Clinical characteristics and long-term evolution of lamellar macular hole in high myopia

Purpose To evaluate the clinical characteristics and evolution of lamellar macular hole (LMH) in high myopia and the parameters associated with structural worsening, defined as the development of foveal detachment or full-thickness macular hole. Methods Patients with high myopia and LMH were retrospectively recruited. The clinical characteristics and various parameters of optical coherence tomography were identified at baseline and during follow-up visits. Cox regression analysis was used to evaluate the hazard ratios for foveal detachment and full-thickness macular hole. Results Among 112 eyes (98 patients), 64.3% were female; the mean axial length of all eyes was 29.6 ± 1.9 mm. The ‘LMH without retinoschisis’ group accounted for 39.3% of the eyes. Forty-two percent developed structural worsening within a median follow-up of 67 months. Multivariable regression on all cases showed elevated tissue inside the LMH (P = 0.003) protected against structural worsening while V-shaped LMH (P = 0.006) predicted it. In the “LMH with retinoschisis group”, ellipsoid zone disruption (P = 0.035), and V-shaped LMH (P = 0.014) predicted structural worsening, while elevated tissue inside the LMH (P = 0.028) protected against it. In the “LMH without retinoschisis group”, no associated factor was identified. Conclusions LMHs in high myopia are unstable, especially those with V-shaped LMH. Elevated tissue inside LMHs have a protective effect against further structural worsening.

Introduction Lamellar macular hole (LMH) is a common macular structural abnormality in patients with high myopia (4.8%-20.7%) [1,2]. In highly myopic eyes, the clinical course of LMH may be more complicated than that in idiopathic LMH because the adherent vitreous cortex and epiretinal membrane exert a complex anteroposterior and tangential traction, and because the internal limiting membrane (ILM) is rigid and posterior staphyloma is often present [3][4][5]. Some LMHs may stay unchanged or show minimal progression for many years, while others may progress to full-thickness macular hole (FTMH) or macular hole with retinal detachment (MHRD), causing severe vision loss [6][7][8]. Studies on myopic traction maculopathy (MTM) in general have shown that, when foveal retinoschisis advances to retinoschisis with foveal detachment (FD), the probability of FTMH formation is very high [2]. However, few investigations have addressed the evolution of LMH in high myopia specifically. Furthermore, in highly myopic eyes, LMH may exist alone or in association with retinoschisis [3,5,9], and it is unclear whether these two different types of LMH exhibit different clinical characteristics and outcomes.
In the present study, we retrospectively reviewed LMH in highly myopic eyes and examined lesion evolution. By analyzing the various parameters related to progression, we aimed to identify the risk and protective factors associated with structural worsening. This may inform decisions regarding follow-up strategy and timing of surgery.

Study population
From January 2010 to December 2017, consecutive cases of high myopia with LMH were retrospectively reviewed. Regardless of whether it was associated with retinoschisis, LMH was diagnosed using the following updated criteria proposed by the International Vitreomacular Traction Study Group: (1) irregular foveal contour; (2) defect in the inner fovea (with or without actual loss of tissue); (3) separation between the inner and outer retinal layers in the fovea; (4) absence of full-thickness foveal defect [10]. High myopia was defined as a spherical equivalent refractive error more severe than -6.0 diopters and/or an axial length � 26 mm. LMHs were identified using spectral-domain optical coherence tomography (SD-OCT) (RTVue Model-RT 100 scanner, version 3.5: Optovue Inc., Fremont, CA, USA). Only patients with a follow-up of at least 2 years were included. Those with a history of RD, vascular or inflammatory disease, or previous intraocular surgery other than cataract operation were excluded. The study was approved by the Institutional Review Board at National Taiwan University Hospital and followed the tenets of the Declaration of Helsinki. The informed consent was waived.

