Optic disc characteristics evaluated with swept source optical coherence tomography angiography (SS-OCTA) in nonarteritic anterior ischemic optic neuropathy (NAION): a cross-sectional study

Abstract

Background

Nonarteritic anterior ischemic optic neuropathy (NAION) is one of common optic neuropathies affect optic disc. But the structure of optic disc in NAION remains poorly understood. Our aim was to identify optic disc characteristics of NAION.

Methods

This cross-sectional study included chronic-stage affected eyes and unaffected eyes from NAION patients and normal eyes from age-matched individuals. Measurements of optic disc and macula from swept source optical coherence tomography angiography (SS-OCTA) were recorded. Besides conventional OCTA parameters, we also manually measured circumpapillary retinal nerve fiber layer (cRNFL) thickness, lamina cribrosa depth (LCD) and prelaminar tissue thickness (PLTT). Additionally, we introduced a new parameter, retinal nerve fiber layer density (RNFLD).

Results

This study finally included 53 NAION-fellow eyes, 26 chronic-stage NAION eyes and 50 normal eyes. NAION-fellow eyes and NAION eyes showed significantly thicker PLTT compared to normal eyes (p < 0.001). NAION fellow eyes showed similar peripapillary retinal nerve fiber layer (pRNFL) thickness but significantly thicker cRNFL (p < 0.001) compared to normal eyes. And NAION eyes showed significantly thinner pRNFL (p < 0.001) but similar cRNFL compared to normal eyes. RNFLD was higher in NAION-fellow eyes compared to the other two groups (p < 0.01). Logistic analyses showed PLTT was independently associated with NAION susceptibility (OR = 1.017, 95% CI: 1.009–1.025, p < 0.001). ROC curve of PLTT showed that the area under the curve (AUC) was 0.905 (95% CI: 0.846–0.964, p < 0.001).

Conclusions

Both affected and unaffected eyes of NAION patients exhibited an increase in non-nerve fiber components within the retinal nerve fiber layer at optic disc. These increased components might be glial tissue or the remnants of the primitive vitreous. Additionally, PLTT was an independent susceptibility factor for NAION.

Background

Nonarteritic anterior ischemic optic neuropathy (NAION) is one of common optic neuropathies and a prevalent cause of blindness in middle-aged and elderly individuals. Studies have shown that the incidence of NAION can reach 2.5–10/100 000 in people over 50 years old [1, 2]. NAION is related to systemic diseases such as hypertension, diabetes, and obstructive sleep apnea syndrome (OSAS) but the etiology and risk factors for NAION remain unclear [3,4,5].

Several studies indicated that crowded optic disc is a predisposing factor for NAION. Due to limitations in previous optical coherence tomography (OCT) technology which had a shallow scanning depth, previous studies primarily focused on two-dimensional parameters such as the cup-to-disc ratio (C/D). It was not possible to comprehensively evaluate crowded optic disc from a three-dimensional perspective in the past. We hypothesized that NAION-risk eyes not only exhibit small C/D but also possess greater tissue volume at optic disc. Therefore, we conducted this study to validate our hypothesis. In this study, we used swept source optical coherence tomography angiography (SS-OCTA) to compare the structural differences of optic disc among the affected and unaffected eyes from NAION patients and normal eyes from age-matched individuals.

In recent years, with the rapid development in OCT technology, we are able to see more details of the fundus. SS-OCTA offers a great scanning depth and clear resolution, allowing for a detailed visualization in the structure of the optic disc and lamina cribrosa [6]. In this study, we measured circumpapillary nerve fiber layer (cRNFL) thickness, prelaminar tissue thickness (PLTT) and lamina cribrosa depth (LCD). In addition to measuring C/D to assess the traditional concept of “crowded optic disc”, we also introduced a new parameter, retinal nerve fiber layer density (RNFLD), which we believed can effectively demonstrate crowded optic disc. Besides these measurements, we also assessed conventional OCT and OCTA parameters of optic disc and macula.

Methods

This cross-sectional study adhered to the tenets of the Declaration of Helsinki and was approved by Ethics Committee of Tianjin Medical University Eye Hospital [2024KY(L)-08]. We included NAION patients and age-matched controls who visited the Department of Medical Retina and Neuro-ophthalmology of Tianjin Medical University Eye Hospital from December 2022 to April 2024.

