Article Text
PDF
Abstract
Aims To determine the relationship between the macular thickness and volume determined by spectral domain optical coherence tomography (SD-OCT) and the refractive error (RE) and axial length (AL).
Methods 48 eyes of 24 healthy Japanese subjects were examined. The REs (spherical equivalent) were measured with the Tonoref RKT-7000 autorefractometer (Nidek Inc, Aichi, Japan), and the ALs were determined by the IOL-Master (Carl Zeiss Meditec Inc, Dublin, California, USA). The thicknesses and volumes of the central, inner, and outer macular areas were measured by Spectralis SD-OCT (Heidelberg Engineering, Heidelberg, Germany). The correlations between the macular thickness and volume and the RE and AL were determined by linear regression analyses.
Results The average central foveal thickness (within 1 mm) was 278.7±10.8 μm and the volume was 0.22±0.01 mm3. The foveal thickness was negatively correlated with the RE and positively correlated with the AL (p<0.01 for both). Similar correlations were also found for the volume of the central macular areas.
Conclusions The significant correlation of the foveal thickness and the RE and AL in healthy eyes should be considered when these parameters are evaluated in eyes with retinal diseases.
- Axial length
- myopia
- retinal thickness
- spectral domain optical coherence tomography
- retina
- macula
- imaging
Statistics from Altmetric.com
Introduction
In highly myopic eyes, degenerative changes cause a thinning of the retina and atrophy of the choroid.1 The chorioretinal atrophy in the macula is one of the causes of the decrease of vision in these eyes, although the pathogenesis of the degenerative changes has still not been completely determined. Optical coherence tomography (OCT) is a non-invasive technique to measure retinal thickness,2 and the OCT images of the retina and optic disc of healthy and diseased eyes have been published.3–5
Higher resolution images of the posterior fundus have been obtained with the spectral domain OCT (SD-OCT).6 Because of the higher resolution with faster scanning speed and image processing, the SD-OCT images have less speckle noise than the conventional time domain OCT (TD-OCT) images.7–10 The Spectralis SD-OCT (Heidelberg Engineering, Heidelberg, Germany) has an eye tracking system that corrects for eye movements made during the scanning process.11 The use of eye tracking leads to better quality retinal images and better reproducible retinal thickness measurements.12
Two studies have reported that the macular thickness was not correlated with the refractive error (RE) or the axial length (AL).13 14 However, other studies have found that the fovea was thicker in myopic eyes.15–20 The purpose of our study was to investigate the relationship between the macular thickness and the RE and AL in myopic eyes with the more reliable Spectralis SD-OCT.
Methods
Subjects
Both eyes of 24 healthy normal Japanese volunteers (10 men and 14 women) who were examined between April and June 2008 at the Matsumoto Dental University Hospital were studied. The RE (spherical equivalent) and AL of both eyes were measured, but only the right eye was used in the statistical analyses. The exclusion criteria were; presence of media opacifications, history of ocular trauma, and any ocular disease affecting the cornea, lens, retina or optic nerve. The subjects were divided into three groups according to their RE: high myopia (<−6.0 diopters (D), n=11); low to moderate myopia (between −0.5 and −6.0 D, n=25); and approximately emmetropic (between −0.5 and +0.75 D, n=12).
Each subject was informed on the purpose of the experiments and the procedures to be used, and an informed consent was obtained. All of the procedures adhered to the tenets of the Declaration of Helsinki.
All subjects underwent a slit-lamp examination, fundus biomicroscopy and indirect ophthalmoscopy. The RE was measured with the Tonoref RKT-7000 autorefractometer (Nidek Inc, Aichi, Japan). The AL was measured with the IOL Master (Carl Zeiss Meditec Inc, Dublin, CA), which is based on laser Doppler interferometry (LDI). The eyes of all of the subjects were ocular disease-free other than the refractive error.
Optical coherence tomography
All OCT scans were performed with the Spectralis OCT (Heidelberg Engineering) with a scan rate of 40 000 A scans/s at a depth of 7 μm. The transverse resolution was 14 μm. The light source of the Spectralis OCT is a superluminescent diode with a peak output at 870 nm. The Spectralis OCT acquires transectional images and volume measurements of the retina, and confocal scanning laser ophthalmoscopic images of the fundus simultaneously. A real time eye tracker couples the SD-OCT scanner to the eye position and stabilises the OCT scan on the retina.
