The effect of internal limiting membrane peeling to normal retinal function evaluated by microperimetry-3

Background surrounding macular holes (MH) by microperimetry-3(MP-3). Methods This is a prospective, cohort study which included 44 eyes of 44 patients with MHs who were treated by 23-gauge 3-port pars plana vitrectomy and ILM peeling with air tamponade. Color fundus photography, retinal optical coherence tomography and MP-3 were used 1 week before and 1, 4 months after operation. In MP-3 examination, a customized follow-up pattern with 45 spots in the central 8° visual field was used. The spots corresponding to the retina surrounding macular holes were selected for comparison of pre- and post-operative function. Results All eyes had achieved an anatomical success at the last follow-up. BCVA (logMAR) significantly improved both in 1 and 4 months after surgery (1.06±0.40 versus 0.53±0.30 and 0.31±0.24, P<0.01). The mean retinal sensitivity (MRS) (in dB) of the retina surrounding macular hole significantly increased 1 and 4 months after ILM peeling: pre-operative 23.46±3.01 dB versus post-operative 26.25±2.31 dB (u=-4.88, P<0.01) in 1 month and 27.14±2.45 dB (t=-6.29, P<0.01) in 4 months. Patients with increased MRS are significantly younger than patients with deceased MRS (59.72±3.22 years versus 65.60±8.19 years, P<0.01). After ILM peeling, the increasing extent of MRS was significantly higher in inferior and nasal retina than that in superior and temporal retina both in 1 and 4 months (P<0.05). compare pre- post-operative retinal In our study, we assessed the functional changes of the normal retina surrounding the MH after ILM peeling using MP–3. In order to ensure the result gives a strong indication for the effect of ILM peeling on the normal retina, we only choose points in the outer two rings which corresponded to the normal retina surrounding the macular hole, and the area within MH was excluded. There are 28 points in the outer two rings, which only occupied 60% of the whole 45 points, but covered more than 75% area of the 8° retina. These points located from 4° to 8°. The diameter of 8° visual field was 2500um (about 1.6PD). During the operation, the ILM we peeled off was at least 2PD, which means the 8° area was completely contained in the ILM peeling area. In the current study, the diameter of the largest MH is 876μm, which corresponded to approximately central 3.5° in visual field. To further excluding the confounding effect of MH on functional analysis for ILM peeling, the points with a distance from the margin of MH less than 0.5PD were also excluded.