Inclusion and exclusion criteria

For NAION-fellow group, the inclusion criteria included fellow eyes from unilateral NAION patients diagnosed by neuro-ophthalmologists. For NAION group, the inclusion criteria included NAION eyes at the chronic stage from NAION patients. The chronic stage was defined as at least six months after disease onset with complete resolution of optic disc edema. In cases of both eyes met the inclusion criteria, one was randomly selected. The control group were normal eyes from age-matched individuals who visited the Department of Medical Retina and Neuro-ophthalmology for routine examination and did not have any neuro-ophthalmologic condition. For each control individual, we randomly selected one of the two eyes.

The exclusion criteria were as follows: (1) NAION eyes within six months after NAION onset; (2) presence of other neuro-ophthalmologic conditions; (3) signal strength index of OCTA image less than 7/10; (4) high myopia (axial length greater than 26 mm); (5) previous intraocular surgeries except cataract surgeries.

Ophthalmic examination and baseline characteristics

All subjects underwent a comprehensive ophthalmic examination including best-corrected visual acuity (BCVA), slit-lamp biomicroscopy, intraocular pressure (IOP) measurement, dilated fundus examination and SS-OCTA examination. Other baseline characteristics such as age, sex, laterality and systemic disorders including high blood pressure (HBP), diabetes and obstructive sleep apnea syndrome (OSAS) were also recorded.

Optical coherence tomography angiography

SS-OCTA (VG200, Intalight, China) was used to scan and image the macula and optic disc of all subjects. OCTA scan protocols included a 6 × 6 mm² scan area at optic disc and a 6 × 6 mm² scan area at macula. Data such as radial peripapillary capillary density (RPC density), peripapillary retinal nerve fiber layer (pRNFL) thickness, Bruch’s membrane opening area (BMOa), cup-to-disc ratio (C/D), vessel density (VD) in superficial vascular complex (SVC) and deep vascular complex (DVC), thickness of ganglion cell-inner plexiform layer (GCIPL) and choroid thickness (CT) were automatically obtained using the built-in algorithm in the device.

For optic disc parameters, RPC density was measured based on Garway-Heath peripapillary grid (Fig. 1A), a 2–4 mm-diameters annulus area centered on optic disc. Peripapillary choroid thickness (pCT) was measured in the same way. pRNFL thickness was measured in a 3.46 mm-diameters circle centered on optic disc.

For macular parameters, macular area was segmented into three concentric circles with diameters of 1 mm, 3 mm, and 6 mm according to Early Treatment Diabetic Retinopathy Study (ETDRS) grid (Fig. 1B). The inner 1 mm-diameter circle region, 1–3 mm anulus region and 3–6 mm anulus region were designated as fovea, parafovea and perifovea respectively. SVC-VD, DVC-VD and macular choroid thickness (mCT) were automatically measured based on this grid through the build-in algorithm. SVC was defined as the vasculature extending from 5 μm above the inner limiting membrane to the interface between the inner two-thirds and outer one-third of the ganglion cell layer. DVC referred to the vasculature between the interface and 25 μm below the lower border of the inner nuclear layer. GCIPL thickness was measured and obtained based on an elliptical annulus centered on fovea. The vertical diameters of the annulus were 1 and 4 mm, while the horizontal diameters were 1.2 and 4.8 mm (Fig. 1C).

Fig. 1

Segmentation of optic disc (A) and macula (B, C). (A) Segmentation of optic disc based on Garway-Heath peripapillary grid. (B) Segmentation of macula based on ETDRS grid. (C) Segmentation of macula to measure GCIPL thickness. GCIPL, ganglion cell-inner plexiform layer

We manually measured circumpapillary retinal nerve fiber layer (cRNFL) thickness, lamina cribrosa depth (LCD) and prelaminar tissue thickness (PLTT) by the built-in software. cRNFL is the retinal nerve fiber layer at Bruch’s membrane opening (BMO) as BMO can be seen as the optic disc margin. It was measured in vertical and horizontal B-scans at optic disc (Fig. 2A). And the average of the measurements in four quadrants was recorded as cRNFL thickness. LCD was calculated by dividing the area of the region formed between anterior lamina cribrosa surface and the reference plane, which connects the two BMO, by the BMO distance (BMOd) in B-scan (Fig. 2B). And the average of the measurements from vertical and horizontal B-scans was recorded as LCD. PLTT was calculated by dividing the prelaminar tissue area (PLTTarea) by BMOd in B-scan (Fig. 2C). And the average of the measurements from vertical and horizontal B-scans was recorded as PLTT.

