Determinants of Quantitative Optic Nerve Measurements Using Spectral Domain Optical Coherence Tomography in a Population-Based Sample of Non-glaucomatous Subjects

Abstract

Purpose.: To evaluate the ocular and systemic factors influencing measurements of optic nerve head (ONH) parameters and retinal nerve fiber layer (RNFL) thickness with spectral-domain optical coherence tomography (SD-OCT) in healthy Chinese adults.

Methods.: Adults ranging in age from 40 to 80 years were consecutively recruited from the population-based Singapore Chinese Eye Study. A SD-OCT instrument was used to measure ONH and RNFL parameters. A total of 542 eyes from 542 non-glaucomatous Chinese subjects were analyzed. Univariable and multiple linear regression analyses were performed to examine the effects of a range of ocular (e.g., intraocular pressure [IOP], axial length [AL], disc area) and systemic (e.g., blood pressure, serum glucose) factors on ONH and RNFL parameters.

Results.: In multiple regression analyses, neuroretinal rim area was independently associated with age (β = −0.006, P < 0.001), disc area (β = 0.183, P < 0.001), IOP (β = −0.009, P = 0.008), AL (β = −0.023, P = 0.004), and lens nuclear color (β = 0.042, P = 0.001). Vertical cup-to-disc ratio was independently correlated with age (β = 0.003, P < 0.001), disc area (β = 0.207, P < 0.001), IOP (β = 0.004, P = 0.014), AL (β = 0.010, P = 0.008), and lens nuclear color (β = −0.017, P = 0.006). Average RNFL thickness was independently related to age (β = −0.204, P = 0.001), disc area (β = 4.218, P < 0.001), signal strength (β = 1.348, P < 0.001), and AL (β = −1.332, P < 0.001). Disc area had the strongest effect on measurements of ONH parameters and RNFL thickness.

Conclusions.: In a non-glaucomatous population, optic nerve measurements with SD-OCT vary with disc area, age, IOP, AL, lens nuclear color, and signal strength, but systemic parameters have little influence. This information may be useful for interpretation of SD-OCT measurements.

Despite years of research, current tools for glaucoma screening remain limited. ^{1 –5} Optical coherence tomography (OCT) offers cross-sectional quantitative assessment of the retinal nerve fiber layer (RNFL) and optic nerve head (ONH), potentially allowing for detection of early glaucomatous optic neuropathy. ^6,7 Li et al. ⁸ recently evaluated the performance of RNFL measurements using time-domain OCT to screen for glaucoma in high-risk populations, and reported a moderate sensitivity (67%) and a high specificity (96%) for glaucoma diagnosis, when RNFL and ONH parameters were combined. 

Studies have shown that RNFL measurements based on time-domain OCT are influenced by age, sex, ethnicity, axial length, and optic disc size in normal subjects. ^{9 –12} These studies, however, have been limited by inclusion of highly selected clinic- or volunteer-based samples, which are prone to selection bias. Furthermore, there have been no studies investigating whether ONH and RNFL measurements are influenced by the same set of ocular or systemic variables. 

The development of spectral-domain OCT (SD-OCT or high-definition OCT [HD-OCT]) now allows faster scanning speed and even higher image resolution compared with that of time-domain OCT. ^13,14 Knowing the range of ocular and systemic parameters that may affect SD-OCT measurements is crucial before SD-OCT can be considered for glaucoma screening, but there are currently no population-based studies that have evaluated the factors that influence SD-OCT measurements. 

The aim of our study was to examine the influences of age, sex, and a range of ocular and systemic factors on the measurements of ONH and RNFL using SD-OCT (Cirrus HD-OCT; Carl Zeiss Meditec, Inc., Dublin, CA) in non-glaucomatous Chinese subjects from a population-based study. 

Methods

Study Population

Data for this analysis were derived from the Singapore Chinese Eye Study (SCES), a population-based cross-sectional study of eye diseases in Chinese adults, ranging in age from 40 to 80 years, residing in Singapore. In brief, subjects were selected, using an age-stratified (by 10-year age groups) random sampling method, from a computer-generated list provided by the Singapore Ministry of Home Affairs. The methodology and objectives of the study, modeled after the Singapore Malay Eye Study, ¹⁵ have been reported in detail elsewhere. ¹⁶ Written informed consent was obtained from each participant. The study adhered to the Declaration of Helsinki, and ethics committee approval was obtained from the Singapore Eye Research Institute Institutional Review Board. Participants underwent a standardized interview, systemic and ocular examinations, and laboratory investigations. 

