Article Text
PDF
Abstract
Background/aim: Retinal and retinal nerve fibre layer (RNFL) thinning with age have been described in histological studies. In vivo techniques like optical coherence tomography (OCT) have shown thinning of optic nerve RNFL and the retina in specific areas. One would expect thinning of the total macula, but so far, no correlation with the quantitative OCT macular map tool and age has been found.
Methods: Sixty-seven healthy individuals underwent three repeated scans in both eyes with the macular thickness map protocol in the Stratus OCT. That protocol divides the macula area into nine ETDRS fields. The RNFL was measured in one specific location close to the optic disc. Correlations between retinal, RNFL thickness, macular volume and age were determined.
Results: We found a statistically significant negative relationship between retinal thickness and age for all ETDRS areas, total macular volume and RNFL thickness. Retinal thickness decreased by 0.26–0.46 μm, macula volume 0.01 mm3 and RNFL 0.09 μm per year.
Conclusion: Retinal thickness within the area covered by the macular map significantly decreases with age. In the area examined in the papillomacular bundle, 20% of the retinal thinning is due to the RNFL, and 80% is due to thinning of other layers of the retina.
Statistics from Altmetric.com
Histological studies of the human retina and optic nerve have shown a decreased density of photoreceptors, ganglion cells, retinal pigmentepithelium and optic nerve fibres with age.123 These findings, however, do not necessarily have to result in retinal thinning, but one would expect some shrinkage of the total retina over time. Non-invasive techniques have made it possible to measure the thickness of retinal structures in vivo. With optical coherence tomography (OCT), both qualitative and quantitative measurements of the retina can be made.4567 A macular mapping technique that has shown a good reproducibility is implemented into the OCT 2000 and OCT3 scanners, which are in clinical use today.8910111213 Observations based on single scan measurements have shown a decrease in retinal and RNFL thickness with age. In a pilot study, Schuman et al reported that the peripapillary retinal nerve fibre layer (RNFL) decreases with age using OCT 1.14 Poinoosawmy and coworkers also demonstrated a progressive reduction in the nerve fibre layer thickness with age using scanning laser polarimetry (GDx).15 In an OCT study by Alamouti and Funk, both retinal and RNFL thickness decreased slightly with age.16 Finally, Kanai and coworkers also found that retinal thickness decreases with age.17 None of these investigators, however, used the OCT mapping technique. Based on the findings in postmortem and in vivo studies, one would expect a slight thinning of the total retina. Surprisingly, studies on normal retinal thickness with the mapping protocol have so far not shown any significant correlation between retinal thickness and age.81112 Therefore, we wanted to examine the relation between retinal thickness and age with the macular map technique.
Subjects
The study was conducted in accordance with the guidelines of the Declaration of Helsinki, and the protocol was approved by the ethics committee of Uppsala University. A total of 134 eyes from 67 healthy Caucasian subjects (43 women (64%) and 24 men (36%), median age 34.4 years (range 12–74)), recruited among the staff and their families at the department of ophthalmology, University Hospital, Uppsala, were included. All subjects underwent three measurements with the map protocol in both eyes. Before inclusion, a thorough medical history was obtained, and all subjects underwent a routine ophthalmological examination including visual acuity, slit-lamp examination and dilated fundus examination with the 60-dioptre lens in the slit lamp. Subjects with a history of any ophthalmological condition or positive findings on a routine eye examination were excluded. A visual acuity below 1.0 (20/20) and a refractive error of more than 6 dioptres spherical and/or more than 3 dioptres cylindrical were exclusion criteria.
Methods
The pupils were dilated with 0.5% tropicamide, and all subjects underwent a dilated slit-lamp examination. The latest commercially available OCT model, Stratus OCT 3 (Carl Zeiss Meditec. Dublin, California) with software 4.0.1, was used. In Stratus OCT, there are three different map protocols. The macular thickness map protocol was used in this study.
