Optic Disc Tilt Direction Determines the Location of Initial Glaucomatous Damage

Abstract

Purpose.: To explore differences in optic disc tilt and torsion between normal control and glaucoma subjects, and to determine whether the direction of optic disc tilt is consistent with the initial location of glaucomatous visual field (VF) defect.

Methods.: Glaucoma patients with isolated superior or inferior hemifield loss (n = 136) and normal controls (n = 99) were analyzed. Disc ovality index and torsion degree were measured on retinal photographs. Imaging of optic disc was obtained using Heidelberg retinal tomography (HRT) III and Cirrus spectral-domain optical coherence tomography (OCT). The degree of temporal disc tilt was assessed using horizontal topographic images and vertical (upward or downward) disc tilt using vertical topographic images, respectively.

Results.: In all subjects, disc ovality was significantly associated with HRT- and OCT-measured temporal disc tilt and axial length (AL) (all P < 0.001), whereas disc torsion degree was associated with HRT- and OCT-measured vertical disc tilt and AL (all P < 0.05). Association of AL with disc ovality and torsion became more evident as the mean deviation increased. When data on glaucoma patients with superior and inferior hemifield defects were compared, the vertical disc tilt (HRT- and OCT-assessed, P < 0.001 and 0.030, respectively) and the torsion degree (P = 0.002) differed significantly. Upon multivariate logistic regression analysis, the HRT-measured vertical disc tilt was an independent factor determining initial location of the VF defect (P = 0.012).

Conclusions.: Measurement of vertical disc tilt may give valuable information about the superior versus inferior regional susceptibilities of glaucoma.

Introduction

Evidence indicates that particular optic disc morphologic parameters, including peripapillary atrophy (PPA), disc torsion, and tilt, are associated with regional susceptibility to glaucomatous damage. ^1–4 Beta-zone PPA, which occurs more commonly in glaucomatous eyes, is associated with both glaucoma development and subsequent disease progression. ^1,2 In a recent study, the location of the largest beta-zone PPA was found to predict rapid deterioration of the visual field (VF). ³ In addition, the disc torsion direction is associated with the location of the glaucomatous damage (superior versus inferior) in patients with myopic normal-tension glaucoma. ⁴ These findings suggest that optic disc characteristics are relevant to glaucoma pathogenesis. 

Advances in imaging techniques have enabled high-level axial resolution of optic disc anatomy, in turn allowing precise examination of the optic disc. New imaging techniques and novel methods of direct optic disc tilt determination have been described. ^5–7 The optic disc tilt direction was divided into horizontal (disc tilt in the temporal direction) and vertical (disc tilt in the upward or downward direction) planes using horizontal and vertical height profiles in confocal scanning laser ophthalmoscopy, and the degree of the disc tilt in each plane was determined. ⁵ The degree of disc tilt in the temporal direction has also been evaluated using a new approach to spectral-domain optical coherence tomography (OCT), known as enhanced depth imaging. ⁶ In both of these studies, the degree of temporal disc tilt was significantly associated with disc ovality, which has been used as a surrogate index of tilt. ^5,6 However, little is known about the clinical significance of vertical disc tilt. 

Disc tilt is thought to result from posterior scleral expansion, particularly in myopic eyes. ⁸ There has recently been considerable interest in understanding the anatomy of the posterior sclera in relation to glaucoma, because the peripapillary sclera of glaucomatous eyes may have different biomechanics compared to healthy eyes. ^9–11 Therefore, a comprehensive evaluation of the optic disc tilt in glaucomatous eyes will help elucidate glaucoma pathophysiology. 

In the present study, we investigated differences in optic disc tilt and torsion between normal control and glaucoma subjects. In addition, we determined whether the direction of the optic disc tilt is consistent with the location of the initial glaucomatous VF defect. 

Patients and Methods

Study Subjects

The medical records of all consecutive patients with primary open-angle glaucoma (POAG) with isolated superior or inferior hemifield defects, examined by a glaucoma specialist (CKP) between August 2010 and September 2011 at the glaucoma clinic of Seoul St. Mary's Hospital (Seoul, Korea), were reviewed retrospectively. When both eyes of a patient met the inclusion criteria, one eye was randomly selected for evaluation. 