Subjects

Participants were consecutively recruited from February 2009 through July 2010. All subjects underwent a full ophthalmic examination, including measurement of logarithmic minimal angle resolution (logMAR), best-corrected visual acuity testing, refraction, intraocular pressure (IOP), gonioscopy, visual field, and fundus examination. Subjects were excluded if logMAR > 0.5, there were evidence of macular disease, previous retinal or refractive surgery, neurologic diseases, or clinical features compatible with a diagnosis of a glaucoma suspect or glaucoma. A glaucoma suspect was defined as having any of the following criteria in the presence of normal visual field: (1) IOP > 21 mm Hg, (2) signs consistent with pseudoexfoliation or pigment dispersion syndrome, (3) narrow angles (posterior trabecular meshwork was seen for <180° during static gonioscopy), and (4) peripheral anterior synechiae or other findings consistent with secondary glaucoma. ¹⁷ Glaucoma was defined based on the presence of visual field defects (as described in the following text), regardless of the structural features, to avoid potential bias in the evaluation of the ocular and systemic factors influencing measurements of ONH and RNFL parameters with SD-OCT. 

Visual Field Examination

Standardized visual field testing was performed with static automated white-on-white threshold perimetry (Swedish interactive threshold algorithm fast 24--2, Humphrey Field Analyzer II; Carl Zeiss Meditec). A visual field was defined as reliable when fixation losses were <20% and false-positive and false-negative rates were <33%. A visual field defect was defined as the presence of three or more significant (P < 0.05) non–edge-contiguous points with at least one at the P < 0.01 level on the same side of the horizontal meridian in the pattern deviation plot, and classified as “outside normal limits” in the Glaucoma Hemifield Test. All subjects included for the final analysis had a reliable and normal visual field (without a visual field defect). 

Imaging

The commercially available SD-OCT instrument (Cirrus HD-OCT; Carl Zeiss Meditec, Inc, Dublin, CA) is an SD-OCT device with a scan speed of 27,000 axial scans per second and an axial resolution of 5 μm. ¹⁸ This SD-OCT instrument images the ONH and peripapillary RNFL in an area of 6 × 6 mm², and samples 200 × 200 data points in <1.5 seconds. With the latest software version 5.0, a series of ONH parameters including disc area, neuroretinal rim area, cup volume, average cup-to-disc ratio, and vertical cup-to-disc ratio, in addition to RNFL thicknesses (average, clock hours, and quadrants), are derived automatically from a single scan. The details of the ONH measurements have been previously described, and good accuracy of the new ONH parameters to differentiate normal from glaucoma eyes has recently been demonstrated by Mwanza et al. ¹⁹  

After pupil dilation using tropicamide 1% and phenylephrine hydrochloride 2.5%, ONH and RNFL scan acquisitions were performed for each participant using an optic disc cube 200 × 200 scan protocol, which generates a cube of data in a 6 × 6 mm² grid with 200 × 200 axial measurements, following the recommended procedure (Cirrus HD-OCT manual). ²⁰ In brief, the subject's pupil was first centered and focused in an iris viewing camera on acquire screen, and the line-scanning ophthalmoscope (LSO) with “auto focus” mode was then used to optimize the view of the retina. The “center” and “enhance” modes were used to optimize the Z-offset and scan polarization respectively for the OCT scan to maximize the OCT signal. After each capture, motion artifact was checked with the LSO image with the OCT en face overlaid. Rescanning was performed if a motion artifact (indicated by discontinuity of blood vessels) or saccades through the calculation circle (3.46 mm diameter around the ONH) were detected. The OCT scans were excluded if there was the presence of RNFL or ONH algorithm segmentation failure. All the OCT scans included in the study had signal strength of at least 6, which is considered as acceptable quality. ²⁰  

We randomly selected one eye from each participant for final analysis because RNFL and ONH parameters are highly correlated between eyes in non-glaucomatous persons. ^21,22  