OCT is a non-invasive, non-contact technique by which to obtain high-resolution images from the anterior and posterior segments of the eye.456 Utilising the principles of ultrasound, OCT measures backscattered low-coherent light from intraocular structures as an A-scan. Cross-sectional tomographic images (B-scans) are constructed from repetitive axial A-scans while the probe beam scans across the retina. The retinal thickness is defined as the distance between the retinal surface (ILM) and the highly reflective band from the outer segment of the photoreceptors in Stratus OCT (fig 1A). These boundaries are automatically identified by the instrument. Quantitative measurements of macular thickness are possible using the macular mapping mode introduced by Hee and coworkers.7 The topographic two-dimensional macular map is built from six radial B-scans. By using an internal fixation, the scans are equally spaced in a spoke-wheel pattern through a common central axis over the macula centre. The linear tomograms are scanned 30° apart (fig 1B), and the thickness of areas in between the scans is interpolated by the instrument’s software. The average thickness/volume is presented as numerical values in nine modified ETDRS areas (fig 1C). The protocol used in this study was the macular thickness map protocol. In that protocol, the B-scans are manually obtained with an acquisition time of around 1 s/scan. Each B-scan is fixed on a 6 mm scan length and consists of 512 A-scans each, thus making the whole map made up from 3072 A-scans (512×6). The central point where all six line crosses is the foveal minimum and is calculated from six A-scans only. The total macular volume within the area examined can be calculated in cubic millimetres. To separate RNFL from total retinal thickness, one out of the six 6 mm scans that create a map was selected. That scan extends horizontally from the nasal to the temporal side of the macula (fig 1B). In this scan, one specific A-scan, number 50 of the 512, counting from the nasal side was used. That A-scan examines the retina in the papillomacular nerve fibre bundle, within ETDRS area 9. The RNFL is clearly defined in this area, which makes artefacts less likely. It is also adjacent to the area used by Alamouti and Funk when they described the relationship between RNFL and total retinal thickness.16 No manual measurement is involved, since the instrument software can calculate the RNFL thickness and the total retinal thickness in the same A-scan (fig 2).
(A) 6 mm scan demonstrating the distinctive boundaries of the retina, the vitreoretinal interface (a) and the high reflective band from the outer segment of the photoreceptors (b) marked by two white lines. The algorithm calculates the distance between these two retinal borders (c). (B) Placement of the six scans used to create the macular map scanned 30° apart. The 0° horizontal scan was used when calculating retinal nerve fibre layer thickness. (C) Macular map, automatically divided into nine areas. In the present study, Area A9 is always the area closest to the optic disc, and the right and left macular maps are thus mirror images.
(A) Retinal nerve fibre layer thickness in one A-scan (symbolised by the yellow line), calculated between the thin arrows (a) and corresponding to the blue line (b) on the thickness chart. (B) Total retinal thickness, calculated in the same A-scan (c) and corresponding blue line on thickness chart (d).
In one scanning session performed by a single examiner (UE), each patient underwent three standard maps in each eye. There was a short rest between every map obtained.
Statistics
Although all measurements were made in both eyes, only one eye, randomly chosen, was used for the statistics when calculating retinal thickness. When loss of thickness with age of the RNFL was analysed, both right and left eyes were analysed, since these measurements are based on only one scan.
Linear regression of retinal thickness and volume with age was based on the mean values of three measurements, in order to reduce the effect of measurement error. A linear regression analysis was also performed for single (first) measurements.
The statistical calculations were made with Statistica 7.1 (StatSoft, Tulsa, Oklahoma).
Results
The relationship between thickness in each ETDRS area including the foveal minimum (central piont) and macular volume with age was analysed. Using linear regression analysis, we found a statistically significant negative relationship between retinal thickness and age for all nine ETDRS areas in the macular thickness map (p = 0.042 for A1, p<0.01 for all other areas). Figure 3 gives an overview of all ETDRS areas together with the 95% CI and shows a negative slope for all of them. Table 1 lists the change in thickness with 95% CI for all nine ETDRS areas, the foveal minimum and the total macular volume. Retinal thickness decreased by 0.26–0.46 μm and the total retinal volume by about 0.01 mm3 per year. The change was significant for the volume and all nine areas but not for the foveal minimum. r Values are also included in the table and the coefficient of determination, r2, shows that only 10–20% of the variation in the measurements is due to age.
Relationship between thickness and age for areas A1 to A9 in one randomly chosen eye in 67 individuals between 12 and 74 years of age. For all nine areas, the y-axis is cut below 150 μm. The regression line and its 95% confidence interval are plotted for each area.
Linear regression of thickness (μm) or volume (mm3) with age in 67 individuals between 12 and 74 years of age
The values are the means of three values determined with the standard protocol and 95% CI.