Each initial patient visit featured a review of medical history; measurement of best-corrected visual acuity and refraction; slit-lamp biomicroscopy; gonioscopy; Goldmann applanation tonometry; dilated stereoscopic examination of the optic disc; disc and red-free fundus photography (Canon, Tokyo, Japan); standard automated perimetry (SAP; 24-2 Swedish Interactive Threshold Algorithm, Humphrey Field Analyzer II; Carl Zeiss Meditec, Inc., Dublin, CA, USA); and measurement of central corneal thickness and axial length (AL) (Tomey Corporation, Nagoya, Japan). Optical coherence tomography using a Cirrus HD-OCT (Carl Zeiss Meditec, Inc.) and confocal scanning laser ophthalmoscopy using Heidelberg Retina Tomograph III (HRT III; Heidelberg Engineering, Heidelberg, Germany) were performed. Patients were followed up in an identical manner, usually at 6- to 12-month intervals. The study was performed in strict accord with the tenets of the Declaration of Helsinki after approval by our Institutional Review Board. 

All included subjects had a best-corrected visual acuity ≥ 20/40 on two or more consecutive VF tests, and normal anterior chamber angles in both eyes on slit-lamp biomicroscopy and gonioscopy. Patients with a neurologic or intraocular disease that could cause a VF defect, eyes yielding consistently unreliable VF results (defined as > 25% false-negative results, > 25% false-positive results, or > 20% fixation losses), and eyes with high myopia (a spherical equivalent [SE] ≤ −10 diopters [D], thus outside the focal range of the HRT III [SE ≥ +12 D]), were excluded. 

Subjects with an intraocular pressure (IOP) ≤ 21 mm Hg, a normal optic disc appearance upon examination of color stereoscopic photographs (an intact neuroretinal rim without peripapillary hemorrhage, thinning, or localized pallor), absence of any retinal nerve fiber layer (RNFL) abnormality visible on red-free fundus photographs, and normal VF test results were included in the normal group. A normal VF presentation was defined as a glaucoma hemifield test result within normal limits and mean and pattern standard deviation values associated with probabilities of normality greater than 5%. 

Glaucoma was defined by the presence of glaucomatous optic neuropathy associated with typical reproducible VF defects evident on SAP. A glaucomatous VF change was defined as a glaucoma hemifield test result outside normal limits and the presence of at least three contiguous points in the pattern deviation plot with P values < 5%, with at least one point associated with a P value < 1% (excluding points directly above or below the blind spot), on two consecutive reliable SAP examinations. Consecutive eyes with VF defect clusters within or outside the central 10° of an isolated hemifield were selected and designated as a central and a peripheral subgroup, respectively. These two VF regions were divided into the superior and inferior sectors, as in a previous study. ¹² Patients with overlapping VF defect clusters in both sectors were excluded from the study. To minimize false-positive results, a VF loss had to be apparent on at least two consecutive examinations. 

Measurement of Disc Ovality and Torsion

Color disc and red-free RNFL photographs were obtained using standard settings of a nonmydriatic retinal camera (Nonmyd 7; Kowa, Tokyo, Japan). The photographs and red-free images were evaluated independently, in random order and in a masked fashion, by two of the authors (JAC and H-YLP), who thus lacked knowledge of all clinical information. Disc ovality and torsion were measured on photographs using National Institutes of Health image analysis software (ImageJ version 1.40; available at http://rsb.info.nih.gov/ij/index.html [in the public domain]; developed by Wayne Rasband, National Institutes of Health, Bethesda, MD, USA). 

Disc ovality index was determined using the tilt ratio, which is the ratio between the longest and shortest disc diameter. ^13,14 Disc torsion was defined as the deviation of the long disc axis from the vertical meridian, which was a vertical line perpendicular to a reference line connecting the fovea and the center of the disc. ⁴ The angle between the vertical meridian and the long axis of the disc was termed the degree of torsion. Positive and negative angles indicated the presence of inferotemporal and supranasal torsion, respectively. 

