Spectral-Domain Optical Coherence Tomographic Angiography in Children With Amblyopia

Key Points

Question  Are there microvascular abnormalities on optical coherence tomographic angiography in children (<18 years) with amblyopia?

Findings  In this case-control study of 63 eyes of 59 patients, reduced capillary density in the superficial and deep plexus in the central 6 × 6 mm of the macula was noted in amblyopic eyes but not in control eyes.

Meaning  This finding suggests amblyopic eyes have reduced retinal capillary density in the macula, but the clinical relevance of this finding remains to be determined.

Importance  Amblyopia is the most common cause of visual impairment in childhood, with a prevalence of 1% to 4% in children in the United States. To date, no studies using noninvasive optical coherence tomographic angiography (OCTA) have measured blood flow in the retinal capillary layers in children with amblyopia.

Objective  To evaluate the retinal and microvascular features using OCTA in children (<18 years) with amblyopia.

Design, Setting, and Participants  This observational case-control study enrolled patients from September 1, 2016, through May 31, 2017, and was conducted from September 1, 2016, through June 30, 2017, at the Stein Eye Institute at UCLA (University of California, Los Angeles). Participants included 59 children (<18 years) with amblyopia and without amblyopia examined at a pediatric ophthalmology clinic or referred to the clinic by coinvestigators. All patients underwent comprehensive ophthalmological examination, including visual acuity, refraction, and ocular motility tests; anterior and posterior segment examination; and OCTA.

Main Outcomes and Measures  Reduced superficial and deep retinal capillary vessel density on OCTA.

Results  Of the 63 eyes evaluated, 13 (21%) were amblyopic and 50 (79%) were control eyes. Of the 59 patients, the mean (SD) age of patients with amblyopia was 8.0 (4.0) years and 10.3 (3.3) years for the controls; 33 patients (56%) were female; and 5 of 13 (39%) and 27 of 46 (54%) patients in the amblyopic and control groups, respectively, were identified as white. The macular vessel density of the superficial capillary plexus was lower in the amblyopic group than in the control group in both 3 × 3-mm and 6 × 6-mm scans. After adjusting for age and refractive error, the mean (SD) difference in the superficial capillary plexus in the 6 × 6-mm scan was statistically significant (49.3% [4.1] vs 51.2% [2.9]; P = .02). Macular vessel density of the deep capillary plexus in the 6 × 6-mm scans was also considerably different between groups: mean (SD) vessel density of the deep retinal capillary plexus was 54.4% (4.7%) in the amblyopia group and 60.1% (3.3%) in the control group, with a difference of 5.7% (95% CI, 3.4%-8.1%; P = .002).

Conclusions and Relevance  The study found that OCTA reveals subnormal superficial and deep retinal capillary density in the macula of patients with amblyopia. Further studies are needed to determine the clinical relevance of this finding.

Amblyopia is the most common cause of visual impairment in childhood, with a prevalence of 1% to 4% in children aged 7 to 17 years in the United States.^1-5 Without timely recognition and treatment, amblyopia may cause irreversible vision loss in adulthood.¹ There have been inconsistent spectral-domain optical coherence tomography (OCT) reports of retinal and macular abnormalities in the amblyopic eye, which include increased thickness of the outer nuclear layer,⁶ although these pathoanatomical findings have been disputed.⁴

The vascular system of the retina is vital for visual function, and its assessment is valuable in the evaluation of macular diseases.⁷ Optical coherence tomographic angiography (OCTA) is a noninvasive modality that provides depth-resolved visualization of the microvasculature of the retina, allowing the identification of the superficial retinal capillary plexus (SCP), the deep retinal capillary plexus (DCP), and the choroid.⁸ To our knowledge, macular OCTA findings in children with amblyopia have not been previously reported. The aim of this study was to assess the density of the retinal capillary and radial peripapillary plexuses as well as the area of the foveal avascular zone (FAZ) in amblyopic eyes. We hypothesized that these vascular layers may have abnormalities not recognized by clinical examination or spectral-domain OCT.^4,6

A total of 59 patients were recruited and 63 eyes (13 [21%] of which were amblyopic eyes and 50 [79%] were control eyes) were evaluated. For participants with amblyopic eyes, the inclusion criteria were age younger than 18 years and an evaluation during a routine visit to the pediatric eye clinic or a referral to the clinic by the coinvestigators (3 of us, F.G.V., J.L.D., and S.L.P.). We enrolled patients from September 1, 2016, through May 31, 2017, and conducted the study from September 1, 2016, through June 30, 2017, at the UCLA Stein Eye Institute. This study was approved by the UCLA Institutional Review Board and adhered to the Declaration of Helsinki⁹ and the Health Insurance Portability and Accountability Act. Written informed consent was obtained before imaging from the parents of the participants and directly from those older than 7 years.

