Abstract
purpose. To study the function and morphology of the macula of the eye before and after the removal of unilateral idiopathic epiretinal membrane (ERM).
methods. Focal macular electroretinograms (fmERGs) elicited by a 15° stimulus were recorded in 37 eyes of 37 patients with a unilateral ERM. The amplitudes of the a- and b-waves and the oscillatory potentials (OPs) were compared with the corresponding waves in the normal fellow eyes before and after removal of the ERM. In 29 eyes followed up for more than 6 months after surgery, the fmERGs and foveal and parafoveal thicknesses, measured by optical coherence tomography (OCT), were evaluated.
results. Before surgery, the mean amplitudes of all components of the fmERGs were significantly smaller than in the fellow eyes, with the decrease largest for the OPs, followed by the b-waves and then the a-waves. The eyes with less severely reduced a-wave amplitude (>70% of the fellow eyes) had significantly lower b-wave to a-wave (b/a) ratios. After surgery, the amplitudes of the b-wave and OPs were still significantly smaller in the affected eyes. The mean foveal and parafoveal thicknesses were significantly less after surgery; however, the thickness was still more in the affected eyes. The decrease of the OPs remained after surgery and correlated with increased parafoveal thickness (r = −0.460, P = 0.011).
conclusions. The decreased fmERGs indicate that macular function is impaired in eyes with ERM. The decrease of the b-wave and OPs in the 29 eyes examined after vitrectomy may be due to the still thickened macular retina.
Abnormalities of macular function have been demonstrated in eyes with an idiopathic epiretinal membrane (ERM), by subjective tests and also by objective methods such as focal macular electroretinogram (fmERG)
1 and multifocal ERGs.
2 The characteristics of the macular dysfunction suggested that the dysfunction arises from an impairment of the inner retinal layers
1 and is similar to that found in eyes with aphakic cystoid macular edema.
3
To date, the effect of the removal of an ERM during vitrectomy on the recovery of the different components of the fmERGs has not been studied in detail.
1 In addition, the morphologic appearance of the macula has not been examined before and after the removal of the ERM, although morphologic changes have been reported to occur.
4 5 6
Optical coherence tomography (OCT) offers a method of obtaining quantitative measurements of macular structure,
7 8 and in eyes with ERM, an increased thickness of the macula has been demonstrated by OCT.
4 5 6
The purpose of this study was to analyze each component of the fmERG before and after the removal of the ERM. In addition, the relationship between the each fmERG component and the visual acuity or the OCT-determined thickness of the macula was analyzed before and after the removal of the ERM.
We studied 37 eyes of 37 patients in whom ERM was diagnosed in one eye, with a normal fellow eye. There were 13 men and 24 women (age range, 53–78 years; mean 66.4 ± 1.0 [±SE]). Patients with a secondary ERM or a significant cataract were excluded. Patients with other ophthalmic diseases in either eye or systemic disease that may influence the electroretinograms were also excluded. None of the patients had vitreomacular traction syndrome.
All the eyes underwent vitrectomy with successful removal of the ERM, and the 29 eyes that were under observation for more than 6 months (6–22 months, 10.9 ± 0.8) were evaluated by pre- and postoperative visual acuity, fmERGs, and measurement of the foveal and parafoveal thickness by OCT.
The surgical technique
9 10 11 consisted of a standard three-port pars plana vitrectomy, and the ERM was completely removed by retinal forceps. Although we did not perform histopathologic examination, the internal limiting membrane was unintentionally peeled off together with ERM in some cases. Of the 29 eyes, 16 eyes underwent concomitant intraocular lens surgery, 8 eyes underwent intraocular lens surgery after the first vitrectomy, and 5 eyes underwent vitrectomy only.
The standard Japanese visual acuity chart was used for visual acuity measurements, and the results were converted to Snellen visual acuity. The visual acuity was also expressed as the logarithm of minimal angle of resolution (logMAR) for statistical analyses.
To assess macular function, fmERGs elicited by a 15° stimulus were recorded in both eyes of the 37 patients.
