PATIENTS AND METHODS

This was a retrospective study with examination of medical records from 61 eyes of 59 consecutive patients who underwent ERM surgery between November 2004 and April 2006, and were followed for 6 months after surgery. This study was approved by the Institutional Review Board and performed in accordance with the ethical standards of the 1989 Declaration of Helsinki. Written informed consent was obtained from all patients enrolled in this study. Patient demographics, preoperative diagnosis, preoperative best-corrected visual acuity (BCVA) were all recorded. BCVA was converted to the logarithm of minimal angle of resolution (logMAR) BCVA for analysis.

All patients underwent standard three-port vitrectomy, removal of the posterior hyaloid and ERM removal by intraocular forceps. After ERM removal, 0.2 ml of 0.25 mg/ml BBG solution was injected into the vitreous cavity for staining the ILM and immediately washed out by irrigation. The residual ILM was dyed blue and visualised clearly as previously reported.12 The residual ILM pattern was divided into three groups: (1) residual type (ILM mostly remained) (fig 1A), (2) half type (approximately half of ILM remaind) and (3) no residual type (ILM mostly removed with ERM) (fig 1B). If ILM remained, residual ILM was removed before the conclusion of surgery in all cases.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

(A) Representative intraoperative findings of group 1: residual type. After epiretinal membrane (ERM) removal with triamcinolone acetonide (TA) (arrows show the ERM peeling area), Brilliant Blue G (BBG)-stained internal limiting membrane (ILM) that dyed blue was observed in almost the same area. ILM mostly remained. (B) Representative intraoperative findings of group 3: no residual type. After ERM removal with TA (arrows show the ERM peeling area), BBG-stained ILM that dyed blue was observed in only a limited area (arrowheads). ILM had been mostly peeled with ERM.

The correlation between the degree of ERM and residual ILM range was evaluated. In this study, ERM grading was evaluated by preoperative BCVA. Recurrence of ERM was defined as biomicroscopic evidence of a recurrent macular pucker or persistent contraction to the ILM and retinal vessels.2

The ILM specimens were immunohistochemically examined. Surgically removed ILMs were immediately fixed in 4% paraformaldehyde at 4°C for 24 h and subjected to whole-mount immunohistochemistry. Fixed ILMs were rinsed with phosphate-buffered saline (PBS) and permeated in 100% methanol at room temperature for 10 min. The samples were blocked using 5% skim milk in PBS for 1 h at room temperature, incubated in primary antibody against glial fibrillar acidic protein (GFAP) (dilution, 1:400, rabbit IgG; Dako, Glostrup, Denmark) at 4°C for 24 h and then incubated with AlexaFluor546 (dilution, 1:100; goat antirabbit IgG, Molecular Probes, Eugene, Oregon) for 30 min at room temperature. Finally, nuclear stain was performed using Hochest33342 (dilution, 1:400; Molecular Probes). The samples were extended on the slide glass using 26G micro hook needle, mounted in Crystal/Mount (Biomedia, Foster City, California) and subjected to fluorescent microscopy (BZ-9000; KEYENCE, Osaka, Japan).

The non-parametric Mann–Whitney U test was used to assess continuous variables between groups 1 and 3. Categorical variables were compared using the Fisher exact test. Group 2 was excluded from analysis because the number of patients was too small (n = 3). A two-sided p value of <0.05 was considered to be statistically significant.

RESULTS

The patients’ ages ranged from 42 to 83 years (mean (SD), 68.0 (7.8) years). There were 33 females (54.1%) and 28 males (45.9%). Forty-six cases (75.4%) of ERM were classified as primary idiopathic ERMs, and 15 cases (24.6%) were secondary (table 1). The cause of secondary ERMs consisted of nine patients with uveitis (15.0%), three patients with diabetic retinopathy who had previous pan retinal photocoagulation (4.9%), two patients with retinal tear (3.3%) and one patient with trauma (1.6%). The mean preoperative logMAR BCVA was 0.42 (0.31) (mean (SD)). No intraoperative or postoperative complications relating to the use of BBG was observed.

After ERM removal, the residual ILM pattern was divided into three groups (fig 1A,B). Twenty-eight eyes (45.9%) were of the residual type. Three eyes (4.9%) were of the half type, and 30 eyes (49.2%) were of no residual type (table 2). The mean preoperative logMAR BCVA was 0.39 (0.32) (mean (SD)), 0.70 (0) and 0.41 (0.31), respectively. There was no significant correlation between the degree of ERM evaluated by preoperative VA and the residual pattern of ILM (Mann–Whitney U test, p = 0.77).

Thick ERMs with severe contraction of the retina were observed in five eyes (17.9%) of the 28 eyes in the residual type, no eye of the three eyes in the half type and nine eyes (30.0%) of the 30 eyes in the no residual type. There was no significant association between the thick ERMs and non-thick ERMs in the residual ILM pattern (Fisher exact test, p = 0.36).

In idiopathic ERMs, the residual type was observed in 20 eyes (43.5%) of the 46 eyes, the half type was observed in three eyes (6.5%), and no residual type was observed in 23 eyes (50.0%). In secondary ERMs, the residual type was observed in eight eyes (53.3%) (uveitis four, retinal tear one, diabetic retinopathy two, trauma one) of the 15 eyes, the half type was not observed, and no residual type was observed in seven eyes (46.7%) (uveitis five, retinal tear one, diabetic retinopathy one). No significant association was found between idiopathic ERM and secondary ERM with the residual ILM pattern (Fisher exact test, p = 0.39).

Two eyes (3.3%) of the 61 eyes had no posterior vitreous detachment (PVD) before operation. One was idiopathic ERM, and another was secondary ERM accompanied by diabetic retinopathy. Both were of the residual type. The former showed a round defect of posterior vitreous cortex after surgical PVD leaving ERM remaining on the macula, and the latter showed PVD along with ERM without a round defect of the posterior vitreous cortex.