The macular thickness and volume were obtained with the volume scan mode. The retinal thickness was determined from a 30×20 degree raster scans consisting of 25 line scans, and the distance between the lines was 240 μm. At each position, 12 B-scans were averaged with the automatic real time mode to reduce speckle noise.
The mean retinal thickness and volume maps were determined for nine sectors in three concentric circles of diameters 1, 3 and 6 mm (see the Early Treatment Diabetic Retinopathy Study Report).21 The inner and outer rings were divided into four quadrants (figure 1). The foveal thickness was defined as the average thickness in the central 1 mm diameter.
Example of a retinal thickness measurement by Spectralis optical coherence tomography (OCT) in a healthy (non-myopic) control. A. Scanning laser ophthalmoscope fundus image showing the Early Treatment Diabetic Retinopathy Study (ETDRS) plot centred on the fovea. The sectors of the ETDRS map are numbered. Map of the fovea (area 1) and four quadrants of the inner (areas 2, 3, 4 and 5) and the outer macula (areas 6, 7, 8 and 9) areas. B. Corresponding retinal thickness (upper) and volume (lower) plot. Values are given in μm and mm3, respectively. C. The rectangle indicates the scanning area. 25 B-line scans were made. D. B-line scan centred on the fovea.
Statistical analyses
The spherical equivalent of the refractive error was used for the statistical analyses. The association between the macular thickness and volume and the RE and AL was determined by linear regression. A p<0.05 was considered to be statistically significant. The differences among the three diagnostic groups were evaluated by one-way analysis of variance. All data were analysed with statistical software (StatView V.5.0).
Results
The mean age of the subjects was 37.6±7.8 years with a range of 27 to 52 years. The mean RE was −3.30±3.07 D with a range of +0.75 to −10.50 D. The mean AL was 25.09±1.37 mm with a range of 22.49 to 28.21 mm. The values of the macular thicknesses measured by SD-OCT are shown in table 1. The mean foveal thickness was 278.7±10.8 μm, and the minimum foveal thickness was 229.4±14.9 μm.
Macular measurements in all subjects, and correlations between macular measurements and refractive error (spherical equivalent)/axial length
A significant correlation was found between the RE and the foveal thickness (r=−0.378, p<0.01) and between the AL and the foveal thickness (r=0.519, p<0.01; figure 2; table 1). Thus, the fovea was significantly thicker in eyes with higher myopia and longer axial lengths. In the highly myopic group, the outer temporal retina was thinner (r=0.285, p<0.05), and the ALs were negatively correlated with the thickness of the outer temporal area (r=−0.298, p<0.05). However in the other retinal sectors, no significant correlation was found between the RE and the AL and the macular retinal thickness.
Scatterplots of average central foveal (1 mm) thickness against refractive error (spherical equivalent) (A) and axial length (B).
In contrast to the time-domain OCT (Stratus OCT: Carl Zeiss Meditec, Inc, Dublin, CA), the minimum and total macular thickness cannot be measured automatically by the Spectralis OCT although they can be measured manually. However, we used the retinal volume measurements because reliable values can be obtained with the Spectralis OCT. The mean total volume for all of the normal eyes was 8.71±0.31 mm3 (table 1). The correlations between the average foveal volume and the RE and AL were similar to that for the foveal thickness. Scatterplots of the REs and the ALs as a function of the total macular volume are shown in figure 3 for all subjects. None of the measures of the macular volume was significantly correlated with the RE or the AL. In contrast, significant correlations were found between the foveal volume and the RE and the AL (figure 3).
Scatterplots of foveal (1 mm) volume against RE (A) and axial length (B).
Although the regional macular thicknesses and macular volumes were compared among the three groups, no significant differences were found in each group.
We then compared the mean foveal thickness and volume in the non-myopic group (n=11) to that in the high myopic group (n=12; table 2). The average fovea in the highly myopic group was significantly thicker than that in non-myopic group (284.8±9.5 μm vs 273.7±8.6 μm; p=0.009). In contrast to the foveal thickness, the nasal inner macula, inferior inner macula, and temporal outer macula in the highly myopic group were significantly thinner than those in the non-myopic group (table 2).