Definitions of Temporal and Vertical Disc Tilt

The degrees of temporal and vertical disc tilt were measured using both HRT III and Cirrus HD-OCT. First, topographic analysis of the optic disc was performed using HRT III. Topographic images were obtained through dilated pupils. The intrascan standard deviation was required to be <30 μm. Disc area was obtained using the HRT III. The extents of temporal and vertical tilt were measured on HRT printouts using the ImageJ software, as described previously. ⁵ The temporal disc tilt was defined as the tilt degree between a horizontal line and a line that was manually drawn to connect the two points where the height profile and the disc margin met (Fig. 1A). The vertical disc tilt was defined as the angle between the vertical line and the line connecting the two points where the height profile and the disc margin met (Fig. 1B). 

Figure 1

 

Measurement of temporal and vertical disc tilt using Heidelberg retina tomography (HRT) III. Horizontal and vertical height profiles and topographic images from HRT printouts are shown. Temporal and vertical disc tilts were measured using the horizontal and vertical height profiles corresponding to horizontal or vertical lines in the topographic images. (A) Temporal disc tilt was the angle (α) between the horizontal line (red) and a line (dashed) connecting the two points where the horizontal height profile and two vertical lines corresponding to the temporal and nasal disc margins met. (B) Vertical disc tilt was the angle (β) between the vertical line (red) and the dashed line described above.

Next, all patients underwent imaging using spectral-domain OCT (Cirrus HD-OCT). Imaging was performed using an optic cube scan featuring 200 × 200 axial scans (pixels) of the optic nerve region. Image quality was assessed by an experienced examiner blinded to patient identity and all test results. Only well-focused, well-centered images without eye movement and with signal strengths of 7/10 or greater were used in the analysis. To measure the degree of temporal and vertical disc tilt, horizontal and vertical tomographic images obtained using the Cirrus HD-OCT were saved and analyzed using the ImageJ software, as described previously. ⁷ On both HRT and OCT, positive and negative temporal disc tilt indicated the presence of temporal and nasal tilt, respectively. Also, positive and negative tilt on vertical meridian indicated the presence of downward and upward tilt, respectively. 

Data Analysis

The distributions of all variables were checked for normality using the Kolmogorov-Smirnov test. Means with interquartile ranges (IQRs) and ranges (minimum to maximum values) are reported for variables with non-normal distributions, rather than means with standard deviations. Intraobserver (two consecutive measurements by JAC) and interobserver (measurements by JAC and H-YLP) reproducibility was assessed by calculation of intraclass correlation coefficients (ICCs). Thirty randomly selected retinal photographs, as well as HRT and SD-OCT tomographic images, were used. 

Student's t-test and the χ² test were used to compare between-group means and percentages derived from independent samples. Associations between disc ovality index and torsion degree and all of age, AL, SE, and the degree of temporal and vertical disc tilt by HRT or OCT were sought using Pearson's correlation approach. Multivariate logistic regression analyses were performed to identify factors defining the initial location of a VF defect. The dependent variable was either the superior or inferior location of the defect. Independent variables were age, mean deviation (MD), AL, disc ovality index, torsion degree, disc size, and HRT-measured temporal and vertical disc tilt. Optical coherence tomography–measured tilt values were excluded from final analysis due to multicollinearity with HRT-measured tilt values. 

All analyses were performed using SPSS for Windows, version 14.0 (SPSS, Chicago, IL, USA), and MedCalc version 9.6 (Mariakerke, Belgium). P values < 0.05 were considered to indicate statistical significance. 

Results

Two hundred thirty-five eyes (of 235 patients) that met the inclusion and exclusion criteria were included in the present study. Table 1 shows the demographics and baseline characteristics of subjects. We examined 99 normal and 136 glaucomatous eyes (77 with superior and 59 with inferior hemifield defects). 