The proportion of this tissue loss with age due to the well-established reduction in RNFL with age was also analysed. As for the total retinal thickness, we found a statistically significant thinning of the RNFL with age (fig 4), indicating a nerve fibre layer thinning of 0.09–0.12 μm per year in this area.
Relationship between thickness and age for 67 right eyes in area 9 (A), one A-scan, nr 50 counting from the nasal side, situated in area 9 (B) and the RNFL thickness of that particular A-scan as calculated by the instrument (C). The regression line and its 95% CI are plotted for each area, and the linear regression equations are shown also.
Discussion
We know from histological studies of the human retina and optic nerve that there is a decrease in density of photoreceptors, ganglion cells and RPE with age.12 Loss of optic nerve fibres has been estimated to be 4000–5000 fibres per year, giving a total reduction of 35% of the optic nerve fibres in a lifetime.18192021 Alamouti and Funk found a significant degree of thinning with age for both total retinal and RNFL thickness.16 Their measurements were performed with three single vertical OCT scans, 2.3 mm long, placed at the temporal edge of the optic disc. They estimated a decrease by 0.53 μm per year for the total retina and 0.44 μm for the RNFL, and concluded that about 80% of retinal thinning is due to shrinkage of the RNFL. Kanai and coworkers also found that retinal thickness decreases significantly with age by measuring at five specified points, one in the foveola and one in each quadrant around the macula with a distance of 1 mm from the central foveola.17 Others have reported a lack of any relationship between retinal thickness and age.81112 To our knowledge, this is the first report of statistically significant retinal thinning with age for the macula map protocol. Depending on the ETDRS area, we found a reduction of 0.26–0.46 μm per year for the total retinal thickness. When we evaluated the proportion of RNFL to retina, we looked at one particular A-scan. One can definitely question whether one can rely on the measurement in only one of the 512 A-scans. However, when we compare A-scan 50 with the value for the whole ETDRS area number 9, in which it is situated, the resemblance is convincing. The loss in retinal thickness per year was 0.46 μm for the single A-scan and 0.44 μm for ETDRS area 9. The loss of RNFL in A-scan 50 was 0.09 μm per year in the right eye (fig 4) and 0.12 μm in the left eye, which indicates that about 20–25% of the retinal thinning in this area is due to the RNFL and 75–80% due to thinning of other layers of the retina. This is not in accordance with Alamouti and Funk, who estimated that 80% of the thinning was due to RNFL shrinkage. However, the proportion between RNFL and the other retinal layers depends on the area of the retina that is examined. They made their measurements at the temporal edge of the optic disc, a location where one would expect a high percentage of nerve fibres. Our location was around 1 mm temporally from the optic disc. Thus, part of the difference in the present study and the study by Alamouti and Funk may be due to different locations of examination. Still, we have chosen a location rather close to that of Alamouti and Funk, and different algorithms for delineating the RNFL between the first OCT instrument and Stratus OCT may well be the main reason. The important finding is that we, like Alamouti and Funk,16 found a significant age-related reduction for both RNFL and the total retina. It is interesting to note that well-performed studies with the same instrument as well as the normal database in Stratus software 4.01 have failed to detect any relationship between macular thickness and age. Our findings are unlikely to be due to chance variations, since we found a significant relationship between macular thickness and age in all nine ETDRS areas and volume in all three measurements with the standard map for both eyes (all data not shown), that is a total of 60 regression analyses. We found no correlation with age and minimum foveal thickness. This may be due to the architecture of the foveola with elongated cones and fewer retinal cell layers compared with the rest of the macula. Also, the foveal minimum is calculated from only six A-scans. It is located at the concave bottom of the foveal pit, and small scan displacements will have a large effect on the measurements. We believe the size of our material, the low range of refractive errors, a quite wide age range and a homogenous Caucasian population explain our findings. Studies on different ethnic groups may well reveal differences in macular thickness. In fact, Poinoosawmy et al reported a difference in nerve fibre layer thickness between Caucasians and Afro-Caribbeans.15
In conclusion, we found a small but significant effect of age on macular thickness that should be kept in mind in clinical studies.
REFERENCE
Footnotes
Funding Supported by Stiftelsen Ögonklinikens vid Akademiska Sjukhuset i Uppsala Forskningsfond.
Competing interests None.
Ethics approval Ethics approval was provided by the ethics committee of Uppsala University.
Patient consent Obtained.