity, CMT: Central macular thickness, META-PM: Meta-analysis for Pathologic Myopia, SFCT: Subfoveal choroidal thickness

Age and sex were analyzed at the participant level using the independent-samples t-test and chi-square test, respectively. Eye-level continuous variables were compared using GEE models accounting for within-patient correlation. Preoperative spherical equivalent values were used for baseline comparison between the groups, whereas postoperative spherical equivalent values were available only for the study group and are presented descriptively; therefore, no between-group statistical comparison was performed for postoperative spherical equivalent. META-PM categories were compared between groups using an ordinal GEE model accounting for within-patient correlation

The mean FAZ area and VD values in different macular regions for both groups are presented in Table 2. In unadjusted GEE analyses accounting for within-patient correlation, the study group demonstrated significantly lower whole-image and parafoveal SCP-VD values, including the temporal, superior, nasal, and inferior parafoveal regions, compared with the control group. DCP-VD values were also significantly lower in the study group across all evaluated regions. These differences remained statistically significant after Bonferroni correction for the 15 primary OCTA comparisons (adjusted significance threshold, p < 0.0033). No statistically significant between-group difference was detected in foveal SCP-VD (β = −1.23; 95% CI, − 3.40 to 0.93; p = 0.265) or FAZ area (β = 0.005; 95% CI, − 0.035 to 0.045; p = 0.808). The corresponding GEE group coefficients (β) and 95% confidence intervals are presented in Table 2.

In univariable GEE analyses, age and axial length showed significant positive associations with BCVA (logMAR), whereas significant negative associations were observed between these variables and CMT, SFCT, and multiple OCTA parameters. No statistically significant associations were observed between the surgical parameters, including CDE and surgery duration, and BCVA, CMT, SFCT, or any OCTA parameter. The detailed analysis is presented in Table 3.

Discussion

The principal finding of the present study was that highly myopic eyes with a history of uncomplicated cataract surgery had lower whole-image and parafoveal SCP-VD and lower DCP-VD across all evaluated regions compared with non-operated highly myopic eyes. The mean surgery-to-imaging interval in the operated group was 18.2 months. Unlike previous studies that primarily evaluated serial short-term postoperative OCTA measurements, the present study provides a cross-sectional comparison between operated and non-operated eyes within a highly myopic population. The inclusion of a highly myopic comparison group reduces, although does not eliminate, the potential confounding influence of high myopia itself when interpreting the observed between-group vascular differences.

A previous study involving 38 eyes compared OCTA findings before and at 1 day, 1 week, 1 month, and 3–6 months after phacoemulsification between a high myopia group (axial length > 26.5 mm) and a control group (22 mm < axial length ≤ 24.5 mm). The mean changes in macular VD were significantly higher in high myopic patients on the first postoperative day compared to the control group [11]. Another study, including 55 eyes, evaluated OCTA findings in high myopia and low myopia groups before and at 1 day, 1 week, 1 month, and 3 months after surgery. Although macular VD values increased in the high myopia group at 1 week and 1 month postoperatively, they were significantly lower at 3 months compared to preoperative levels [12]. In the present study, at a mean surgery-to-imaging interval of 18.2 months, whole-image and parafoveal SCP-VD values, as well as DCP-VD values across all evaluated regions, were lower in the operated highly myopic group than in the non-operated highly myopic control group. Although previous studies have reported early postoperative increases in VD, the present cross-sectional study identified lower VD values in operated highly myopic eyes compared with non-operated highly myopic controls at a later postoperative assessment. Additionally, the use of a non-operated highly myopic comparison group provides a clinically relevant reference for interpreting the observed between-group differences. These findings provide additional insight into the retinal microvascular differences observed between operated and non-operated highly myopic eyes at a later postoperative assessment. In highly myopic eyes, ocular magnification may influence OCTA-derived vessel density measurements, potentially leading to lower values independent of true microvascular loss. Although the inclusion of a highly myopic control group partially mitigates this effect, the lack of axial length-based magnification correction should be considered when interpreting these findings. In addition, OCTA measurements in highly myopic eyes may demonstrate lower reproducibility due to factors such as segmentation errors and fixation instability. Previous studies have reported reduced repeatability of vessel density parameters in eyes with long axial length. To minimise image quality-related bias, only OCTA scans with acceptable signal strength indices were included in the analysis. Moreover, the use of a fixed 3 × 3 mm scan area in eyes with long axial length may correspond to a larger true retinal area, potentially introducing measurement heterogeneity across individuals.