Introduction
Internal limiting membrane (ILM) peeling has been considered as an important procedure in surgeries of macular diseases such as macular hole or epi-retinal membrane to increase the anatomic success rate. It may release the tangential traction to the retina of macular area, and activate Müller cells, stimulating the secretion of collagen, basement membrane components, inflammatory factors which may stimulate glial cell-mediated closure of macular holes (MH) [1].However, the use of ILM peeling in 3 macular surgery is still controversial.
Controversy focused on the potential injury of ILM-peeling. Major side effects of ILM peeling have been reported as potential mechanical or functional injury to retina [2][3][4][5][6][7][8][9][10]. Previous studies about retinal functional changes that caused by ILM peeling are conflicting: a few authors found no changes after peeling [11,12], others showed decrease of retinal sensitivity [13,14]. Tongren Hospital approved the study protocol, and all participants gave their written informed consent. Color fundus photography, retinal optical coherence tomography (OCT) (Carl Zeiss, Dublin, CA, USA) and microperimetry-3 (NIDEK, Gamagori, Japan) were performed for each patient 1 week before and 1,4 months after operation. MH was ensured by OCT. We defined the minimum diameter as the diameter of a MH.
Patients with glaucoma, myopia<-3.0 diopters (D), severe cataract, or other ocular diseases that could interfere with the measurements were excluded. The opacities of the patients' lens should be under N3C2P1 grade assessed by Lens Opacities Classification System III (LOCSIII).
A standard 23-gauge 3-port pars plana vitrectomy was performed by the same experienced surgeon (W. L.). Phacoemulsification and IOL implantation were performed if necessary. A subtotal vitrectomy was performed followed by internal limiting membrane peeling without staining. The posterior hyaloid 4 was elevated and trimmed in all patients. The ILM was peeled off with forceps in an area of about 2disc diameter around the MH. A fluid-gas exchange was carried out, and the vitreous was filled with air. All surgery was performed without any serious postoperative complications. Patients were asked to stay in a prone position for 5-7 days after surgery. One and four months after surgery, patients returned for a follow-up visit. Color fundus photography, optical coherence tomography (OCT) (confirming the closure of the MH) and microperimetry-3 were performed for each patient.
Microperimetry (MP) was selected for retinal function evaluating. MP is a subjective, quantitative, noninvasive diagnostic exam aimed at assessing retinal functionality and to put it in strict correlation with retinal morphology. Microperimetry-3 (MP-3) is the newest generation of microperimetry. It has a wider range of stimulus intensity, from 0 to 34 dB, compared to the MP-1. The MP-3 measures perimetric threshold values, even for normal eyes. A maximum stimulus luminance of 10,000asb allows evaluation of low-sensitivity. The MP-3 device features faster tracking, increased automation and a broader dynamic range compared with the MP-1. 9 Another important feature of this microperimeter is that target light is projected onto the retina rather than a screen. The position of the retina is therefore tracked so that target presentations can be automatically aligned, and the exact same location is stimulated at each target presentation. In this manner, we would expect to observe highly reproducible measurements of retinal sensitivity. 10 The microperimetry examination was performed in a dark room. All patients underwent a dark adaptation for at least half an hour until the pupil size reached 4mm or larger. The infrared fundus image was registered, and the central fixation point was aligned to the,center of MH in pre-operative examination. The follow-up pattern was used to make sure the pre-and post-operative examinations and comparisons were point to point perfectly matched. A customized pattern with 45 spots in central 8° visual field was used. The 45 test points in the MP-3 are shown in Figure 1.
The fixation target was a 1° diameter red circle, and the background luminance was set at 31.4asb.
This pattern gives a suitable evaluation of macular sensitivity and enables the detection of small visual field defects in the macular area. Only reliable VFs were used in analyses, which were defined 5 as fixation loss (FL) rate<20% and a false-positive (FP) rate<15%. We used a Goldman size III stimulus with duration of 200ms. Using the obtained retinal sensitivities, the mean sensitivity at the fovea, within two degrees, four degrees, six degrees and eight degrees were calculated. Four regions, superior nasal, inferior nasal, inferior temporal, superior temporal, are divided and shown in Figure 2.
When calculating, the points located on X-axis or Y-axis will be excluded. For example, when comparing the retinal sensitivity between superior and inferior retina, the points located on X-axis (point B1,2,3 and D1,2,3, figure 2) will be excluded. On the same way, when comparing the retinal sensitivity between nasal and temporal retina, the points located on Y-axis (point A1,2,3 and C1,2,3, figure 2) will be excluded. When calculating, we only choose 28 points in the outer ring zone instead of all 45 points, which located in the normal retina instead of MHs area ( Figure 3).
These 28 points located in the outer two rings, which only occupied 60% of the whole 45 points, but covered more than 75% area of the 8° retina. These points located from 4° to 8°. The diameter of 8°v isual field was 2500um (about 1.6PD). During the operation, the ILM we peeled off was at least 2PD, which means the 8° area was completely contained in the ILM peeling area. If the distance from the margin of MH to the selected points was less than 0.5°, the points will also be excluded.

