Anterior Segment Dimensions in Asian and Caucasian Eyes Measured by Optical Coherence Tomography

Abstract

BACKGROUND AND OBJECTIVE

To compare Asian and Caucasian anterior segment dimensions measured by optical coherence tomography (OCT).

PATIENTS AND METHODS

Anterior segment OCT images were obtained in normal subjects. Four line scans were acquired at the 90°, 45°, 0°, and 135° meridians of each eye. Computer calipers acquired anterior segment dimensions of corneal diameter, anterior chamber width, corneal vault, and anterior chamber depth on OCT images. Univariate and multivariate analyses were performed to assess correlations.

RESULTS

Corneal diameter was 0.5 mm narrower (P < .01), anterior chamber width was 0.46 mm narrower (P < .01), and corneal vault was 0.22 mm lower (P < .01) in Asian eyes. All anterior segment dimensions decreased with age.

CONCLUSION

Asian eyes had smaller anterior segments compared to Caucasian eyes. Regardless of race, anterior segment dimensions were smaller in older subjects. Age-related changes may affect the tolerability of long-term implants such as phakic intraocular lenses.

INTRODUCTION

The dimensions of the ocular anterior segment affect the sizing of ophthalmic devices such as contact lenses, microkeratomes, suction rings, corneal trephines, corneal implants, and intraocular lenses (IOLs). Thus, accurate measurements of the anterior segment dimensions are important. Anterior segment anatomical structures vary among different races,^1,2 and our hypothesis is that they are smaller in Asians compared to Caucasians.

Previous ultrasonographic studies reported no racial differences between Asian and Caucasian eyes in anterior chamber (AC) depth.^3,4 Nevertheless, ultrasound imaging requires contact with the cornea and the results are limited to two-dimensional scans. Optical coherence tomography (OCT) is a non-contact scanning system that is capable of visualizing the cornea, anterior lens, and AC angle structures and measuring anterior segment dimensions based on three-dimensional scan patterns.

Because of its high spatial resolution and non-contact nature, OCT accurately measures the anatomical dimensions of the anterior segment.⁵ Our previous study showed that it was capable of measuring the scleral spur and angle recess.⁶ The purpose of this study was to measure the anterior segment of the eye by OCT and to identify and assess the factors that affect its size and shape.

PATIENTS AND METHODS

Subjects

All subjects of this prospective observational study were recruited at the Doheny Eye Institute, University of Southern California, Los Angeles, California. This study was approved by the Institutional Review Board of the University. It followed the tenets of the Declaration of Helsinki and was in accord with the Health Insurance Portability and Accountability Act of 1996. Informed consent was obtained from each of the study subjects.

All participants were normal volunteers residing in the greater Los Angeles area and were recruited by posting. Each received a screening examination including manifest refraction, uncorrected and best spectacle-correction visual acuity measurement, and slit-lamp examination before enrollment. Exclusion criteria included corneal pathology, previous corneal surgery, best spectacle-corrected visual acuity worse than 20/25, contact lens wear within the past 3 months, and age younger than 18 or older than 60 years.

OCT Calibration and Imaging

For this study, we used a prototype anterior segment, time-domain OCT instrument (Optovue Inc., Fremont, CA). It operated with a 1,310 nm-wave-length light source. The scanning speed was 2,000 axial scans per second. The axial and transverse resolutions were 17 and 28 µm, respectively.

The OCT scan dimensions were calibrated with a specially designed calibration block composed of an aluminum plate with engraved crosshair grooves (depth = 2.493 mm) and covered with a fine metal grid mesh (30 per inch; SPI Supplies, West Chester, PA). The base plate was manufactured at the Cleveland Clinic Prototype Laboratory (Cleveland, OH) with a super precise vertical machining center (SV-400, Mori Seiki Co., Nagoya, Japan) and inspected with a coordinate measuring machine (LK-CMM G-90CG; LK Metrology Systems, Inc., Brighton, MI) that had 5 µm accuracy along the z-axis.

OCT scans were performed during 11 AM to 5 PM on weekdays. During OCT scanning, the operator used a cotton tip to hold the eyelids of the subject open. The eyelids were pressed against the periorbital bones to avoid pressure on the eyes while the subject gazed at an internal fixation target. To compensate for kappa, the angle between the visual and pupillary axes, the operator adjusted the target position to level the iris plane on the real-time OCT display. The scan was centered on the corneal vertex as visualized by the strong vertex reflection on OCT display. Scans were performed along the 90°, 45°, 0°, and 135° meridians in time sequence. Each scan line was 18 mm in length and consisted of 256 axial scans acquired in 0.128 second. Two consecutive scans were performed for each meridian. The scans were then repeated on the contra-lateral eye.

