Confocal Scanning Laser Ophthalmoscopy in High Myopic Eyes in a Population-Based Setting

Abstract

Purpose.: To compare the parameters of confocal scanning laser ophthalmoscopy (Heidelberg Retina Tomograph [HRT] II; Heidelberg Engineering, Heidelberg, Germany) in high myopia with those in age-matched emmetropia.

Methods.: A population-based study in which HRT II data were analyzed from 135 healthy subjects with high myopia (spherical equivalent [SE] from −6 to −12 D) and 135 age-matched subjects with emmetropia (SE from −1 to +1 D). The HRT parameters, the correlation between disc area and ovality, and asymmetry between the right and left eyes were evaluated.

Results.: High myopia was associated with greater disc ovality, smaller cup areas and cup volumes, higher rim volumes, height variation contour (HVC), and the mean retinal nerve fiber layer (RNFL) thickness compared with those parameters in emmetropia (P ≤ 0.003). The intergroup differences in the rim volume and mean RNFL thickness remained significant (P = 0.003) after adjustment for the disc area and ovality. The disc area was correlated significantly (P ≤ 0.002) with most parameters in both groups. The disc ovality was significantly (P = 0.005) negatively correlated with the disc area only in high myopia and significantly (P ≤ 0.003) positively correlated with the rim volume, HVC, and mean RNFL thickness in both groups. Asymmetry of the mean RNFL thickness was significantly (P = 0.003) greater in high myopia than in emmetropia.

Conclusions.: The characteristics of the HRT parameters in highly myopic eyes involved smaller cup parameters and greater rim and RNFL parameters compared with emmetropic eyes. The effects of disc area and ovality on the HRT parameters in highly myopic eyes differed from those in emmetropic eyes.

Myopia is a significant risk factor for development of primary open-angle glaucoma (POAG) in cross-sectional population-based studies. ^1–5 In a Japanese population-based study, ⁴ low myopia (spherical equivalent [SE], −1.0 to −3.0 D) and moderate to high myopia (≤ −3.0 D) were significant risk factors for POAG with odds ratios (ORs) of 1.85 and 2.60, respectively. In other populations, an association was found between myopia (SE ≤ −0.5 D) and glaucoma, including suspected glaucoma, with an OR of 1.5 in the Barbados Eye Study. ² In the Blue Mountains Eye Study in Australia, ¹ low myopia (−1.0 to −3.0 D) had an OR of 2.3 and moderate to high myopia (≤ −3 D) an OR of 3.3. In the Beaver Dam Eye Study in the United States, ³ individuals with myopia, defined as an SE of ≤ −1.0 D, were 60% more likely to have glaucoma than were emmetropic individuals (OR, 1.6). These results suggested that the prevalence of POAG is relatively higher in myopic eyes in an elderly population. In highly myopic eyes, however, the diagnosis of glaucoma and the evaluation of glaucoma progression are often difficult because of myopia-induced changes in the retina and optic disc, such as oblique insertions or the presence of a tigroid fundus. 

With the development of optic disc imaging devices, objective and quantitative measurements of the shape of the optic disc for evaluating glaucomatous optic neuropathy have become available. A confocal scanning laser ophthalmoscope, the Heidelberg Retina Tomograph (HRT) or HRT II (Heidelberg Engineering, Heidelberg, Germany), allows three-dimensional topographic analysis of the optic disc and provides a quantitative measure of a variety of optic disc parameters. ^6–8  

Regarding the effect of myopia on HRT parameters, several studies reported a significant effect of the refractive error on some HRT parameters, ^9–11 whereas others failed to show a significant relationship. ^12–14 Among these studies, only the report by Leung et al. ¹⁰ included a large number of eyes with high myopia; the authors found a significant correlation between the disc area measured by HRT with myopia. However, they did not evaluate the effect of myopia on other HRT parameters. Thus, little information is available regarding the characteristics of the HRT parameters in high myopia. Because the prevalence of POAG is especially high in populations with higher degrees of myopia, it is imperative to understand the characteristics of the HRT parameters in normal eyes with high myopia to evaluate and diagnose POAG in these eyes with the help of the HRT. 