Clinical characteristics and OCT parameters
Demographics and ocular biometry were recorded. Best-corrected visual acuity (BCVA) was measured using a Snellen chart and converted to logarithm of the minimal angle of resolution values for statistical analysis. The post-operative visual acuity was measured at least 6 months after the surgery. Structural worsening was defined as the development of FD or FTMH. The time to structural worsening was documented, as were the surgical methods and results. Macular structures were regularly examined using slit-lamp biomicroscopy, fundus photography, and OCT every 3 to 6 months.
Standard 8-mm horizontal and vertical OCT scans, centered on the fovea, were taken using the registration function. The following OCT parameters were documented or measured using the manual caliper function of the built-in software: central retinal thickness (CRT), subfoveal choroidal thickness, maximal horizontal and vertical LMH diameters, and minimal residual foveal thickness. The minimal residual foveal thickness was measured as the shortest distance from the base of the LMH to the Bruch's membrane at the fovea. The OCT images were also examined for the presence of lamellar hole-associated epiretinal proliferation [12], macular retinoschisis, epiretinal membrane, staphyloma, vitreomacular traction, and elevated tissue inside the LMH (Fig 1A and 1B); the integrity of the ellipsoid zone was also assessed.
The presence of retinoschisis at the macula was recorded at baseline and at follow-up visits. The type of LMH was determined at baseline. The 'LMH with retinoschisis' group had both LMH and macular retinoschisis (Fig 1C-1I), while the 'LMH without retinoschisis' group had no macular retinoschisis at baseline (Fig 1J-1N). The extent of retinoschisis was categorized as follows, according to Shimada's classification [13], as either parafoveal retinoschisis, which involved foveal retinoschisis that did not cover the entire macula (Fig 1h), or extensive retinoschisis, which involved the entire macula (Fig 1C and 1F). Macular retinoschisis was further classified into inner or outer types. Eyes with outer macular retinoschisis had intraretinal splitting in or outer to the outer plexiform layer. Eyes with inner macular retinoschisis had intraretinal splitting inner to the outer plexiform layer [8].
LMHs were classified as either V-shaped ( Fig 1C, 1H and 1J) or A-shaped (Fig 1F and 1M). V-shaped LMHs had a smaller diameter intraretinally than at the retinal surface, while Ashaped LMHs had a larger diameter intraretinally.

Surgical criteria and technique
During follow-up, operation would be suggested to patients who suffered worse vision associated with progression of the foveal structural changes, FD, FTMH, or RD. If the patient agreed to receive operation after pros and cons of the procedures were explained, vitrectomy would be arranged within one month based on the severity of the condition. In brief, a standard three-port 23-gauge or 25-gauge pars plana vitrectomy was performed by three retinal specialists (C.M.Y. & T.C.H. & C.H.Y). Triamcinolone acetonide-assisted posterior hyaloid separation and removal were performed, followed by epiretinal membrane removal with microforceps if identified. Indocyanine green dye-assisted ILM peeling was then performed within the arcade. For cases without FTMH, fovea-sparing ILM peeling technique was adopted [14]. For cases with FTMH, temporal inverted ILM flap technique [15] or inverted ILM flap insertion technique [16] was performed depending on the condition. The flap was stabilized by copious amount of Viscoat (Alcon Laboratories, Fort Worth, TX). Finally, air-fluid exchange was done. The vitreous cavity was flushed with 15% perfluoropropane (C3F8) in those cases with FTMH. Rarely, silicone oil tamponade was used at the surgeons' discretion.

Statistical analysis
All statistical analyses were performed using RStudio (version 3.6.0; RStudio, Inc, Boston, MA, USA). For descriptive statistics, mean and standard deviation were calculated for parametric data, while median and range were calculated for non-parametric data. Percentages were calculated for categorical variables. Changes in clinical parameters between enrollment and final follow-up were evaluated using paired t-tests. To identify predictive factors for the development of FTMH or FD, Cox proportional hazards regression analysis was performed.
Schoenfeld residuals were obtained to test the proportional hazard assumption of the Cox regression. Variable with P-values < 0.1 in univariable Cox proportional hazards regression analysis was included in multivariable Cox proportional hazards regression analysis. For the subgroup analysis of eyes with different extent of retinoschisis and LMH shapes, Chi-squared test or Fisher's exact test was used for categorical variables and independent t-test was used for numerical variables. P-values < 0.05 were considered statistically significant.

