Central Retinal Thickness and Visual Acuity Outcomes After Vitrectomy for Epiretinal Membrane With and Without Internal Limiting Membrane Peel

Abstract

Purpose:

This work compares best-corrected visual acuity (BCVA) and central retinal thickness (CRT) outcome measures following pars plana vitrectomy (PPV) with and without internal limiting membrane (ILM) peel for epiretinal membrane (ERM).

Methods:

A retrospective cohort study was conducted of 114 eyes of 114 patients with visually significant ERM undergoing PPV with membrane peel (PPV/MP) at a single institution. CRT measurements were collected from Heidelberg and Zeiss spectral-domain optical coherence tomography imaging units. Results compared preoperative and postoperative CRT and BCVA among patients with at least 60 days of follow-up.

Results:

A total of 114 eyes with ERM met the inclusion criteria. Eighty-one eyes (71%) underwent PPV/MP with ILM peel and 33 eyes (29%) underwent PPV/MP without ILM peel. There was no statistically significant difference between preoperative CRT between the 2 study groups (P = .95). The mean follow-up time of eyes undergoing PPV/MP without ILM peel and with ILM peel was 391 days and 319 days, respectively. There was no statistical difference in the mean reduction of CRT between the 2 groups (P = .470). Both groups had improvement of BCVA following surgery and there was no statistical difference in final logMAR visual acuity when comparing the 2 groups (P = .738).

Conclusions:

There was an overall improvement of CRT and BCVA among patients undergoing PPV/MP for ERM with or without ILM peel. There was no statistically significant difference in the final BCVA or CRT between the 2 groups.

Introduction

Epiretinal membranes (ERMs) are proliferative fibrocellular membranes found on the inner retinal surface. ¹ Optical coherence tomography (OCT) has improved our ability to identify the preretinal membrane and associated retinal structure changes, including increased central retinal thickness (CRT), loss of foveal contour, foveal cystoid changes, lamellar macular holes, macular pseudoholes, and photoreceptor defects. ² Increased CRT has been shown to negatively affect best-corrected visual acuity (BCVA) in eyes with ERM. ^2,3 In addition to decreased VA, patients with ERM may present with metamorphopsia, photopsia, macropsia, and less commonly, monocular diplopia. ⁴ Pars plana vitrectomy with membrane peel (PPV/MP) is the primary intervention for patients with vision loss associated with metamorphopsia.

A variety of surgical techniques have been used. The basic procedure involves PPV with peeling of the ERM from the inner surface of the retina to release tractional forces on the retina to preserve the structural architecture of the underlying retina. Many surgeons have adopted the use of vital dyes, such as indocyanine green (ICG), to identify and peel the internal limiting membrane (ILM) in addition to the ERM. This technique has been reported to reduce ERM recurrence by eliminating the scaffold used for proliferation by the fibrocytic membrane and further reduce preretinal traction. ^{5 -7} Studies have reported better final VA outcomes in addition to decreased recurrence of ERM when peeling the ILM with the ERM, compared with ERM peeling alone. ^{5 -7} In contrast, Ahn et al demonstrated better visual outcomes at 1 month with ERM removal without ILM peeling, but no significant difference in VA outcomes or recurrence rates at subsequent visits were found. ⁸ Currently, there is a lack of convincing evidence that favors VA and recurrence outcomes between the 2 surgical subgroups.

The purpose of this study is to determine whether a significant difference in VA outcome exists in patients undergoing ERM peeling with or without ILM peeling, and to determine if there is a difference in postoperative CRT between the 2 groups.

Methods

A retrospective review of medical records was performed to identify consecutive patients who underwent PPV with and without ILM peel for idiopathic ERM between January 1, 2010, and May 31, 2018. Exclusion criteria included individuals with fewer than 60 days of postoperative follow-up and those with lack of preoperative or postoperative OCT. Patients were also excluded if they had a previous diagnosis of macular hole, tractional retinal detachment, diabetic macular edema, or if they had any previous retinal surgical procedures. Surgery was performed with 23- or 25-gauge PPV, and the decision to stain and peel the ILM was at the discretion of the surgeon. In all cases that underwent ILM peel, the ILM was stained with a 0.1% concentration of ICG diluted in dextrose solution. Trypan blue was used in some cases to help identify the ERM when not peeling the ILM. No cases with postoperative complications, such as retinal detachment, were included in this analysis.