Image Processing and Anatomical Measurement

Subjects were excluded if an image on any of the four meridians was eccentric due to poor centering on the vertex or contained landmarks that were shadowed due to an eyelash or eyelid. Acceptable images were “dewarped” to compensate for the refractive index change at the air/tear interface. All images from left eyes were flipped horizontally in a mirror image fashion before measurements. Corneal diameter, AC width, corneal vault, and AC depth were measured by computer calipers on each line scan (Fig. 1). The scleral spurs were manually identified first. The AC width was measured as the distance between the two scleral spurs. The anterior scleral sulci were located on the anterior eye surface and connected with the scleral spurs by lines perpendicular to the anterior surface. The corneal diameter was taken as the distance between the two anterior scleral sulci. The AC depth was measured as the distance from the anterior corneal vertex to the anterior lens capsule along the bisector of the corneal diameter. The corneal vault, the distance between corneal apex and anterior surface of the lens, was measured from the anterior corneal vertex to the line connecting the scleral sulci. Measurements from the four meridians were averaged. All inspections and measurements were done by a single experienced ophthalmologist (BQ) masked to subject information.

Figure 1.

Horizontal optical coherence tomography section of the anterior eye. Anterior chamber (AC) width was measured between the scleral spurs.

Patient Information, Refractive Error, Axial Length, and Corneal Power

Body height was measured with a medical scale. Racial heritage, age, and gender were established by subject self-reporting. Refractive error was acquired by manifest refraction and was converted to spherical equivalent refraction. Axial length and corneal power were measured by IOL Master (software version 5.0; Carl Zeiss Meditec Inc., Dublin, CA), an instrument that combines a partial coherence interferometer and an automated keratometer.

Statistical and Data Analysis

All subjects were divided into two racial groups, Asian and Caucasian, according to their self-reported racial heritage. The values of corneal diameter, AC width, corneal vault, and AC depth were first averaged from the two consecutive line scans from each meridian and then averaged from the four meridional scans.

Assuming that the limbal shape is elliptical and the AC width, W, varied with meridional angle θ, then

$$W(\mathit{2}\theta )=W_{\mathit{0}}+d(\mathit{2}\theta )$$

where W₀ is the AC width averaged over four meridians (W(0°), W(45°), W(90°), and W(135°), and d(2θ) is the deviation from circularity (Fig. 2). Further assuming that d(2θ) = A cos(2θ) + B sin(2θ), where A = W(0°) – W(90°), and B = W(45°) – W(135°), then

$$d(\mathit{2}\theta )=C\mathit{\text{ cos}}(\mathit{2}\theta -\mathit{2}\alpha ),$$

where $\mathrm{C}=\sqrt{{\mathrm{A}}^2+{\mathrm{B}}^2}$ and α is the angle of the widest meridian of AC width representing the orientation of the limbal ellipse, which was calculated using the following formula:

$$\alpha =\mathit{0.5}\times \mathit{\text{Arctan}}(B/A)+{\mathit{90}}^{°}(A<\mathit{0})$$

$$\alpha =\mathit{0.5}\times \mathit{\text{Arctan }}(B/A)+\mathit{180}°(A>\mathit{0})$$

$$\alpha =\mathit{45}°(A=\mathit{0}\mathit{\text{ and }}B>\mathit{0})$$

$$\alpha =135°(A=\mathit{0}\mathit{\text{ and }}B<\mathit{0}).$$

Figure 2.

Parameters of anterior chamber shape calculation. The ellipse and the circle share the same center. W₀ is the anterior chamber width averaged over 4 meridians, and α is the angle of the widest meridian.

