This study followed the ethical standards established in the Declaration of Helsinki and was approved by the institutional review board of Inje University Busan Paik Hospital. We retrospectively reviewed the medical records of consecutive patients who underwent vitrectomy for idiopathic ERM between July 2007 and April 2010.

Patients were included in the study if they were followed-up for at least 12 months. The patients with previous history of vitrectomy, other accompanying macular diseases, retinal detachment, retinal vascular disease, endophthalmitis, or diabetic retinopathy were excluded from this study. Initially, 103 eyes of 99 patients who underwent vitrectomy for ERM during the study period were reviewed. Finally, 43 eyes of 43 patients who met the criteria were included for data analysis.

All procedures were performed by a single surgeon (IHY) at the Inje University Busan Paik Hospital, Busan, Korea. A standard 20-gauge pars plan vitrectomy was performed. Phacoemulsification and intraocular lens implantation through a 3.0 mm temporal clear corneal incision was performed before vitrectomy, if indicated.

After core vitrectomy, triamcinolone acetonide (TA) was injected into the vitreous cavity to determine if posterior vitreous detachment had occurred. For cases without posterior vitreous detachment, a pick or a vitreous cutter switched to the vacuum mode was used to induce posterior vitreous detachment. Then, TA was injected onto the ERM to better visualize the boundary of the peeled ERM. Intraocular forceps were used to grasp and remove the ERM.

Since November 2009, we began to use a double staining technique using ICG dye (Diagnogreen; Daiichi Pharmaceutical, Tokyo, Japan) to completely peel the ILM from underneath the ERM. This technique involved injecting 0.5% ICG dye onto the macula right after the previously described procedure to determine if any residual ILM remained. In the cases that it was present, we removed the residual ILM, which was about two disc diameters in size.

Best-corrected Snellen visual acuity and optical coherence tomography (OCT) data were collected at baseline and postoperative 3, 6, and 12 months. Stratus OCT (Carl-Zeiss Meditec, Dublin, CA, USA) or Cirrus HD-OCT (Carl-Zeiss Meditec) was used to acquire central macular thickness.

Preoperative and postoperative multifocal ERG data were also compared. Multifocal ERG was performed using RETIscan (Roland Consult Elektrophysiologische Diagnostik Systeme, Wiesbaden, Germany) following the manufacturer's instructions. Briefly, pupils were maximally dilated, and an ERG jet was used as an electrode. The stimulus consisted of an array of 61 hexagons scaled with eccentricity. The viewing distance was 28 cm, which covers approximately 30° of visual angle. Signals were band pass filtered (5-100 Hz) and amplified (gain, 100,000). The mean simultaneous response component for the first-order kernel was recorded. The averages of responses recorded during eight cycles were calculated for each subject with 61 hexagons and analyzed using the RETIscan software, version 3.15. The amplitude and the latency of the first positive peak (P1) of the ring 1 area were collected for comparison. Perioperative and postoperative complications including punctate retinal hemorrhages during ERM peeling were reviewed.

Patients were divided into two groups. Group A consisted of patients where simple ERM removal using only triamcinolone was performed. Group B consisted of patients where the double staining technique using ICG dye after the TA-assisted ERM peeling and complete ILM peeling was performed. Wilcoxon's signed rank test was used to compare the baseline and the postoperative data at each time point. Mann-Whitney's U-test was used to compare numerical data between the two groups. A p-value of <0.05 was considered statistically significant.

Of the 43 patients included in this study, 17 were male (39.5%) and 26 were female (60.5%). The average age of the patients was 64.6 years (range, 51 to 75 years). Nine eyes (20.9%) were pseudophakic at baseline (Table 1). Vitrectomy alone was performed in five pseudophakic eyes and in 18 eyes with a clear lens in group A and in four and 16 such eyes, respectively, in group B. Combined vitrectomy and cataract surgery was performed in one eye with lens opacity in group A.

