Abstract
Purpose.:
To study the effect of epiretinal membrane (ERM) removal on the function and structure of the retina, and to determine whether the functional changes were correlated with the changes in the thickness of different retinal layers.
Methods.:
Focal macular electroretinography (FMERG) and spectral-domain optical coherence tomography (SD-OCT) were performed on 17 eyes of 15 patients before and after ERM surgery. The parafoveal retina was divided into an inner layer, a middle layer, and an outer layer in the OCT images. The thickness of each layer was measured before and after the ERM surgery. The a-wave, b-wave, and oscillatory potentials (OPs) of the FMERGs were analyzed before and after the ERM surgery.
Results.:
The thickness of the inner and middle retinal layers was significantly reduced after surgery (by 39% and 23%, respectively). The mean amplitudes of the b-waves and OPs at 6 months postoperatively were significantly larger than those recorded preoperatively (by 21% and 61%, respectively). The ratios of the pre- to postoperative b-wave and OP amplitudes were correlated with the thickness reduction of the middle retinal layer (b-wave, r = −0.51, P < 0.05; OPs, r = −0.82, P < 0.01).
Conclusions.:
The significant correlations between the reduction in the thickness of the middle retinal layer and increase in the amplitude of the b-waves and OPs suggest that the improvement of macular function after ERM peeling is due to the decrease in the thickness of the middle retinal layer.
An idiopathic epiretinal membrane (ERM) is a relatively common macular disease in aging patients.
1 An ERM on the macula causes traction on the retina, leading to a distortion of vision and/or a decrease of visual acuity. The surgical treatments for an ERM include vitrectomy with ERM peeling, and successful removal results in improved visual function.
To study an ERM or examine the effects of ERM removal on the retina, clinicians have used optical coherence tomography (OCT), which can obtain cross-sectional images of the retina with micrometer resolution and good repeatability. Optical coherence tomography is performed noninvasively and quantitative measurements of the different layers of the retina can be made. Earlier OCT studies have shown the morphologic features of an ERM.
2–12 The results of these studies have shown that the visual dysfunction in eyes with an ERM is due to a thickening of the macular area, and a reduction in the thickness after surgery is accompanied by improvements of macular function.
2,3,5,6,9 Most of the earlier studies use subjective tests, such as visual acuity,
3,5–8 Amsler grid,
12 and M-CHARTS,
11 to assess the retinal function of ERM patients. Other studies use objective methods including multifocal electroretionograms
13–15 and focal macular electroretionograms (FMERGs).
16–19
Because our laboratory believed that examination of the different components of the FMERGs was an informative way to evaluate the pathophysiology and function of the macular area, we have used FMERGs to analyze the functions of many retinal diseases.
16,17,19–25 We have found that the degree of reduction of the oscillatory potentials (OPs) of the FMERGs is relatively greater than the reduction of the a- and b-waves in eyes with an ERM.
16,17,19 If the sensory retina is divided into “inner” and “outer” layers, we have suggested that an ERM impairs the function predominantly of the inner retinal layer.
16,17,19
However, the OCT instruments at the time of those studies did not have enough spatial resolution to detect finer changes in the retinal structure. Recent advancements of OCT technology, for example, spectral-domain OCTs (SD-OCTs), have made it possible to view and measure the retinal structures more accurately with better resolution. This has allowed clinicians to measure the thicknesses of the different retinal layers.
8–12,26,27
The results of these earlier studies indicate that the visual acuity in eyes with an ERM is significantly associated with the thickness of the inner nuclear layer (INL).
9,12 However, other studies report that the visual acuity is associated with alterations of the outer retinal layer or the photoreceptors.
8,27 In addition, others have demonstrated that the visual function is associated with both the INL and photoreceptor misalignment.