Table 1

 

Baseline Characteristics of the Studied Population

Characteristics	Normal vs. Glaucoma			Glaucomatous Eyes
	Normal Control, n = 99	Glaucomatous Eyes, n = 136	P Value*	Superior Hemifield Defect, n = 77	Inferior Hemifield Defect, n = 59	P Value*
Age, y	52.9 ± 5.5	54.3 ± 13.8	0.464	52.39 ± 13.65	55.80 ± 13.69	0.163
Sex, % female	47.1	57.8	0.113	46.7	47.5	0.536
Spherical equivalent, D	−2.20 ± 2.04	−1.92 ± 2.93	0.404	−2.09 ± 2.97	−1.50 ± 2.72	0.315
Axial length, mm	24.78 ± 1.13	24.66 ± 1.60	0.591	24.61 ± 1.78	24.71 ± 1.41	0.772
Baseline IOP, mm Hg	14.82 ± 2.58	16.22 ± 3.36	<0.001	16.48 ± 3.41	15.90 ± 3.30	0.585
Number of medications	-	1.3 ± 0.5	NA	1.3 ± 0.5	1.3 ± 0.6	0.874
Mean RNFL thickness, μm	98.21 ± 8.70	75.20 ± 9.27	<0.001	75.96 ± 9.32	74.05 ± 9.81	0.253
Mean MD, dB	0.28 ± 0.62	−3.01 ± 1.83	<0.001	−2.91 ± 1.86	−3.28 ± 1.86	0.250
Mean PSD, dB	1.38 ± 0.30	4.54 ± 2.63	<0.001	4.49 ± 2.45	4.73 ± 2.96	0.596

 All data are expressed as the mean ± standard deviation. PSD, pattern standard deviation.

*  Independent t-test.

In all study subjects, the median values (with IQRs) of AL and SE were 24.58 mm (23.80–25.49 mm) and −1.5 D (−3.5 to 0.5 D). The median disc ovality index was 1.14 (1.06–1.20) and the median torsion −1.25° (−8.9° to 1.8°). The median temporal disc tilt yielded by HRT and OCT was 7.55° (4.49°–12.70°) and 3.60° (1.61°–6.40°), respectively, and the vertical disc tilt 2.12° (−0.95° to 6.16°) and 0.00° (−1.03° to 1.62°). Measures of disc ovality index, torsion degree, and temporal and vertical disc tilt all exhibited excellent reproducibility (Table 2). Pearson's correlation coefficient for measures of temporal disc tilt by HRT and OCT was 0.613 (P < 0.001), and that for vertical disc tilt was 0.183 (P = 0.042). 

Table 2

 

Reproducibility of the ONH Measurements Using Retinal Photographs, HRT III, and SD-OCT

	Retinal Photographs		HRT III		SD-OCT
	Disc Ovality	Disc Torsion	Horizontal Tilt Angle	Vertical Tilt Angle	Horizontal Tilt Angle	Vertical Tilt Angle
Intraobserver ICC (95% CI)	0.921	0.922	0.947	0.965	0.882	0.873
	(0.839–0.961)	(0.743–0.976)	(0.825–0.984)	(0.885–0.989)	(0.476–0.973)	(0.435–0.971)
Interobserver ICC (95% CI)	0.885	0.844	0.909	0.911	0.801	0.848
	(0.623–0.965)	(0.567–0.923)	(0.702–0.972)	(0.707–0.973)	(0.117–0955)	(0.325–0.966)

 ONH, optic nerve head; CI, confidence interval.

Table 3 shows the results of univariate analyses seeking factors potentially affecting disc ovality index and torsion degree in the study samples, consisting of 99 eyes from control subjects and 136 eyes from glaucoma patients. Upon such analysis, age (P = 0.010), AL (P = 0.026), SE (P = 0.005), and OCT- and HRT-measured temporal disc tilt (both P values < 0.001) all exhibited statistically significant associations with disc ovality index. In terms of disc torsion, AL (P = 0.004), SE (P < 0.001), and OCT- and HRT-measured vertical disc tilt (P = 0.011, P < 0.001) all exhibited significant associations. The associations between AL and disc ovality index and between AL and disc torsion, evaluated by calculation of Pearson's correlation coefficients (R values), were stronger in glaucomatous eyes than in normal control, and became stronger as the severity of the VF defect (assessed using MD) increased (Fig. 2). 