Limbal ellipticity (E) was defined as the ellipticity of the limbus and was calculated using the following formula:

$$\mathrm{E}=\sqrt{\frac{{(1+\frac{\mathrm{C}}{{\mathrm{W}}_0})}^2-{(1-\frac{\mathrm{C}}{{\mathrm{W}}_0})}^2}{{(1+\frac{\mathrm{C}}{{\mathrm{W}}_0})}^2}}$$

The repeatability of anterior segment dimensions was evaluated by pooled standard deviation of repeated measurements and coefficient of variation. Univariate analyses of eye dimensions and corneal power against race and gender were performed using two sample t tests. Univariate analyses of eye dimensions and corneal power against age, body height, and refractive error were performed using linear regression. Factors that significantly affected eye dimensions or corneal power were set as covariates in multivariate analyses. Multivariate regression was used to evaluate the effects of covariates on eye dimensions. The chi-square statistic was used to assess the relative importance of the independent covariates. The generalized estimating equation⁷ was used to account for the inter-eye correlation. The level of significance was set at a P value of less than .05. All analyses were done in SAS 9.1 software (SAS Institute Inc., Cary, NC).

RESULTS

Fifty-four normal subjects were enrolled in the study. After image quality inspections, 33 subjects (66 eyes) were used for data analyses. They included 18 men and 15 women, 14 Caucasians of European origin and 19 Asians including 12 Chinese, 4 Japanese, and 3 Koreans. The mean ± standard deviation age of all subjects was 33.0 ± 7.3 years (range: 22 to 51 years). The mean height was 171.2 ± 8.6 cm (range: 155 to 188 cm).

The repeatability of all OCT measurements was better than 4% coefficient of variation (Table 1). As shown in univariate analyses, race (Table 2), gender (Table 2), age (Table 3), and height (Table 3) significantly affected anterior segment dimensions. For Asian eyes, corneal diameter was 0.5 mm narrower (P < .01), anterior chamber width was 0.46 mm narrower (P < .01), and corneal vault was 0.22 mm lower (P < .01) compared to Caucasian eyes (Table 2). There was no racial difference in AC depth (P = .11) or axial length (P = .47). For all subjects, age was negatively correlated with corneal diameter (0.032 mm per year, P = .013, Table 3), AC width (0.025 mm per year, P < .01), corneal vault (0.013 mm per year, P < .01), and AC depth (0.023 mm per year, P < .01). For all subjects, gender and height also affected all anterior segment dimensions in univariate analyses (P < .01).

TABLE 1.

Measurements and Repeatability of Anterior Segment Dimensions

Anterior Segment Dimensions	Mean (mm) (Range)	Pooled SD	CV
Corneal diameter	12.87 (9.03–14.21)	0.16	1.24%
AC width	11.75 (10.75–13.07)	0.15	1.28%
Corneal vault	2.98 (2.33–3.37)	0.11	3.56%
AC depth	3.77 (2.37–4.49)	0.14	3.74%

SD = standard deviation; CV = coefficient of variation; AC = anterior chamber.

TABLE 2.

Univariate Analyses of Eye Dimensions and Corneal Power Against Race and Gender

Eye Parameters	Mean ± SD		P	Mean ± SD		P
	Asian	Caucasian		Male	Female
Corneal diameter (mm)	12.73 ± 0.43	13.23 ± 0.44	< .01	13.09 ± 0.45	12.60 ± 0.82	< .01
AC width (mm)	11.58 ± 0.40	12.04 ± 0.39	< .01	11.90 ± 0.42	11.48 ± 0.73	< .01
Corneal vault (mm)	2.89 ± 0.23	3.11 ± 0.24	< .01	3.05 ± 0.18	2.83 ± 0.33	< .01
AC depth (mm)	3.68 ± 0.35	3.87 ± 0.35	.11	3.93 ± 0.36	3.57 ± 0.35	< .01
Corneal vault/diameter	0.23 ± 0.013	0.24 ± 0.014	.071	0.23 ± 0.010	0.22 ± 0.015	< .01
Axial length (mm)	24.23 ± 1.18	24.02 ± 1.14	.47	24.58 ± 1.28	23.61 ± 0.73	< .01
Limbal ellipticity (mm)	0.13 ± 0.04	0.11 ± 0.03	.072	0.12 ± 0.03	0.12 ± 0.04	.28
Corneal power (D)	43.86 ± 1.56	44.20 ± 1.65	.37	44.06 ± 1.49	44.14 ± 1.21	.80

SD = standard deviation; AC = anterior chamber; D = diopters.

TABLE 3.