We obtained a large sample of HRT data from the population-based Japanese Tajimi Study. ^15,16 The characteristics of the HRT parameters and related factors among normal subjects with an SE refractive error between −5 and +5 D have been reported elsewhere. ¹¹ To establish the fundamental data for early detection or diagnosis of glaucoma in highly myopic eyes, we assessed the HRT parameters in more than 135 eyes with high myopia (SE between −12 and −6 D) and compared the findings with the results in age-matched emmetropic eyes (SE between −1 and +1 D). Moreover, we compared the intereye asymmetry between highly myopic and emmetropic eyes and the effect of the disc area and disc ovality on the HRT parameters. 

Methods

Population Sampling and Diagnostic Procedures

The Tajimi Study, a population-based eye study of Japanese subjects aged 40 years or older, was conducted in Tajimi City, Japan, the details of which have been published elsewhere. ^15,16 Briefly, from 54,165 residents aged 40 years or older, 4,000 subjects were selected randomly and encouraged to participate. After 48 subjects who died and 82 subjects who were not Tajimi residents or had moved were excluded, 3870 subjects were eligible and 3021 participated in the screening examinations. The investigation adhered to the tenets of the Declaration of Helsinki and the municipal statutes. The study protocol was approved by the local ethics committee. Written informed consent was obtained from all participants. 

Screening examinations included autokeratorefractometry (KP-8100PA; Topcon, Tokyo, Japan), visual acuity (VA), slit lamp examination, applanation tonometry, central corneal thickness measurement with a specular-type pachymeter (SP-2000P; Topcon), digital color fundus photography through an undilated pupil (IMAGEnet digital fundus camera system, TRC-NW6S; Topcon) with angles of 30° and 45°, visual field screening with a frequency-doubling technology screener (FDT; Carl Zeiss Meditec, Inc., Dublin, CA), and optic disc measurements using HRT II (software version, 1.4.1). 

During screening, three glaucoma specialists independently evaluated the color fundus photographs (nonstereoscopic photographs) for any abnormal findings, including glaucomatous optic disc appearance and nerve fiber layer defects. The criteria for suspected glaucoma were a vertical cup-to-disc ratio of 0.6 or more, a difference in the vertical cup-to-disc ratio of 0.2 or more between both eyes, a rim width at the superior (11–1 clock hours), and inferior portions (5–7 clock hours) of 0.2 or less of the disc diameter, a nerve fiber layer defect, or a splinter disc hemorrhage. When at least one examiner noted any of these findings, suggesting the presence of an abnormality or glaucoma, the subjects were recruited for a definitive examination, during which stereoscopic photographs were taken through dilated pupils, and four independent glaucoma specialists who were not part of the screening process evaluated the optic disc findings. Eyes with a congenital disc anomaly or a suspected abnormality also were identified. In the present study, only eyes diagnosed as normal in the screening and definitive examinations were included. 

HRT II Optic Disc Measurements

During screening, topographic images of the optic disc were obtained with the HRT II through undilated pupils. According to the instrument's operating manual, ¹⁷ a refractive error (SE) up to ±12 D was adjusted by using the adjustable lens on the instrument without adding corrective glasses. When a reliable quality image could not be obtained due to a refractive error exceeding 12 D or astigmatism exceeding 2 D, the measurement was performed with corrective glasses. The results in eyes in which the HRT data were obtained with corrective lenses were excluded from the present study. Reliable HRT-II image quality was defined by appropriate focus, brightness, and clarity; minimal ocular movement; an optic disc centered in the image; and a standard deviation of the mean topographic image less than 40 μm. When a reliable quality image was not obtained, the HRT II measurements were repeated up to three times. If the repeated measurements were again unreliable, we abandoned the measurements. The same experienced operator (TT), who was masked to the other ocular information, outlined the optic disc margin around the inner margin of the peripapillary scleral rings while viewing nonstereo color fundus photographs. The contour line was reviewed in the topography and reflectance images and the height profile graph included in the instrument. Ten HRT parameters obtained during routine examination were analyzed: disc area, cup area, rim area, cup-to-disc area ratio, cup volume, rim volume, mean cup depth, height variation contour (HVC), cup shape measure (CSM), and mean retinal nerve fiber layer (RNFL) thickness. Magnification errors were corrected by the HRT II software according to the subject's SE refractive error and corneal curvature (the average of the steepest and flattest meridians) based on autokeratorefractometry. 

