Introduction

Glaucoma is the leading cause of irreversible blindness worldwide, accounting for an estimated 2.8–4.4% of all cases of blindness and 3.2–5.2% of moderate to severe visual impairment in individuals aged 50 years and older1. Projections indicate that the prevalence of glaucoma will increase from 76 million in 2020 to 111.8 million by 2040, primarily due to the global aging population2. This trend highlights the urgent need to address glaucoma as a major public health concern.

Glaucoma often develops asymptomatically during its early stages, characterized by the degeneration of retinal ganglion cells and changes in the optic nerve head1. Evaluating the peripapillary retinal nerve fiber layer (RNFL) thickness is essential for diagnosing and managing patients with glaucoma and other optic nerve disorders. This is because RNFL damage is one of the earliest signs of glaucoma, occurring before any detectable alterations in visual fields. Studies suggest that up to 50% of retinal ganglion cells may be lost before visual field defects become apparent3. This parameter has aided in diagnosing neurodegenerative conditions and their association with depression4, as well as in diagnosing and monitoring the progression and treatment of Alzheimer’s disease, Parkinson’s disease, and multiple sclerosis5,6,7.

The RNFL thickness reference values are determined for each specific measurement device, and these values vary among different devices8. The importance of racial factors has led to research examining the variations in RNFL thickness among different racial groups9,10. Research indicates that RNFL thickness diminishes at an average rate of 0.36% annually, corresponding to a reduction of approximately 0.365 microns (µm)11. Although age-related variations in RNFL thickness have been documented, there is a notable scarcity of studies focusing specifically on middle-aged and older populations12,13,14,15,16,17,18,19,20. A limited proportion of participants in previous studies were middle-aged or elderly13, and those studies concentrating solely on these age groups often involved small sample sizes21. This is while the prevalence of glaucoma and other optic nerve disorders tends to rise with age, with reported prevalence rates increasing from 0.9% among individuals aged 40–44 years to 3.55% for those aged 60–64 years22. This trend underscores the necessity of investigating RNFL thickness within this demographic.

Optical coherence tomography (OCT) is a biomedical instrument that uses non-invasive low-coherence interferometry to generate quantitative cross-sectional images of the retina, optic nerve, and associated structures in vivo. This technology enables the early and objective differentiation between normal and glaucomatous eyes by evaluating RNFL thickness, a crucial parameter in OCT analysis23. The Heidelberg Spectralis device is a spectral-domain (SD) OCT system that operates at a speed 60 times greater and with an accuracy 5 times higher than that of time-domain (TD) OCT24,25. In our previous research, we documented changes in the RNFL thickness over five years within the same study cohort utilizing Cirrus HD-OCT26. Our primary objective is to evaluate the normal distribution of RNFL thickness in individuals aged 50 to 74 years, along with various associated factors, using the Heidelberg Spectralis system.

Materials and methods

This report is part of the third phase of the Shahroud Eye Cohort Study, conducted cross-sectionally in 2019. This particular phase targeted individuals aged 50 to 74 years. Detailed descriptions of the sampling methods and methodologies related to the first phase of the study have been presented in earlier publications27, with a concise summary provided in this section.

In this research, participants were obtained from Shahroud city (urban region) through a multi-stage stratified cluster sampling technique. A total of 300 clusters were identified within Shahroud, with each cluster comprising 20 individuals. Given that the primary health care system in Iran encompasses 9 health centers serving the Shahroud population, each health center was treated as a distinct stratum, and the number of clusters was allocated in accordance with the population distribution across these centers. The heads of the clusters for each center were systematically determined using the household file numbers maintained at the health centers. Following obtaining informed consent from each participant, demographic data, information regarding tobacco use, and medical and ophthalmological histories were collected. Subsequently, optometric and ophthalmological assessments, along with imaging procedures, were conducted for the participants.

