Age-Related Change in Retinal Nerve Fiber Layer Thickness Measured With Spectral Domain Optical Coherence Tomography

Abstract

Purpose.: We investigated age-related change in peripapillary retinal nerve fiber layer thicknesses (RNFLT) measured with spectral domain optical coherence tomography (SD-OCT) in healthy individuals.

Methods.: In this cross-sectional study, peripapillary RNFL measurement was performed using a Cirrus SD-OCT device. Linear regression analysis and Spearman's correlation analysis were performed to investigate the level of difference in age-related change in the overall mean, 4 quadrants mean, and 12 clock-hour sectors mean RNFLT.

Results.: Mean age of the 121 healthy participants was 39 years (range, 20–59 years). Mean 4 quadrant and mean 12 clock-hour sector RNFLT in the 121 randomly selected eyes in 121 healthy participants decreased significantly with age in the overall group (Spearman's correlation analysis, P < 0.05). Overall mean RNFLT decreased by 0.365 μm (95% confidence interval [CI], 0.47–0.26; linear regression analysis, P < 0.001) for every one year increase in age. Overall, the greatest decrease in mean RNFLT was in the lower quadrant (0.575 μm/y; 95% CI, 0.733–0.416), versus the least decrease in the nasal quadrant (0.141 μm/y; 95% CI, 0.272–0.010), and the greatest decrease in mean RNFLT was in the 6 clock-hour sector (0.656 μm/y; 95% CI, 0.939–0.374), versus the least decrease in the 3 clock-hour sector (0.119 μm/y; 95% CI, 0.266–0.029).

Conclusions.: Age-related decrease in RNFLT varied according to sector. Age-related change should be accounted for in any assessment of RNFLT. Regional age-related change is accounted for by Cirrus SD-OCT in its normative database.

Introduction

Glaucoma is an optic neuropathy characterized by slow progression and apoptosis of the retinal ganglion cells (RGCs), resulting in irreversible thinning of the retinal nerve fiber layer (RNFL), and then visual field (VF) abnormalities. ^1–4 Previous studies support the notion that damage to the RNFL often can be identified before alterations in the VF are detected. ^5–9 Age-related change in RNFL thickness (RNFLT) should be taken into consideration before concluding that loss of RNFLT is due to glaucoma. ^3,10  

Many studies on RNFL change with age have been reported. ^1,3,11–39 Various cross-sectional histologic studies on the retina and optic nerve have shown that the number of RGCs can decrease with age in healthy individuals. ^11–16 The RNFL, as an axonal extension of RGCs, can be observed using various digital imaging techniques, including optical coherence tomography (OCT) and scanning laser polarimetry. ^17–29 While many studies have reported an age-related decrease in RNFLT, ^{11–15,17–25,28,29} others ^16,26,27 do not support this finding. Various studies have shown that peripapillary RNFLT measured with time domain OCT (TD-OCT) is reliable and reproducible. ^30–35 The consistency and reliability of RNFLT measurement is increasing with the advent of spectral domain OCT (SD-OCT), with higher scanning speed and resolution. ³⁶  

The aim of our study was to investigate age-related change in peripapillary RNFLT measured with the Cirrus SD-OCT device in healthy volunteers without glaucoma. 

Materials and Methods

Study Participants

This study followed the tenets of the Declaration of Helsinki. Informed consent was obtained from the healthy participants after they were provided information about the nature and possible consequences of the study. The study protocol was approved by the Erciyes University Institutional Review Board (IRB). Each participant underwent a full ophthalmic examination at Nigde State Hospital, Ophthalmology Clinic, as follows: visual acuity and refractive error examination, noncontact IOP measurement (NT-2000 noncontact tonometer; Nidek Co., Ltd., Tokyo, Japan), VF testing using the 30-2 SITA algorithm (Humphrey Field Analyzer II; Carl Zeiss Meditec, Inc., Dublin, CA), and anterior segment and posterior segment examinations, including the optic nerve head (ONH) with a 90-diopter (D) lens. Then, peripapillary RNFLT was measured using a Cirrus SD-OCT device (Carl Zeiss Meditec, Inc.). 