Measurement of Other Ocular Factors

IOP was measured with a Goldmann applanation tonometer (GAT; Haag-Streit, Bern, Switzerland) before pupil dilation. The static refraction of each eye was measured using an autorefractor (Canon RK 5 Auto Ref-Keratometer; Canon Inc. Ltd., Tochigiken, Japan). Central corneal thickness (CCT) was measured with an ultrasound pachymeter (Advent; Mentor O & O, Norwell, MA), the mean of five measurements was used in the analysis. Axial length (AL) and anterior chamber depth (ACD) were measured with noncontact partial coherence laser interferometry (IOL Master V3.01, Carl Zeiss Meditec AG, Jena, Germany), and the mean of five measurements was used in the analysis. Lens opacity was assessed by two trained ophthalmologists (RW and YZ) during the visit using Lens Opacities Classification System (LOCS) III ²³ with a slit-lamp microscope (Haag-Streit model BQ-900) in comparison with standard photographic slides for nuclear opalescence (NO), nuclear color (NC), and cortical and posterior subcapsular (PSC) cataract. 

Measurement of Systemic Factors

Systolic and diastolic blood pressures were measured using a digital automatic blood pressure monitor (Dinamap model Pro Series DP110X-RW, 100V2; GE Medical Systems Information Technologies, Inc., Milwaukee, WI), after subjects were seated for at least 5 minutes. Body mass index (BMI) was calculated as body weight (in kilograms) divided by body height (in meters) squared. Current smokers were defined as those currently smoking any number of cigarettes (i.e., current versus past/never). The number of packs smoked per week was recorded. Nonfasting venous blood samples were analyzed at the National University Hospital Reference Laboratory for biochemical testing of serum total cholesterol, high-density lipoprotein (HDL) cholesterol, low-density lipoprotein (LDL) cholesterol, triglycerides, glycated hemoglobin (HbA1c), and glucose. 

Statistical Analysis

Statistical analysis was performed using commercial analytic software (SPSS version 17.0; SPSS, Inc., Chicago, IL). Ocular variables (IOP, spherical equivalent, AL, ACD, CCT, and LOCS III score), systemic variables (age, systolic blood pressure, diastolic blood pressure, BMI, smoking, number of packs of cigarette per week, serum glucose, HbA1c, HDL cholesterol, LDL cholesterol, and triglycerides,) and SD-OCT (Cirrus HD-OCT) optic nerve parameters (ONH parameters and RNFL thicknesses) between males and females were compared with independent t-test or χ² test. Univariable and multiple regression analyses were performed to determine ocular and systemic factors (independent variables) associated with SD-OCT (Cirrus HD-OCT) measurements (dependent variables). We constructed two linear regression models in the multiple regression analysis. First, factors significant at P < 0.1 with any optic nerve parameters from univariable analysis and potential confounders (age, sex, and OCT signal strength) were included in Model 1. We then used stepwise regression with a backward selection procedure to include those dependent variables that contributed significantly only at P < 0.1 to the model (Model 2). We compared standardized regression coefficients (sβ), with higher sβ values indicating stronger associations with optic nerve parameters. Because the calculations for rim area and average cup-to-disc ratio involve disc area in SD-OCT (Cirrus HD-OCT), the null hypotheses in the regression models are invalid and may cause mathematical coupling. ²⁴ We thus corrected the linear regression coefficients of disc area in the multiple regression models according to a method suggested by Moreno et al. ²⁵ We assumed that an upper bound on the measurement error in the disc area is 3% and the error is proportional to the measurement. 

Results

We excluded 23 subjects for the following reasons: motion artifact (n = 11), unreliable RNFL (n = 3), optic disc (n = 3), or cup segmentation (n = 6) scans due to algorithm failure. A total of 542 eyes from 542 healthy Chinese subjects were included in the final analysis: 300 (55.4%) subjects were male and 242 (44.6%) were female. The mean (SD) age was 53.0 (6.40) years (range, 44 to 73 years). The mean (SD) optic disc area, rim area, cup volume, average cup-to-disc area ratio, vertical cup-to-disc ratio, and average RNFL thickness were 2.00 mm² (0.39), 1.29 mm² (0.23), 0.19 mm³ (0.16), 0.56 (0.13), 0.51 (0.13), and 97.6 μm (9.1), respectively. Table 1 shows the demographics, ONH, and RNFL measurements. Females were more likely to have a higher IOP, a shorter AL, and a shallower ACD (all P ≤ 0.013). Males were more likely to have a higher diastolic blood pressure, BMI, serum glucose, triglyceride levels, and lower HDL cholesterol level (all P ≤ 0.002), and were current smokers (P < 0.001), although no significant difference in number of packs of cigarettes per week between males and females among the smokers (P = 0.061). Males were likely to have a larger cup volume, average, and vertical cup-to-disc ratio (all P ≤ 0.019). There was no significant difference in RNFL parameters between males and females. 