Optometric examinations included the measurement of both uncorrected and best-corrected visual acuities, non-cycloplegic refraction through autorefractometer and retinoscopy, as well as lensometry. Biometric data were collected using the Lenstar device, followed by a slit-lamp examination performed by an ophthalmologist for each participant. After administering tropicamide 1% drops for pupillary dilation, fundoscopic and ophthalmoscopic examinations were conducted for all individuals. Retinal imaging was subsequently performed in a dark room by a trained nurse utilizing the Heidelberg Spectralis OCT system (Heidelberg Engineering, Heidelberg, Germany). The RNFL optic disc protocol was employed to assess the RNFL thickness through a circular scan centered on the optic disc. This circular scan consists of 1536 A-scan points arranged in a 12-degree circle, which corresponds to a retinal diameter of 3.5–3.6 mm in eyes exhibiting normal corneal curvature. The device operates at an acquisition rate of 40,000 A-scans per second, achieving an axial resolution of approximately 3.9 mm and a lateral resolution of 6 mm. A real-time eye-tracking system is integrated to adjust for involuntary eye movements that may occur during imaging. Furthermore, automatic segmentation of the upper and lower borders of the RNFL is provided to compute the global RNFL thickness. The RNFL thickness values are also detailed in four distinct quadrants: superior, inferior, nasal, and temporal. The accompanying software generates a quality score (Q score) that reflects signal strength, ranging from 0 (indicating poor quality) to 40 (indicating excellent quality). A 20 × 20° macular volume scan was also applied to measure macular thickness and macular volume.

Definitions

Refractive errors were defined according to the spherical equivalent (SE) refraction. Myopia was identified as an SE of less than − 0.5 diopters (D), while hyperopia was characterized by an SE exceeding + 0.5 D28,29,30.

Exclusion criteria

In this report, the exclusion criteria were the inability to measure the RNFL thickness or any missing data in its sections, a history of retinal surgeries or diseases, previous glaucoma surgeries or glaucoma conditions, any ocular surgeries other than cataract procedures, best-corrected visual acuity worse than 20/25, myopia exceeding − 6.00 D, OCT scans exhibiting artifacts (motion artifacts, inadequate segmentation, signal loss, and a quality score below 15), as well as data classified as outliers (those deviating more than three standard deviations from the mean).

Statistical analysis

The average values and 95% confidence intervals (CI) for the global RNFL thickness and RNFL thickness measurements across different sectors, were recorded for all participants and categorized by study variables. The cluster sampling method was considered in calculating the standard error. Furthermore, the 1% and 5% percentiles were included to depict the distribution of the global and sectoral RNFL thicknesses. To evaluate normality, the skewness and kurtosis of the global RNFL thickness were computed, and the normal range was calculated as the mean ± 2 standard deviations.

Since data from both eyes were included in the analysis, statistical models were applied to examine the relationships between variations in RNFL thickness and other influencing factors. Simple and multiple generalized estimating equations (GEE) models were employed to analyze these associations. The final multiple GEE model was developed using the backward method, with coefficients reported alongside their respective 95% CIs.

Ethical considerations

The Shahroud Eye Cohort Study was approved by the Ethics Committee of Shahroud University of Medical Sciences (ethics code: IR.SHMU.REC. 1398.039). All methods were carried out in accordance with the Helsinki tenets and other relevant guidelines and regulations. Written informed consent was obtained from all participants.

Results

Among the 4,394 individuals involved in this research, ‎5266 eyes from 2648‎ participants underwent retinal imaging using Heidelberg Spectralis OCT.

A total of 1,566 eyes were excluded from the analysis for various reasons. Specifically, 188 eyes were omitted due to the lack of RNFL thickness measurements or incomplete data. Additionally, 173 eyes were excluded because of retinal diseases and/or related surgical procedures. Furthermore, 357 eyes were removed from the analysis due to glaucoma and its associated surgical treatments. Moreover, 482 eyes were excluded as their best-corrected visual acuity was worse than 20/25. Fifteen eyes were omitted for having myopia exceeding − 6.00 D, while 74 eyes were excluded due to a prior history of ocular surgery. Lastly, 150 eyes were removed because their signal quality was below 15, and 127 eyes were classified as outliers. Ultimately, 3,700 eyes from 2,102 participants were analyzed in this report.