Determination of Optic Disc Ovality

On a digital color fundus photograph with a 45° viewing angle of the eyes whose HRT data were used for the current analysis, the dimensions of the optic disc were determined with image analyzing software that was originally developed and installed in the IMAGEnet system. Seven points were plotted on the inner margin of the scleral ring on the fundus photographs with the 45° viewing angle by an examiner (AH) masked to the demographic data and final diagnosis of the subject, and an ellipse was drawn automatically by using the least mean square method that produces a contour of the optic disc. The ovality index was calculated as the ratio of the longest diameter to the shortest diameter of the optic disc. 

Data Analysis

Eyes for which reliable results of the HRT II and color fundus photographs were available were included in the analyses. Reliable image quality of a color fundus photograph was defined by appropriate focus, brightness, and clarity sufficient to identify the fovea and the optic disc contour. 

The eyes that were considered normal based on the screening and definitive examinations were included in the analysis. Normal subjects had a best corrected VA of 20/30 or more, a normal IOP of 21 mm Hg or less, an optic disc with a normal appearance, a normal visual field based on the FDT screener or Humphrey perimetry, no previous laser surgery or intraocular surgery, and no substantial ocular disease. Eyes with a refractive error (SE) of ±12 D or more or astigmatism greater than 2 D were excluded because of the possibility of use of corrective lenses when performing HRT according to the operating manual of the HRT II. ¹⁷ Subjects with glaucoma or suspected glaucoma, POAG, ocular hypertension, pseudoexfoliation, congenital disc anomalies, or other ocular diseases in at least one eye were excluded. 

Based on the inclusion and exclusion criteria, 3819 normal eyes of 2157 subjects were eligible. The reasons for exclusion of the 2223 eyes were as follows: screening was performed in the subjects' home (88 eyes); HRT II and/or fundus photography could not be completed at the screening sites because of ocular or physiologic problems (905 eyes); the SD of the HRT II measurements was 40 μm or more (541 eyes); eyes with definitive glaucoma, suspected glaucoma, pseudoexfoliation, POAG, ocular hypertension or the fellow eyes of these eyes (400 eyes); eyes with other ocular diseases including congenital disc anomalies that could affect the disc shape (98 eyes); pseudophakic eyes (45 eyes); eyes with excessive myopia (SE < −12 D) or astigmatism (>2 D, 15 eyes); and eyes with best-corrected VA (BCVA) worse than 20/30 (131 eyes). Of the 2157 subjects, 135 subjects had high myopia defined as an SE between −12 and −6 D in at least one eye, and 1223 subjects had emmetropia defined as an SE between −1 and +1 D in at least one eye. No subject had high myopia and emmetropia bilaterally. The refractive error was determined based on the results of corrective lenses that provided the BCVA in eyes in which the uncorrected VA was worse than 20/20 and based on the results of autokeratorefractometry in eyes in which the uncorrected VA was 20/20 or better. If both eyes were eligible, data from the randomly chosen eye were used for analyses. To establish the age-matched control group of emmetropia, 135 eyes of 135 age-matched subjects with emmetropia were chosen randomly as follows: first, the 135 highly myopic eyes were stratified into four age groups (40–49, 50–59, 60–69, and 70+ years of age), and the numbers of eyes were counted in each age group; the 1223 emmetropic eyes then were stratified into the same age groups and the same numbers of eyes in each age group were randomly chosen by using a random number table. Bilateral differences in the HRT parameters were assessed in 21 subjects with bilateral high myopia and 62 subjects with bilateral emmetropia, in whom the difference in the SE between the right and left eyes was 0.25 D or less. 