Figure 2

 

Associations of axial length (AL) with disc ovality index and torsion degree in terms of the severity of visual field defect assessed by mean deviation (MD) in 235 eyes from 136 glaucoma patients and 99 normal controls.

Table 3

 

Associations of Potential Clinical and Biometric Parameters With the Optic Disc Ovality Index and Torsion Degree Using Univariate Analyses in 235 Eyes From 136 Glaucoma Patients and 99 Normal Controls

Variables	Disc Ovality Index		Disc Torsion Degree
	Coefficients (95% CI)	P Value*	Coefficients (95% CI)	P Value*
Age, per y	−0.002 (−0.004 to −0.001)	0.010	−0.062 (−0.180 to 0.056)	0.302
Axial length	0.017 (0.002 to 0.032)	0.026	1.452 (0.476 to 2.427)	0.004
Spherical equivalent	−0.009 (−0.016 to −0.003)	0.005	0.980 (−1.463 to −0.496)	<0.001
Mean deviation	−0.002 (−0.009 to 0.004)	0.530	−0.441 (−0.980 to 0.099)	0.109
Disc tilt on horizontal meridian
HRT-measured	0.010 (0.007 to 0.013)	<0.001	0.208 (−0.097 to 0.513)	0.180
OCT-measured	0.017 (0.011 to 0.022)	<0.001	0.357 (−0.079 to 0.794)	0.108
Disc tilt on vertical meridian
HRT-measured	−0.002 (−0.006 to 0.003)	0.409	0.422 (0.097 to 0.747)	0.011
OCT-measured	0.007 (−0.002 to 0.015)	0.118	1.214 (0.643 to 1.784)	<0.001

*  Linear regression analysis.

Table 4 compares optic disc parameters in glaucomatous eyes with superior and inferior hemifield defects. In eyes with superior hemifield defects, the torsion degree was significantly higher, and the proportion of eyes exhibiting inferior torsion significantly larger, than in eyes with inferior hemifield defects (P = 0.002 and P < 0.001 for degree and direction of torsion, respectively). The OCT- and HRT-measured vertical disc tilts were significantly larger in eyes with superior rather than inferior hemifield defects (P < 0.001, P < 0.030 for HRT and OCT data, respectively), while the OCT- and HRT-measured temporal disc tilts were not significantly different between groups. 

Table 4

 

Comparisons of the ONH Parameters Between Superior and Inferior Hemifield Groups in Glaucoma Patients

Characteristics	Superior Hemifield Defect, n = 77	Inferior Hemifield Defect, n = 59	P Value*
Optic disc parameters using retinal photographs
Disc ovality index	1.17 ± 0.14	1.15 ± 0.11	0.382
Torsion degree, deg	1.53 ± 11.25	−4.23 ± 9.36	0.002
Direction of torsion, % of inferior	46.6	19.0	<0.001
HRT-measured parameters
Disc size, mm2	2.31 ± 0.52	2.16 ± 0.45	0.089
Temporal disc tilt, deg	8.7 ± 7.0	8.8 ± 6.2	0.872
Vertical disc tilt, deg	4.3 ± 5.1	1.0 ± 4.7	<0.001
Cup shape measure	−0.08 ± 0.06	−0.09 ± 0.07	0.482
OCT-measured parameters
Temporal disc tilt, deg	4.25 ± 4.50	4.76 ± 4.21	0.523
Vertical disc tilt, deg	1.12 ± 3.54	−0.11 ± 2.49	0.030

*  Independent t-test.

A comparison of the clinical characteristics and optic disc parameters between eyes with superior versus inferior hemifield defect in central and peripheral VF defect subgroups is shown in Table 5. When eyes in the peripheral VF defect subgroup exhibiting superior and inferior hemifield defects were compared, the degree of torsion, the proportion of eyes exhibiting inferior torsion, and the HRT-measured vertical disc tilt all differed significantly (P = 0.002, P < 0.001, and P < 0.001, respectively). The values of these parameters did not differ significantly between eyes in the central VF defect subgroup with superior and inferior hemifield defects (P = 0.573, P = 0.444, and P = 0.080). 