10,11 One difficulty in interpreting the results of these earlier studies is the fact that the ERM surgeries are performed simultaneously with cataract surgery, and it is not possible to eliminate the additional effects of a clearer optical pathway. Thus, FMERGs became a valuable method to assess the retina after any type of treatment. Because the a- and b-waves, OPs, and other components of the FMERGs originate from different retinal layers, they can be used not only to analyze the macular function objectively but also to evaluate the function of the retina layer by layer.
Thus, the purpose of this study was to determine the relationship between the thickness of the different retinal layers and macular function after ERM surgery. To accomplish this, we recorded FMERGs before, and 3 and 6 months after, surgery and measured the thickness of the different retinal layers in the SD-OCT images recorded at the same times.
Fifteen consecutive patients who underwent surgery to remove an idiopathic ERM at Nagoya University Hospital between October 2010 and July 2012 by a single surgeon (HT) were recruited for this study. Patients with a secondary ERM, significant cataracts, glaucoma, and excessive myopia (more than −6.0 diopters or an axial length > 25 mm) were excluded. Seventeen eyes of the 15 patients (7 men and 8 women) who agreed to participate in this study and were willing to be followed up for at least 6 months after the surgery were enrolled. The mean age ± standard deviation was 68.9 ± 7.6 years with a range from 59 to 88 years.
Standard 3-port pars plana vitrectomy was performed on 6 eyes with a 23-gauge system and on 11 eyes with a 25-gauge system. After core vitrectomy, the ERM and internal limiting membrane were peeled with assistance of triamcinolone acetonide. In all cases, the membrane-peeling procedure was performed without the use of indocyanine green. Phacoemulsification with aspiration and intraocular lens implantation were performed during the vitrectomy in all of the eyes.
All of the patients signed an informed consent for the surgery and agreed to the recording of the visual acuity, OCT, and FMERGs during the follow-up examinations. The procedures used in this study were approved by the Institutional Review Board Committee of Nagoya University Graduate School of Medicine (approval No. 2013-0009). All of the procedures conformed to the tenets of the Declaration of Helsinki. A written informed consent was obtained from all the patients after they were provided with information on the procedures to be used.
The best-corrected visual acuity was measured before, and 3 and 6 months after, the surgery. A standard Japanese visual acuity chart was used, and the decimal BCVA was converted to the logarithm of the minimum angle of resolution (logMAR) for the statistical analyses.
The macular thickness was measured on the SD-OCT (Cirrus HD-OCT; Carl Zeiss Meditec, Dublin, CA) images before, and 3 and 6 months after, the surgery. The methods we used for the SD-OCT examinations have been described in detail.
28 Five-line vertical and horizontal raster scans of 6-mm length were made, and the scan through the fovea was used to measure the macular thickness. We measured the foveal thickness and parafoveal thicknesses. The parafoveal thickness was calculated as the mean thickness at 1.2 mm nasal, temporal, superior, and inferior to the fovea (
Fig. 1A). We selected these points because the foveal pit and the immediate surrounding area lacked the inner retinal layers, which made it not suitable to analyze the different retinal cellular layers. Segmentation of the retinal layers was done manually by one experienced masked operator. We divided the retinal layers of parafovea into 3 layers: (1) inner retinal layer consisting of the ERM, inner limiting membrane, and retinal nerve fiber layer (RNFL); (2) a middle retinal layer consisting of the ganglion cell layer (GCL), inner plexiform layer (IPL), inner nuclear layer (INL), and outer plexiform layer (OPL); and (3) an outer retinal layer consisting of the photoreceptor layer and retinal pigment epithelium (RPE). The total retinal thickness was defined as the distance from the vitreoretinal interface to the outer border of the RPE layer.
The inner segment/outer segment (IS/OS) junction line is a hyperreflexive line in the photoreceptor layer of SD-OCT images as shown in
Figure 1B. We assessed the integrity of the IS/OS line in the OCT images preoperatively in all of the patients. We examined the horizontal and vertical images through the fovea. When the IS/OS line in both images was detected to be continuous, we classified the IS/OS line to be “intact,” otherwise we classified it to be “disrupted.” We analyzed the relationship between preoperative IS/OS line status and the FMERG components.