Comparisons of the mean values between the right and left eyes and between groups were analyzed with paired and unpaired t-tests, respectively. The effects of the disc area and ovality were adjusted by using the analysis of covariance model that included the disc area and ovality as the covariance factor and group (high myopia or emmetropia) as a fixed factor. The correlation between the two values was analyzed with Pearson's correlation coefficient. Because 10 HRT parameters were analyzed, P < 0.005 (0.05/10) were considered significant according to Bonferroni's method for multiple comparisons (SPSS 15.0J for Windows; SPSS Japan, Inc., Tokyo, Japan). 

Results

There were no significant differences in the ratio of men to women (P = 0.222, χ² test) and that of right to left eyes (P = 0.222), the central corneal thickness (P = 0.276, unpaired t-test), and the IOP (P = 0.139) between 135 highly myopic subjects and 135 age-matched emmetropic subjects (Table 1). 

Table 1.

 

Demographic Data from 135 Eyes with High Myopia and 135 Age-Matched Eyes with Emmetropia

	High Myopia	Emmetropia	P
Male/female	65/70	70/65	0.543*
Right/left eye	68/67	57/78	0.222*
Age, y	48.7 ± 6.8	49.5 ± 7.0	Matched
Refractive error, D	−7.3 ± 1.3	−0.2 ± 0.5	<0.001†
Central corneal thickness, mm	0.516 ± 0.034	0.520 ± 0.033	0.276†
IOP, mmHg	14.9 ± 2.5	14.4 ± 2.6	0.139†

 Data are expressed as the mean ± SD.

 Significance of comparisons of the highly myopic eyes and the emmetropic eyes determined by the χ² test (*) or the unpaired t-test (†).

Table 2 shows the averages of the ovality index and the HRT parameters in the two groups. The ovality index was significantly greater in the highly myopic group than in the emmetropic group (P < 0.001, unpaired t-test). The disc area tended to be smaller in high myopia than in emmetropia (P = 0.020). The cup area and the cup volume were significantly (P = 0.003, P < 0.001, respectively) smaller in high myopia than in emmetropia. The rim volume, HVC, and mean RNFL thickness were significantly (P < 0.001) greater in high myopia. The intergroup differences in rim volume and mean RNFL thickness were still significant after adjustment for the disc area and ovality using analysis of covariance (P ≤ 0.003). 

Table 2.

 

Average Ovality Index and HRT Parameter between 135 Highly Myopic Eyes and 135 Age-Matched Emmetropic Eyes

Parameter	High Myopia	Emmetropia	P *	P Adjusted for Disc Area and Ovality†
Ovality index	1.248 ± 0.156	1.155 ± 0.074	<0.001‡	—
Disc area, mm2	2.001 ± 0.525	2.142 ± 0.461	0.020	—
Cup area, mm2	0.38 ± 0.318	0.507 ± 0.377	0.003‡	0.065
Rim area, mm2	1.621 ± 0.406	1.635 ± 0.301	0.754	0.065
Cup-to-disc area ratio	0.177 ± 0.127	0.221 ± 0.128	0.005	0.067
Cup volume, mm3	0.066 ± 0.083	0.12 ± 0.145	<0.001‡	0.005
Rim volume, mm3	0.536 ± 0.197	0.442 ± 0.134	<0.001‡	0.001‡
Mean cup depth, mm	0.192 ± 0.085	0.205 ± 0.089	0.216	0.323
Height variation contour	0.483 ± 0.152	0.394 ± 0.089	<0.001‡	0.009
Cup shape measure	−0.202 ± 0.071	−0.203 ± 0.073	0.900	0.373
Mean RNFL thickness, mm	0.323 ± 0.099	0.264 ± 0.064	<0.001‡	0.003‡

 Data are expressed as the mean ± SD. The ovality index indicates the ratio of the longest to the shortest disc diameters.