Inclusion criteria were a visual acuity of 20/25 or better without glasses or with a spherical equivalent between −1 and +1; IOP of ≤22 mm Hg; normal VF test findings (no associated patterns of glaucomatous VF defects, such as paracentral scotoma; partial arcuate, arcuate, or Bjerrum scotoma; nasal step, altitudinal defect, and temporal wedge, as well as test reliability indices values, such as false positive; false negatives; and fixation loss less than 15%); no glaucomatous ONH change (ONH abnormality was based on the presence of an optic rim notch, diffuse/generalized loss of optic rim tissue, vertical cup/disk diameter ratio asymmetry unexplained by side differences in optic disk size, and disk hemorrhage ⁴⁰ ); no structural optic nerve abnormalities (tilted disk, and so forth); no retinal pathology; negative history of previous eye disease, eye trauma, and eye surgery; and negative history of neurologic or systemic disease (i.e., hypertension, diabetes). 

The RNFL measurements were made without dilating the pupils and using the 200 × 200 optical disc cube Cirrus HD-OCT protocol. The ONH was brought to the center of the simultaneously scanned image and a 200 × 200-pixel resolution axial scan was made over an area of 6 × 6 mm, including the ONH and its surroundings. Images with a signal power ≥7 were used for analysis. In all 4 groups overall mean, 4 quadrants mean, and 12 clock-hour sectors mean RNFLT and ONH thickness were recorded for each participant using the optic disc cube 200 × 200 glaucoma OU analysis mode in Cirrus HD-OCT software v.4.5.1.11 (Carl Zeiss Meditec, Inc.). The participants were divided into 4 age groups: 20 to 29, 30 to 39, 40 to 49, and 50 to 59 years. In each group the 4 quadrants mean and 12 clock-hour sectors mean RNFLT and overall mean RNFLT were calculated. 

Statistical Analysis

Data were analyzed using SPSS v.16.0 for Windows (SPSS, Inc., Chicago, IL). The Kolmogorov-Smirnov test was used to determine 4 quadrants and 12 clock-hour sectors RNFLT normal distribution overall and for the 4 groups (P > 0.05). These measurements were normally, but not homogeneously, distributed; therefore, Spearman's correlation analysis (a nonparametric test) was used to determine the level of significance of the relationship between age and RNFLT. Linear regression analysis was used to calculate the mean decrease in RNFL associated with age (absolute slope) in the 4 quadrants and 12 clock-hour sectors. Annual rates of change in RNFLT are shown as percentage/y. The annual rate of decrease (percentage/y) was calculated in 2 stages. Firstly, the differences in quadrant/sector values between each of the younger (20–29) and older (50–59) age groups were divided by the quadrant/sector values of each of the younger (20–29) age groups. Secondly, the results acquired from the first stage calculations were divided by 3 decades to obtain the annual rate of decrease. The association between increasing age, and overall mean, 4 quadrants, and 12 clock-hour sectors mean RNFLT change was evaluated statistically at the 95% confidence interval (CI), P < 0.05 significance level. 

Results

The study included 121 randomly selected eyes in 121 healthy volunteers (72 female and 49 male), with a mean age of 38.89 ± 11.19 years (range, 20–59 years). Lower quadrant RNFL mean thickness was highest (129.32 μm), followed by upper (119.21 μm), nasal (74.79 μm), and temporal (64.72 μm) quadrant mean RNFLT. Overall mean, 4 quadrants mean, and 12 clock-hour sectors mean RNFLT values are shown in Table 1. Participants were divided into four age groups as follows: 20 to 29, 30 to 39, 40 to 49, and 50 to 59 years. Mean RNFLT was highest in the 20 to 29-year-old age group (103.44 μm), versus 92.21 μm in the 50 to 59-year-old age group. Mean RNFLT values for other age groups are shown in Table 1. In the overall group, the decrease in average RNFLT was 0.365 μm for every 1 year increase in age (95% CI, 0.47–0.26; linear regression analysis, P < 0.001). This decrease also was significant (Spearman's correlation analysis, P < 0.05). The age-related decrease in mean RNFLT is shown in Figure 1. The most significant decrease was observed in the lower quadrants (0.575 μm/y; 95% CI, 0.733–0.416; linear regression analysis, P < 0.001), followed by the upper quadrants (0.488 μm/y; 95% CI, 0.646–0.330; linear regression analysis, P < 0.001), temporal quadrants (0.253 μm/y; 95% CI, 0.350–0.156; linear regression analysis, P < 0.001), and nasal quadrants (0.141 μm/y; 95% CI, 0.272–0.010; linear regression analysis, P = 0.035). Age-related RNFLT changes in the 4 quadrants are shown in Figure 2. 