12 The technique of recording fmERGs under direct fundus observation has been described in detail.
13 14 Briefly, an infrared television fundus camera, equipped with a stimulus light, background illumination, and fixation target, was used to stimulate and monitor the position of the stimulus on the macula. The size of the stimulus spot was adjustable, and we selected a 15° spot centered on the fovea. The background light was delivered to the eye from the fundus camera at a visual angle of 45°. Additional background illumination outside the central 45° produced homogeneous background illumination for nearly the entire visual field.
A Burian-Allen bipolar contact lens electrode was used for the ERG recordings and allowed not only a very low noise level but also permitted a clear view of the fundus displayed on a television monitor. The intensities of the white stimulus light and background light were 29.46 and 2.89 cd/m2, respectively.
After the patients’ pupils were fully dilated with 0.5% tropicamide and 0.5% phenylephrine hydrochloride, fmERGs were elicited by 5-Hz square-wave stimuli. A total of 512 responses were averaged by a signal processor. A time constant of 0.03 second (5.3-Hz low-cutoff) with a 100-Hz high-cutoff filter on one amplifier was used to record the a- and b-waves, and a time constant of 0.003 second (53-Hz low-cutoff) and a 300-Hz high-cutoff filter was used on a second amplifier to record the oscillatory potentials (OPs).
The amplitudes and implicit times of the a- and b-waves and the OPs of the fmERGs were analyzed. The a-wave amplitude was measured from the isoelectric line to the trough of the a-wave, and the amplitude of the b-wave was measured from the trough of the a-wave to the peak of the b-wave. The amplitude of each oscillatory potential (O1, O2, and O3) was measured from a line that was drawn as a first-order approximation between the troughs of successive wavelets. The amplitude of the sum of the O1, O2, and O3 amplitudes were used in the statistical analyses. Similarly, the implicit time of a- and b-wave were analyzed. And the b- to a-wave (b/a) ratio was also calculated in all eyes.
OCT was used to measure foveal and parafoveal thicknesses. The images were cross-sectional scans passing through the fovea horizontally and vertically. The foveal thickness was calculated as the mean of the vertical and horizontal thicknesses at the fovea. The parafoveal thickness was calculated as the mean thickness at four points, 1 mm nasally, temporally, superiorly, and inferiorly from the fovea in the OCT images that passed through the fovea.
This research was conducted in accordance with institutional guidelines and conformed to the tenets of the World Medical Association Declaration of Helsinki. After providing sufficient information on other treatment options including observation only, an informed consent for surgery, the fmERG, and the OCT examination was obtained from each patient.
The mean preoperative best corrected visual acuity (BCVA) in logMAR units was 0.38 ± 0.04 (mean ± SE, 20/48 Snellen acuity) in the 37 eyes with ERM. Of the 29 eyes under observation for more than 6 months after the removal of the ERM, 18 (62.1%) eyes had an improvement of logMAR by more than 0.2 log unit, and 11 (37.9%) eyes had no change. No eyes had decreased BCVA after surgery. The mean preoperative logMAR BCVA was 0.39 ± 0.04 (20/49 Snellen acuity), and the mean postoperative visual acuity was 0.09 ± 0.03 (20/24 Snellen acuity). This improvement in visual acuity was significant (P < 0.0001, Wilcoxon signed rank test).
The mean amplitude of the a-waves, b-waves, and OPs of the fmERGs recorded in the 37 eyes with ERM were significantly smaller than the corresponding waves in the normal fellow eyes (
P < 0.0001, Wilcoxon signed rank test;
Table 1 ). The preoperative relative amplitudes (affected eye and normal fellow eye) of the a-waves, b-waves, and the OPs were 75%, 69%, and 45%, respectively. These ratios showed that the OPs were the most affected component of the fmERGs.