Fig. 2

Measurements of cRNFL, RNFLD, LCD and PLTT. (A) Measurements of cRNFL and RNFLD, RNFLD=(cRNFL1 + cRNFL2)/BMOd. (B) Measurement of LCD, LCD = LCDarea/BMOd. (C) Measurement of PLTT, PLTT = PLTTarea/BMOd. cRNFL, circumpapillary retinal nerve fiber layer; RNFLD, retinal nerve fiber layer density; LCD, lamina cribrosa depth; PLTT, prelaminar tissue thickness

We also created a new parameter: retinal nerve fiber layer density (RNFLD). It was calculated by dividing the sum of the two cRNFL thickness by BMOd in B-scan (Fig. 2A). We thought it can demonstrate the degree of optic disc crowding. Like LCD and PLTT, RNFLD was measured through both vertical and horizontal B-scans. The average of the two measurements in vertical and horizontal B-scans was recorded as RNFLD.

For manually measured parameters, two experienced ophthalmologists who were masked to the subjects’ clinical information performed the measurements. Each ophthalmologist measured each parameter three times, and the mean of these measurements were used for analysis.

Statistical analysis

All statistical analyses were performed using SPSS version 25.0. Quantitative data following a normal distribution were presented as mean ± standard deviation, and differences among the three groups were assessed using one-way analysis of variance (ANOVA). Quantitative data not following a normal distribution were presented as median (p25%, p75%), and Kruskal-Wallis test was used for comparison. Categorical data were presented as percentages, and Chi-square test was used for comparison. Bonferroni correction or Tamhane’s T2 test were used for all-pairwise post hoc comparisons. Intraclass correlation coefficient (ICC) was used to evaluate the consistency of manually measured parameters both intraoberver and interobserver. Univariate and multivariate logistic regression analyses between the NAION-fellow group and the control group were conducted to identify independent susceptibility factors, and the receiver operating characteristic curve (ROC) analysis for these susceptibility factors were also generated. A p-value of less than 0.05 was considered statistically significant.

Results

Demographic and clinical information

A total of 129 subjects were finally included in this study. NAION-fellow group included 53 fellow eyes from 53 unilateral NAION patients, NAION group included 26 chronic-stage NAION eyes from 26 NAION patient, and control group included 50 eyes from 50 age-matched individuals.

The mean age of NAION-fellow group, NAION group and control group were 57.5 ± 10.7 years, 58.3 ± 10.3 years and 59.0 ± 11.1 years, respectively. There were 21 (39.6%), 14 (53.8%) and 23 (46.0%) were male, respectively. NAION group showed significantly worse BCVA (logMAR) than the other two groups (p < 0.001). There were no statistical differences in terms of age, sex, laterality, IOP, and systemic disorders among the three groups. Details of the demographic and clinical information were shown in Table 1.

Table 1 Comparisons of demographic and clinical information among NAION-fellow, NAION and control groups

Comparisons of VD in SVC/DVC, mCT and GCIPL

NAION group showed significantly lower SVC-VD and DVC-VD in fovea, parafovea and perifovea than the other two groups (all p < 0.01). Post hoc comparisons showed that there were no significant differences in SVC-VD and DVC-VD between NAION-fellow group and control group. For GCIPL, NAION group was thinner than NAION-fellow group and control group (p < 0.001). Post hoc comparisons showed that there were no significant differences in GCIPL thickness between NAION-fellow group and control group (p = 0.861). There were no differences in mCT among the three groups (all p > 0.05). Details of SVC-VD, DVC-VD, GCIPL and mCT were shown in Supplemental Table S1.

Comparisons of optic disc parameters

The intraobserver and interobserver measurements of manually measured parameters (cRNFL, LCD, PLTT and RNFLD) all exhibited excellent agreement (all ICC > 0.9, all p < 0.05, Supplemental Table S2).

NAION-fellow group and NAION group showed significantly lower C/D than control group (p < 0.001, Fig. 3A). For pRNFL, NAION group was significantly thinner than the other two groups (p < 0.001, Fig. 3B). For cRNFL, NAION group and control group were significantly thinner than NAION-fellow group (p < 0.001, Fig. 3C). NAION group and control group also showed significantly lower RNFLD than NAION-fellow group (p < 0.001, Fig. 3D). For PLTT, NAION-fellow group and NAION group were thicker compared to control group (p < 0.001, Fig. 3E). There were no significant differences in LCD among the three groups (p = 0.389, Fig. 3F).

For RPC density, NAION group was significantly lower than NAION-fellow group and control group (p < 0.001, Fig. 3G). There were no significant differences in BMOa and pCT among the three groups (all p > 0.05, Fig. 3H and I). Details of optic disc parameters were shown in Fig. 3 and Supplemental Table S3.