Figure 1

 

Scatterplot showing overall group's mean RNFLT decrease with aging.

Figure 2

 

Scatterplot showing four different quadrant's mean RNFLT decrease with aging.

Table 1

 

Mean ± SD RNFLT in the 4 Quadrants and 12 Clock-Hour Sectors in Each Age Group and in Overall Group

	Overall	20–29 y	30–39 y	40–49 y	50–59 y
N of eyes	121	32	31	30	28
Mean RNFL	97.01 ± 7.42	103.44 ± 7.04	97.16 ± 6.75	94.47 ± 5.07	92.21 ± 5.45
4 quadrants
Upper	119.21 ± 11.18	128.38 ± 12.01	118.35 ± 8.57	116.03 ± 8.69	113.11 ± 8.80
Nasal	74.79 ± 8.24	77.44 ± 7.98	75.13 ± 9.11	72.97 ± 8.25	72.32 ± 7.09
Lower	129.32 ± 11.72	138.75 ± 11.53	129.68 ± 10.66	126.83 ± 7.52	120.82 ± 9.16
Temporal	64.72 ± 6.37	69.06 ± 5.90	65.45 ± 5.40	62.40 ± 6.37	61.43 ± 6.30
Clock-hour
12	121.71 ± 20.16	130.28 ± 23.12	120.45 ± 20.70	116.97 ± 17.10	116.39 ± 16.65
1	107.00 ± 17.21	115.69 ± 20.87	105.94 ± 11.62	103.50 ± 16.82	102.00 ± 15.21
2	94.83 ± 14.09	97.72 ± 12.76	96.45 ± 15.45	89.57 ± 14.56	85.39 ± 12.56
3	60.37 ± 9.22	63.25 ± 9.55	62.42 ± 7.89	59.67 ± 9.47	57.86 ± 9.75
4	69.19 ± 12.21	73.41 ± 12.16	69.55 ± 12.60	66.80 ± 12.16	66.64 ± 11.05
5	107.29 ± 17.44	116.00 ± 15.01	105.97 ± 18.08	103.80 ± 18.81	103.25 ± 15.00
6	142.98 ± 18.94	154.00 ± 20.85	143.16 ± 19.21	140.53 ± 13.31	132.79 ± 15.36
7	137.79 ± 21.85	146.31 ± 23.55	142.32 ± 22.50	134.73 ± 14.80	126.32 ± 20.75
8	64.74 ± 10.50	70.06 ± 10.70	64.65 ± 9.44	61.10 ± 8.43	60.64 ± 11.43
9	51.85 ± 5.50	54.06 ± 5.24	52.74 ± 5.54	49.77 ± 5.07	49.57 ± 5.30
10	77.48 ± 10.27	82.91 ± 10.52	79.16 ± 8.56	76.17 ± 10.46	70.82 ± 7.53
11	129.00 ± 19.76	139.34 ± 22.66	128.68 ± 15.70	127.57 ± 18.99	119.07 ± 15.92

In terms of 12 clock-hour sectors analysis, as observed in the lower segment, the most significant decrease was observed in the 6 clock-hour quadrants (0.656 μm/y; 95% CI, 0.939–0.374; P < 0.001), and the least significant decrease was in the 3 clock-hour quadrants (0.119 μm/y; 95% CI, 0.266–0.029; P = 0.011), as noted in the nasal quadrant. Differences in the age-related decreases observed between these clock-hour sectors were significant (Spearman's correlation analysis, P < 0.05). The RNFLT decreases in the 12 clock-hour sectors are shown in Figure 3. Age-related absolute and normalized decreases in overall mean, 4 quadrants mean, and 12 clock-hour sectors mean RNFLT are summarized in Table 2. Applying linear regression analysis to the data obtained in this study, normal expected values for the 4 quadrants mean, 12 clock-hour sectors mean, and overall mean RNFLT were calculated by inserting the age variable in the regression equation model. Regression equations for the overall, each quadrant's, and each clock hour sectors are summarized in Table 3. 