In the 37 eyes, the mean b/a ratio in the affected eyes (2.49 ± 0.13) was not significantly different from that in the fellow eyes (2.64 ± 0.09). However, a careful examination of the b/a ratios showed that the size of the ratio depended on the a-wave amplitude. The ratios significantly decreased with increasing the a-wave amplitude (
r = −0.548,
P = 0.0003). To analyze this relationship more carefully, the eyes were divided into two groups: eyes with a-wave amplitude was more than 70% of the fellow eyes (22 eyes) and eyes with a-wave amplitude less than 70% of the fellow eyes (15 eyes). Typical examples of the two groups are shown as case 1 and case 2, respectively, in
Figure 1 . In the first group, there was a significant lower b/a ratio of 2.15 ± 0.15 in the affected eyes than the 2.61 ± 0.13 in the fellow eyes (
P = 0.003, Wilcoxon signed rank test;
Fig. 2 ). In the second group, the difference in the b/a ratio was not significant between the affected and normal eyes (2.99 ± 0.18 vs. 2.67 ± 0.10;
P = 0.08, Wilcoxon signed rank test).
We wanted to determine whether the preoperative visual acuity correlates with the degree of amplitude reduction of the a-wave, b-wave, or OPs and found that the correlations were not significant. However, the preoperative a-wave amplitude correlated significantly with postoperative visual acuity (
r = 0.543,
P = 0.0019;
Fig. 3 ).
The preoperative implicit times of the a- and b-waves in the affected eyes were significantly longer than those in the fellow eyes (P < 0.0001, Wilcoxon signed rank test).
In the 29 eyes that underwent surgery and were observed for more than 6 months, the mean amplitudes of the a-waves, b- waves, and the OPs were significantly larger after surgery (P = 0.0003, P = 0.0051, and P < 0.0001, respectively, Wilcoxon signed rank test). In these eyes, the amplitude of the affected eye versus the fellow eyes was 102% for the a-wave, 81% for the b-wave, and 71% for the OPs. The difference in the mean amplitude of a-wave of the affected eyes from that of the normal eye was not significant (P = 0.596, Wilcoxon signed rank test). However, the postoperative b-waves and OPs were still significantly smaller than their corresponding waves in the normal fellow eyes (P < 0.0001, P = 0.0001, respectively, Wilcoxon signed rank test). Only three eyes showed more than 90% recovery of the amplitudes of all components of the fmERGs.
In 17 of the 29 eyes under observation for more than 6 months and with less severely reduced a-wave amplitude, the mean b/a ratio in the fellow eyes was 2.77 ± 0.24, which was significantly higher than that in the affected eyes before surgery at 2.06 ± 0.19. After surgery, the b/a ratio was still significantly lower at 2.14 ± 0.17. (
P = 0.0004, Wilcoxon signed rank test). In the other 12 eyes with markedly reduced a-wave amplitudes, the mean b/a ratio before surgery was 2.96 ± 0.22 which then decreased after surgery to 2.21 ± 0.13. This ratio was significantly lower than that in normal fellow eyes (2.74 ± 0.14;
P = 0.0047, Wilcoxon signed rank test). The recovery of the fmERGs and macula morphology in the representative two cases in
Figure 1 are shown in
Figure 4 .
The postoperative implicit times of the a- and b-waves were shorter than the preoperative ones, although the differences were not significant. However, the implicit time of each component was still longer than that in normal fellow eyes (
P = 0.0002,
P = 0.0111, respectively, Wilcoxon signed rank test;
Table 1 ).
The OCT-determined foveal and parafoveal retinal thicknesses are shown in
Table 2 . The mean preoperative foveal and parafoveal retinal thicknesses in the affected eyes were significantly greater than in the normal fellow eyes (
P < 0.0001,
P < 0.0001, respectively, Wilcoxon signed rank test). After surgery, the mean foveal and parafoveal retinal thicknesses decreased significantly (
P < 0.0001,
P < 0.0001, respectively, Wilcoxon signed rank test); however, both the foveal and parafoveal thicknesses after surgery were still thicker than the normal fellow eyes (
P < 0.0001,
P < 0.0001, Wilcoxon signed rank test).
The foveal thickness of the affected eyes versus the fellow eyes were 255% before surgery and 166% after surgery. The comparable relative parafoveal thickness was 148% before surgery and 116% after surgery.
The most severely reduced component of the fmERGs after surgery was the OPs, and the amplitude of the OPs correlated with the postoperative parafoveal thickness (
r = −0.460,
P = 0.011;
Fig. 5 ).