Table 1.

 

Demographics and Optic Nerve Parameters with Respect to Sex

Factor	All (n = 542)		Male (n = 300)		Female (n = 242)		P Value*
	Mean	SD	Mean	SD	Mean	SD
Age, y	52.97	6.42	53.28	6.53	52.59	6.27	0.217
Intraocular pressure, mm Hg	14.51	2.78	14.24	2.85	14.83	2.66	0.013
Spherical equivalent, D	−0.93	2.42	−0.86	2.44	−1.01	2.40	0.464
Axial length, mm	24.10	1.20	24.36	1.21	23.78	1.11	<0.001
Anterior chamber depth, mm	3.32	0.34	3.37	0.35	3.26	0.32	<0.001
Central corneal thickness, μm	555.10	33.70	554.20	34.10	556.10	33.20	0.522
LOCS III nuclear opalescence	1.71	0.79	1.72	0.79	1.70	0.80	0.692
LOCS III nuclear color	1.83	0.77	1.83	0.77	1.83	0.77	0.990
LOCS III cortical	0.76	0.94	0.75	0.93	0.78	0.94	0.647
LOCS III PSC	0.24	0.40	0.25	0.45	0.22	0.32	0.287
Systolic blood pressure, mm Hg	131.03	16.34	132.01	14.68	129.82	18.15	0.122
Diastolic blood pressure, mm Hg	78.57	9.69	81.45	8.73	75.01	9.66	<0.001
Body mass index, kg/m2	23.36	3.36	23.76	3.12	22.87	3.59	0.002
Serum glucose, mmol/L	6.04	2.20	6.37	2.65	5.63	1.34	<0.001
HbA1c, %	5.93	0.75	5.95	0.87	5.92	0.57	0.643
HDL cholesterol, mmol/L	1.31	0.37	1.17	0.33	1.47	0.36	<0.001
LDL cholesterol, mmol/L	3.41	0.88	3.43	0.89	3.37	0.86	0.396
Triglycerides, mmol/L	1.88	1.36	2.20	1.43	1.49	1.17	<0.001
Current smoking, %	14.2%		24.7%		1.24%		<0.001†
Number of packs of cigarettes, per wk	4.66	3.06	4.81	3.06	1.88	0.63	0.061
Optic nerve head parameters
Disc area, mm2	2.03	0.39	2.02	0.41	1.96	0.37	0.079
Rim area, mm2	1.29	0.23	1.28	0.23	1.30	0.22	0.326
Cup volume, mm3	0.19	0.16	0.21	0.18	0.17	0.13	0.001
Average cup-to-disc ratio	0.56	0.13	0.57	0.12	0.54	0.13	0.019
Vertical cup-to-disc ratio	0.51	0.13	0.52	0.13	0.49	0.13	0.014
Retinal nerve fiber layer thickness, μm
Superior	122.97	15.87	122.84	16.43	123.13	15.19	0.833
Nasal	69.16	10.79	69.84	10.04	68.32	11.63	0.102
Inferior	126.77	16.16	125.95	16.59	127.78	15.59	0.190
Temporal	71.57	11.15	70.90	10.42	72.40	11.95	0.120
Average	97.62	9.10	97.38	9.38	97.91	8.77	0.506

  P values are for comparison between males and females.

*  Independent t-test.

†  Chi-squared test.

Table 2 shows the univariable analyses between ocular variables (IOP, spherical equivalent, AL, ACD, CCT, and LOCS III score) with ONH parameters and average RNFL thickness. Both ONH parameters and average RNFL thickness were correlated with spherical equivalent and AL (all P ≤ 0.020). Disc area (r = −0.200, P < 0.001), rim area (r = −0.167, P < 0.001), and average RNFL thickness (r = −0.155, P < 0.001) were significantly correlated with ACD. Disc area (r = −0.087, P = 0.043) and rim area (r = −0.102, P = 0.018) were significantly related to IOP. Rim area (r = −0.088, P = 0.042) and average RNFL (r = −0.100, P = 0.021) were significantly related to LOCS III cortical score; however, the magnitudes of the correlations were weak. There were no significant correlations between optic nerve parameters and CCT (all P > 0.1). 