Figure 3

 

Scatterplot showing 12 different clock-hour sector's mean RNFLT decrease with aging.

Table 2

 

Absolute Slopes and Annual Rates of Decrease for Mean, 4 Quadrants, and 12 Clock-Hour Sectors RNFLT

	Absolute Slope, μm/y	95% CI	P	Annual Rate of Decrease, %/y
Average	0.365	0.465, 0.264	<0.001	0.36
4 quadrants
Upper	0.488	0.646, 0.330	<0.001	0.40
Nasal	0.141	0.272, 0.010	0.035	0.22
Lower	0.575	0.733, 0.416	<0.001	0.43
Temporal	0.253	0.350, 0.156	<0.001	0.37
Clock-hour
12	0.399	0.717, 0.081	0.014	0.36
1	0.469	0.734, 0.204	0.001	0.39
2	0.137	0.363, 0.090	0.023	0.42
3	0.119	0.266, 0.029	0.011	0.18
4	0.171	0.366, 0.025	0.046	0.31
5	0.442	0.712, 0.171	0.002	0.37
6	0.656	0.939, 0.374	<0.001	0.28
7	0.636	0.970, 0.302	<0.001	0.29
8	0.231	0.395, 0.066	0.006	0.45
9	0.128	0.214, 0.042	0.004	0.28
10	0.403	0.553, 0.254	<0.001	0.49
11	0.597	0.897, 0.296	<0.001	0.48

Table 3

 

Regression Equations for Average, 4 Quadrants, and 12 Clock-Hour Sectors

	Regression Equations, y = a + b × X	P
Average	111.19 − 0.365 × y	<0.001
Upper	138.19 − 0.488 × y	<0.001
Nasal	80.26 − 0.141 × y	0.035
Lower	151.68 − 0.575 × y	<0.001
Temporal	74.57 − 0.253 × y	<0.001
12	137.24 − 0.399 × y	0.014
1	125.23 − 0.469 × y	0.001
2	100.16 − 0.137 × y	0.023
3	64.99 − 0.119 × y	0.011
4	75.82 − 0.171 × y	0.046
5	124.48 − 0.442 × y	0.002
6	168.50 − 0.656 × y	<0.001
7	162.54 − 0.636 × y	<0.001
8	73.71 − 0.231 × y	0.006
9	56.83 − 0.128 × y	0.004
10	93.17 − 0.403 × y	<0.001
11	152.20 − 0.597 × y	<0.001

Discussion

Variation in age-related RNFL loss was reported by cross-sectional histologic studies of the optic nerve. ^11–16 Numerous histologic studies in which an age-related decrease in RNFLT was observed reported that the decrease was 500 to 7000 axons per year with increasing age. ^13–15 Currently, measurement of age-related axonal loss, known as RNFL thinning, has become easier due to the advent of imaging techniques, such as scanning laser polarimetry ^19–22 and OCT. ^{17,18,23–36} The OCT has been reported to be a reliable and reproducible method for measuring RNFLT. ^30–32 Before the end of 2006, the OCT instruments used in clinical practice for measuring RNFLT were based on TD-OCT. ^30–34,41 The RNFLT measured using first-generation OCT (OCT1; Humphrey-Zeiss Systems, Dublin, CA, and OCT-2000; Humphrey Instruments, Inc., San Leandro, CA) in normal individuals averaged 106 to 153 μm. ^{18,32,38,42–45} Based on second-generation OCT (OCTII; Carl Zeiss Meditec, Inc.), mean RNFLT was 105 μm. ⁴⁶ Studies that used third-generation OCT (Stratus OCT; Carl Zeiss Meditec, Inc.) reported RNFLT of 98 to 123 μm. ^{23,24,34,35,44,47–59} The SD-OCT is a newer technology that facilitates unprecedented ultra high–resolution, ultra high–speed RNFL imaging. ^36,51–53 Although some studies have shown that RNFLT values do not differ between TD-OCT and SD-OCT machines, or between different SD-OCT machines, ^52,54 others have shown that SD-OCT devices have been shown to produce RNFL measurements that are significantly thinner than Stratus TD-OCT. ^{17,36,52,55–58} The discrepancy in the measurements could be related to an intrinsic difference between the instruments and software edge-detection algorithms for measuring the RNFL, and the precise location of the RNFL measured. It is likely that SD-OCT, by providing scans at a higher resolution than TD-OCT, also may provide measurements that reflect a more accurate delineation of the RNFL margins. ⁵⁵  