Notably, 56% (1,178) of the participants were female. The average age of the individuals analyzed was 60.4 ± 5.96 years (range: 50 to 74 years). The mean SE for the examined eyes was 0.19 ± 1.25 D (range: -5.75 to 5.88 D). Table 1 shows the distribution of demographic characteristics, medical history, and ocular indices of the participants in this study.

Table 1 The distribution of demographic characteristics, medical history, and ocular indices of the participants in this study.
Full size table

Table 2 presents the mean values and 95% CIs for the global and sectoral RNFL thicknesses, categorized by age and sex and among all study participants. The findings indicate that the average global RNFL thickness in this study was 100.3 ‎µm‎ (95% CI: 99.9-100.7). Figure 1 illustrates the distribution of the global RNFL thickness among the study participants.

Table 2 Mean and 95% confidence interval (CI) of the overall and sectoral retinal nerve fiber layer ‎thickness (‎µm‎) by age and sex.
Full size table
Fig. 1
figure 1

The distribution of the average retinal nerve fiber layer thickness (µm) in the adult population of Shahroud aged 50 to 74 years.

Full size image

The skewness and kurtosis values for the global RNFL thickness in this study were − 0.85 and 0.014, respectively. The average global RNFL thickness for male participants was 99.5 ‎µm‎ (95% CI: 98.9-100.1), while female participants exhibited an average thickness of 100.9 ‎µm‎ (95% CI: 100.3-101.5). As shown in Table 2, the changes in global RNFL thickness did not follow a linear pattern concerning age, with the lowest values observed in individuals aged 70 to 74 years.

Additionally, Fig. 2 illustrates a significant interaction between age and sex concerning the average overall RNFL thickness, revealing an increase in thickness for women and a decrease for men as age progresses.

Fig. 2
figure 2

The interaction of age and gender in the average retinal nerve fiber layer thickness (µm).

Full size image

The normal ranges for RNFL thickness values in the examined population were categorized by age and sex, employing the mean and two SD as benchmarks (Table 3). The global RNFL thickness is determined to fall within the normal range of 81.26 to 119.32 ‎µm‎.

Table 3 The normal range (mean ± 2 standard deviations) of the overall and sectoral retinal nerve fiber layer thickness ‎‎(µm‎) ‎ ‎by age and sex.
Full size table

Table 4 illustrates the 1st and 5th percentiles for both global RNFL thickness and RNFL thickness across different sectors. In this cohort, the 1st and 5th percentiles for global RNFL thickness were noted at 77.01 and 85 ‎µm‎, respectively. The temporal-inferior sector demonstrated the highest RNFL thickness at 144.8 ‎µm‎, whereas the temporal sector recorded the lowest thickness at 69.5 ‎µm‎. A repeated measures analysis of variance (ANOVA) revealed a statistically significant difference in RNFL thickness among the various sectors (p < 0.001).

Table 4 The percentiles of the overall and sectoral retinal nerve fiber layer thickness (‎µm‎) ‎in this ‎‎study.
Full size table

The average of global RNFL thickness in emmetropic, myopic, and hyperopic individuals was recorded as 100 μm (95% CI: 99.4-100.5), 98.3 ‎µm‎ (95% CI: 97.4–99.1), and 101.9 ‎µm‎ (95% CI: 101.2-102.5), respectively. Notably, myopes exhibited a significantly reduced global RNFL thickness. As illustrated in Fig. 3, the average RNFL thickness varies with the severity of refractive errors, revealing that individuals with high myopia and high hyperopia exhibited the thinnest RNFL measurements.

Fig. 3
figure 3

The distribution of the average retinal nerve fiber layer thickness (µm) by severity of spherical equivalent.