Fig. 3

Comparisons of C/D (A), pRNFL (B), cRNFL (C), RNFLD (D), PLTT (E), LCD (F), RPC density (G), BMOa (H) and pCT (I) among NAION-fellow, NAION and control groups. C/D, cup-to-disc ratio; pRNFL, peripapillary retinal nerve fiber layer; cRNFL, circumpapillary nerve fiber layer; RNFLD, retinal nerve fiber layer density; PLTT, prelaminar tissue thickness; LCD, lamina cribrosa depth; RPC density, radial peripapillary capillary density; BMOa, Bruch’s membrane opening area; pCT, peripapillary choroid thickness. *p < 0.05, **p < 0.01, ***p < 0.001

Univariate and multivariate logistic regression analysis

We included NAION-fellow group and control group to perform a univariate logistic regression. The results showed that C/D, cRNFL, RNFLD and PLTT were associated with a higher susceptibility for NAION (all p < 0.01, Table 2). C/D, cRNFL, RNFLD and PLTT were all susceptibility factors for NAION.

Table 2 Univariate logistic regression analysis of C/D, cRNFL, RNFLD and PLTT

Then C/D, cRNFL, RNFLD and PLTT were included in the multivariate logistic regression to identify the independent susceptibility factor for NAION. The result showed that thicker PLTT was associated with a higher susceptibility for NAION (OR = 1.017, 95% CI: 1.009–1.025, p < 0.001, Table 3).

Table 3 Multivariate logistic regression analysis of C/D, cRNFL, RNFLD and PLTT

ROC analysis of PLTT

According to the ROC analysis, the area under the curve (AUC) was 0.905 (95% CI: 0.846–0.964, p < 0.001, shown in Fig. 4). The optimal cut-off value of PLTT was 661.9 μm, with a sensitivity of 77.4% and specificity of 92.0%.

Fig. 4

ROC curve of PLTT. PLTT, prelaminar tissue thickness

Discussion

Currently, the pathogenesis of NAION remains unclear. The prevailing view suggests that in the presence of crowded optic disc, systemic conditions such as hypertension, diabetes and obstructive sleep apnea syndrome (OSAS) lead to an acute circulatory insufficiency of the optic disc. This results in optic disc ischemic edema, and the subsequent development of compartment syndrome, which ultimately triggers NAION. Our demographic data showed no significant differences in the prevalence of systemic conditions such as diabetes, hypertension, and OSAS between NAION patients and age-matched individuals. This may be related to the relatively small sample size of our population or our method of obtaining medical histories, as we relied solely on inquisitions without conducting further examinations.

Our study showed that pRNFL of chronic-stage NAION eyes were thinner compared to normal eyes but cRNFL were similar. Since pRNFL represents the RNFL at some distance from the optic disc, whereas cRNFL represents the RNFL at the optic disc margin. This result indicated that although the retinal nerve fibers in chronic-stage NAION eyes had decreased, there were still substantial non-nerve fiber components in retinal nerve fiber layer at optic disc in chronic-stage NAION eyes. Besides, cRNFL of NAION-fellow eyes were thicker compared to the normal eyes but pRNFL were similar. This result also indicated that the unaffected eyes of NAION patients had non-nerve fiber components in retinal nerve fiber layer at optic disc compared to normal eyes. Furthermore, our study showed that the prelaminar tissue thickness (PLTT) in both affected and unaffected eyes of NAION patients were significantly increased, while the lamina cribrosa depth (LCD) did not differ among the three groups. This result indicated that both affected and unaffected eyes of NAION patients had more prelaminar tissue compared to normal eyes, and this was not due to the difference in lamina cribrosa depth. Prelaminar tissue is an extension of the retinal nerve fiber layer towards the lamina cribrosa. Our findings on pRNFL, cRNFL, and PLTT indicated that both affected and unaffected eyes of NAION patients had more tissue in nerve fiber layer at optic disc. But the increased tissue consisted of non-nerve fiber components.