Differences in overall mean RNFLT between our study and earlier studies might be due to ethnic differences between study populations. Studies on normal Chinese eyes, ⁵ Latino eyes, ⁴⁵ and Taiwanese eyes ⁵⁹ reported greater overall mean RNFLT values than noted in Caucasian eyes. ^23,30,34,44 Mean RNFLT was 97.01 μm (based on SD-OCT) in the homogeneously distributed Turkish Caucasian population in our study, and overall mean RNFLT was slightly lower than the 98.1 to 101.5 μm range reported based on TD-OCT. ^23,49,60,61 RNFLT values were reported to be higher in Asians (range, 100–105.8 μm) than in Caucasians (range, 98.1–104.57 μm). ^23,62,63 Hirasawa et al. ¹⁷ reported that mean RNFLT was 102 μm in 251 normal Japanese participants based on SD-OCT. Girkin et al. ⁶⁴ used SD-OCT to investigate racial differences in RNFLT, and reported that those of Hispanic and African descent had had higher overall mean RNFLT than those of Indian, Japanese, and European descent. In our study, the RNFLT of Turkish Caucasian race was thinner than that of African Americans and Hispanics. A German study using SD-OCT reported a mean RNFLT of 97.2 μm, ⁵⁸ which is similar to that observed in our study. In an effort to eliminate the effect of race and ethnicity of RNFL measurements, all participants in our study were Caucasian. To the best of our knowledge, our study is the first from Turkey to use Cirrus SD-OCT to measure 4 quadrant and 12 clock-hour sector RNFLT in Caucasians. Our findings showed that RNFLT decreased significantly with age, according to the 4 quadrant and 12 clock-hour sector-specific values. 

Other factors that possibly might account for the differences in RNFLT in earlier studies are refractive error and axial length. The RNFLT was reported to decrease as the severity of myopia increased, based on TD-OCT ²³ and SD-OCT. ⁵⁸ Leung et al. ⁶⁵ reported that elongation of the globe in myopic eyes leads to mechanical stretching and thinning of the retina, which is associated with thin RNFL values. To minimize the effect of severe myopia and long axial length (>25 mm), all participants in our study were selected based on a spherical equivalent range of −1 to +1 D. 

The overall mean RNFLT decrease with age is in agreement with earlier histologic studies. ³⁷ Although the normal adult optic nerve has approximately 1.0 to 1.2 million RGC axons, the human RNFL loses approximately 5000 axons per year starting from birth. ^7,15 In our study normal mean RNFL thinning was 0.36% per year (annual rate of decrease shown in Table 2). Annual loss of 5000 axons from a total of 1,200,000 axons is equivalent to a loss of 0.42% per year. A normal individual theoretically may lose 0.42% of axons from the optic nerve annually; the annual percentage of decrease in RNFLT in our study was similar. 

The reported decrease in mean RNFLT measured with OCT varies widely, as follows: 0.24 μm/y (95% CI, 0.31–0.18), ³ 0.16 μm/y (95% CI, 0.29–0.02), ²⁴ 0.26 μm/y (95% CI, 0.44–0.07), ²⁵ 0.21 μm/y (95% CI, 0.24–0.17), ²⁹ and 0.44 μm/y. ³⁸ In our study, the decrease in mean RNFLT was 0.365 μm/y (95% CI, 0.465–0.264). The lowest overall mean RNFLT in the study of Parikh et al. ²⁴ from current and previous reported studies could depend on the age distribution of the sample size in their study. Mean age in their study sample was 33.0 ± 19.7 years, which is lower than that in our study (38.89 ± 11.19 years). 