Table 2.

 

Univariable Analysis between Factors with Optic Nerve Head Parameters and Average Retinal Nerve Fiber Layer (RNFL) Thickness

Factor	Disc Area (mm2)	Rim Area (mm2)	Cup Volume (mm3)	Average Cup-to-Disc Ratio	Vertical Cup-to-Disc Ratio	Average RNFL Thickness (μm)
Intraocular pressure, mm Hg	−0.087 (P = 0.043)	−0.102 (P = 0.018)	0.009 (P = 0.838)	0.004 (P = 0.923)	0.007 (P = 0.870)	−0.054 (P = 0.206)
Spherical equivalent, D	0.384 (P < 0.001)	0.151 (P = 0.001)	0.198 (P < 0.001)	0.269 (P < 0.001)	0.246 (P < 0.001)	0.229 (P < 0.001)
Axial length, mm	−0.310 (P < 0.001)	−0.212 (P < 0.001)	−0.115 (P = 0.007)	−0.141 (P = 0.001)	−0.100 (P = 0.020)	−0.255 (P < 0.001)
Anterior chamber depth, mm	−0.200 (P < 0.001)	−0.167 (P < 0.001)	−0.080 (P = 0.064)	−0.070 (P = 0.104)	−0.075 (P = 0.082)	−0.155 (P < 0.001)
Central corneal thickness, μm	−0.026 (P = 0.541)	0.040 (P = 0.355)	−0.047 (P = 0.270)	−0.046 (P = 0.286)	−0.061 (P = 0.158)	0.015 (P = 0.729)
LOCS III nuclear opalescence	0.007 (P = 0.876)	0.062 (P = 0.152)	−0.049 (P = 0.254)	−0.038 (P = 0.383)	−0.024 (P = 0.581)	0.033 (P = 0.450)
LOCS III nuclear color	0.008 (P = 0.856)	0.088 (P = 0.042)	−0.078 (P = 0.073)	−0.055 (P = 0.202)	−0.041 (P = 0.341)	0.016 (P = 0.713)
LOCS III cortical	0.004 (P = 0.928)	−0.088 (P = 0.042)	0.001 (P = 0.976)	0.079 (P = 0.068)	0.084 (P = 0.052)	−0.100 (P = 0.021)
LOCS III PSC	0.005 (P = 0.912)	0.032 (P = 0.462)	−0.021 (P = 0.634)	−0.019 (P = 0.663)	−0.011 (P = 0.808)	−0.014 (P = 0.750)

  P values in bold indicate statistical significance.

Associations of systemic variables (age, systolic blood pressure, diastolic blood pressure, BMI, smoking, number of packs of cigarette per week, serum glucose, HbA1c, HDL cholesterol, LDL cholesterol, and triglycerides) and SD-OCT parameters were also examined by univariable analyses. In general, ONH parameters and average RNFL thickness were not significantly associated with systemic variables (all P > 0.05), except for age, smoking, serum glucose, and HbA1c levels. Rim area (r = −0.115, P = 0.007), average cup-to-disc ratio (r = 0.125, P = 0.004), vertical cup-to-disc ratio (r = 0.154, P < 0.001), and average RNFL thickness (r = −0.154, P < 0.001) were significantly correlated with age. Disc area was significantly associated with smoking (r = 0.088, P = 0.041), serum glucose (r = 0.118, P = 0.006), and HbA1c (r = 0.107, P = 0.014) levels (data not shown). 