Amblyopia was defined as best-corrected visual acuity (BCVA) between 20/40 and 20/200 in one or both eyes without organic cause for the decreased vision. Only amblyopia due to strabismus or anisometropia was included. In patients with bilateral amblyopia, the eye with poorer vision was selected.

For the normal control group, we selected the fellow eye of patients with unilateral amblyopia if the BCVA was 20/20 or better. Normal patients presenting to the eye clinic with normal visual acuity and no evidence of any ocular abnormality that could cause amblyopia were also recruited. All control patients underwent a complete eye examination and were required to have 20/20 BCVA. In controls, we selected the right eye only for analysis to eliminate similarities of the measurements in the same person as a confounding factor. Furthermore, we performed 2 additional comparisons: amblyopic group vs fellow eyes in patients with amblyopia only and amblyopia group vs normal controls without amblyopia in the fellow eye. Snellen BCVA was converted to logMAR for analysis.

The exclusion criteria were the presence of deprivation amblyopia, nystagmus, retinal detachment, retinal disease, intraocular inflammation, and media opacity, such as corneal disease or cataract. Patients with a history of prematurity, neurologic disease, or systemic conditions that could alter the microvasculature (including diabetes, hypertension, cardiovascular disease, and renal disease) were excluded. In addition, patients with head, neck, or other injuries preventing proper positioning or with an inability to maintain retinal fixation on a specified target (owing to poor vision or cooperation) were excluded.

All patients underwent comprehensive ophthalmological examination, including visual acuity, refraction, and ocular motility tests; anterior and posterior segment examination; and OCTA, which was performed using a spectral-domain device with software (RTVue XR Avanti with AngioVue software; Optovue, Inc). Macular 3 × 3-mm and 6 × 6-mm scans as well as optic nerve 4.5 × 4.5-mm scans were performed in every patient (Figure 1). Each scan was automatically segmented by the software to visualize the SCP and DCP of the retina (Figure 2).⁸ The SCP OCTA image was segmented, with an inner boundary 3 μm below the internal limiting membrane and an outer boundary 15 μm below the inner plexiform layer.⁸ The DCP image was segmented with an inner boundary 15 μm below the inner plexiform layer and an outer boundary 71 μm below the inner plexiform layer.⁸ Analyzed features included vessel density (as a percentage) of the SCP and DCP of both macular scans, FAZ area (in millimeters squared) measured in the 3 × 3-mm scan, papillary and peripapillary vessel density, and foveal and parafoveal thickness in the 3 × 3-mm scan. All features were analyzed using the AngioVue automated software (Figure 3). Central macular thickness was calculated by an automated algorithm available in the machine that measures foveal and parafoveal thickness from the inner limiting membrane to the retinal pigment epithelium.

A 2-tailed, unpaired t test was used to compare mean values. However, in cases of skewed data, we performed a nonparametric Kruskal-Wallis test to compare values. In addition, logistic regression was used to evaluate the effect of age and refractive error. Data were analyzed with JMP Pro software, version 13 (SAS Institute Inc). A 2-sided P = .05 was considered statistically significant. To guarantee the quality of the images, we excluded scans that had a signal strength lower than 40 or, if greater than 40, had excessive motion artifacts not correctable by the AngioVue software.¹⁰

