A three-step approach versus the inverted internal limiting membrane flap technique in large full thickness macular hole surgery: a comparative study

Abstract

Objective

To evaluate the anatomical and functional outcomes of our novel surgery (a three-step approach) and the conventional inverted internal limiting membrane flap technique (IFT) in treating large full-thickness macular holes (FTMHs).

Methods

This was a retrospective, consecutive, nonrandomized comparative study of patients who underwent either the novel surgery (n = 27, Group A) or IFT (n = 27, Group B). The main outcomes of MH closure rates and the best corrected visual acuity (BCVA) at 1-, 3-, and 6-month follow-up were compared between the two groups.

Results

At 6 months postoperatively, MH closure was achieved in 24/27 patients in Group A and 22/27 patients in Group B (88.89% vs. 81.48%, P = 0.704) with U-shaped closure rates being significantly higher in Group A (P = 0.029). The average BCVA at month 6 was 0.69 ± 0.38 (LogMAR) in Group A and 0.91 ± 0.39 in Group B (P = 0.015) with the improvement in BCVA being significantly higher in Group A (0.50 ± 0.59 vs. 0.31 ± 0.59, P = 0.045). The recovery rates of ELM were significantly higher in Group A (P = 0.026).

Conclusions

Our three-step approach greatly improves anatomical and functional outcomes compared with IFT. This novel surgery has a dominant advantage in earlier and higher ultimate closure rate, U-type closure rate, and ELM recovery rate, and more importantly, better recovery of BCVA.

Key messages

What is known: the postoperative visual acuity outcomes of large full-thickness macular holes (FTMHs) patients are not satisfactory. What is new: our three-step approach greatly improves anatomical and functional outcomes compared with the inverted internal limiting membrane flap technique (IFT) in the treatment of large FTMHs.

Introduction

A full-thickness macular hole (FTMH) is an anatomic defect in the fovea featuring interruption of all neural retinal layers from the internal limiting membrane (ILM) to the retinal pigment epithelium (RPE) [3], which affects central visual acuity. Various factors, including trauma and myopia, are associated with its occurrence, but the vast majority are idiopathic [2]. Its prevalence increases with age, particularly among elderly women [12], thereby imposing a significant economic burden on public health systems. Despite the recent development of surgical techniques, which have increased the anatomic success rate of macular hole repair [4, 7, 11], the visual acuity outcomes are not always satisfactory in eyes with surgical success [10].

Here, we present a novel surgical technique for PPV combined with ILM peeling and a three-step approach: first, gently releasing the adhesion between the MH edges and RPE to decrease the size of the hole; second, approximating the edges of the hole; and finally, tapping the edges flat to promote attaching the edges to the RPE. We call this technique the releasing-closing-tapping approach, which has been elaborated in our previous study [14].The aim of this study was to evaluate the anatomical and functional outcomes of our novel surgery (a three-step approach) versus the conventional inverted internal limiting membrane flap technique (IFT) in the treatment of large FTMHs.

Materials and methods

This was a retrospective, consecutive, nonrandomized comparative study based on reviewing the medical records concerning fifty-four eyes of 54 patients with large stage 4 MHs (minimum diameter > 400 μm) who underwent surgical treatment by a skilled surgeon (QQ) in the Department of Ophthalmology, Shanghai General Hospital, the Affiliated Hospital of Shanghai Jiao Tong University School of Medicine between January 2019 and November 2022. The patients were informed in advance of the purpose, procedure and possible complications of the operation, and informed consent was thus obtained. The study was approved by the Ethics Committee of Shanghai General Hospital, and all the procedures conformed to the tenets of the Declaration of Helsinki.

The inclusion criteria were as follows: FTMH, minimum MH > 400 μm, age > 18 and at least 6 months of follow-up. The exclusion criteria were as follows: patients (1) with any vitreoretinal diseases; (2) with a history of any previous ophthalmic surgery except for cataract surgery; (3) with a cloudy cornea; and (4) with poor OCT image quality.