Focal macular electroretinograms (ER-80; Kowa, Nagoya, Japan) were recorded before, and 3 and 6 months after, the surgery. The technique of recording FMERGs under direct fundus observation has been described in detail.
29,30 Briefly, after the patients' pupils were fully dilated with 0.5% tropicamide and 0.5% phenylephrine hydrochloride, a Burian-Allen bipolar contact lens electrode (Hansen Ophthalmic Development Laboratories, Iowa City, IA) was used to record the FMERGs. The size of the stimulus spot was 15° (
Fig. 1A), and the background light from the fundus camera (CF-60DSi; Canon, Tokyo, Japan) illuminated nearly the entire visual field. The luminance of the stimulus was 30 cd/m
2, and the background luminance was 1.5 cd/m
2. The position of the spot on the fundus was monitored during the recording with a modified infrared fundus camera (ER-80; Kowa). The responses were digitally bandpass filtered from 5 to 500 Hz for the a- and b-waves and from 50 to 500 Hz for the OPs. Five hundred responses were averaged at a stimulation rate of 5 Hz (Neuropack S1 MEB-9400; Nihon Kohden, Tokyo, Japan). The amplitudes of the a-waves, b-waves, and OPs and the implicit times of a-waves and b-waves of the FMERGs were analyzed. The a-wave amplitude was measured from the baseline to the first negative trough, and the amplitude of the b-wave was measured from the trough of the a-wave to the positive peak of the b-wave (
Fig. 1C). Similarly, the implicit time of the a-waves and b-waves was analyzed at the time of the trough of the a-wave and peak of the b-wave. For the OP amplitudes, we measured the amplitude of each OP from the trough to the peak (O1∼O3), and the sum of the O1, O2, and O3 amplitudes was used for the statistical analyses (
Fig. 1C).
31
Statistical analyses were performed by using the Spearman rank correlation tests, Wilcoxon signed rank tests, and Student's t-tests. The data were analyzed with Statcel software (2nd edition), which is an add-in module for Microsoft Excel (Statcel; OMS, Tokyo, Japan). A P < 0.05 was considered to be statistically significant.
The mean preoperative BCVA was 0.41 ± 0.05 logMAR units (mean ± SE). The BCVA was 0.13 ± 0.03 logMAR units at 3 months and 0.10 ± 0.03 logMAR units at 6 months. A visual acuity of 0.0 logMAR units corresponds to 20/20 vision on the Snellen chart, and smaller logMAR units indicate better visual acuity. The improvement in the BCVA was significant at both times (P < 0.001, Wilcoxon signed rank test).
The horizontal OCT images of 5 representative patients obtained preoperatively and at 6 months postoperatively are shown in
Figure 2A. The thickness of the macula decreased in all cases but the degree of reduction varied widely. Examination of the pre- and postoperative SD-OCT images showed that the structures of the outer retina were mainly preserved, but the shape of the inner and middle retinal layers were distorted even after the surgery.
The mean foveal thickness was decreased significantly 6 months after the surgery (
P < 0.05, Wilcoxon signed rank test;
Table 1). The mean parafoveal thickness of each retinal layer of the preoperative eyes, the normal fellow eyes, and eyes at 3 and 6 months postoperatively are presented in
Figure 3. All retinal layers in the preoperative eyes were significantly thicker than the corresponding layers of the normal fellow eyes (
P < 0.05, Student's
t-test).
After the surgery, the total, inner, and middle retinal layers were significantly thinner at 3 and 6 months than the preoperative thicknesses (
Fig. 3 and
Table 1;
P < 0.01, Wilcoxon signed rank test). The reduction occurred mainly during the first 3 months. The thickness of the outer retinal layer did not change until 6 months after the surgery (
Fig. 3 and
Table 1). There were no significant differences among the thicknesses between the 2 postoperative times.