Table 5

 

Comparisons of Clinical Characteristics and Optic Disc Parameters Between Eyes With Superior Versus Inferior Hemifield Defect in Central and Peripheral Visual Field Defect Subgroups

	Central VF Defect			Peripheral VF Defect
	Superior Central VF Defect, n = 50	Inferior Central VF Defect, n = 25	P Value*	Superior Peripheral VF Defect, n = 27	Inferior Peripheral VF Defect, n = 34	P Value*
Age, y	54.88 ± 13.73	57.00 ± 13.65	0.530	49.00 ± 13.90	56.20 ± 13.44	0.045
Spherical equivalent, D	−1.57 ± 2.41	−1.22 ± 2.19	0.577	−3.10 ± 3.66	−1.66 ± 2.83	0.105
Axial length, mm	24.57 ± 1.81	24.63 ± 1.29	0.898	24.80 ± 1.87	24.64 ± 1.42	0.756
Ovality index	1.15 ± 0.13	1.14 ± 0.13	0.892	1.15 ± 0.19	1.11 ± 0.13	0.392
Torsion degree, deg	1.40 ± 11.28	−0.11 ± 8.94	0.573	1.77 ± 11.40	−7.14 ± 8.63	0.002
Direction of torsion, % of inferior	40.0	28.0	0.444	52.0	8.8	<0.001
HRT-measured disc tilt
Temporal disc tilt, deg	7.31 ± 6.34	8.97 ± 6.94	0.317	11.35 ± 7.55	8.82 ± 5.82	0.152
Vertical disc tilt, deg	3.28 ± 5.24	1.12 ± 3.80	0.080	6.23 ± 4.57	1.01 ± 5.37	<0.001
OCT-measured disc tilt
Temporal disc tilt, deg	4.06 ± 3.35	4.17 ± 4.78	0.911	4.59 ± 6.10	5.17 ± 3.76	0.659
Vertical disc tilt, deg	1.02 ± 3.36	−0.33 ± 3.21	0.115	1.26 ± 3.91	0.03 ± 1.87	0.122

*  Independent t-test.

The HRT-measured vertical disc tilt was an independent factor defining the initial location of a VF defect upon both univariate (P = 0.002) and multivariate analyses (P = 0.012) after controlling for age, MD, AL, disc ovality index, torsion degree, disc size, and HRT-measured temporal disc tilt (Table 6). 

Table 6

 

Associations of Potential Clinical and Biometric Parameters With the Initial Location of VF Defect (Superior Versus Inferior) on Univariate and Multivariate Logistic Regression Analysis

Characteristics	Univariate		Multivariate
	Odds Ratio (CI)	P Value*	Odds Ratio (CI)	P Value*
Age	0.98 (0.95–1.00)	0.085	0.96 (0.91–1.00)	0.046
Mean deviation	1.11 (0.92–1.34)	0.251	1.06 (0.75–1.50)	0.730
Axial length	0.94 (0.82–1.08)	0.391	0.73 (0.49–1.08)	0.732
Disc ovality index	3.38 (0.21–52.39)	0.383	16.50 (0.12–2257.01)	0.264
Disc torsion degree	1.05 (1.01–1.09)	0.002	1.07 (1.01–1.12)	0.012
HRT-measured parameters
Disc size	1.47 (0.71–3.01)	0.298	0.51 (0.15–1.70)	0.271
Temporal disc tilt	1.00 (0.95–1.04)	0.871	0.94 (0.85–1.05)	0.270
Vertical disc tilt	1.13 (1.05–1.22)	<0.001	1.22 (1.08–1.38)	0.001
OCT-measured parameters
Temporal disc tilt	0.97 (0.89–1.05)	0.520
Vertical disc tilt	1.14 (1.00–1.30)	0.025

 The dependent variable was the initial location of the VF defect (superior versus inferior). The independent variables of the multivariate logistic regression analysis were age, mean deviation, axial length, disc ovality index, disc size, HRT-measured horizontal, and vertical tilt angle.