Our findings confirm those in an earlier report from our laboratory that the preoperative a- and b-waves and the OPs in the affected eyes were significantly reduced, and the implicit times of all components were significantly delayed compared with those in the normal fellow eyes.
1 In the earlier study, we found that the degree of b-wave amplitude reduction correlated significantly with visual acuity, but in this study, visual acuity did not correlate with the degree of reduction in b-wave amplitude. This difference is probably due to the use in the present study of a larger spot size that covered a greater portion of the retina.
In both studies, the largest reduction was in the amplitude of the OPs. The similarity of the changes in the different waves of the fmERGs in this study support the earlier hypothesis that the pathophysiological changes in eyes with ERM may have resulted from macula edema, as in aphakic cystoid edema.
3
There was a greater reduction in b-wave than in a-wave amplitudes in both studies. However, we found a significant reduction of only the b/a ratio when the eyes with ERM were segregated into those with a relatively preserved a-wave (>70%). In contrast, the b/a ratio in the eyes with a greater reduction of the a-wave (<70%) was not significantly different from that in the normal fellow eyes. These results suggest that the reduction in the b-wave occurs before the reduction of the a-wave, and the a-wave then decreases as the disease progresses. Thus, there appears to be a dissociation in the course of reduction of the a- and b-waves
(Fig. 4) .
After surgery, the recovery of the a- and b-waves occurred in the reverse order as was reported in eyes with aphakic cystoid macular edema
3 and central serous chorioretinopathy.
15 After surgery, the b/a ratio in the eyes with relatively preserved a-waves were still significantly reduced and did not recover, but in the eyes with a more reduced a-wave, the b/a ratio was lower after surgery, because the recovery of the a-wave was more prominent than that of the b-wave. Again, these results suggest a dissociation of the a- and b-waves during the recovery processes.
Because our stimulus for the fmERGs was of long duration, the b-wave was recorded without contamination by the d-wave. According to the model by Bush and Sieving,
16 and Sieving et al.,
17 the a-wave originates mainly from the combined activity of the photoreceptors and the off-bipolar cells, and the b-wave is mainly shaped by the on-bipolar cells. Thus, our results suggest that in eyes with ERM, the inner retinal layer is predominantly impaired initially, followed by an impairment of the outer retinal layers. An alternative hypothesis is that the on-bipolar system may be impaired initially, followed by an impairment of the off-bipolar system and photoreceptor system. In any case, the degree of reduction of the a-wave appears to be a good indicator for the severity of the disease, and in fact, the preoperative a-wave reduction correlated with postoperative visual acuity.
In the postoperative analysis of 29 eyes, mean postoperative visual acuity improved significantly from 20/48 to 20/24. The OCT-determined foveal and parafoveal thickness decreased significantly, although both did not recover to that of the normal fellow eyes. A previous OCT study of eyes also reported that there were a few eyes that recovered to the normal concave foveal configuration after removal of an ERM,
6 and postoperative visual acuity was good. Our patients also showed good recovery of visual acuity. However, when the macular retinal function was assessed by fmERG, there were only three patients in whom amplitudes recovered to 90% or more of that in the normal fellow eye.
The mean amplitude of the a-wave after surgery in the ERM-affected eye was almost the same as that in fellow eyes, whereas the recovery of the b-wave amplitude was incomplete and the OPs were much more reduced. The percentage reduction of the OPs correlated significantly with the percentage increase in parafoveal thickness. However, further recovery of macular function may be expected after a longer follow-up period because the recording of the fmERGs after surgery was performed, on average, 10 months (range: 6–22 months) after surgery.
In conclusion, eyes with ERM have greater impairment of the inner retinal layers. The OCT-determined changes and fmERG-determined functional alterations after surgery suggest a delay or incomplete recovery of macular morphology and function, although good postoperative visual acuity was attained. This incomplete recovery may explain patients’ reports of blurred vision despite relatively good Snellen acuity. Future development of treatment for inner neural components should be helpful for quicker and better functional recovery.