Although the neuroretinal rim ISNT rule is used classically for optic nerve examination, ⁶⁶ in our study mean RNFLT in the four quadrants, based on SD-OCT, adhered to the ISNT rule. As reported previously, ^23,66–68 mean RNFLT in our study also was greatest in the inferior quadrant (129.32 ± 11.72 μm), followed by the superior (119.21 ± 11.18 μm), nasal (74.79 ± 8.24 μm), and temporal (64.72 ± 6.37 μm) quadrants. The RNFL ISNT rule may apply in many normal eyes, ^57,68,69 although eyes that do not follow the RNFL ISNT rule should not always be considered abnormal. ⁵⁷ Normal eyes probably exhibit some variation in the number and distribution of RGC axons, which may be why some normal eyes do not follow the RNFL ISNT rule. ⁵⁷  

Mean RNFLT decreased with age in the upper, nasal, and temporal quadrants in our study, which is in agreement with earlier histologic studies. ³⁷ The reported findings of fourquadrants analysis are inconsistent. Parikh et al. ²⁴ reported an upper quadrants RNFLT decrease of 0.23 μm/y (95% CI, 0.37–0.09) and a lower quadrants RNFLT decrease of 0.09 μm/y (95% CI, 0.23–0.06); they also reported that lower quadrants RNFLT is least affected by age and that in patients with suspected glaucoma, the decrease in the 6 clock-hour quadrants should be monitored carefully. In their study, some patients were in the +5.0 to −5.0 D range of spherical error. Myopic thinner RNFLT decrease was less with aging, ⁵⁷ which may be why RNFLT differed between our study and that of Parikh et al. ²⁴ Feuer et al. ³ reported a decrease in RNFLT in the upper quadrant of 0.43 μm/y (95% CI, 0.53–0.33) and in the nasal quadrant of 0.29 μm/y (95% CI, 0.39–0.18), which are similar to the values reported by Parikh et al. ²⁴ The decrease in upper quadrant RNFLT reported by Feuer et al. ³ was similar to that in our study. Sung et al., ²⁵ however, reported a statistically significant reduction in the lower quadrants RNFLT (0.36 μm/y; 95% CI, 0.54–0.18) in addition to a decrease in the upper quadrants (0.35 μm/y; 95% CI, 0.53–0.16), and that they were different than those reported by Parikh et al., ²⁴ perhaps because the latter study included participants aged ≤18 years. The findings of Lee et al. ²⁹ were similar to those of Sung et al., ²⁴ and Lee et al. reported that the decrease in the lower quadrants was noteworthy, and that clinicians should be aware of such decreases. In our study, the decrease in RNFLT was 0.488 μm/y (95% CI, 0.646–0.330), 0.575 μm/y (95% CI, 0.733–0.416), 0.253 μm/y (95% CI, 0.350–0.156), and 0.141 μm/y (95% CI, 0.272–0.01) in the upper, lower, temporal, and nasal quadrants, respectively. The limited decreases in the temporal quadrant may have been the result of a less dense papillomacular band in this quadrant. Again, racial differences might be the cause for the variability of RNFLT in the four quadrants. Girkin et al. ⁶⁴ reported that there were regional differences in the variation of RNFLT across racial strata; Hispanics and Indians have greater RNFLT in the superior and inferior quadrants than the European descent, and Africans and Indians have significantly thinner RNFLTs compared to the European descent in the temporal nerve corresponding to the papillomacular bundle. Knight et al. ⁷⁰ reported that individuals of European descent had the thinnest RNFL values based on SD-OCT, except in the temporal quadrant, in which those of African descent had the thinnest RNFL values. The findings of Kim et al. ⁷¹ suggested that Asians may have a thicker RNFL in the nonnasal regions than Caucasians and Hispanics. 