At baseline and at every follow-up visit (1 month, 3 months and 6 months postoperatively), all patients received a complete examination consisting of the best corrected visual acuity (BCVA) after refraction, intraocular pressure (IOP), slit-lamp examination and indirect ophthalmoscopy. The BCVA measurement with a Snellen chart was converted to the logarithm of the minimum angle of resolution (LogMAR) for statistical analysis. Delta BCVA (LogMAR) was defined as LogMAR BCVA at the 6-month follow-up - preoperative LogMAR BCVA. The images of fundus photography and OCT were captured after pupil dilation, the exclusion of which yielded poor quality and a signal strength index less than 50.

The size of the MHs was measured according to the International Vitreomacular Traction Study Group Classification System [3]. The narrowest hole point in the mid-retina was defined as the minimum inner diameter and was measured using the OCT caliper function with a line drawn roughly parallel to the RPE; the base diameter was measured at the RPE level; and the hole height was defined as the greatest height of the hole from the vitreoretinal interface to the RPE. Full-thickness macular holes are classified based on their minimum diameter, with large holes defined as those exceeding 400 μm according to the International Vitreomacular Traction Study Group Classification System. The RPE conditions of all the patients were filtered as normal at baseline without chorioretinal atrophy involving the fovea, which included a hypertransmission region at least 250 μm in diameter on SD-OCT [1], a disruption or attenuation zone at least 250 μm in diameter [8], photoreceptor degeneration [15], scrolled RPE or RPE tears [13]. The ratio of the hole height to the base diameter, serving as the perpendicular and horizontal dimensions of the hole, was referred to as the macular hole index (MHI), and the patient with an MHI < 0.5 was considered to have a poorer BCVA postoperatively than the patient with an MHI > 0.5 [9].

PPV with a releasing-closing-tapping approach combined with C2F6 gas (perfluoroethane) tamponade was implemented in 27 eyes (Group A: novel surgery group) between January 2021 and November 2022, while PPV with an inverted ILM flap (IFT) technique combined with C2F6 gas tamponade was performed in 27 eyes (Group B: ILM flap group) between January 2019 and November 2020. Combined phacoemulsification and intraocular lens (IOL) implantation were conducted in 12 phakic eyes (5 eyes in Group 1 and 7 eyes in Group 2).

All participants were followed up at 1, 3, and 6 months after surgery, and anatomical and visual outcomes were assessed and compared at each follow-up. Anatomical closure referred to flattening of the hole with the neurosensory retina covering the fovea shown by SD-OCT. The diverse postoperative closure types were divided into U-shaped, V-shaped, W-shaped (irregular), and bridge-like closures based on their cross-sectional structure (Fig. 1). Bridge-like closures had a U-shape in the inner retinal layers with a cystoid space in the outer retinal layers. The recovery of the continuous back-reflection lines corresponding to the ellipsoid zone (EZ) and the external limiting membrane (ELM) implied restoration of the photoreceptor layer. We defined the continuous and intact EZ and ELM lines as recovered and absent or disrupted EZ and ELM lines as unrecovered [5].

Fig. 1

Closure type after macular hole surgery. (A) Bridge-like closure; (B) V-shaped closure; (C) U-shaped closure; (D) W-shaped closure. (A) A female patient with a minimal inner diameter of 568 μm underwent conventional IFT surgery and received bridge-like closure 1 month after surgery. (B) A female patient with a minimal inner diameter of 737 μm underwent novel surgery with a three-step approach and received V-shaped closure 1 month after surgery. (C) A female patient with a minimal inner diameter of 566 μm underwent novel surgery with a three-step approach and received U-shaped closure 3 months after surgery. (D) A male patient with a minimal inner diameter of 612 μm who underwent conventional IFT surgery received W-shaped closure 3 months after surgery

Surgical techniques

A three-step approach of massage

A standard three-port vitrectomy was performed using a 23 G vitrectomy system (Constellation^®, Alcon, Fort Worth, TX, USA), and ILM peeling was conducted about 1-to-2-disc diameters around the MH after staining with indocyanine green (ICG). In the first step, the retinal surface was gently scraped from an approximately 2-to-3-disc diameter around the MH to the center with a membrane loop (FINESSE Flex loop; Alcon, Fort Worth, TX, USA) to induce release of the adhesion between the MH edges and the RPE. Next, the round hole was converted into a transverse slit through massaging the loosened retina toward the center with the membrane loop to achieve MH closure. In the third step, the edges were gently tapped flat with no space left using the blunt end of a vitreous cutter to obtain the reattachment of retinal neuroepithelium to RPE. Fluid-air exchange was performed, and gas tamponade (12% C3F8) was carried out at the end of the surgery. Patients were required to keep their heads facing downward for 5–7 days according to clinical experience.