Results
One hundred and twelve eyes of 98 patients (26 men and 72 women) were enrolled. The patients' mean age was 61.8 ± 9.4 years. Their mean axial length was 29.6 ± 1.9 mm and their BCVA at baseline was 0.42 ± 0.35. At presentation, 44 eyes (39.3%) were in the 'LMH without retinoschisis' group, while 68 (60.7%) were in the 'LMH with retinoschisis' group. There were more women in the 'LMH with retinoschisis' group. Eyes in the 'LMH with retinoschisis' group had worse baseline BCVA, more staphyloma, longer axial length, thicker CRT, thinner subfoveal choroidal thickness, and deeper LMH vertical extension ( Table 1). The shape of LMH showed no association with the presence of retinoschisis (P = 0.223 by Chi-squared test), Fig 1. (a-b) Two eyes with elevated tissue inside a lamellar macular hole (LMH) (arrow). The evolution of LMHs in the 'LMH with retinoschisis group': (c-e) The first case had a V-shaped LMH with extensive retinoschisis. The LMH progressed into foveal detachment after 6 months, and further progressed into fullthickness macular hole (FTMH) about 4 months later. (f-g) The second case had an A-shaped LMH with retinoschisis. The LMH remained stable during the 44-month follow-up. Lamellar-associated epiretinal proliferation was noted around the opening of LMH (arrow). (h-i) The third case had parafoveal retinoschisis with a V-shaped LMH. The LMH remained stable during the 44-month follow-up. The evolution of LMHs in the 'LMH without retinoschisis group': (j-l) The fourth case had a V-shaped LMH. Lamellar-associated epiretinal proliferation developed around the LMH (arrow). During the 15-month follow-up, the LMH deepened progressively and developed into a FTMH with a wide opening. (m-n) The fifth case had an A-shaped LMH. Optical coherence tomography showed evidence of previous vitreomacular traction with a piece of avulsed retinal tissue. It progressed to FTMH 21 months later. This narrowopening pattern was similar to an idiopathic FTMH.

Factors associated with foveal detachment or full-thickness macular hole
Univariable Cox proportional hazards regression analysis on all cases showed that the following conditions were positively associated with structural worsening: LMH with retinoschisis (P = 0.017), the presence of inner retinoschisis (P = 0.006) and outer retinoschisis (P = 0.029), CRT (P = 0.003), extensive retinoschisis (P < 0.001), V-shaped LMH (P < 0.001), and increased vertical diameter of LMH (P = 0.005). Elevated tissue inside the LMH (P < 0.001) was protective against structural worsening (Table 3). Eyes in the 'LMH without retinoschisis' group had no clinical features associated with structural worsening (Table 3). In eyes with LMH and retinoschisis, the presence of elevated tissue inside the LMH (P = 0.003) had a protective effect against structural worsening. Shorter axial length (P = 0.025), V-shaped LMH (P = 0.023), and extensive retinoschisis (P = 0.005) were associated with structural worsening (Table 3).
Multivariable Cox proportional hazards regression analysis on all cases showed that elevated tissue inside the LMH (P = 0.003, hazard ratio = 0.20) was protective against structural worsening and that V-shaped LMH (P = 0.006, hazard ratio = 3.77) predicted it (Table 4). Eyes in the 'LMH without retinoschisis' group had no clinical features associated with structural worsening (Table 4). In eyes with LMH and retinoschisis, ellipsoid zone disruption (P = 0.035, hazard ratio = 2.28), and V-shaped LMH (P = 0.014, hazard ratio = 4.56) were associated with structural worsening, while elevated tissue inside the LMH (P = 0.028, hazard ratio = 0.25) was protective against structural worsening (Table 4).