Full size image

The association between global RNFL thickness and the examined variables was analyzed using simple and multiple GEE models, with the findings presented in Table 5. The multiple GEE model revealed a statistically significant interaction between age and sex in relation to global RNFL thickness. Specifically, women aged 70 to 74 years exhibited a greater global RNFL thickness [β = 3.15 (95% CI: 0.00; 6.29); p = 0.050] when compared to their male counterparts within the same age group. Furthermore, this model indicated that global RNFL thickness is significantly directly associated with body mass index [β = 0.11 (95%CI: 0.02; 0.19); p = 0.013], corneal radius of curvature [β = 5.27 (95%CI: 3.58; 6.96); p < 0.001], and macular volume [β = 8.30 (95%CI: 7.47; 9.13); p < 0.001)], while exhibiting a significant inverse relationship with axial length [β=-2.29 (95%CI: -2.84; -1.74); p < 0.001)] and central macular thickness [β=-21.39 (95%CI: -35.05; -7.73); p = 0.002)]. Moreover, the analysis indicated that global RNFL thickness was significantly lower in individuals with diabetes compared to those without the condition [β=-1.09 (95%CI: -1.91; -0.27); p = 0.009)].

Table 5 The association of average retinal nerve fiber layer thickness with studied variables according to simple and multiple generalized estimating equations models.
Full size table

Discussion

This study investigated the distribution of RNFL thickness and its associated factors in a normal population aged 50 to 74 years, employing the Spectralis SD-OCT device. As previously noted, changes in RNFL thickness over a five-year period have already been documented in a different sample from this cohort using the Cirrus device26. The current study specifically focuses on the RNFL thickness distribution within the 50 to 74 age range, thereby providing significant contributions to existing databases.

The results of this study show that the average global RNFL thickness was 100.3 ‎µm‎, with a range of 99.9 to 100.7 ‎µm‎. A summary of findings from previous studies is provided in Table 6. Specifically, the influence of race and genetic factors should not be overlooked, as prior research has indicated that RNFL thickness measurements and various ocular parameters exhibit significant differences among racial groups9,52.

Table 6 Overview of additional research on retinal nerve fiber layer(RNFL) thickness from various global ‎studies.‎.
Full size table

The literature reveals a significant variability in the mean RNFL thickness, with reported values ranging from 87.3 to 126.98 μm. This inconsistency can largely be attributed to several factors, notably the type of OCT device used. Additionally, variations in age distribution among study populations likely contribute to these discrepancies, as RNFL thickness is known to decrease with age. In comparison to East and South Asian populations, the average RNFL thickness in the Iranian cohort is lower, measuring around 92.47 ± 8.56 μm, although some studies report values as high as 98.6 μm depending on age and methodology. In contrast, African American populations demonstrate higher RNFL thickness, with averages reaching up to 101.5 μm. These findings underscore the necessity for population-specific normative data in the cross-ethnic screening of glaucoma and other optic neuropathies, as neglecting ethnic differences may result in misdiagnosis or inadequate clinical management. Furthermore, the variability in measurement techniques across different studies may also play a role in the observed differences.

In the present study, the RNFL thickness was observed to be greatest in the inferior sector, measuring 129.75 μm, followed by the superior sector at 126.05 μm, the nasal sector at 75.9 μm, and the temporal sector at 69.5 μm. These findings align with previous research conducted on younger populations and can be explained by the “Inferior-Superior-Nasal-Temporal” rule, suggesting a “double-humped” configuration53,54.