ILM flap technique

A standard three-port vitrectomy was performed using a 23 G vitrectomy system (Constellation^®, Alcon, Fort Worth, TX, USA). After being stained with indocyanine green (ICG), the ILM was peeled (1 to 1.5 diameter of the MH) with the center left attached to the edges of the hole. The ILM flap, approximately 1 to 1.5 diameter of the MH, was lifted, inverted, and covered to the MH gently using forceps (25-gauge Eckardt End-gripping Forceps, Dutch Ophthalmic, USA) to fill the hole. Fluid-air exchange was implemented using passive aspiration with low intraocular pressure to avoid flap displacement caused by turbulence. Gas tamponade (12% C3F8) was performed at the end of the surgery. Patients were required to keep their heads facing downward for 5–7 days according to clinical experience.

Statistical analysis

All analyses were performed using SPSS 26.0 statistical software (SPSS, Chicago, IL, USA) to illustrate the differences in baseline characteristics, anatomical outcomes including MH closure and functional outcomes of BCVA between the novel surgery group (Group 1) and ILM flap group (Group 2). Continuous variables are here expressed as the mean ± standard deviation (range), and categorical variables are expressed as percentages. The comparison of continuous variables between two groups was analyzed with two-tailed standard t tests or Mann–Whitney U tests, while the comparison of categorical variables between two groups was analyzed with Fisher’s exact test or the chi-square test. The Wilcoxon rank test was used to compare the difference in pre-, postoperative and delta BCVA. A P value less than 0.05 was considered statistically significant.

Results

Seventy-five consecutive medical records of eyes with FTMHs that underwent PPV between January 2019 and November 2022 were reviewed. 21 eyes with a follow-up duration of less than 6 months or irregular visits were excluded. A total of 54 eyes were included in this study, with 27 eyes in the novel surgery group (Group A) and 27 eyes in the IFT group (Group B). Forty eyes were pseudophakic, and the remaining 14 eyes underwent combined phacoemulsification and IOL implantation.

The comparison of the preoperative data between the two groups is shown in Table 1. There was no significant difference in sex (P = 1), age (P = 0.630), illness duration (P = 0.219), IOP (P = 0.883), preoperative BCVA (P = 0.693), minimal inner diameter (P = 0.897), base diameter (P = 0.361), hole height (P = 0.186), or MHI (P = 0.558). In this study, phacoemulsification combined with intraocular lens (IOL) implantation was performed in 7 patients (25.93%) in Group A and an identical proportion of 7 patients (25.93%) in Group B. Consequently, all participants were pseudophakic at the onset of the follow-up period.

Table 1 Demographic and baseline characteristics of two groups

The postoperative data of anatomic recovery are shown in Table 2. The closure rate was 85.19% (23/27) in Group A and 70.37% (19/27) in Group B at 1 month of follow-up, while SD-OCT showed ultimate MH closure in 24 cases (88.89%) in Group A and 22 cases (81.48%) in Group B at 6 months of follow-up, which showed earlier and higher closure rates in Group A. In the comparison of closure types postoperatively (Table 2), Group A demonstrated a pronounced prevalence of U-shaped closures, starting at 40.74% (11/27) at 1 month and markedly increasing to 62.96% (17/27) at 3 months, and further to 66.67% (18/27) at 6 months. This trend in Group A contrasts with Group B, where U-shaped closures were less dominant, peaking at only 33.33% (9/27) at 6 months. While both groups initially showed similar rates of U-shaped closures at 1 month (P = 0.387), significant differences emerged at the 3- and 6-month intervals (P = 0.029). The presence of other closure types like V-shaped, bridge-like, and W-type closures, though recorded, showed less significant differences and trends over time.