In a study by Liu et al. involving 58 eyes with normal axial length, perifoveal DCP-VD significantly increased at 1 month postoperatively compared to preoperative OCTA findings, but this effect disappeared at 3 months [13]. Similarly, Pilotto et al. reported an increase in intermediate capillary plexus and DCP perfusion 1 day after uncomplicated cataract surgery, while SCP perfusion remained unchanged. By 90 days, all parameters had returned to baseline levels, which the authors attributed to the inflammatory nature of the surgery [14]. The present study did not assess early postoperative changes; however, lower VD was observed in the operated highly myopic group compared with non-operated controls at a later postoperative assessment. One possible explanation for the observed between-group difference in retinal vessel density is that highly myopic eyes may have reduced microvascular reserve as a consequence of axial elongation, retinal and choroidal stretching, tissue thinning, and chronic structural remodeling. This raises the possibility that the vascular response to cataract surgery in highly myopic eyes may not be uniform over time. Although transient increases in macular vessel density have been reported during the early postoperative period, the lower vessel density observed at a later assessment may reflect the interaction between the acute postoperative response and the pre-existing microvascular vulnerability of highly myopic eyes. Thus, rather than proposing a single mechanism, our findings raise the hypothesis that pre-existing myopic microvascular remodeling may modify the longer-term vascular response to cataract surgery. Alternatively, the observed differences may partly reflect the underlying chronic microvascular characteristics of highly myopic eyes rather than a direct long-term effect of cataract surgery. These proposed mechanisms remain hypothetical in the present study, because inflammatory, hemodynamic, and neurovascular parameters were not directly measured. However, standardized GCL and RNFL measurements were not available in the present dataset, precluding direct assessment of the relationship between retinal structure and vascular density. Accordingly, the observed between-group differences should not be attributed to a specific mechanism on the basis of the present data.

Given that cataract density and phacoemulsification energy may influence postoperative inflammation and retinal microvascular status, the relationships between intraoperative surgical parameters and OCTA outcomes were additionally evaluated. Neither CDE nor surgery duration showed significant associations with SCP-VD, DCP-VD, FAZ area, or other structural retinal parameters. These findings suggest that differences in intraoperative energy exposure or operative duration are unlikely to fully explain the observed between-group differences in VD. Nevertheless, the potential contribution of cataract severity and other unmeasured perioperative factors cannot be completely excluded.

A study comparing diabetic and non-diabetic patients undergoing cataract surgery found that SCP-VD was significantly higher in the diabetic group at 3 months postoperatively [15]. Although the findings of that study differ from those of the present study, direct mechanistic comparisons between diabetic and highly myopic eyes should be interpreted cautiously because the underlying vascular and inflammatory characteristics of these conditions differ and were not specifically evaluated in the present study.

Previous studies comparing FAZ areas between high myopia and normal axial length groups found no significant differences [16]. Similarly, another study found no significant difference in FAZ area in diabetic and non-diabetic patients within the first 3 months after uncomplicated cataract surgery [15]. Additionally, OCTA imaging before and 2 weeks after cardiopulmonary bypass surgery revealed no significant change in FAZ area [17]. Consistent with the literature, this study also found no significant difference in FAZ area between the two groups. The lack of a statistically significant between-group difference in FAZ area in the present study is consistent with previous reports showing no significant differences in FAZ area across different clinical settings.