Univariate Analyses of Eye Dimensions and Corneal Power Against Age, Height, and Refractive Error

Eye Parameters	Age (Y)		Height (cm)		Refractive Error (D)
	Slope	P	Slope	P	Slope	P
Corneal diameter (mm)	−0.032	.013	0.052	< .01	−0.011	.75
AC width (mm)	−0.025	< .01	0.042	< .01	−0.006	.85
Corneal vault (mm)	−0.013	< .01	0.022	< .01	−0.021	.14
AC depth (mm)	−0.023	< .01	0.034	< .01	−0.029	.16
Corneal vault/diameter	0.00	.15	0.001	< .01	−0.001	.022
Axial length (mm)	−0.030	.18	0.11	< .01	−0.36	< .01
Limbal ellipticity (mm)	0.00	.71	0.006	.24	−0.021	.087
Corneal power (D)	0.01	.83	−0.042	.092	−0.12	.085

AC = anterior chamber; D = diopters.

Multivariate analyses showed that race, height, and age were significantly correlated with some anterior segment dimensions (Table 4). Asian eyes had 0.32 mm narrower corneal diameter (P = .035), 0.35 mm narrower AC width (P = .015), and 0.11 mm borderline lower corneal vault (P = .06) compared with those of Caucasian eyes. Age was negatively correlated with all anterior segment dimensions, and height was positively correlated with corneal diameter and corneal vault. The effect of gender was no longer present in any of the anterior segment dimensions after adding other covariates in multivariate analyses.

TABLE 4.

Multivariate Analyses of Anterior Segment Dimensions

Anterior Segment Dimensions	Covariates	Slope	Chi-square	P
Corneal diameter (mm)	Caucasian	0.32	3.46	.035
	Height (cm)	0.039	2.82	< .01
	Age (y)	−0.033	6.69	< .01
	Female	−0.055	0.08	.78
AC width (mm)	Caucasian	0.35	4.29	.015
	Height (cm)	0.033	2.09	.24
	Age (y)	−0.031	6.76	< .01
	Female	−0.037	0.04	.85
Corneal vault (mm)	Caucasian	0.11	2.93	.06
	Height (cm)	0.024	5.71	< .01
	Age (y)	−0.016	7.31	< .01
	Female	0.018	0.04	.84
AC depth (mm)	Caucasian	0.071	0.57	.45
	Height (cm)	0.023	2.94	.056
	Age (y)	−0.025	6.22	< .01
	Female	−0.013	0.01	.94

AC = anterior chamber.

The meridianal distribution of AC width was compared between Asian and Caucasian eyes. The average asymmetries in the cardinal meridians (ie, the difference between 0° and 90° meridians) in Asian and Caucasian eyes were −340 ± 329 µm and −316 ± 274 µm, respectively (P = .75). The average asymmetries in the oblique meridians (ie, the difference between 45° and 135° meridians) in Asian and Caucasian eyes were −570 ± 248 and −502 ± 249 µm, respectively (P = .28). As shown in Figure 3, most of the widest AC meridians were located along the superotemporal-inferonasal meridian (100° to 140° for the right eye and the mirror image for the left eye). For Asians, the angle of the widest AC meridian was 119.63° ± 13.00°, and for Caucasians it was 120.14° ± 14.91° (P = .88).

Figure 3.

Distribution of the angle of the widest anterior chamber (AC) meridian in all eyes. Left eyes were analyzed in a mirror image.

DISCUSSION

There have been several studies with good repeatability using anterior segment OCT to measure the AC.^6,8,9 In this study, good repeatability was also achieved in all measured dimensions. Apart from AC depth, we also measured the corneal vault. AC depth might vary as lens thickness changes during accommodation¹⁰ or aging.¹¹ We found that corneal vault measurements eliminate this effect by avoiding the need for lens measurements. Thus, vault measurements better describe the characteristics of corneal shape.

To our knowledge, our study is the first to use OCT to compare anterior segment dimensions between different racial groups. We did not find racial difference in AC depth, which further confirmed previous ultrasonographic findings. Asian eyes in our study had a smaller corneal diameter, AC width, and corneal vault than did Caucasian eyes. These conclusions were confirmed after accounting for the effect of age, height, and gender. Asian eyes did not have a shorter axial length compared with Caucasian eyes, in agreement with another report.⁴ Therefore, Asian eyes are smaller than Caucasian eyes only in the anterior segment.