Representative FMERGs from 5 patients recorded before surgery and 6 months postoperatively are shown in
Figure 2B. In all cases, the amplitude of the b-waves and OPs increased after surgery, especially in case 5. However, in most cases the FMERGs did not recover to the normal range. The mean amplitudes of the a-waves, b-waves, and sum of OPs and the mean implicit times of the a-waves and b-waves, recorded before surgery and 3 and 6 months postoperatively, are shown in
Table 1. The implicit time of b-waves was significantly shorter at 3 and 6 months after the surgery (Wilcoxon singed rank test). The amplitudes of the b-waves and OPs increased significantly at 6 months after surgery (Wilcoxon signed rank test) but that of the a-waves did not change significantly. The mean amplitude of the OPs at 6 months postoperatively was 1.6 times larger than the preoperative amplitude, and the mean amplitude of the b-wave was 1.2 times larger than that of the preoperative b-wave.
From the preoperative SD-OCT images, we classified 7 eyes (41%) as having an intact IS/OS line, while 10 eyes (59%) had disrupted IS/OS lines. There were no significant differences in visual acuity and amplitude of the a-wave between the 2 groups (Student's
t-test,
Table 2). However, there were significant differences in the amplitudes of the b-wave and OPs and the implicit times of the a- and b-waves between the 2 groups (Student's
t-test,
Table 2).
Correlation Between Retinal Layer Thickness and FMERG Components After ERM Surgery
We determined whether the change of each retinal layer thickness was significantly associated with the improvement of the BCVA and with each component of the FMERGs. We determined the correlation between preoperative and 6 months' postoperative values. An improvement of the BCVA was not correlated with the ratios of the 6 months' postoperative to the preoperative thicknesses at the parafovea for all retinal layers (
Supplementary Fig. S1). The 6 M-post/preoperative ratios of the b-wave amplitude were correlated with the post/preoperative ratios of the parafoveal thickness of the total retina (
r = −0.55,
P = 0.028, Spearman rank correlation test) and that of the middle retina (
r = −0.51,
P = 0.042;
Fig. 4). The 6 M-post/preoperative ratios of the sum of the OPs amplitude were also significantly correlated with the post/preoperative ratios of the total retinal thickness (
r = −0.74,
P = 0.003) and that of the middle retina (
r = −0.82,
P = 0.001;
Fig. 5). The coefficient of correlation was higher for the OPs than for the b-waves. In addition, the change of the b-wave implicit time was significantly correlated with the post/preoperative ratios of the middle retinal thickness (
r = 0.50,
P = 0.046;
Fig. 6).
Our SD-OCT results showed that the parafoveal thickness of the inner and middle retinal layers was significantly decreased but the outer retina did not change significantly after the ERM surgery. The mean amplitudes of the FMERG b-waves and OPs were significantly larger postoperatively than those recorded preoperatively. We then determined whether the changes in the thickness of each retinal layer were significantly correlated with the increase in the amplitudes and implicit times of the FMERG. Our analysis showed that the change of thickness in the middle layer, but not inner and outer layer, was significantly correlated with the increase of b-wave and OP amplitude and shortening of b-wave implicit time after ERM peeling. Thus, we confirmed that the improvement of FMERGs by ERM peeling was associated mainly with a reduction of the thickness of the middle retinal layer. Because the coefficient of correlation was higher for the relative amplitudes of the OPs than for the b-waves, improvements in the amplitude of the OPs were more strongly correlated with the change in the middle retinal layer thickness than were those of b-waves.
We have reported that the reduction in the amplitude of OPs was significantly greater than that of a-waves and b-waves of the FMERGs of patients with ERM. From these results, we suggest that the ERM probably induces damage mainly to the inner retinal neurons, including the retina from the INL to GCL, because these FMERG components are similarly changed in eyes with cystoid macular edema, which is known to damage the inner retina.