A study investigating long-term (12-month) changes in SFCT after cataract surgery found no significant difference compared to preoperative values [18]. The same study suggested that uncomplicated phacoemulsification might cause subclinical changes in CMT, likely due to surgical inflammation, but that CMT eventually returned to baseline. Similarly, no statistically significant between-group differences in CMT or SFCT were detected between highly myopic eyes with and without previous cataract surgery.

Fan et al. evaluated participants under 40 years old with different refractive statuses and found a negative correlation between axial length and macular VD using OCTA [19]. Similarly, the present study demonstrated negative associations between axial length and several OCTA vessel density parameters, supporting the potential contribution of axial elongation to reduced retinal microvascular density in highly myopic eyes. Importantly, axial length did not differ significantly between the operated and non-operated groups, suggesting that the observed between-group differences in VD were unlikely to be explained solely by differences in baseline axial length. However, because serial axial length measurements were not available, the present study could not determine whether longitudinal myopic progression contributed to the observed vascular findings. The associations between OCTA parameters and clinical variables, although evaluated within the GEE framework to account for inter-eye correlation, should therefore be interpreted as exploratory rather than as evidence of a causal relationship.

The principal limitation of this study is its cross-sectional design and the absence of preoperative OCTA measurements. Therefore, the observed differences between operated and non-operated eyes cannot be interpreted as postoperative changes, persistence of a previous change, or a causal effect of cataract surgery. Another limitation of this study is that no correction formula for ocular magnification was applied to OCTA measurements. Although the inclusion of a highly myopic control group partially mitigates this issue, the lack of axial length-based magnification correction should be considered when interpreting these findings. In addition, differences in lens status between phakic and pseudophakic eyes may have introduced measurement bias in OCTA-derived parameters and cannot be completely excluded. Because participants were not randomly assigned to the operated and non-operated groups, residual confounding related to unmeasured clinical or perioperative characteristics cannot be excluded. Moreover, the exclusion of eyes with advanced pathologic myopia-related complications may limit the generalizability of the findings to the broader highly myopic population. Furthermore, standardized GCL and RNFL thickness measurements were not available, limiting our ability to evaluate the relationship between inner retinal structural changes and retinal microvascular density. Serial axial length measurements were also unavailable, preventing assessment of whether longitudinal myopic progression contributed to the observed differences in retinal vessel density. The single-center design and use of a single OCTA platform may have improved measurement consistency; however, these factors do not eliminate the potential for residual confounding or other design-related biases.

Conclusion

The present study demonstrated lower macular vessel density in highly myopic eyes with a history of uncomplicated cataract surgery compared with non-operated highly myopic controls at a later postoperative assessment, while no statistically significant between group difference in FAZ area was detected. Given the cross-sectional design and absence of preoperative OCTA measurements, these findings should be interpreted as between-group differences rather than postoperative changes or causal effects of cataract surgery. Further longitudinal studies with preoperative and serial postoperative OCTA assessments are warranted to clarify the temporal course and clinical significance of these findings.

Abbreviations

BCVA

Best corrected visual acuity

CDE

Cumulative dissipated energy

CMT

Central macular thickness

DCP

Deep capillary plexus

FAZ

Foveal avascular zone

GEE

Generalized Estimating Equations

logMAR

Logarithm of the Minimum Angle of Resolution

OCT

Optical coherence tomography

OCTA

Optical coherence tomography angiography

SCP

Superficial capillary plexus

SFCT

Subfoveal choroidal thickness

VD

Vessel density

Author contributions

Material preparation, data collection and analysis were performed by MEK, DG, MD, STD. The first draft of the manuscript was written by MEK and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. All patients had provided informed written consent. Approval was granted by the Clinical Research Ethics Committee of Şişli Hamidiye Etfal Training and Research Hospital. (2020/No 2640).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.