*  Logistic regression analysis.

Discussion

We found that AL-related changes in disc ovality and torsion differed between glaucomatous and normal control eyes. The temporal disc tilt was associated with disc ovality, whereas the vertical disc tilt was related to disc torsion. Furthermore, the vertical disc tilt was associated with the initial location of VF defects in glaucomatous eyes. This association was more prominent in glaucoma patients with peripheral than with central VF defects. Representative examples are shown in Figure 3. 

Figure 3

 

Representative cases showing the impact of vertical disc tilt on the initial location of glaucomatous damage. (A) Images from a 40-year-old male with glaucoma. He had moderate myopia (axial length [AL] = 26.14 mm) with a mean deviation (MD) of −2.73 dB in standard automated perimetry. The ovality index of optic disc was 1.17, and torsion degree was 15.3°. The HRT-measured vertical and temporal disc tilt was 15.33° and 15.35°, indicating the presence of inferior and temporal side disc tilt, respectively. He had retinal nerve fiber layer (RNFL) defect on the inferotemporal side and superior nasal step scotoma. (B) Images from a 52-year-old female with glaucoma with a MD of −3.97 dB in standard automated perimetry. Her AL was 23.17 mm. The ovality index of optic disc was 1.05, and torsion degree was −19.5°. The HRT-measured vertical and horizontal disc tilt was −4.57° and 5.6°, indicating superior and temporal side disc tilt. This patient had RNFL defect on the superotemporal side and inferior nasal step scotoma.

To investigate disc ovality and torsion, we determined the degree of disc tilt using both HRT III and Cirrus HD-OCT to minimize errors associated with the use of either modality alone. The HRT III yields horizontal and vertical height profiles in association with topographic images defined by the horizontal or vertical plane. The HRT, which is a confocal scanning laser ophthalmoscope, is capable of detecting regional variation in the deformation of the optic disc ¹⁵ and known to be less affected by optic disc tilt, compared to OCT technique. ^16,17 We determined the disc tilt between the disc plane and the retinal surface in the horizontal and vertical planes according to Takasaki et al. ⁵ In addition, we evaluated the temporal and vertical tilt using extracted horizontal and vertical B-scan images provided by Cirrus HD-OCT. The disc margin was taken as the end point of the RPE/Bruch's membrane intersection as defined by Reis et al., ¹⁸ whereby the end point of the Bruch's membrane was more consistent in anatomical terms than the clinically noted disc margin. The degrees of temporal and vertical disc tilt obtained were each correlated only moderately between HRT III and Cirrus HD-OCT (r = 613 and r = 0.183, respectively), which may be related to between-system differences in the delineation of the disc margin and the retinal plane. However, in both forms of measurement the degree of temporal disc tilt was significantly associated with the ovality index, which is consistent with previous studies. ^5,6 In addition, we found that the vertical disc tilt was significantly associated with the torsion degree (Table 3). The temporal disc tilt is influenced by the extent of the postnasal expansion of the posterior sclera on the temporal side. ⁸ Because the postnatal expansion of the peripapillary sclera can occur in both the temporal and vertical directions, our study suggests that topographic evaluation of the disc tilt in both directions will help clarify the overall shape of the peripapillary sclera. 

In the present study, the relationship between AL-related changes in disc ovality and torsion differed with the severity of glaucoma as assessed using MD (Fig. 2). The changes were greater in eyes with more advanced glaucoma, whereas minimal or no correlation was evident between the disc tilt and torsion and AL in normal controls. Disc tilt, torsion, and PPA are generally accepted to be associated with myopic presentation of the optic disc. ^19,20 Hosseini et al. ⁶ recently reported that the degree of temporal disc tilt was associated with both the stage of glaucoma (MD-assessed) and the AL. The human posterior sclera supports the posterior pole and is subject to regional variation in mechanical strain. In particular, the peripapillary sclera is subjected to a greater degree of tensile strain than is the adjacent midperipheral sclera. ⁹ Recently, glaucomatous eyes were reported to exhibit different biomechanics in the peripapillary sclera compared to normal eyes. ^9–11 In this regard, greater changes in disc ovality and torsion of glaucomatous compared to normal control eyes in the present study suggest that these optic disc parameters reflect biomechanical changes in the posterior peripapillary region of glaucomatous eyes. 