In our study, the decrease in RNFLT in the 6 clock-hour sectors (0.656 μm/y; 95% CI, 0.939–0.374; corresponding to the lower quadrant) was greater than in other clock hour sectors. The decrease in the 3 clock-hour sectors (0.119 μm/y; 95% CI, 0.266–0.029) and 9 clock-hour sectors (0.128 μm/y; 95% CI, 0.214–0.04), corresponding to the nasal and temporal quadrants, was limited. Studies on RNFL thinning in the 12 clock-hour sectors report inconsistent findings. ^3,24,25,29 Parikh et al. ²⁴ reported that the greatest decrease in RNFLT was in the 8 clock-hour sectors (0.35 μm/y; 95% CI, 0.41–0.14), versus the least decrease in the 6 clock-hour sectors (0.022 μm/y; 95% CI, 0.17–0.012). Sung et al. ²⁵ reported the greatest decrease in RNFLT was in the 1 clock-hour sectors (0.45 μm/y; 95% CI, 0.63–0.26), versus the least decrease in the 7 clock-hour sectors (0.31 μm/y; 95% CI, 0.49–0.13). Feuer et al. ³ reported that the greatest decrease was in the 1 clock-hour sectors (0.54 μm/y; 95% CI, 0.68–0.41) and that the least decrease was in the 6 clock-hour sectors (0.09 μm/y; 95% CI, 0.27–0.07). Lee et al. ²⁹ observed the greatest decrease in the 6 clock-hour sectors (0.45 μm/y; 95% CI, 0.59–0.32) and the least decrease in the 3 clock-hour sectors (0.02 μm/y; 95% CI, 0.08–0.03). In our study, the greatest decrease was observed in the 6 clock-hour sectors (0.656 μm/y; 95% CI, 0.939–0.374), whereas the least decrease was in the 3 clock-hour sectors (0.119 μm/y; 95% CI, 0.266–0.029), followed by the 9 clock-hour sectors (0.128 μm/y; 95% CI, 0.214–0.042). The findings from various studies on RNFLT, including our study, are shown in Table 4. Differences in the decrease in RNFLT between the 4 quadrants and 12 clock-hour sectors may be due to differences in ethnicity. ²⁴ Kim et al. ⁷¹ suggested that OCT data obtained from Asians might differ significantly in various clock-hour sectors from data obtained in other races. The main retinal blood vessels are located superotemporal (11 clock-hour sector) and inferotemporal (7 clock-hour sector) to the optic nerve, where the largest RNFL decreases were seen. ⁷² In our study, the second and third greatest age-related decreases in RNFLT were noted in the 7 clock-hour (0.636 μm/y) and 11 clock-hour (0.597 μm/y) sectors, respectively; therefore, the development of new algorithms that can subtract blood vessels from total RNFLT automatically may be clinically important to further study on age-related changes in RNFLT in these clock-hour sectors. 

Table 4

 

RNFLT Decrease in Our Study and in Other Studies (μm/Decade, 95% CI)

	Current Study	Lee et al.29	Feuer et al.3	Sung et al.25	Parikh et al.24
Mean	3.65 (4.65, 2.64)	2.1 (2.4, 1.7)	2.4 (3.1, 1.8)	2.6 (4.4, 0.7)	1.6 (2.9, 0.2)
4 quadrants
Upper	4.88 (6.46, 3.30)	3.4 (4.1, 2.7)	4.3 (5.3, 3.3)	3.5 (5.3, 1.6)	2.3 (3.7, 0.9)
Nasal	1.41 (2.72, 0.10)	1.0 (1.4, 0.5)	2.9 (3.9, 1.8)	2.5 (4.3, 0.6)	1.1 (2.5, 0.4)
Lower	5.75 (7.33, 4.16)	2.9 (3.6, 2.2)	1.5 (2.5, 0.4)	3.6 (5.4, 1.8)	0.9 (2.3, 0.6)
Temporal	2.53 (3.50, 1.56)	1.1 (1.5, 0.6)	1.0 (1.8, 0.2)	0.7 (2.5, 1.2)	2.0 (3.4, 0.7)
Clock-hour
12	3.99 (7.17, 0.81)	3.2 (4.5, 1.8)	4.5 (6.0, 3.0)	3.6 (5.5, 1.8)	1.9 (3.3, 0.5)
1	4.69 (7.34, 2.04)	4.1 (5.2, 2.9)	5.4 (6.8, 4.1)	4.5 (6.3, 2.6)	3.2 (4.6, 1.8)
2	1.37 (3.63, 0.90)	2.1 (3.0, 1.2)	3.9 (5.2, 2.5)	3.6 (5.5, 1.8)	1.1 (2.5, 0.4)
3	1.19 (2.66, 0.29)	0.2 (0.8, 0.3)	2.0 (3.0, 0.9)	1.8 (3.7, 0)	0.5 (1.9, 1.0)
4	1.71 (3.66, 0.25)	0.6 (1.1, 0)	2.7 (3.9, 1.5)	1.9 (3.7, 0)	0.9 (2.3, 0.5)
5	4.42 (7.12, 1.71)	2.3 (3.2, 1.3)	1.1 (2.6, 0.4)	3.9 (5.7, 2.0)	0.4 (1.9, 1.0)
6	6.56 (9.39, 3.74)	4.5 (5.9, 3.2)	0.9 (2.7, 0.7)	3.9 (5.7, 2.0)	0.2 (1.7, 1.2)
7	6.36 (9.70, 3.02)	1.8 (2.9, 0.7)	2.4 (3.8, 1.0)	3.1 (4.9, 1.3)	2.3 (3.7, 0.8)
8	2.31 (3.95, 0.66)	0.9 (1.6, 0.2)	1.4 (2.3, 0.4)	1.1 (2.9, 0.8)	2.8 (4.1, 1.4)
9	1.28 (2.14, 0.42)	0.8 (1.2, 0.4)	0.6 (1.2, 0)	0.4 (2.2, 1.5)	1.5 (2.9, 0.1)
10	4.03 (5.53, 2.54)	1.5 (2.1, 0.9)	1.1 (2.1, 0)	0.1 (2.5, 1.2)	1.8 (3.2, 0.4)
11	5.97 (8.97, 2.96)	3.0 (4.2, 1.8)	3.0 (4.3, 1.6)	2.3 (4.2, 0.5)	1.6 (3.0, 0.2)