Overall, the mean best-corrected Snellen visual acuity improved from 0.35 logMAR at baseline to 0.52 logarithm of minimum angle resolution (logMAR) at postoperative month 12. This change was significant (p = 0.001). In group A, visual acuity changed from 0.35 logMAR at baseline to 0.40, 0.43, and 0.50 logMAR at postoperative months 3, 6, and 12, respectively, and in group B from 0.44 to 0.56, 0.46, and 0.54 logMAR, respectively. Postoperative visual acuity improvement at postoperative month 12 was statistically significant in both groups (Wilcoxon's signed rank test, p = 0.011 in group A and 0.017 in group B). The difference between the groups at postoperative month 12 was not significant (Mann-Whitney's U-test, p = 0.740) (Table 2).

Decrease in the visual acuity at postoperative month 12 was observed in four patients (two in group A and two in group B). The cause of the decrease in visual acuity was attributed to the development of a cataract. Of the four patients, three received cataract operation and showed an increase in visual acuity of more than two lines compared to the baseline acuity.

Overall, central macular thickness measured by OCT changed from 471.5 ± 101.0 µm at baseline to 307.8 ± 70.3 µm, 298.0 ± 59.1 µm, and 289.6 ± 57.2 µm at postoperative months 3, 6, and 12, respectively. In group A, it changed from 461.0 ± 99.7 µm at the baseline to 310.4 ± 71.2 µm, 293.2 ± 56.2 µm, and 285.4 ± 51.4 µm at postoperative months 3, 6, and 12, respectively. In group B, it changed from 483.7 ± 103.7 µm at baseline to 304.8 ± 71.0 µm, 303.8 ± 63.7 µm, and 294.9 ± 64.8 µm at postoperative months 3, 6, and 12, respectively (Table 3). The change from baseline to postoperative month 12 was statistically significant in both groups (p = 0.000 in group A and 0.000 in group B). However, the difference between the groups at each time point was not statistically significant.

Multofocal-electroretinogram (mfERG) data were available for 23, 12, and 13 patients at baseline and postoperative months 6 and 12, respectively, in group A. In group B, data was available for 19, 8, and 9 patients at baseline and postoperative months 6 and 12, respectively. In group A, the P1 amplitude (nV/deg²) at ring 1 changed from 112.6 ± 38.8 to 122.3 ± 34.2 and 123.7 ± 46.2 at postoperative months 6 and 12, respectively. In group B, it changed from 106.8 ± 40.6 to 139.5 ± 55.2 and 121.1 ± 38.5 at postoperative months 6 and 12, respectively. Comparison of the P1 amplitude within the group at each time point and between the groups did not show any significant differences (Table 4).

Overall, in group A, the P1 latency (ms) at ring 1 changed from 40.1 ± 3.5 to 41.5 ± 4.7 and 41.5 ± 2.9 at postoperative months 6 and 12, respectively. In group B, it changed from 39.9 ± 6.9 to 43.6 ± 4.6 and 39.8 ± 5.7, respectively. Comparison of the P1 latency within the group at each time point and between the groups did not show any significant differences (Table 5).

In the 20 eyes from group B in which ICG staining was used to determine if any residual ILM existed after removal of the ERM, 12 eyes (60%) demonstrated no staining of ICG at the site of ERM peeling. This was highlighted by the surrounding ICG stained area.

Overall, punctate retinal hemorrhage was observed right after ERM peeling in 27 eyes (62.8%); 14 eyes (60.9%) in group A and 13 eyes (65.0%) in group B. In group B, all 12 eyes in which the ILM had already been removed together with the ERM, as confirmed by ICG double staining, showed punctate retinal hemorrhage (Table 6).

Serious complications including retinal detachment were not observed in either of the groups (Table 7). Vitreous hemorrhage occurred in one patient in group B and was absorbed without sequelae during the first week after surgery. A cataract developed in five and six eyes in groups A and B, respectively. None of the patients in this study exhibited recurrence of the ERM.