Age is recognized as a significant factor affecting RNFL thickness. Prior research has established an inverse correlation between age and RNFL thickness, indicating an annual reduction ranging from 0.15 to 0.63 μm55,56,57. Nevertheless, several studies failed to demonstrate a relationship between age and RNFL thickness, potentially attributable to the younger age of their participants58,59. Notably, the most pronounced loss of nerve fibers occurs after the age of 50 years, as evidenced by histological investigations conducted by Quigley et al.60, which reveal that individuals over 50 years of age experience a threefold increase in the loss of optic nerve axons compared to their younger counterparts. The decline in RNFL thickness associated with aging may be linked to a reduction in venous caliber, leading to cytotoxic and vasogenic edema, which ultimately results in arteriolar stenosis21. In the present study, we found a significant interaction between age and sex concerning the RNFL thickness. Specifically, RNFL thickness increased with age in females, whereas it decreased in males, with women exhibiting a higher global RNFL thickness. The research conducted by Wanger et al. highlighted distinct RNFL profiles between genders61. While numerous studies have indicated that women typically have a thicker RNFL49, Hashemi et al. reported findings suggesting that women may have a thinner RNFL in their analysis10. Additionally, several investigations did not find a significant relationship between sex and RNFL thickness57. Alpogan et al. observed that, after controlling for age, a later onset of menopause correlated with increased RNFL thickness62. Moreover, Jeong et al. noted that the decline in RNFL thickness with age was less marked in women63. These findings are consistent with our results, which demonstrate a significant interaction between age and sex concerning RNFL thickness—an area that has not been extensively examined in existing literature. Our findings indicate that RNFL thickness generally increases with age in women, whereas it tends to decrease in men. In contrast, research conducted by Rougier et al. revealed that women aged 80 years and older exhibited lower RNFL thickness compared to their younger counterparts under 80 years64.

In the present study, an indirect association was observed between axial length and the RNFL thickness, which is consistent with previous research findings65. Specifically, various studies have demonstrated that for each millimeter increase in axial length, the average RNFL thickness decreases by approximately 2.2 to 3.6 μm13. Lamparter et al. identified an independent relationship between myopia and axial length with RNFL thickness; however, the potential impact of the magnification effect on this association remained unclear66. Vernon et al. reported no significant correlation between age, refractive errors, and axial length, a finding that may be influenced by the small sample size and limited age range of the participants67. The significant correlation between axial length and RNFL thickness suggests that axial length should be included as a measurable variable in normative databases. This inclusion could enhance the accuracy and relevance of assessments related to ocular health68.

The present study corroborates the findings of Choi et al., revealing a significant association between myopia and a reduction in RNFL thickness69. Research by Zhao et al. has established that retinal elongation plays a role in the observed decrease in RNFL thickness70. Additionally, Grudzinska et al. have identified that diminished retrobulbar blood flow in individuals with myopia contributes to reduction in RNFL thickness71. The magnification effect associated with OCT scans results in larger projection rings on the retina in eyes with elongated axial lengths, which in turn leads to a decrease in RNFL thickness as the distance from the optic disc increases21,72. Hwang et al. have shown that when this magnification effect is taken into consideration, the relationship between myopia and RNFL thickness loses its statistical significance73. Importantly, higher levels of myopia are linked to a thinner RNFL in the superior, nasal, and inferior regions, whereas the temporal region tends to exhibit a thicker RNFL74,75. The mechanical stretching and thinning of the RNFL due to increased axial length in myopia is especially evident, except for the temporal sector76. The relative consistency of the temporal quadrant in response to varying degrees of myopia may be due to its naturally thinner structure and the limited direct influence of myopia on this area, or it may result from the retina being dragged toward the temporal side. In the study conducted by Hoh et al.77. , no significant correlation was found between RNFL thickness and myopia or axial length77. Furthermore, Hoh et al.77. indicated that varying degrees of myopia and axial length did not influence RNFL thickness when a 4.5 mm scan diameter was used. Conversely, Alsail et al. reported that for each diopter increase in refractive error towards hyperopia, there was an average increase of 1.4 μm in RNFL thickness57. Consistent with the present study, studies by Yasir et al. and Wu et al. demonstrated that greater severity of myopia was linked to a notable reduction in RNFL thickness78,79. Additionally, Lee et al. observed that the decline in RNFL thickness among the elderly occurred more rapidly in individuals with high myopia compared to those in the control group56. Consequently, particularly at elevated levels of myopia, it is crucial to consider differential diagnoses for reduced RNFL thickness to avoid misdiagnosis with open-angle glaucoma, which is more prevalent among myopic individuals80. Utilizing normative databases while accounting for myopia significantly mitigates the discrepancies in RNFL thickness observed among individuals with varying degrees of myopia81.