A total of 63 eyes of 59 patients were included in the study, 13 (21%) of which were amblyopic eyes and 50 (79%) were control eyes. Of the 50 control eyes, 46 (92%) were eyes of normal patients and 4 (8%) were normal fellow eyes of patients with amblyopia. The sample included 33 female children (56%), and the mean (SD) age was 8.0 (4.0) years (range, 4-17 years) for patients with amblyopia and 10.3 (3.3) years (range, 4-17 years) for the controls; 5 of 13 (39%) and 27 of 46 (54%) patients in the amblyopia and control groups, respectively, identified as white. No difference in age, sex, or race/ethnicity was observed between groups (Table 1). As expected, visual acuity was significantly different between both groups. The mean (SD) BCVA was 20/64 (logMAR, 0.51 [0.17]) with a range of 20/40 to 20/126 (logMAR, 0.30-0.80) in the amblyopia group and was 20/18 (logMAR, −0.03 [0.05]) with a range of 20/15 to 20/20 (logMAR, −0.12 to 0.00) in the control group, and a difference of 0.54 (95% CI, 0.45 to 0.63; P < .001). No significant refractive difference in the sphere, cylinder, or spherical equivalent was found between the 2 groups (Table 1), although amblyopic eyes demonstrated more than 3 diopters (D) hyperopia.

In the 3 × 3-mm scans, the mean (SD) macular vessel density of the SCP was 51.5% (3.4%) in the amblyopia group and 53.5% (2.5%) in the control group, with a difference of −1.95 (95% CI, −3.94 to 0.04; P = .07). In the 6 × 6-mm scans, the mean (SD) density of the SCP was 49.3% (4.1%) in the amblyopic group and 51.2% (2.9%) in the control group, with a difference of −1.84 (95% CI, −4.29 to 0.62; P = .20; Table 2). After adjusting for age and spherical equivalent, the difference in the SCP in the 6 × 6-mm scan was statistically significant (−1.84 [95% CI, −4.29 to 0.62]; P = .02). The mean (SD) macular vessel density of the DCP in the 3 × 3-mm scan was 55.9% (5.5%) in the amblyopic group and 59.6% (2.14%) in the control group, with a difference of −3.68 (95% CI, −6.75 to −0.60; P = .04; Table 2. After accounting for age and spherical equivalent, this association was no longer statistically significant (3.68 [95% CI, −6.75 to −0.60]; P = .20). The mean (SD) macular vessel density of the DCP in the 6 × 6-mm scans was 54.4% (4.7%) in the amblyopia group and 60.1% (3.3%) in the control group, with a difference of −5.67 (95% CI, −8.53 to −2.80; P = .02; Table 2 and Figure 2]. The skewness of the macular vessel density of the DCP in the 6 × 6-mm scan was found to be moderately skewed with a skewness value of 0.5 for patients with amblyopia and −2 for the control patients. A Kruskal-Wallis test was performed for this comparison given that the data were moderately skewed, and the mean difference was found to be statistically significant (54.4% [4.7%] vs 60.1% [3.3%]; P < .001). After accounting for age and spherical equivalent, this association remained statistically significant (54.4% [4.7%] vs 60.1% [3.3%]; P < .001) (Figure 3).

The mean (SD) foveal thickness was 245 (19.9) μm in the amblyopic group and 247 (16.5) μm in the control group, with a difference of −2.34 (95% CI, −14.46 to 9.78; P = .70). The mean (SD) parafoveal thickness in the amblyopic group was 318 (17.9) μm and 315 (18) μm in the control group, with a difference of −2.74 (95% CI, −8.93 to 3.44; P = .60).

The area of the FAZ of the SCP was 0.28 mm² in the amblyopic group and 0.27 mm² in the control group, with a difference of 0.02 (95% CI, −0.05 to 0.08; P = .60). For the DCP, the area of the FAZ was 0.37 mm² in the amblyopic group and 0.34 mm² in the control group, with a difference of 0.03 (95% CI, −0.05 to 0.10; P = .50).

The vessel density in the 4.5 × 4.5-mm whole optic nerve scan was 55.6% for the amblyopic group and 57.6% for the control groups, with a difference of 2% (95% CI, −0.05% to 4%; P = .10). Inside the optic disc, the mean (SD) vessel density was 57.3% (4.3%) in the amblyopic group and 58.4% (3.2%) in the control group, with a difference of 1.2% (95% CI, −1.4% to 3.8%; P = .40); in the peripapillary area, the density was 58.9% and 60.2% in the amblyopic and control groups, respectively, with a difference of 1.2% (95% CI, −0.7% to 3.2%; P = .20).