Table 3 summarizes the multiple linear regression analysis of the effects of age, sex, signal strength, disc area, IOP, AL, LOCS III nuclear color (NC) score, LOCS III cortical score, smoking, and serum glucose on (1) rim area, (2) vertical cup-to-disc ratio, and (3) average RNFL thickness. Rim area was independently associated with age (β = −0.006, P < 0.001), disc area (β = 0.183, P ≤ 0.001), IOP (β = −0.009, P = 0.008), AL (β = −0.023, P = 0.004), and LOCS III NC score (β = 0.042, P = 0.001). Vertical cup-to-disc ratio was independently correlated with age (β = 0.003, P < 0.001), disc area (β = 0.207, P < 0.001), IOP (β = 0.004, P = 0.014), AL (β = 0.010, P = 0.008), and LOCS III NC score (β = −0.017, P = 0.006). Average RNFL thickness was independently related to age (β = −0.204, P = 0.001), signal strength (β = 1.348, P < 0.001), disc area (β = 4.218, P < 0.001), and AL (β = −1.332, P < 0.001). Cup volume was independently correlated with sex (β = −0.033, sβ = −0.100, P = 0.004), OCT signal strength (β = −0.011, sβ = −0.068, P = 0.050), disc area (β = 0.254, sβ = 0.608, P < 0.001), IOP (β = 0.006, sβ = 0.095, P = 0.007), and LOCS III NC score (β = −0.023, sβ = −0.107, P = 0.002), whereas average cup-to-disc area ratio was associated with age (β = 0.002, sβ = 0.119, P = 0.001), disc area (β = 0.199, sβ = 0.611, P < 0.001), IOP (β = 0.004, sβ = 0.084, P = 0.017), and LOCS III NC score (β = −0.018, sβ = −0.110, P = 0.002) (data not shown). The strongest association with all optic nerve parameters was optic disc area (had highest standardized β values). 

Table 3.

 

Multiple Linear Regression Analyses of (A) Rim Area, (B) Vertical Cup-to-Disc Ratio, and (C) Average Retinal Nerve Fiber Layer (Dependent Variables) with Age, Sex, Signal Strength, Disc Area, Intraocular Pressure, Axial Length, LOCS III Nuclear Color Score, LOCS III Cortical Score, Smoking, and Serum Glucose (Independent Variables)

A. Rim Area (mm2)
Variable	Model 1				Model 2
	β	(95% CI)	Standardized β	P Value	β	(95% CI)	Standardized β	P Value
Age, y	−0.004	(−0.008 to −0.001)	−0.123	0.008	−0.006	(−0.009 to −0.003)	−0.162	<0.001
Sex*	0.020	(−0.024 to 0.063)	0.043	0.372	—	—	—	—
Signal strength	0.013	(−0.007 to 0.032)	0.056	0.197	—	—	—	—
Disc area, mm2	0.189	(0.140 to 0.238)	0.323	<0.001	0.183†	(0.136 to 0.230)	0.313†	<0.001
Intraocular pressure, mm Hg	−0.008	(−0.015 to −0.002)	−0.102	0.015	−0.009	(−0.015 to −0.002)	−0.108	0.008
Axial length, mm	−0.016	(−0.033 to 0.0004)	−0.086	0.056	−0.023	(−0.038 to −0.007)	−0.121	0.004
LOCS III nuclear color score	0.042	(0.016 to 0.067)	0.139	0.001	0.042	(0.018 to 0.067)	0.144	0.001
LOCS III cortical score	−0.009	(−0.031 to 0.012)	−0.038	0.388	—	—	—	—
Smoking‡	0.008	(−0.037 to 0.054)	0.016	0.719	—	—	—	—
Serum glucose, mmol/L	0.003	(−0.006 to 0.011)	0.025	0.547	—	—	—	—
Adjusted R 2	0.168				0.169

B. Vertical Cup-to-Disc Ratio
Variable	Model 1				Model 2
	β	(95% CI)	Standardized β	P Value	β	(95% CI)	Standardized β	P Value
Age, y	0.003	(0.001 to 0.004)	0.128	0.001	0.003	(0.002 to 0.005)	0.154	<0.001
Sex*	−0.008	(−0.029 to 0.014)	−0.029	0.487	—	—	—	—
Signal strength	−0.002	(−0.012 to 0.007)	−0.017	0.661	—	—	—	—
Disc area, mm2	0.204	(0.180 to 0.228)	0.617	<0.001	0.207	(0.183 to 0.231)	0.624	<0.001
Intraocular pressure, mm Hg	0.004	(0.001 to 0.008)	0.094	0.010	0.004	(0.001 to 0.007)	0.086	0.014
Axial length, mm	0.008	(−0.0001 to 0.017)	0.076	0.053	0.010	(0.003 to 0.018)	0.095	0.008
LOCS III nuclear color score	−0.016	(−0.029 to −0.004)	−0.096	0.010	−0.017	(−0.029 to −0.005)	−0.100	0.006
LOCS III cortical score	0.005	(−0.005 to 0.016)	0.039	0.314	—	—	—	—
Smoking‡	−0.004	(−0.026 to 0.019)	−0.013	0.734	—	—	—	—
Serum glucose, mmol/L	−0.001	(−0.005 to 0.003)	−0.017	0.641	—	—	—	—
Adjusted R 2	0.371				0.379