In the present study, a direct relationship was observed between the corneal radius of curvature and RNFL thickness. Previous research has explored the association between corneal parameters and RNFL thickness. For instance, Liu et al. reported that eyes with higher levels of corneal astigmatism demonstrated reduced RNFL thickness in the temporal sector, a finding attributed to variations in OCT reflectivity and backscattering82. It has been shown that corneas with steeper keratometric values tend to exhibit thinner central corneal thickness83. The study conducted by Arranz-Marquez et al. indicated a direct relationship between RNFL thickness and central corneal thickness84. Moreover, findings from Wang et al. suggested that a flatter anterior corneal curvature was linked to increased RNFL thickness33.

The present study observed a significant direct association between the RNFL thickness and body mass index. Prior investigations have suggested that a higher BMI may confer a protective effect against glaucoma85,86. For example, research conducted by Leske et al. indicated that an elevated BMI is linked to a decreased risk of developing open-angle glaucoma, suggesting potential shared genetic factors87. However, the nature of this association appears to differ among various studies. For instance, Khawaja et al. reported an inverse correlation between RNFL thickness and body mass index, however, this finding was limited to male participants, presenting a contrast to our findings that could be influenced by racial differences88. Furthermore, some studies did not find a significant link between RNFL thickness and body mass index; notably, the study performed by Lamparter et al.66 did not utilize multiple regression analysis, while the study conducted by Cheung et al.89 had a small sample size.

In the current investigation, an inverse correlation was observed between the central macular thickness and the RNFL thickness. This association may be linked to the aging process, as indicated by the findings of Chua et al., who noted that RNFL thickness at the fovea, along with the inner and outer segments of photoreceptors, tend to increase with age due to the activity and thickening of Müller glial cells90. The research conducted by Chauhan et al. reported no significant changes in RNFL thickness within the macula as individuals age91. Furthermore, the present study revealed a direct relationship between RNFL thickness and macular volume. Ojima et al. documented a similar direct correlation in the early stages of glaucoma, where both macular volume and RNFL thickness were found to be diminished92. Additionally, the work of Lederer et al. highlighted that as glaucoma progresses, there is an escalation in ganglion cell death, leading to a reduction in macular cellularity and, consequently, a decrease in macular volume93.

In alignment with the present study, the research conducted by Ezhilvendhan et al. demonstrated a significant association between diabetes and a reduction in RNFL thickness94. Diabetes induces apoptosis in ganglion and Müller cells, leading to a decrease in RNFL thickness before the onset of diabetic retinopathy95. The neurodegenerative processes associated with diabetes are attributed to ischemic conditions and metabolic interactions that precede vascular alterations96. Furthermore, RNFL thickness may serve as a predictive marker for the onset of diabetes97. The duration of diabetes plays a critical role; for instance, the findings of Lee et al. indicated that patients with a diabetes duration exceeding ten years exhibited markedly lower RNFL thickness compared to those with a shorter duration of the disease98. Conversely, the study by Demir et al. did not identify a significant relationship between type 2 diabetes and RNFL thickness, although it is important to note that their sample size was relatively limited99.

This study acknowledges certain limitations that warrant attention. Due to the cross-sectional design, causal relationships between variables such as myopia, age, and ‎other factors cannot be definitively established‎. Furthermore, the findings are specific to the Iranian population, and caution is advised when attempting to generalize these results to other groups, particularly patient populations. One significant limitation of this study is the potential amplification effect arising from variations in axial length, which could affect the measurements of RNFL thickness, particularly in myopic eyes. To address this issue, future investigations should incorporate axial length-based correction techniques to reduce bias and improve the comparability of results across diverse populations. Nonetheless, a notable strength of this research lies in examining the RNFL thickness within a substantial sample of the normal population, utilizing a random sampling approach for individuals aged 50 years and older.

In conclusion, applying standard databases, while taking into account a range of variables, proves to be more effective for the early diagnosis of glaucoma compared to traditional methods that rely exclusively on age. This underscores the significance of incorporating demographic, ocular, and systemic factors that influence RNFL thickness.