 * Comparisons between highly myopic eyes and emmetropic eyes (unpaired t-test).

 † Comparisons between highly myopic eyes and emmetropic eyes after adjustment for disc area and disc ovality (analysis of covariance).

 ‡ P < 0.005 (0.05/10, with Bonferroni's correction for multiple comparisons).

The correlation coefficients of the HRT II parameters with the disc area and ovality are summarized in Table 3. The HVC correlated negatively with the disc area only in high myopia (P = 0.002). The mean RNFL thickness significantly (P ≤ 0.002) negatively correlated in both groups. All other parameters correlated positively with the disc area in both groups (P < 0.001). Disc ovality significantly (P = 0.005) negatively correlated with the disc area only in high myopia and significantly (P ≤ 0.003) positively correlated with the rim volume, HVC, and mean RNFL thickness in both groups. 

Table 3.

 

Correlation between HRT Parameters with Disc Area and Ovality Index

Parameter	Correlation with Disc Area		Correlation with Disc Ovality
	High Myopia	Emmetropia	High Myopia	Emmetropia
Disc area			−0.24 (0.005)	−0.03 (NS)
Cup area	0.64 (< 0.001)	0.76 (< 0.001)	−0.16 (NS)	−0.13 (NS)
Rim area	0.80 (< 0.001)	0.58 (< 0.001)	−0.18 (NS)	0.11 (NS)
Cup-to-disc area ratio	0.40 (< 0.001)	0.58 (< 0.001)	−0.12 (NS)	−0.13 (NS)
Cup volume	0.45 (< 0.001)	0.67 (< 0.001)	−0.16 (NS)	−0.13 (NS)
Rim volume	0.17 (NS)	0.13 (NS)	0.25 (0.004)	0.26 (0.003)
Mean cup depth	0.23 (NS)	0.56 (< 0.001)	0.03 (NS)	−0.07 (NS)
Height variation contour	−0.26 (0.002)	−0.12 (NS)	0.60 (< 0.001)	0.29 (0.001)
Cup shape measure	0.31 (< 0.001)	0.47 (< 0.001)	−0.08 (NS)	−0.06 (NS)
Mean RNFL thickness	−0.37 (< 0.001)	−0.26 (0.002)	0.45 (< 0.001)	0.27 (0.002)

  n = 135 eyes in both age-matched groups. Data are Pearson's correlation coefficients (P). Ovality index, the ratio of the longest to shortest disc diameters; NS, not significant (P < 0.005; 0.05/10, with Bonferroni's correction for multiple comparisons).

In 21 subjects with bilateral high myopia with an intereye difference in SE of 0.25 D or less and 62 subjects who had bilateral emmetropia with an intereye difference in SE of 0.25 D or less, all HRT parameters correlated significantly between the right and left eyes, with correlation coefficients exceeding 0.4 (P < 0.005, Pearson's correlation coefficient; Table 4). Absolute differences in the mean RNFL thickness between the right and left eyes were significantly larger in highly myopic subjects than in emmetropic subjects (P = 0.003, unpaired t-test), whereas the absolute difference in the SE between the highly myopic subjects and the emmetropic subjects was not significant (0.13 ± 0.08 D vs. 0.14 ± 0.10 D, P = 0.80). 

Table 4.

 

Correlation Coefficients and Absolute Differences in the HRT Parameters in the High Myopia and Emmetropia Groups