It is possible that the racial differences we found might not be validated in the world population. Although we assume that the differences between Asian and Caucasian subjects are largely of genetic origin, we cannot exclude an environmental component. Although our study was conducted in the United States, we do not know whether all of the subjects were raised there. Thus, this same study should be done in Asian countries to see whether there is an additional environmental effect. If the anterior segment is also found to be smaller for Asians in Asian countries, the finding would support the development of smaller contact lens, corneal implants, and intraocular implants for use in Asian countries.

Asian corneas were smaller than Caucasian corneas, but the shape and orientation of the limbal ellipse was the same for the two races. We also found that the AC width was widest in most subjects, both Asian and Caucasians, along the superotemporal-inferonasal meridian. However, others reported that the corneal diameter is wider horizontally than vertically.¹² In our study, the AC width was an internal AC measurement that was different from the external measurement of corneal diameter. In another study, Scheimpflug imaging indicated that the corneal radius was different along different meridians and that the widest meridian was between 90° and 180°.¹³ Similarly, others using OCT indicated that the vertical AC width was larger than the horizontal one.^9,14 The difference was 300 µm,⁹ which was close to our finding of 330 µm. Therefore, the usual practice of placing an angle-supported AC phakic IOL with the haptics in the horizontal meridian might not be ideal because the haptics could rotate toward the widest meridian.

We found that older subjects have shallower ACs. This has been reported by several studies using ultrasonography, ^4,15 Orbscan,¹⁶ and OCT.¹⁴ The rate of decrease in AC depth was reported to be 0.017 mm/year in an OCT study of Chinese subjects¹⁴ and 0.021 mm/year in an ultrasonographic study of Eskimos, Asians, and Caucasians.⁴ These rates are similar to our finding of 0.025 mm/year. We also found smaller AC widths in older subjects. A study using white-to-white measurement showed that age is inversely correlated with corneal diameter and AC depth.¹⁶ When measured by OCT, AC width was smaller in older subjects.¹⁷ The reason for these findings might be corneal cross-linking that is part of the aging process.^18,19 The cornea undergoes glycosylation-dependent cross-linking with aging²⁰ that leads to shrinkage and stiffening of the corneal tissue. These findings could also be explained by non–age-related factors because the study was cross-sectional. ^16,17 Our subjects grew up in different eras and possibly in different parts of the world, and they might have experienced different nutritional or environmental effects during development.

If our finding of smaller corneal size in older subjects is a result of age-related anatomical changes, it might have implications for phakic IOL implantation. One of the most serious long-term concerns after phakic IOL implantation is progressive endothelial loss caused by the proximity between the IOL and the cornea.^21–23 Shrinking of the anterior segment over time might reduce the distance between the lens and the cornea, thus potentially speeding the loss of endothelial cells. Because most phakic IOL surgeries were conducted within the past 20 years, potential complications might not become evident until the patients get older. Our preliminary results suggested that the AC width might shrink by 0.031 mm per year. The change after 30 years of aging would be 0.93 mm, which might cause more compression of angle-supported IOL through haptic–angle contact and progressive changes in the anterior vault. After 30 years, corneal vault and AC depth would also be reduced by 0.47 and 0.75 mm, respectively. The combination of these changes could greatly decrease the distance between the corneal endothelium and all types of phakic IOLs. Currently, the AC depth requirement is 3.0 mm or greater for the Visian Implantable Collamer Lens and 3.2 mm or greater for the Artisan/Verisyse phakic IOL.²⁴ However, this does not allow for possible greater age-related change in younger patients. Our results suggested that patients with phakic IOLs need to be monitored for decades after implantation, and long-term studies of the safety of phakic IOL are needed.

There are several limitations to our study. First, the scleral spurs were located by a human grader and subject to human errors. Second, our study was cross-sectional, and a longitudinal study is needed to establish that the age-related correlation we found was indeed caused by aging rather than generational differences. Third, only 61.1% of the OCT images were used for data measurement and analysis. The main reason for the large number of unusable scans was the difficulty in scanning beyond the limbus, especially in the vertical meridian. The subjects’ effort of opening eyes played an important role. Imaging artifact introduced by instrument malfunction during a certain period of the study also contributed to the high rejection rate. Nevertheless, the results did show an interesting trend that might be useful for future studies. The reasons for unusable scans were not related to instrument used, ocular structure, race, and age.

The current study of measuring anterior segment dimensions using OCT suggested that the anterior segments of the eye are smaller in Asian Americans compared to Caucasian Americans. Older subjects tend to have smaller anterior segment dimensions. These differences should be considered in the development of instruments and implants for the eye.