Flat-mount immunohistochemical examination of residual ILMs after ERM removal revealed that many cells remained on the residual ILMs in all 10 specimens examined. There were both immunopositive and immunonegative cells against GFAP antibody in all 10 cases examined (fig 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Representative fluorescent microscopic photograph of surgically excised residual internal limiting membrane (ILM). Nuclei were recognised as red, and glial fibrillar acidic protein was recognised as green. Many cells were observed on residual ILM after epiretinal membrane removal. (original magnification, ×4). Scale bar = 500 μM. A magnified view is shown in the upper-left corner (scale bar = 50 μM).

There was no recurrence needing surgical treatment within 6 months of surgery.

DISCUSSION

The recurrence of ERM sometimes becomes problematic because it causes visual disturbance and occasionally requires reoperation. There are several reports indicating that the peeling of ILM is useful to reduce the recurrence rate of ERM.24 In this study, widespread residual ILM after ERM peeling was observed in 51% combining the residual type and half type. In half of the cases, ILM remained broadly after ERM removal. In a previous study, Gibran et al reported that ILM specimens after ERM peeling contained ERM in six (37.5%) of 16 or cellular remnants in the six of 16.13 In ILM specimens after ERM peeling, Kwok et al reported that 11 (61.6%) of the 18 eyes showed various amounts of microscopic ERM.6 However, there are no reports demonstrating the residual ILM after ERM peeling with flat-mount immunohistochemistry as demonstrated in this study.

Our flat-mount immunohistochemical examination of residual ILMs after ERM removal revealed that, in all cases examined, many cells remained on the surface of the residual ILM. Our findings indicated that remnant cells appeared to be present on the residual ILM more frequently as compared with previous reports.613 There were both immunopositive and immunonegative cells against GFAP antibody. In previous studies, it was suggested that the main component of ERMs was glial cells on the grounds of morphological analyses.14 Okada et al reported that idiopathic ERM specimens involved ILM in 10/15 cases, indicating that ILM is left after ERM peeling in many cases in consistence with our results. They also demonstrated that GFAP positive cells were found in 3/8 specimens examined.15 However, our result suggested that GFAP negative cells also might play a role in the formation of ERM. We previously reported that hyalocytes, considered to be one of the macrophage lineage and mainly existing in the posterior cortical vitreous,16 appeared to be associated with the pathogenesis of various vitreoretinal interface diseases including proliferative vitreoretinopathy with its potential of myofibroblastic transdifferentiation.17181920 It is theoretically possible that the hyalocytes involved in residual posterior hyaloid after PVD might be one of the cells of importance in idiopathic ERM. Sebag et al indicated that vitreoschisis, the separation of posterior vitreous cortex, in the macula area causes ERM formation. If the vitreous separation level occurs in the anterior area, the remaining layer of vitreous will involve hyalocytes and be relatively thick, hypercellular and have contractile potential.21222324 Okada et al also reported that type II collagen was involved in idiopathic ERM specimens, indicating that the vitreous component contributes to the formation of idiopathic ERM.15 We also postulate that hyalocytes might be involved in the remnant cells on residual ILM and be associated with the formation of recurrent ERM. Maguire et al reported that myofibroblasts were present with increased frequency in recurrent membranes compared with studies of idiopathic ERMs.5 Fibrocytes, myofibroblasts, fibrous astrocytes and retinal pigment epithelial cells were present in recurrent ERMs. Newly synthesised collagen was also observed frequently.5 ERM recurrence therefore appears to result not only from cellular proliferation and/or migration but also from active deposition of an extracellular matrix.5 ILM peeling is thus theoretically beneficial in preventing ERM recurrence by removing the scaffold for cell proliferation.

In this study, we also compared the residual ILM pattern, the preoperative logMAR BCVA, the grade of ERM and the type of ERM (idiopathic or secondary). However, none of these had a statistically significant correlation with the residual ILM pattern.

PVD was observed in almost all eyes with ERM. Two eyes (3.3%) of the 61 eyes had no PVD before operation. In eyes without PVD, one was idiopathic ERM, and another was secondary ERM accompanied by diabetic retinopathy. Both had extensive retained ILM after ERM removal. The former had a round defect of posterior vitreous cortex after surgical PVD leaving ERM on the macula, the latter showing PVD along with ERM without a similar defect of posterior vitreous cortex. Yamashita et al reported in ERM that had no PVD that three patterns were observed when PVD was created during vitrectomy. The relationship between the posterior vitreous cortex and the epiretinal membrane was (A) a round defect in the posterior vitreous cortex after surgical PVD, leaving an epiretinal membrane on the macula, (B) a complete detachment of the vitreous cortex along with the epiretinal membrane and (C) a detachment of the posterior vitreous cortex without a round defect, leaving an epiretinal membrane on the macula.25 Types A and B were also observed in our study.

We used BBG to stain ILM, since BBG is considered less toxic to retinal cells as compared with ICG.1011 As we previously reported, BBG stained ILM clearly and facilitated ILM peeling. In addition, no intraoperative or postoperative complications relating to the use of BBG were observed.

In the present study, we only examined recurrence within 6 months of surgery. Although there were no cases of recurrent ERM, a longer follow-up is necessary to fully evaluate the risks of recurrence.

In summary, approximately half of all patients undergoing ERM removal have evidence of residual ILM with remnant GFAP positive and negative cells. This residual ILM is readily visualised intraoperatively with BBG staining. Intraoperative staining using this technique appears helpful in identifying the extent of residual ILM and confirming complete removal, thus possibly reducing the risks of subsequent ERM recurrence.