Parameter	Correlation Coefficient between Right and Left Eyes*		Absolute Difference between Right and Left Eyes†		P ‡
	High Myopia	Emmetropia	High Myopia	Emmetropia
Disc area	0.76	0.91	0.131 ± 0.084	0.137 ± 0.100	0.008
Cup area	0.86	0.87	0.162 ± 0.240	0.052 ± 0.041	0.854
Rim area	0.56	0.69	0.257 ± 0.201	0.153 ± 0.132	0.027
Cup-to-disc area ratio	0.73	0.86	0.154 ± 0.119	0.159 ± 0.124	0.793
Cup volume	0.77	0.77	0.285 ± 0.214	0.193 ± 0.140	0.433
Rim volume	0.48	0.52	0.076 ± 0.066	0.072 ± 0.053	0.032
Mean cup depth	0.67	0.82	0.044 ± 0.044	0.058 ± 0.076	0.414
Height variation contour	0.56	0.47	0.172 ± 0.167	0.110 ± 0.086	0.017
Cup shape measure	0.59	0.52	0.051 ± 0.047	0.043 ± 0.039	0.428
Mean RNFL thickness	0.53	0.47	0.110 ± 0.066	0.071 ± 0.062	0.003§

 Data are shown for the right and left eyes with bilateral high myopia (n = 21 subjects), with an intereye difference of SE ≤ 0.25, and for eyes with bilateral emmetropia (n = 62) with an intereye difference of SE ≤ 0.25 D.

 * Pearson's correlation coefficient. All correlation coefficients in the high myopia and emmetropia groups were significant (P < 0.005).

 † Data are expressed as the mean ± SD.

 ‡ Comparisons between highly myopic eyes and emmetropic eyes by unpaired t-test.

 § P < 0.005 (=0.05/10 with Bonferroni's correction for multiple comparisons).

Discussion

In the present study, the HRT parameters as well as the disc area in the highly myopic eyes were characterized for the first time in a large number of highly myopic eyes and compared with those in age-matched emmetropic eyes. Regarding the disc area, previous studies ^10,18,19 have reported that highly myopic eyes had a larger disc area measured on stereoscopic ¹⁸ or monoscopic ¹⁹ fundus photographs or by HRT III. ¹⁰ However, in the present study, the disc area tended to be smaller in the highly myopia group (2.00 ± 0.53 mm² vs. 2.14 ± 0.46 mm²) with marginal significance (P = 0.020), with Bonferroni's method for multiple comparisons). 

There was no significant correlation between the disc area and the refractive error in either study group (Fig. 1). There are several possible reasons for this discrepancy (i.e., the difference in the definition of high myopia, racial differences, the differences in the magnification correction of the imaging devices, and differences in the subjects' selection procedures). High myopia was defined as an SE less than −8 D in previously published studies, ^18,19 ; in the present study we used an SE of −6 D or less as in other previous population-based studies on refractive errors. ^5,20–23 When the disc area was compared between eyes with an SE of −6 to −8 D and those with an SE of −8 to −12 D in the present study, the latter had a larger disc area (2.13 ± 0.78 mm²) than the former (1.97 ± 0.45 mm²), suggesting a trend similar to that shown in other studies. ^10,18,19 Regarding racial differences, a study in which the HRT III was used in normal Chinese subjects ¹⁰ showed results opposite those of the present study. Although it is unclear whether there are racial differences in ocular or refractive conditions between Chinese and Japanese subjects, differences in the procedure for selecting the study subjects (hospital-based versus population-based) and/or differences in the parameters between the HRT II and the HRT III ²⁴ may have played a role in the discrepancy. Issues regarding the magnification correction in fundus photography and the HRT also should be discussed. The HRT magnification correction is reportedly as accurate as the correction methods that use all refraction, keratometry, and axial length data and is more accurate than the method of Littmann that uses refraction and keratometry. ²⁵ In previous studies in which fundus photography was used, ^18,19 the magnification correction was done with Littmann's method using refraction and keratometry ¹⁸ or refraction alone. ¹⁹ Thus, the differences in the magnification correction between the HRT and fundus photographs also may be related to the discrepancy in the association between the disc area and the refractive error. In addition, the HRT II may not compensate fully for the magnification of the fundus image, especially in highly myopic eyes, as a possible reason for the discrepancy regarding the differences in the optic disc area between emmetropic and highly myopic eyes. 

Figure 1.