Table 2 Postoperative anatomic outcomes of two groups

The postoperative BCVA and improvement in the BCVA (delta BCVA) are shown in Table 3. The mean BCVA improved significantly after surgery in Group A and Group B at the 6-month follow-up. The postoperative BCVA and delta BCVA increased gradually in both groups, with the postoperative BCVA and delta BCVA in Group A being larger than those in Group B throughout the follow-up period, but there was no statistically significant difference at 1 and 3 months of follow-up. The mean BCVA at 6 months was 0.69 ± 0.38 LogMAR in Group A and 0.91 ± 0.39 LogMAR in Group B (P = 0.015), while delta BCVA at 6 months was − 0.50 ± 0.59 LogMAR in Group A and − 0.31 ± 0.59 Log MAR in Group B (P = 0.045). The significant differences indicated better BCVA recovery in Group A at the 6-month follow-up.

Table 3 Postoperative functional outcomes of two groups

SD-OCT examination of the photoreceptor layers in Group A revealed ELM recovery in 3 (11.11%), 7 (25.93%), and 15 (55.56%) patients at 1, 3 and 6 months, respectively, and EZ recovery in 1 (3.70%), 2 (7.41%), and 10 (37.04%) patients at 1, 3 and 6 months, respectively (Table 3). In contrast, in Group B, ELM recovery was observed in 1 (3.70%), 3 (11.11%), and 6 (22.22%) patients at 1, 3 and 6 months, respectively, and EZ recovery was observed in 1 (3.70%), 2 (7.41%), and 5 (18.52%) patients at 1, 3 and 6 months, respectively (Table 3). In both groups, rates of EZ and ELM recovery progressively increased over time, with ELM restoration typically preceding that of the EZ. Notably, complete EZ recovery was not observed prior to full ELM recovery. Throughout the follow-up period, Group A consistently showed higher rates of both ELM and EZ restoration compared to Group B. This difference was statistically significant for ELM restoration (55.56% in Group A vs. 37.04% in Group B, P = 0.026) at the 6-month mark, although the difference in EZ recovery rates (25.93% vs. 18.52%, P = 0.224) was not statistically significant. Figure 2 highlights a case of gradual ELM and EZ reconstruction following the novel three-step approach of massage.

Discussion

This was a retrospective, consecutive, nonrandomized comparative study to evaluate the anatomical and functional outcomes of our novel surgery (a three-step approach) versus the conventional inverted internal limiting membrane flap technique (IFT) in the treatment of large full-thickness macular holes (FTMHs). The 6-month follow-up showed that, compared with the current recommended IFT, our novel surgery could achieve an earlier and higher ultimate closure rate, U-type closure rate, and ELM recovery rate, and more importantly, better recovery of BCVA.

The closure rate of macular holes is an important prognostic index of FTMHs, which directly affects the postoperative visual acuity recovery of patients. The International Vitreomacular Traction Study (IVTS) Group divided FTMH into small (< 250 μm), medium (250–400 μm) and large (> 400 μm) FTMH according to the size of the macular hole (horizontally measured linear width across hole at the narrowest point) [3]. Previous studies have shown that large FTMHs are often associated with a relatively low postoperative closure rate, and up to 44% of large macular holes remain open after the first surgery [11]. In 2010, Michalewska et al. introduced the inverted ILM flap technique as an effective alternative to treat large MHs, and in their study, macular hole closure was observed in 98% of patients (45/46) in the inverted ILM flap group [11]. In our study, the closure rate of large MHs is not that high, and macular hole closure was observed in only 81.48% of patients (22/27) in the IFT group. However, this number seems acceptable in clinical practice, where patients may not be able to achieve perfect postoperative management. In addition, the inverted ILM flap has the risk of spontaneous detachment, and a learning curve is required by the surgeon. However, in our novel surgery group, SD-OCT showed a prominently earlier and higher closure rate: 85.19% (23/27) at 1 month after surgery and 88.89% (24/27) at the 6-month follow-up. We are not surprised by this result. In our new method, we reduce the size of macular holes factitiously during surgery by massaging the holes and flattening the edges, rather than waiting for glial cells to fill the photoreceptor defects.