Discussion
ILM peeling has been considered as a useful technique in surgeries for vitreomacular interface diseases. It has been reported that the macular hole closure rate was 90-100% when treated with 7 vitrectomy and ILM peeling, while it was only 60-90% without ILM peeling [15][16][17][18]. However, potential damages to retinal function caused by ILM peeling was considered as a side effect of this technique.
Previous studies about influence of ILM peeling on retinal function were controversial. Some studies evaluated a dissociated optic nerve fiber layer (DONFL) in ILM-peeling area and found the retinal function in this area is not changed after surgery. Yasuki et al [7]. compared the retinal sensitivity of DONFL area and non-DONFL area in twenty ILM-peeled eyes with MH more than 4 months after the vitrectomy by scanning laser ophthalmoscopy (SLO) microperimetry. Yoshinori et al. [11] performed In our study, we assessed the functional changes of the normal retina surrounding the MH after ILM peeling using MP-3. In order to ensure the result gives a strong indication for the effect of ILM peeling on the normal retina, we only choose points in the outer two rings which corresponded to the normal retina surrounding the macular hole, and the area within MH was excluded. There are 28 points in the outer two rings, which only occupied 60% of the whole 45 points, but covered more than 75% area of the 8° retina. These points located from 4° to 8°. The diameter of 8° visual field was 2500um (about 1.6PD). During the operation, the ILM we peeled off was at least 2PD, which means the 8° area was completely contained in the ILM peeling area. In the current study, the diameter of the largest MH is 876μm, which corresponded to approximately central 3.5° in visual field. To further excluding the confounding effect of MH on functional analysis for ILM peeling, the points with a distance from the margin of MH less than 0.5PD were also excluded.
Patients with severe cataract, which may interfere with the MP-3 measurements (The opacities of all patients' lens under LOCSIII N3C2P1 grade), were excluded. Phacoemulsification and IOL implantation were performed in 35 eyes. MRS were increased in both groups. The increasing extent of MRS has no different between the patients with phacoemulsification and those without, which suggests the opacity of lens was not severe in patients with phacoemulsification and this extra procedure did not influence the results.
In the current study, the retinal sensitivity in ILM peeling area increased both 1 and 4 months after surgery. The reason for this unexpected result in our research might be as following. Firstly, this is a short-term study. We only observed the changes of retinal function for 4 months after surgery. The ILM peeling procedure itself could be an injury to motivate retinal neural protection and lead to the release of neural protective factors [20,21]. These factors may promote the retinal function in a short-term. Secondly, in the current study, the retinal function was evaluated by MP-3. Compared with MP-1 and MP-2, MP-3 has auto tracking and auto aligment, fixation test, wider measurement range, higher resolution non-mydriatic fundus camera and a better system to accomplish the images for preand post-treatment comparison. These techniques enable us to do more accurate assessment of macular function.
The post-operative MRS in the selected area increased in 37 patients and deceased in 7 patients.
Patients with decreased MRS were significantly older than other patients. We think the reason may be related with the retinal recovery ability. Patients with younger age may have a better recovery ability in RS than aged patients. While the sample was not big enough. If the sample enlarged, the result might be different.
The pre-operative MRS had no difference between superior and inferior retina or between nasal and temporal retina pre-operation. While the increasing extent of retinal sensitivity in superior retina was significantly higher than that in inferior retina. When performing ILM peeling, the surgeon used to start from superior retinal area. The initiation of ILM peeling may bring more mechanical injury to the superior retina. It may be the reason of this phenomenon. We also found the increasing extent of retinal sensitivity in temporal retina was significantly lower than that in nasal retina. Takayuki et al.
[22] had the similar result. They performed vitrectomy and ILM peeling on 39 eyes with MH, and found the retinal sensitivity was significantly lower in the temporal area than in the other areas 3 and 6 months after surgery. The reason for this restricted change to the temporal retina might be as following. Firstly, the removal of the ILM started from the temporal superior retina to the fovea.
Secondly, the nerve fiber layer has been reported to be thinnest in temporal quadrant around fovea [23]. Thirdly, the density of ganglion cells at the temporal retina is less than that at the nasal retina within 2 mm from fovea [24].
The limitations of the current study included lack of a control group. A prospective randomized control study is indicated in the future to draw more definitive conclusion.

Conclusion
ILM peeling in normal retina do not decrease the retinal function in a short-term after surgery, except in some patients with older age. We didn't use any dye during surgery, the retinal toxicity of specific dye needs further study. ILM peeling alone is a safe and useful technique in surgeries for macular hole.   Figure 1 A customized pattern was used in 8° of the visual field, with 45 spots.

Figure 2
The area was divided into four regions, superior nasal(point A1-8), inferior nasal(point B1-8), inferior temporal(point C1-8), and superior temporal(point D1-8). When calculating, the points located on X-axis or Y-axis will be excluded. For example, when comparing the retinal sensitivity between superior and inferior retina, the points located on X-axis (point B1,2,3 and D1,2,3) will be excluded. On the same way, when comparing the retinal sensitivity between nasal and temporal retina, the points located on Y-axis (point A1,2,3 and C1,2,3) will be excluded.  Left eye of a 63-year-old Chinese woman with a huge macular hole (MH). Points a, b, c, d was excluded for the distance from these points to the margin of MH was less than 0.5°.
When calculated for this patient, we only chose 24 points in the outer ring zone.

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download. STROBE_checklist_cohort.doc