Spectral-domain OCT imaging was obtained on the Heidelberg Spectralis SD-OCT (Heidelberg Engineering) and Zeiss Cirrus HD-OCT (Carl Zeiss Meditec Inc). CRT values were obtained from the Early Treatment Diabetic Retinopathy Study’s average central subfield thickness cube scan or from a horizontal raster image through the fovea if the cube scan’s Early Treatment Diabetic Retinopathy Study plot was not available. BCVA was recorded at each visit and converted to logMAR for statistical analysis.

All statistical populations were tested for normativity and the F test was performed to determine whether the variances were equal. In cases of paired data with equal variance, the paired t test was performed; in cases of unequal variance the Wilcoxon signed rank test for paired data was performed. For comparison between populations with unpaired data and equal variances, the unpaired t test was performed; in cases of unequal variances the Mann-Whitney Wilcoxon test for unpaired data was performed. An experiment-wise level of 0.05 was used to define positive findings.

Results

A total of 114 individuals met our inclusion and exclusion criteria; 81 underwent ILM peel (71%) and 33 did not (29%). The baseline CRT of the ILM peel group was 497.32 µm, while that of the group without the ILM peel was 498.30 (P = .949). The baseline BCVA of the ILM peel group was 0.48 logMAR, while that of the no–ILM peel group was 0.51 logMAR (P = .425). Fifty of the 81 patients who underwent ILM peel were phakic at baseline (62%), while 22 of the 33 participants who did not undergo ILM peel were phakic at baseline (67%).

Postoperative improvement in mean CRT was demonstrated within both groups. The ILM peel group’s CRT improved to 381.18 µm (P < .001), while that of the group that did not have an ILM peel improved to 374.56 µm (P < .001). There was no significant difference between these 2 groups’ final CRT (P = .979). There was no significant difference in the change in CRT from preoperative to postoperative measurements between the ILM peel group (–122.8 µm) and the group that did not receive an ILM peel (–116.9 µm) (P = .470). There was no significant difference in the postoperative interval at which final CRT was compared. The mean number of days between surgery and when the CRT was last measured was 319 days for the ILM peel group, while that of the group that did not receive an ILM peel was 391 days (P = .343).

Significant improvement in BCVA was observed within each group. The BCVA of the group that received an ILM peel improved to 0.29 logMAR (P < .001), and that of the group that did not receive an ILM peel improved to 0.31 logMAR (P < .001). There was no significant difference in the final BCVA between the 2 groups (P = .738). In addition, there was no significant difference in the change in BCVA from preoperative to postoperative measurements between the ILM peel group (–0.19 logMAR) and the group that did not receive an ILM peel (–0.22 logMAR) (P = .578). Fifty-nine individuals were pseudophakic (73%) at follow-up among the participants who underwent ILM peel, compared with 29 individuals (88%) who did not undergo ILM peel. There was no significant difference in the postoperative interval at which final BCVA was compared. The mean number of days from surgery to when the BCVA was last measured was 392 days for the ILM peel group, while that of the group that did not receive an ILM peel was 399 days (P = .176).

The final visit at which VA was assessed was different from when the final CRT was assessed because BCVA was checked at every visit, whereas an OCT was not obtained at every visit. There were no cases of a second PPV/MP for recurrent ERM in our study.