Another factor influencing the postoperative outcome is the variation in postoperative macular hole closure types. OCT images of the repaired macular holes were categorized into three patterns. The U type showed mild to moderate backscattering layers with a smooth circular surface covering the RPE. In eyes with the V type, the RPE was covered with moderate backscattering layers with a notch. The W type showed abruptly or gradually terminating sensory retinal layers to expose the surface of the RPE [6]. The U-shaped closure pattern is associated with the best visual outcomes among the different closure types due to its anatomical characteristics. This pattern reflects the optimal reapproximation of the photoreceptor layer to the RPE, facilitating the restoration of photoreceptor integrity and visual function. In contrast, other patterns, such as V-shaped or W-shaped closures, may involve less effective reattachment or glial cell proliferation, which can impair functional recovery [6]. In our study, U-shaped closure rates, which were associated with the best functional results, were significantly higher in the novel surgery group at 6 months (66.67% vs. 33.33%, P = 0.029). The inverted ILM flap technique is based on the assumption that the remaining ILM may induce glial cell proliferation, resulting in macular hole filling with proliferating cells that enhance closure. However, our new approach does not preserve the ILM. We massage the holes into a transverse slit that follows the course of the optic nerve fiber, and then we flatten the edges by tapping to attach them to the RPE, which reduces the size of the macular hole factitiously during surgery. We speculate that this approach may be more conducive to the reconnection of photoreceptor cells in the macular region rather than the formation of glial scars with delayed healing, which has little effect on the restoration of visual acuity.

However, there are concerns about this new approach, including whether reducing the size of macular holes factitiously during surgery will cause unknown damage to the physiological structure and function of the retina and whether the patient’s visual acuity can be further improved after surgery. In our study, the results were encouraging. The mean BCVA at 6 months was 0.69 ± 0.38 LogMAR in Group A and 0.91 ± 0.39 LogMAR in Group B (P = 0.015), while delta BCVA at 6 months was − 0.50 ± 0.59 LogMAR in Group A and − 0.31 ± 0.59 LogMAR in Group B (P = 0.045). The significant differences indicated better recovery of BCVA in Group A at the 6-month follow-up. In addition, SD-OCT examination of the photoreceptor layers in the two groups showed that, compared with Group B, restoration of both the ELM and the EZ was higher in Group A throughout the follow-up period (Table 3; Fig. 2), with a significant difference in the ELM recovery rate at the 6-month follow-up (55.56% vs. 22.22%, P = 0.026). These results further confirm our hypothesis.

Fig. 2

60-year-old patient presented with a large FTMH. (A) Fundus photography at baseline. (B) SD-OCT at baseline showing a minimal inner diameter of 566 μm, combined with a preoperative BCVA of 0.05; (C) SD-OCT at 1 week after surgery showing a U-shape closure of the MH after conducting the novel three-step approach surgery. The foveal hyperreflective tissue by SD-OCT is shown to move closer to the center in the same direction as the operation with the early appearance of a continuous ELM line. (D) SD-OCT at the 3-month follow-up showing a persistent U-shape closure of the MH. Disruption of the EZ (between red arrowheads) remained, while the ELM line was almost complete. (E) SD-OCT at 6 months follow-up showing a persistent U-shape closure of the MH with complete restoration of both the ELM and the EZ lines. The BCVA increased to 0.6

This study has several limitations. The study population was single, the sample size was small, and the follow-up period was short. Besides, in our study, we selected patients with at least 6 months of follow-up to ensure long-term monitoring of the macular hole closure. We excluded patients who had undergone multiple surgical interventions to minimize potential confounding factors, which may cause bias. In addition, as mentioned above, maybe because patients are not able to achieve perfect postoperative management in clinical practice, the closure rates and final BCVA in our ILM flap technique group are worse than previously published studies. Therefore, a cohort study with a larger sample size and longer follow-up time might enable us to obtain more complete results.

In conclusion, this study presents a novel surgical technique (a three-step approach) that could achieve an earlier and higher ultimate closure rate, U-type closure rate, and ELM recovery rate, and more importantly, better recovery of BCVA than the conventional inverted internal limiting membrane flap technique (IFT) in the treatment of large full-thickness macular holes.