C. Average RNFL Thickness (μm)
Variable	Model 1				Model 2
	β	(95% CI)	Standardized β	P Value	β	(95% CI)	Standardized β	P Value
Age, y	−0.221	(−0.354 to −0.089)	−0.152	0.001	−0.204	(−0.319 to −0.088)	−0.144	0.001
Sex*	−0.137	(−1.885 to 1.612)	−0.007	0.878	—	—	—	—
Signal strength	1.370	(0.590 to 2.150)	0.152	0.001	1.348	(0.592 to 2.104)	0.149	<0.001
Disc area, mm2	4.633	(2.649 to 6.617)	0.198	<0.001	4.218	(2.314 to 6.122)	0.183	<0.001
Intraocular pressure, mm Hg	−0.164	(−0.436 to 0.108)	−0.050	0.238	—	—	—	—
Axial length, mm)	−1.310	(−1.993 to −0.628)	−0.172	<0.001	−1.332	(−1.977 to −0.686)	−0.175	<0.001
LOCS III nuclear color score	0.702	(−0.317 to 1.721)	0.059	0.177	—	—	—	—
LOCS III cortical score	−0.250	(−1.116 to 0.6162)	−0.025	0.570	—	—	—	—
Smoking‡	−0.414	(−2.251 to 1.423)	−0.020	0.658	—	—	—	—
Serum glucose, mmol/L	−0.061	(−0.398 to 0.276)	−0.015	0.722	—	—	—	—
Adjusted R 2	0.151				0.143

 Model 1 includes age, sex, OCT signal strength, and the factors significant at P < 0.1 with any optic nerve parameters from univariable analysis. Model 2 includes the factors that contributed significantly only at P < 0.1 after stepwise regression with a backward selection procedure. CI, confidence interval.

*  Male as the reference group.

†  Linear regression coefficients are corrected. ²⁵

‡  Current smoking as the reference group.

Discussion

We report the ocular and systemic determinants of ONH and RNFL parameters measured by SD-OCT in a population-based sample of Chinese subjects. We showed that optic nerve parameters were significantly and independently associated with age, sex, signal strength, disc area, AL, IOP, and lens nuclear color. In these non-glaucomatous eyes, disc area had the strongest effect on ONH and average RNFL thickness measurements. There were no relationships between most systemic factors, ONH parameters, and average RNFL thickness. 

Our results provide the first population-based data on optic nerve measurements obtained with the latest analysis software (Cirrus HD-OCT, version 5.0) in non-glaucomatous subjects selected by random sampling. In the multiple regression analysis, we showed disc area was independently associated with all the optic nerve measurements. A smaller disc area was related to a thinner neuroretinal rim, a smaller cup-to-disc ratio, a thinner RNFL thickness, a smaller cup volume, and a smaller average cup-to-disc ratio. An older age was independently associated with a thinner neuroretinal rim area, a larger vertical cup-to-disc ratio, a thinner average RNFL thickness, and a larger average cup-to-disc ratio. A longer eyeball was independently associated with a thinner neurioretinal rim area, a larger vertical cup-to-disc ratio, a thinner average RNFL thickness, and a larger average cup-to-disc ratio. Elevated IOP was related to a thinner neuroretinal rim area, larger vertical cup-to-disc ratio, larger cup volume, and larger average cup-to-disc ratio. A lower OCT signal strength was associated with a thinner RNFL thickness and larger cup volume. Increased lens nuclear color score was related to a thicker neuroretinal rim area, smaller vertical cup-to-disc ratio, smaller cup volume, and smaller average cup-to-disc ratio. CCT was not related to any optic nerve parameters. It is noted that the magnitudes of the correlations between ocular and systemic factors with optic nerve parameters were weak despite being statistically significant. Our findings were consistent with previous studies conducted in clinic- or volunteer-based samples largely measured by the time-domain OCT device, adding to the weight of evidence regarding the effects of age, ocular factors, and OCT signal strength on quantitative optic nerve measurements. ^{9 –12,26,27}  