Macular thickness in non-myopic and high myopic group
Discussion
Our results showed that the central macula was significantly thicker in eyes with higher myopia and longer ALs. However, there have been reports that myopic eyes have thinner retina and sclera although the findings have not been consistent.13–20 Thus, Wakitani et al13 and Kelty et al14 have reported that the macular thickness was not significantly correlated with the AL, but several other studies have shown that the macula was thicker and was correlated with the RE in myopic eyes.15–20 This discrepancy is probably due to the more precise instruments that have been available only recently. The retina is scanned faster with the Spectralis SD-OCT, and the real-time eye tracking system reduced the noise caused by eye movements. A very good intra-observer reproducibility of retinal thickness measurements can be obtained by Spectralis OCT.12 Wolf-Schnurrbusch et al reported that the coefficient of variation for repeated measurements with the Spectralis OCT instrument was lower than that of other SD-OCT systems.22 Because of these features, the retinal thicknesses measured by the two ophthalmologists (AS and EF) in our study had excellent reproducibility.
LDI is a non-contact optical method that uses partial coherence interferometry to measure the AL, and it has higher resolution and is more reliable than conventional A-mode echography.23 However, the LDI measurements of the AL are reported to be approximately 0.2 mm longer than that with A-mode ultrasonographic measurements.24
To the best of our knowledge, our study is the first study using both Spectralis SD-OCT to measure retinal thickness and the IOL Master to measure the AL. Because of the improvements of these newer instruments, we were able to obtain more reliable results that are comparable to those reported for a smaller number of subjects.
The average foveal thickness measured by Spectralis SD-OCT was 278.7±10.8 μm. However, several studies have shown that the macula measured by SD-OCT was significantly thicker than that measured by TD-OCT.7–10 Thus, Legarreta8 showed that the average foveal thickness obtained by Cirrus SD-OCT was 266.2±22.7 μm which was significantly thicker than the 197.2±17.8 μm measured by the Stratus TD-OCT. The difference between the SD-OCT and TD-OCT measurements can be explained by the difference in the definition of retinal thickness. In TD-OCT, including the Stratus OCT, the retinal thickness is defined as the distance from the surface of the inner limiting membrane to the boundary of the inner segment/outer segment photoreceptor interface of the photoreceptor layer. In contrast, it is possible to delineate of the inner segment/outer segment interface and the retinal pigment epithelium by SD-OCT,25 and the definition of the posterior retinal boundary varies according to the kind of instrument. While the retinal pigment epithelium is set as the posterior retinal boundary in Cirrus SD-OCT, the Spectralis SD-OCT defines Bruch's membrane as the posterior boundary. This may explain why the macula in our subjects was thicker than that in other subjects determined by other SD-OCT system. The image obtained by the Spectralis SD-OCT is stabilised so that off-foveola fixation cannot explain the thicker macula in myopic eyes.
Our findings that the fovea was significantly thicker in eyes with longer axial lengths are in agreement with recent studies.15–20 Thus, Lim et al showed that the inner regions of the macula were thinner and the fovea thicker in patients with myopia.16 Our results also showed a similar tendency (table 2), but the reason for this relationship is unclear. Wu et al hypothesised that one possible mechanism was that the increased AL would result in a mechanical stretching of the retina and sclera,17 and the stretching of the retina would result in a panretinal thinning. On the other hand, the stretching and flattening tendency of the internal limiting membrane and the traction of the posterior vitreous result would elevate the fovea. The prevalence of macular holes and retinoschisis in highly myopic eyes is consistent with this hypothesis.
Another study showed that the photoreceptor outer segments were longer in a form-deprivation myopia model of tree shrews.26 Luo reported that the increase in foveal thickness might be due to pathological changes.18 Although the OCT images are not comparable to the histological images, further studies on myopic degeneration, macular retinoschisis, and macular hole with the Spectralis OCT may reveal the pathogenesis of myopia-related diseases.
The coefficients of correlation of the foveal thickness and axial length and RE were relatively low as opposed to that of the macular thickness. The reason for this is unclear, but the wide age range (27–52 years) might have lowered the correlation coefficients because Kanai et al reported that macular thickness decreases with age.15
Possible limitations of this study include the small sample size (48 eyes of 24 subjects), heavily weighted to younger subjects, lack of statistical power, gender and a single ethnicity.
In summary, there was a significantly positive correlation between average foveal thickness and volume and the refractive error, and a positive correlation between the refractive error and axial length and the foveal thickness. These variations in foveal thickness should be considered in evaluating the macular thickness in eyes with retinal diseases. Moreover, the type of OCT instruments used should be considered in evaluating the data, because the measurement of retinal thickness varies according to the type of OCT system used.
Acknowledgments
The authors thank Professor Duco Hamasaki for editing this manuscript.
References
Footnotes
Competing interests None.
Ethics approval This study was conducted with the approval of the Matsumoto Dental University.
Provenance and peer review Not commissioned; externally peer reviewed