 

Scatterplots of the relationship between the SE refractive error and disc area determined by the HRT II in highly myopic (top, n = 135 eyes) and age-matched emmetropic eyes (bottom, n = 135 eyes). There is no significant correlation between the refractive error and the disc area in highly myopic eyes (Pearson's correlation coefficient = −0.098, P = 0.259) and in emmetropic eyes (R = −0.047, P = 0.590).

In the present study, based on population-based data in Japan, the disc area averaged 2.00 and 2.14 mm² in highly myopic and emmetropic eyes, respectively. No previous population-based study using the HRT has been published. As the result of population-based studies using fundus photography, the following values have been reported as mean disc areas: 2.42 mm² in the Rotterdam Study (the Netherlands), ²⁶ 2.94 (black subjects) and 2.63 (white subjects) mm² in the Baltimore Eye Survey (United States), ²⁷ 2.65 mm² in the Beijing Eye Study (China), ²⁸ and 2.58 mm² in the Vellore Eye Study (India). ²⁹ Thus, the mean disc area determined by the HRT in the Tajimi Study population is apparently smaller than that determined by fundus photography in other populations. Because the magnification correction procedures differ between fundus photography and the HRT, further studies that analyze the results of fundus photography in the Tajimi Study or those using the HRT in other populations are needed to compare the optic disc characteristics among different populations. 

The effect of refractive error on intradisc parameters (i.e., HRT parameters other than disc area) has been reported. Nakamura et al. ⁹ found that the mean and maximum cup depths increased significantly with increasing myopia in normal Japanese subjects (SE between −5 and +4.13 D). Abe et al. ¹¹ recently reported a significant positive correlation between the cup parameters and the SE and a significant negative correlation between the rim parameters, HVC and RNFL thickness, and the SE in approximately 1800 normal subjects between −5 and 5 D in the Tajimi Study population. In contrast, Bowd et al. ¹² and Durukan et al. ¹³ found no significant associations between the refractive error and HRT parameters in normal subjects (SE between −5 and +5 D and SE between −4.75 and +4.25 D, respectively). Tong et al. ¹⁴ found no significant association between the refractive error and HRT parameters in children in Singapore with an SE between −8.6 and +3.6 D. However, no or few eyes with high myopia were included in those studies, which might have limited the power of the study to detect changes attributable to high myopia. The present study analyzed a large number of highly myopic eyes and found for the first time that these eyes had smaller cup areas and cup volumes and higher rim volumes, HVCs, and RNFL thicknesses determined by HRT. 

In highly myopic eyes, oblique insertion of the optic disc is seen often, which could affect the three-dimensional configuration of the optic disc. In fact, in the present study, disc ovality, as a surrogate parameter for the oblique insertion, ^30,31 was significantly correlated with the disc area and some intradisc parameters of HRT, especially in the highly myopic eyes. The present study and a previous study ¹¹ in which HRT was used reported a positive correlation between the RNFL thickness and myopia. In contrast, the opposite relationship (i.e., myopic eyes have a thinner RNFL, has been reported in studies using optical coherence tomography ³² or scanning laser polarimetry. ^33–35 In eyes with oblique insertion of the optic disc, because HRT or HRT II could place the reference plane at a falsely posterior position, the distance between the reference plane and the nerve fiber layer surface is increased, resulting in an artifactual increase in the three-dimensional parameters, including the mean RNFL thickness, which implies that oblique insertion should be considered when analyzing HRT parameters, especially in highly myopic eyes. 

In the present study, one comparison using the unpaired t-test showed significant differences in several HRT parameters between highly myopic and emmetropic eyes (Table 2). Because disc area and ovality were correlated significantly with the HRT parameters, we further adjusted the disc area and ovality in the comparisons using an analysis of covariance model. After adjusting for disc area and ovality, most differences in the HRT parameters between high myopia and emmetropia were no longer significant, whereas only the rim volume and the mean RNFL thickness remained significantly different between high myopia and emmetropia (P < 0.003). This suggests the presence of a difference in the three-dimensional disc configuration between high myopia and emmetropia after excluding the effect of disc area and oblique insertion. 