An important parameter that had the strongest influence on ONH and RNFL measurements was the optic disc size (Table 3), an effect independent of age, sex, signal strength, IOP, AL, cataract, smoking, and serum glucose level. The disc size showed variability in this non-glaucomatous Chinese population, with a range of 0.89 to 3.49 mm². Previous findings also demonstrate the effect of disc size on neuroretinal rim and RNFL measurements, ^{9 –11} consistent with our findings. In contrast, a recent study ²⁸ and a previous histologic study ²⁹ did not find any relationship between disc size and RNFL thickness. Detection of abnormal ONH and RNFL measurements often requires comparisons with normative reference values. In fact, all commercially available OCT devices have acquired normative data sets collected from normal healthy subjects. The cutoffs for abnormalities are typically the bottom fifth percentiles of the normal distribution for a classification of “borderline,” and the first percentile for a classification of “outside normal limit.” In the current analysis software (Cirrus HD-OCT, version 5), individual measurements of each patient are compared with the reference values from a normative database adjusted only for age. ²⁰ Disc area and other factors (e.g., AL, signal strength, and ethnicity) have not been taken into consideration. In addition, it has been reported that disc size affects the diagnostic accuracy of OCT for glaucoma detection. ^30,31 The strong effect of disc area on ONH and RNFL measurements suggests that optic disc size should also be considered in the assessment of such measurements and for comparisons with the normative database. 

Except for age, systemic variables did not influence rim area, cup volume, cup-to-disc ratio, and RNFL measurements with SD-OCT in this study cohort. Our findings were consistent with some previous studies regarding the insignificant associations between systemic variables (e.g., blood pressure, lipids) and optic nerve in non-glaucoma persons. ^{32 –34} Although some studies have reported that optic nerve parameters are influenced by systemic variables such as BMI and diastolic blood pressure, ^{35 –38} we have been unable to reproduce these results. Some of these studies were conducted in non-Chinese persons, and the optic nerve was measured by confocal scanning laser ophthalmoscopy or retinal fundus photographs manually in those studies. More studies are therefore needed to further confirm whether there is a link between systemic variables and optic nerve parameters. Moreover, it is noted that some results of the study should be seen as exploratory ones, and confirmatory studies are necessary to evaluate the findings that were considered statistically significant. 

Glaucoma screening in the general population has been challenging, partly due to the lack of simple and cost-effective screening methods. ^{1 –5} Using time-domain OCT, Li et al. ⁸ reported that a moderate sensitivity and a high specificity for glaucoma diagnosis can be obtained when RNFL and ONH parameters are combined, suggesting that OCT may be useful for glaucoma screening. The present study provides insights into the determinants of optic nerve measurements in a normal population, which are crucial to know before SD-OCT devices can be used for glaucoma screening in the general population. Further studies to evaluate the use of SD-OCT devices for glaucoma screening and the influence of the above-mentioned determinants on the diagnostic performance for glaucoma need to be undertaken. 

The strengths of this study include its large unselected population-based sample, standardized assessment of systemic and ocular factors, and laboratory investigations. Our study had some limitations. First, optic nerve measurements with SD-OCT do not correct for ocular magnification, which may affect the accuracy of ONH and RNFL measurements. Second, optic nerve measurements were obtained from subjects ranging in age from 40 to 80 years. Their associations with sex and ocular factors may not be extrapolated to other age groups. Third, similar to other studies, the causal relationships between optic nerve and the factors are unclear, due to the cross-sectional nature of our data. Fourth, only ethnic Chinese were examined in this study and the findings may vary in other ethnic groups because ethnicity is also a factor that influences optic nerve measurements. ^9,10 Moreover, the measurements of serum glucose and lipids were not from fasting venous samples. Furthermore, lens opacity for each participant was assessed subjectively during the clinical evaluation by a single observer. Finally, there may be residual confounding factors that we have not controlled for, and that could have biased or modified the associations observed in our sample. 

In summary, a variety of ocular factors such as disc area, AL, and signal strength influence SD-OCT optic nerve measurements in non-glaucomatous Chinese subjects. In contrast, few systemic parameters other than age and sex influence these measurements. 

Footnotes

 Supported in part by the National Medical Research Council and the National Research Foundation, Singapore.

Footnotes

 Disclosure: C.Y. Cheung, None; D. Chen, None; T.Y. Wong, None; Y.C. Tham, None; R. Wu, None; Y. Zheng, None; C.Y. Cheng, None; S.M. Saw, None; M. Baskaran, None; C.K. Leung, Carl Zeiss Meditec (F, C, R), Optovue (F, C, R), Heidelberg Engineering (F, C, R); T. Aung, Carl Zeiss Meditec (F, C, R)