A histologic study by Mikelberg et al. ³⁹ showed that the quantity of axons in the RNFL vary by quadrant; they reported that axons temporally located in the ONH are thinner and fewer in number, and that axons in the inferotemporal region (7 clock-hour sector) are greater in number those the in other regions. The decrease in RNFLT in the temporal quadrant, where fibers with smaller axon diameter are located, may be less than the decrease in the upper and lower quadrants, which have larger diameter axons. In our study, the RNFLT in overall group was second greatest (137.79 μm) in amount in the 7 clock-hour sector. The decrease in RNFLT in the 4 groups in our study was greater in the 7 clock-hour sector (0.636 μm/y; 95% CI, 0.970–0.302) than in the 9 clock-hour sector (temporal quadrant, 0.128 μm/y; 95% CI, 0.214–0.042), which had the fewest RNFL axons (and lowest RNFL density); however, due to its proximity to the fovea, the RNFL in the temporal quadrant is protected to a greater degree and, thus, we think that central vision should be protected as well. 

Based on our findings, we can predict the level decrease in RNFLT neighboring the ONH in any clock-hour sectors or in any of the four quadrants during a given period of time. The regression equation (y = a + b × X) for each of the 4 quadrants and the 12 clock-hour sectors is used to determine RNFLT decrease, where “a” is the constant obtained using regression analysis (constant or intercept), and “b” is the regression coefficient (absolute slope) multiplied by “X” which is the independent variable (age, Table 3). We can calculate the level of RNFL loss in an individual aged X years using the b × X part of the formula. For instance, the mean RNFL loss in a 40-year-old individual is 0.365 × 40 = 14.6 μm; this is well below the value of 47.535 μm (50% loss) in the mean RNFLT value of a 40-year-old individual associated with functional VF loss measured by perimetry (based on a 40-year-old individual's normal RNFL value of 95.07 μm). ¹  

What is needed in Turkey, as elsewhere, is the establishment of a national archive for RNFL data, which can be used to create population-specific regression equations based on the collected epidemiologic data for determining if an individual has glaucomatous ONH RNFLT changes based on comparison with a national normal RNFLT value. The development of a normative database using data derived from each race is needed to increase the utility of OCT in different racial populations. Manufacturers should develop their reference database for progression adjusting for ethnic differences. The age-appropriate regression equation from the national archive provides important information about the degree of a person's age-related normal RNFLT value deviation. 

As a result, age-related RNFLT measured with OCT in healthy individuals does not indicate the same slope in each sector/quadrant. The primary limitation of our study is its cross-sectional design; as such, the present findings must be confirmed by larger scale longitudinal studies. Age-related change should be accounted for in any assessment of RNFLT. Regional age-related change is accounted for by Cirrus SD-OCT in its normative database. 

Acknowledgments

Presented in part as a rapid-fire oral presentation at the European Society of Ophthalmology (SOE) Congress, Copenhagen, Denmark, June 8–11, 2013. 

Disclosure: A.R.C. Celebi, None; G.E. Mirza, None