Retinal nerve fiber layer contains not only nerve fibers but also non-nerve fiber components, which includes blood vessels and various types of glial tissues such as astrocytes, microglia, and Müller cells. The proportion of glial tissue within the retinal nerve fiber layer is substantial. Histological studies on the retinal nerve fiber layer of primates had shown that the glial content of the retinal nerve fiber layer can reach 20-30%, with even greater variability in different locations, ranging from 18–42% [7]. The glial content increased closer to the optic disc, with levels ranging from 18 to 30% at 1.5–2 mm from optic disc and reaching 42-45% at 0.5–1 mm from optic disc [7]. Similar histological studies on human indicated that at least 20% of the retinal nerve fiber layer consisted of non-nerve fiber components, with increased amounts of astrocytes closer to optic disc [8]. These findings suggested great variability in the glial content within the retinal nerve fiber layer at optic disc, with greater variation of the glial content closer to the optic disc. Combined with our findings that both affected and unaffected eyes of NAION patients exhibited increased non-nerve fiber components within the nerve fiber layer at optic disc, we hypothesized that NAION-risk eyes have substantial amounts of glial tissue in retinal nerve fiber layer at optic disc compared to normal eyes. We attempted to detect differences in the composition of retinal nerve fiber layer between eyes of NAION patients and normal eyes by analyzing the signal intensity of retinal nerve fiber layer in B-scans. However, this method was not fully effective due to the high signal intensity of the RNFL, which caused attenuation of signals in deeper RNFL. Currently, there were no reliable methods for in vivo assessment of the tissue composition of the retina. Future possibilities included adaptive optics optical coherence tomography (AO-OCT), which had shown potential in visualizing microstructures and cellular morphology within internal limiting membrane and RNFL [9]. AO-OCT might become a viable technique for in vivo microscopic examination of retinal tissue composition in the future.

Additionally, we hypothesized that the increased prelaminar tissue may also be due to incomplete regression of the primitive vitreous. One patient, who initially presented with unilateral NAION, and then developed NAION in the fellow eye five months later. This patient’s fellow eye exhibited a Bergmeister’s papilla, which was considered as the remnant of hyaloid artery fibrous sheath that has not completely regressed [10]. And its signal intensity in OCT is similar to RNFL. Furthermore, measurements such as the PLTT, cRNFL and RNFLD in this patient were all above the normal range (mean + 2SD).

Regarding other parameters of optic disc, eyes of NAION patients showed smaller C/D compared to the normal eyes. This finding was consistent with previous studies on susceptibility factors for NAION as a small C/D is also considered a hallmark of NAION-risk eyes [11, 12]. There was no significant difference in Bruch’s membrane opening area (BMOa) between the eyes of NAION patients and normal eyes, consistent with the results from previous studies [13]. We did not find significant differences in peripapillary choroidal thickness (pCT) between the eyes of NAION patients and normal eyes. Our finding differed from previous studies. Some previous studies found that both NAION eyes and their fellow eyes had a thicker pCT compared to normal eyes [14, 15]. Another study found some segments of pCT were thicker in NAION-affected eyes than in normal eyes, while pCT of the fellow eyes showed no difference compared to normal eyes [16]. It is still unclear whether choroidal thickness is related to NAION. In terms of RPC density, our study found that chronic-stage NAION eyes was lower compared to NAION-fellow eyes and normal eyes. This was consistent with most previous studies [17, 18]. And it has become a consensus that chronic-stage NAION eyes exhibit capillary loss in RPC.

As for parameters in macula, our studies showed that chronic-stage NAION eyes have thinner GCIPL, lower SVC-VD and DVC-VD compared to NAION-fellow eyes and normal eyes. This aligned with most previous studies [19,20,21]. Additionally, there were also no differences in mCT among the three groups. This was also consistent with some previous studies [16, 22].

Additionally, through multivariate logistic regression analysis, we found that PLTT was an independent susceptibility factor for NAION. Furthermore, ROC curve analysis revealed that PLTT has high evaluation efficiency. It can be used to identify NAION-risk eyes. PLTT holds promise as a biomarker for screening high-risk populations.

The strength of our study was the comprehensive evaluation of optic disc through three-dimensional measurements. Parameters including cRNFL, RNFLD and PLTT all derived from three-dimensional measurements. Due to the limitations in previous OCT technology, most previous studies focused solely on two-dimensional parameters such as C/D and optic disc diameter, with little consideration given to potential difference in the vertical dimension of crowded optic discs.

The main limitation of our study was the small sample size but it was comparable to the sample sizes of previous OCTA studies on NAION. Additionally, some parameters in our study such as cRNFL, RNFLD and PLTT were manually measured. This might cause potential measurement bias. But we believe this did not impact our conclusions, as the intraobserver and interobserver ICC were high and differences were very significant. However, if the built-in software could automatically measure these parameters, it would lend greater credibility to our findings.

Conclusion

Both affected and unaffected eyes of NAION patients exhibited an increase in non-nerve fiber components within the retinal nerve fiber layer at optic disc. These increased components might be glial tissue or the remnants of the primitive vitreous. Additionally, PLTT was an independent susceptibility factor for NAION.