We demonstrated previously that the disc torsion direction was associated with the location of glaucomatous damage in the eyes of myopic normal-tension glaucoma patients. ⁴ However, in the present study, the degree of vertical disc tilt was independently associated with the initial location of VF defects in glaucoma patients (Table 6; Fig. 3). The association between the direction of vertical disc tilt and the location of VF defect (upward disc tilt with inferior defect and downward tilt with superior defect) suggests that the vertical disc tilt independently affects the superior versus inferior regional susceptibility to glaucomatous damage. It appears that the vertical disc tilt (upward or downward disc tilt) closely reflects the asymmetric (superior versus inferior) postnasal expansion of the posterior sclera, which is associated with superior versus inferior regional susceptibilities to glaucoma. It is hypothesized that exaggerated inferior scleral expansion further stresses the inferior RNFL, whereas superior expansion stresses the superior RNFL. The optic disc tilt and torsion, as features of the myopic optic disc, are not normally prominent in nonmyopic eyes. ^19,20 Therefore, determination of the vertical disc tilt may serve as a quantitative tool for the vertical asymmetry of the peripapillary sclera in glaucomatous eyes that are not severely myopic. Further investigation is needed to elucidate the relationship between the degree of vertical disc tilt and the amount of VF loss. 

Intriguingly, the vertical disc tilt and torsion degree differed between eyes with superior and inferior peripheral VF defects, but not between eyes with central VF defects (Table 5). The risk factors for glaucoma in eyes with central VF defects differ from those associated with peripheral VF defects. ^21,22 Patients with central VF defects had lower maximum values of untreated IOP, a higher frequency of disc hemorrhage, more systemic risk factors (including hypotension, migraine, Raynaud's phenomenon, and sleep apnea), ²¹ and a greater ocular pulse amplitude determined using dynamic contour tonometry ²³ than patients with eyes displaying peripheral VF defects. This suggests that a particular pathogenetic mechanism operating either in isolation or in combination with mechanical factors may play a role in central VF loss in glaucoma. In addition, in eyes with peripheral glaucomatous VF defects, the arcuate nerve fiber bundles may be more susceptible to peripapillary scleral expansion-induced damage than the papillomacular bundles, because the former are located more vertically. 

Our study had several limitations. First, we included glaucomatous eyes with only isolated VF defects. In addition, to increase the validity of the subgroup analysis, we excluded eyes with overlapping central and peripheral VF defects. A number of patients, including advanced glaucoma cases, were excluded from analysis because of these rather strict inclusion criteria. The effects of such strict patient selection on our findings remain unknown. Second, we excluded eyes exhibiting an extremely high myopia of ≥ −10 D, and our patients exhibited mild myopia with a refraction of −1.5 D (IQR; −3.5 to 0.5 D). However, high-myopic discs, some of which were included in this study, are known to be associated with a range of VF defects other than the temporal VF defect. This could act as a confounding factor. The group of glaucoma patients in this study was relatively young. This may be associated with the rapid myopic shift in the young generation in an East Asian population, because myopia is one of the risk factors for glaucoma. ^24,25 Finally, we used a cross-sectional and retrospective study design to investigate the effect of optic disc parameters on the initial location of glaucomatous damage assessed at a single time point. Therefore, the association discovered here may not imply a causal relationship. 

In conclusion, the current study documents the clinical significance of vertical disc tilt. Measurement of vertical disc tilt may give valuable information about the superior versus inferior regional susceptibilities of glaucoma. Topographic evaluation of the disc tilt in both directions will help clarify the overall shape of the peripapillary sclera. Further longitudinal studies will be required to determine the causative relation. 

Acknowledgments

Disclosure: J.A. Choi, None; H-Y.L. Park, None; H-Y. Shin, None; C.K. Park, None