Two analysis programs are included in the HRT II that classify the optic discs as normal or glaucomatous by a combination of the topographic parameters. Iester et al. ³⁶ reported the discriminant function that uses a combination of cup volume, HVC, and age-corrected CSM to classify the optic discs. Moorfields Regression Analysis (MRA) distinguishes between normal and glaucomatous eyes based on the logarithm of rim area adjusted by the disc area. ³⁷ Because HVC is affected strongly by disc ovality in highly myopic eyes, as found in the present study, the discriminant function of Iester et al. ³⁶ may be affected noticeably by disc ovality especially in highly myopic eyes. However, since the disc area was correlated with the ovality in the current series of highly myopic eyes, the MRA in highly myopic eyes also may be affected by the disc ovality. If an index of disc ovality is added as an explanatory variable in such glaucoma classification programs, the performance in highly myopic eyes may be improved further. 

The intereye difference (asymmetry) in the neuroretinal rim configuration between the right and left eyes was included in the criteria for the diagnosis of glaucoma in a population-based study ³⁸ and was reportedly a risk factor for progression from ocular hypertension to glaucoma. ³⁹ Regarding the intereye difference of the HRT parameters, Harasymowycz et al. ⁴⁰ reported that the ratio of the rim area to disc area asymmetry (RADAAR) was correlated significantly with the IOP and the stages of glaucoma in 140 patients with glaucoma. Hawker et al. ⁴¹ reported that neither RADAAR nor asymmetries in the rim parameter was affected significantly by age or sex in 459 normal subjects. In the present study, the intereye differences of some parameters including the mean RNFL thickness were significantly greater in subjects with high myopia than in subjects who were emmetropic, suggesting that refractive errors should be important when an intereye difference in the neuroretinal rim configuration is used as a criterion for glaucoma diagnosis or a risk factor for development of glaucoma. 

Highly myopic eyes with an SE between −6 and −12 D were analyzed in the present study. According to the HRT II operating manual, ¹⁷ refractive errors (SE) up to ±12 D can be adjusted using the adjustable lens on the instrument, and this was followed during the Tajimi Study. Moreover, a study using a model eye found that the HRT had a constant factor of magnification correction (i.e., a telecentric construction) with the range of ametropia between +10 and −10 D. ⁴² Thus, we believe that the values of the HRT parameters obtained in the present study should be reliable. However, other factors such as oblique insertion, disc ovality, or peripapillary atrophy that may affect the HRT measurements in highly myopic eyes are not understood fully and deserve further investigation. 

A limitation of the present study was that all participants in the Tajimi Study were not included in the analyses. In the Tajimi Study, 3021 residents were screened, but only 2157 subjects were eligible for the current HRT study because of several exclusion criteria, such as the fact that they were screened in their own homes, the unavailability of the HRT results and/or fundus photographs, ocular diseases including glaucoma, and excessively high myopia or astigmatism. Between the 2157 subjects who were included and the 864 subjects who were excluded, there were significant age differences (55.6 ± 10.0 vs. 65.4 ±13.1 years old, P < 0.001) and in the male/female ratio (979/1178 vs. 355/509, P = 0.032), suggesting possible but unrecognized selection bias in the present study. 

In conclusion, the present study, based on population-based data obtained with the HRT II, revealed that highly myopic eyes had smaller cup parameters and greater rim and RNFL parameters than did age-matched emmetropic eyes. The correlation between HRT parameters and disc ovality was significant only in high myopia, suggesting that disc ovality and disc area at least should be considered when analyzing the HRT results in high myopia. The asymmetry of several HRT parameters including the rim parameters in highly myopic eyes was significantly greater than in emmetropic eyes. 

Footnotes

 Supported by the Japan National Society for the Prevention of Blindness, Tokyo, Japan, and the Japan Ophthalmologists Association, Tokyo, Japan.

Footnotes

 Disclosure: T. Tsutsumi, None; A. Tomidokoro, None; H. Saito, None; A. Hashizume, None; A. Iwase, None; M. Araie, None

Footnotes

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.