16 We have reported in another study that the reduction of the OPs of the FMERGs does not recover to the normal range after ERM surgery.
17 We assumed that the functional alterations in eyes with an ERM occurs mainly from greater impairment of the inner retina than the outer retina. However, in these studies, we have not been able to confirm the relationship between the thicknesses of the different retinal layers and the FMERG components because of the lower resolution of the OCT instruments at that time. The advancements of OCT technology have enabled us to confirm our earlier assumptions in this study.
A study of the origins of FMERG components, using pharmacologic techniques in monkey retinas, has shown that the b-waves of the FMERGs originate mainly from the ON bipolar cells with additional contribution from the OFF bipolar cell pathway.
32 On the other hand, the origin of the OPs seems to be more from the inner retina than the ON bipolar cells. The results of several full-field ERG studies have suggested that the retinal amacrine cells are the origin of OPs.
33–35 The cell bodies of the ON bipolar cells and amacrine cells are both located in INL, with those of the ON bipolar cells located more in the outer section of the INL and those of the amacrine cells located in the inner section of the INL.
36 We assumed that this difference of location would affect their susceptibility to ERM damage.
One of the commonly used methods to assess macular function is visual acuity. Several studies have examined the relationship between the retinal cell layers and the BCVA in eyes with an ERM.
8,9,12 Kim et al.
9 report that the preoperative parafoveal thicknesses of the GCL+IPL and INL are significantly associated with the preoperative BCVA. Other studies have shown that the degree of metamorphopsia is associated with the macular thickness of the INL.
10–12 These results suggest a similar pathophysiology in eyes with an ERM, indicating that an ERM affects the inner retina more than the outer retina. However, our results showed that the improvement of the BCVA postoperatively was not significantly correlated with the changes in the parafoveal thickness of any of the retinal layers. Even though the BCVA is a relatively easy way to assess the physiology of the fovea, it can be affected by other conditions, such as astigmatism, corneal haze, and cataracts. Because all of our patients underwent cataract surgery, the improvements in the BCVA may have been partially caused by the cataract surgery. This would then mask the relationship with the retinal structure. In addition, the number of patients may have been too low to detect a significant correlation between the thicknesses of the retinal layers and the BCVA.
Other studies have shown that the visual function is significantly associated with the alignment of the photoreceptors.
8,27 Our results also showed that the macular function of eyes that had disrupted IS/OS lines was impaired compared to that of eyes that had intact IS/OS lines. Because the ERMs are located on the surface of the retina, the ERM may disturb the inner retina predominantly in the early stages but the damage may progress to the outer retina in advanced stages.
Our results showed that the amplitudes of the FMERG components were not significantly correlated with the inner retinal thickness. However, this does not mean that the function of the retinal ganglion cells was preserved in eyes with an ERM because the FMERG components analyzed do not originate from the retinal ganglion cell activity.
There were 2 limitations in this study. The first limitation was that we could not measure the macular thickness by macular thickness map owing to the limitation of software. A macular thickness map may allow us to analyze the retinal cell layer 3-dimensionally and to evaluate the retinal structure more precisely. The second limitation was that we did not obtain microperimetry data. It might have helped us to get more information about the sensitivity of each point of retina analyzed by OCT.
In conclusion, the significant correlations between the thickness of the middle retinal layer and the amplitude of the b-waves and OPs suggest that the improvement of macular function after ERM peeling is mainly due to the decrease in the thickness of the middle retinal layer.
We thank Duco I. Hamasaki for discussions and editing the manuscript.
Supported by Grant-in-Aid for Scientific Research C (No. 25462709 [SU]) from the Ministry of Education, Culture, Sports, Science and Technology (
http://www.jsps.go.jp/). The authors alone are responsible for the content and writing of the paper.
Disclosure: N. Hibi, None; S. Ueno, None; Y. Ito, None; C.-H. Piao, None; M. Kondo, None; H. Terasaki, None