The mean (SD) BCVA for amblyopic eyes was 20/64 (logMAR, 0.51 [0.17]; range, 20/40 to 20/126 [logMAR, 0.30 to 0.80]) and 20/27 for the fellow eyes (logMAR, −0.12 [0.05]; range, 20/15 to 20/60 [logMAR, −0.12 to 0.48]), with a mean difference of logMAR, −0.36 (95% CI, −0.5 to −0.2; P < .001). In the 3 × 3-mm scan, the mean vessel density of the SCP was 51.5% for amblyopic eyes and 54.1% for fellow eyes, with a difference of 2.5% (95% CI, 0.2% to 5.2%; P = .05). The mean (SD) DCP in the 3 × 3-mm scan was 56.0% (5.6%) for amblyopic eyes and 59.6% (2.3%) for fellow eyes, with a difference of 3.6% (95% CI, 0.5% to 7.6%; P = .05). All the other analyzed features did not differ considerably between the 2 groups.

Finally, we compared amblyopic eyes with normal control eyes that did not have amblyopia in the fellow eye. We found an important difference between groups in the mean cylinder: 1.44 D in amblyopic eyes and 0.46 D in control eyes, with a difference of 0.98 D (95% CI, 0.3 to 1.7; P = .04). The mean (SD) BCVA was 20/64 (logMAR, 0.51 [0.17]) in the amblyopia group (range, 20/40 to 20/126; logMAR, 0.30 to 0.80) and 20/18 (logMAR, −0.03 [0.05]) in the control group (range, 20/15 to 20/20; logMAR, −0.12 to 0.00), with a difference of 0.53 (95% CI, 0.47% to 0.58%; P < .001). The mean (SD) vessel density of the DCP in the 3 × 3-mm scan was 55.9% (5.5%) in the amblyopia group and 59.6% (2.14%) in the control group, with a difference of 3.6% (95% CI, 1.5% to 5.7%; P = .003). In the 6 × 6-mm scan, the mean (SD) vessel density of the DCP was 54.5% (4.8%) in the amblyopia group and 60.3% (2.9%) in the control group, with a difference of 5.7% (95% CI, 3.4% to 8.1%; P = .002).

In this study, we found that OCTA illustrates a significantly lower vessel density of the DCP, with a mean of a 5.7% reduction in the 6 × 6-mm scan in children with amblyopia. This reduction was a mean of 1.8% in the SCP. We did not find any statistically significant difference in the FAZ or foveal thickness between groups. The macular findings by OCT in patients with amblyopia, including foveal thickness and differences in the retinal layers such as the retinal nerve fiber layer, have been widely studied.^4,11-15 Nevertheless, these findings have been inconsistent and controversial. Li and Yu⁴ performed a meta-analysis to clarify the retinal changes in unilateral amblyopia and found a thicker foveola in the amblyopic eyes than in visually normal control eyes; however, inner macular thickness was not considerably different from outer macular thickness. Our results are not consistent with the findings of Li and Yu because we found no differences between amblyopic and control eyes.

Pineles and Demer,¹⁶ using magnetic resonance imaging, concluded that unilateral amblyopia is associated with subclinical bilateral hypoplastic optic nerves. Furthermore, using photography, Lempert¹⁷ found the optic nerve to be smaller in amblyopic eyes, suggesting that subclinical optic nerve hypoplasia may explain vision loss in amblyopia. On the other hand, the Pediatric Eye Disease Investigator Group¹² used OCT and reported no difference in global or quadrantic retinal nerve fiber layer thickness compared with the fellow nonamblyopic eyes. These results indicate that optic atrophy is not the cause of moderate anisometropic or strabismic amblyopia.¹² We found no abnormality in papillary and peripapillary vessel density, which is consistent with the findings of the Pediatric Eye Disease Investigator Group.