Conclusions

The results of this study show no significant difference in the change in CRT or VA of patients whether they underwent PPV/MP with or without ILM peel. Ahn and colleagues showed similar results with no difference in final postoperative VA or recurrence rate between the groups; however, there was better VA at the 1-month postoperative visit in eyes that underwent PPV/MP without ILM peeling. ⁸ Eyes that underwent ILM peeling were found to have structural changes in the outer retina on OCT during the 1-month visit, likely contributing to this observed difference in VA. However, the structural differences noted on OCT between the 2 groups were no longer present at subsequent visits. ⁸

Chang et al demonstrated no statistical difference in postoperative BCVA or final CRT between the 2 groups but did note that patients who underwent ERM peeling without ILM peeling had a greater proportional decrease in central macular thickness compared with the group with ILM peeling. ⁹ Additionally, Sivalingam et al showed that eyes had worse visual outcomes when there was histological confirmation of ILM with ERM specimens. ¹⁰ In contrast, multiple studies have shown better visual outcomes when the ILM is removed. ^5,6

The PPV/MP procedures included in this study were performed by multiple surgeons of the University of Wisconsin Department of Ophthalmology and Visual Sciences Retina Service. All vitrectomies were performed using 23- or 25-gauge instrumentation, but the decision to peel the ILM was the individual surgeon’s decision. In all cases in which the ILM was peeled, ICG staining was used to identify the ILM.

Owing to the retrospective nature of this study, selection bias may exist, as the decision to peel the ILM was not randomized in this study. It is unknown whether surgeons were inclined to peel the ILM in more advanced cases of ERM, where fundus examination or OCT may have demonstrated more severe traction, and reserved MP alone without ILM peel for milder cases of ERM. There was no statistical difference in preoperative CRT between each group, suggesting that there was no obvious difference in the preoperative severity.

In vitro studies have demonstrated ICG retinal and retinal pigment epithelial toxicity and have suggested that use of ICG intraoperatively may be associated with poorer visual outcomes. ¹¹ Clinical studies have investigated the effects of ICG-assisted ILM peeling on functional and structural outcomes. Haritoglou and colleagues demonstrated poorer VA and visual field outcomes in patients with ICG-assisted peeling compared with patients who underwent ILM peeling without the use of ICG. ¹² They also noted increased amounts of cellular elements adjacent to the retinal surface of the ILM in patients with ICG-assisted peeling. ¹² All ILM peeling in our study used ICG to stain the ILM. Our study does not suggest clinically evident toxicity involved with ILM peeling or the use of ICG, as there was no difference in VA or CRT outcomes when comparing the 2 surgical groups.

Cataract formation is an important factor in evaluating final visual outcomes in surgical studies. In our study 12% of patients without ILM peel were phakic at their final visit, whereas 27% of patients with ILM peel were phakic at their final visit. Owing to the higher percentage of phakic patients in the group who underwent ILM peeling, the final VA of this patient subset may have been affected by the presence of cataract.

Previous studies have shown less ERM recurrence following PPV/MP with ILM peeling compared with no ILM peeling. ^{4 -6} Park et al showed a recurrence rate of 21% in eyes without ILM peeling at the initial surgery, whereas none of the eyes with ILM peeling had recurrence. ⁶ Kwok et al showed similar findings, with recurrence in 3 of 17 eyes without ILM peeling (17.6%) vs recurrence in with none of the 25 eyes that underwent ILM peeling. ¹³ In both studies, recurrence was defined as postoperative recurrent macular pucker or persistent contraction of the ILM on biomicroscopy. OCT was not used to quantify or confirm the recurrence. In our study, we were unable to determine from postoperative notes whether there was biomicroscopic evidence of recurrence; however, there were no eyes that underwent a second MP or were found to have a greater than 10% increase in CRT postoperatively that would suggest ERM recurrence.

In conclusion, there was an overall improvement of CRT and BCVA among patients undergoing PPV/MP for ERM with or without ILM peel. There was no statistical difference in the final BCVA or CRT changes between the 2 groups. Peeling of the ILM with adjuvant ICG does not appear to be either dangerous or advantageous in the surgical treatment of ERM.

Although there were no differences in the final BCVA, change in BCVA, or CRT outcomes between each study group, noninferiority cannot be determined based on these measurements alone. This study is further limited by its retrospective nature. Surgeon outcomes may differ because of individual surgeon’s preferences and the specific needs of each patient.