In our population, OCTA demonstrated a statistically significant lower vessel density in both the SCP and the DCP in patients with amblyopia than in normal controls. The retinal capillary plexus comprises 3 main laminar or planar vascular layers in the parafoveal region: SCP, DCP, and the intermediate retinal capillary plexus.¹⁸ The DCP may be at a greater risk of hypoxic retinal injury because of its more distal location from the retinal arterial and choroidal circulations and the higher metabolic demand in the middle retinal layers.^19-21 We speculate that the lower vessel density of the SCP and DCP in amblyopic eyes may be associated with abnormal development due to the lack of a normal visual experience, although animal models present no evidence of this. The arrested development due to abnormal visual stimulus could correlate with a thicker retinal nerve fiber layer that many other investigators have found, which could be associated with an arrest in maturation. These findings were not seen on the 3 × 3-mm scans, which could indicate that amblyopic changes are not centrally located in the fovea; this indication is further corroborated by the lack of difference in the FAZ area between the 2 groups. We also did not find any differences in the foveal or parafoveal retinal thickness.

This study has several limitations, including a small sample size and lack of additional tests at the time of OCTA imaging. The importance of the small sample size cannot be understated, and it is unclear whether our findings would be generalizable to a larger group of patients. Given the small sample size, it is unclear whether the skewed nature of our data (positive skew for those with amblyopia and negative skew for controls) is an artifact of the small sample size or is truly representative. Moreover, because split-spectrum, amplitude-decorrelation angiographic measurements are derived from changes in reflections and backscattering of light, the larger vessels of the SCP often can result in projection artifacts at the level of the DCP⁸; artifacts can result from poor fixation leading to excessive movement that cannot be corrected by the AngioVue software. Although we excluded patients with obvious nystagmus or inability to fixate for the scan (because the OCTA tracks eye movement), we cannot rule out the possibility that subclinical nystagmus in patients with amblyopia may have affected our results. In addition, this study included multiple comparisons; therefore, we cannot rule out the possibility of statistical artifact. However, the identification of a significantly decreased vessel density in patients with amblyopia is novel. Further studies using a larger sample size are encouraged, including the evaluation of the SCP, DCP, and intermediate capillary plexus separately. Additional studies will clarify our findings and determine their clinical relevance.

Young patients with amblyopia have reduced superficial and deep retinal capillary density on OCTA. These vascular changes may be associated with decreased visual acuity. The clinical relevance of this finding remains to be determined.

Corresponding Author: Stacy L. Pineles, MD, UCLA Stein Eye Institute, 100 Stein Plaza, Los Angeles, CA 90095 (pineles@jsei.ucla.edu).

Accepted for Publication: July 27, 2017.

Published Online: September 14, 2017. doi:10.1001/jamaophthalmol.2017.3423

Author Contributions: Drs Lonngi and Pineles had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Lonngi, Velez, Pineles.

Acquisition, analysis, or interpretation of data: Lonngi, Velez, Tsui, Davila, Rahimi, Chan, Sarraf, Demer.

Drafting of the manuscript: Lonngi, Velez, Tsui, Rahimi, Chan, Pineles.

Critical revision of the manuscript for important intellectual content: Velez, Davila, Sarraf, Demer.

Statistical analysis: Lonngi, Velez, Pineles.

Administrative, technical, or material support: Davila, Rahimi, Pineles.

Study supervision: Velez, Chan, Sarraf, Pineles.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Sarraf reported receiving grants, speaking or consulting fees, and/or nonfinancial support outside of this study from Allergan, Amgen, Bayer, Genentech, Heidelberg, Novartis, Optovue, and Regeneron. No other disclosures were reported.

Funding/Support: This study was funded in part by an unrestricted grant (NIH/NEI K23EY021762) to the Department of Ophthalmology at UCLA (University of California, Los Angeles) from Research to Prevent Blindness, Inc, and by grant EY0089313 from the National Eye Institute (Dr Demer). The Colombian Association of Pediatric Ophthalmology and Strabismus, Pan-American Ophthalmological Foundation, and Retina Research Foundation provided salary support to Dr Lonngi.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Meeting Presentation: This study was presented in part at the 2017 Annual Meeting of the American Association for Pediatric Ophthalmology and Strabismus; April 3, 2017; Nashville, Tennessee; and at the 2017 Annual Meeting of the Association for Research in Vision and Ophthalmology; May 8, 2017; Baltimore, Maryland.

Additional Contributions: Melinda Chang, MD; Rui Zhang, MD; Juan Jose Duque, MD; Nicholas Iafe, MD; Nopasak Phasukkijwatana, MD; and Sarah Harmon, BA, Stein Eye Institute, UCLA, collaborated with the authors on this study. They were not compensated for their contributions.