Discussion
In non-myopic eyes, LMH in most patients has a stable clinical course, and only rarely progresses to FTMH (1%-4%) [5,17]. However, in high myopic eyes, complex tractional forces may induce various macular structural changes, which have been collectively called MTM. LMH with retinoschisis is a specific type of MTM with inner retinal defect, which further weakens the macular structure. Only a few studies have investigated LMH in high myopia, and these yielded conflicting data [5,9,18]. In the present study, we found that LMHs in high myopia were unstable and took different evolution routes from those in non-myopic eyes. We identified certain factors associated with the development of FD or FTMH, which was widely regarded as the major lesions associated with structural worsening. In the 'LMH without retinoschisis' group, we documented the following LMH evolution processes in highly myopic eyes: (1) stable, (2) development of macular retinoschisis, (3) development of FD, (4) direct progression to FTMH. Among these pathways, the risk of FD was the lowest, because FD occurred mostly after retinoschisis formation, and most eyes in the 'LMH without retinoschisis' group did not develop retinoschisis during follow-up. Even in eyes that did eventually develop macular retinoschisis during follow-up, we found that the retinoschisis was limited in extent, and that none of the eyes developed subsequent FD. Most eyes (11 out of 14 eyes, 78.6%) with structural worsening developed FTMH. The risk of developing FTMH (25%) was much higher than patients with idiopathic LMH (1%-4%) [5,17]. Similar as our previous report [4], the progression of LMH to FTMH may have achieved through different mechanisms in this group; it may come from (1) tangential premacular traction due to epiretinal membrane (Fig 1L), (2) abnormal vitreofoveal anteroposterior traction (Fig 1N), and (3) macular atrophy with underlying scar.
Patients in the 'LMH and retinoschisis' group were more likely to be women and to have a longer axial length, thicker CRT, thinner subfoveal choroidal thickness, and deeper vertical extension of LMH. These features were similar to those found in MTM generally [1,19,20]. In this group, we found that several evolution sequences occurred: (1) stable, (2) progression to FD, or (3) development of FTMH. Importantly, the present study revealed that LMH with retinoschisis had a more unstable clinical course compared to MTM (progression rate: 48.5% vs. 17%-28%) [13,21]. As described in the literature [4,6], the evolution of LMH to FD or FTMH in the present study can be further classified into two patterns: (1) formation of focal elevation of the outer retinal layer, leading to FD, with subsequent FTMH formation after rupture of the outer retina; (2) widening and/or deepening of the LMH, leading to FTMH directly. Patients who developed FD were prone to progression into FTMH, so most of them (12 out of 16 eyes, 75%) underwent surgery at this stage. It follows that this early surgical intervention may explain the relatively preserved visual function in the 'LMH and retinoschisis' group. We aimed to find predictive factors for structural worsening to guide follow-up planning and the timing of surgery. Multivariable regression analysis on all cases showed that V-shaped LMH was associated with structural worsening, while elevated tissue inside the LMH was a strong protective factor. The results seemed to suggest that the floor of V-shaped LMHs bore more direct traction than that of A-shaped LMHs. On the other hand, elevated tissue on the floor of the LMH may provide cushion to tractional force. The origin of this elevated tissue is unclear. In some cases, it may be the residual outer retina, or it may contain the same cellular components as lamellar hole-associated epiretinal proliferation, which represent either Müller cell proliferation or glial proliferation driven by Müller cells [12]. The presence of this elevated tissue within the LMH rendered the floor thicker and may prevent a full-thickness floor defect. One recent study involving maculopathy in high myopia reported that thinner macular choroidal thickness may be associated with myopic atrophic and neovascular maculopathy, but not with MTM [11]. The present data further showed that the baseline subfoveal choroidal thickness failed to predict the development of FD or FTMH.
In the 'LMH with retinoschisis' group, V-shaped LMH and elevated tissue inside the LMH were also important predictive factors. Subgroup analysis showed that the effect of V-shaped LMH on structural worsening was more significant in eyes with extensive retinoschisis. We postulated that the presence of extensive retinoschisis signified severe traction and weakened structure; these factors amplified the effect of the shape of LMH on the macular structure. Similarly, V-shaped LMH that demonstrated a narrower opening and an more acute angle were more likely to develop FD or FTMH, possibly because the tractional force was concentrated more directly on the LMH floor. Conversely, LMHs that initially presented with an A-shaped configuration, or those that transitioned from V-shaped to A-shaped in the retinoschisis group, were much less likely to progress to FD or FTMH. We hypothesized that in A-shaped LMH, either the disrupted schistic column partially released the traction or the tractional forces were distributed away from the retinal floor. Multivariable regression analysis also revealed that disrupted ellipsoid zone was associated with structural worsening. Eyes with disrupted ellipsoid zone initially had more weakened floor and was prone to structural worsening.
In contrast, among eyes in the 'LMH without retinoschisis' group, no clinical features or OCT parameters could predict structural worsening. Elevated tissue inside the LMH showed marginal protection from structural worsening. These patients might have less risk of progression. However, all patients should be monitored frequently to detect possible complications and to ensure timely surgical intervention.
Govetto et al. had classified idiopathic LMH into tractional and degenerative types based on their shapes and clinical characteristics [22]. In the current study population of high myopia, although some cases had "top hat" or "moustache" configuration, we found that other clinical characteristics (ellipsoid zone disruption, the presence of lamellar hole-associated epiretinal proliferation or epiretinal membrane) were different from what Govetto et al. have described. We believed that LMH in highly myopic eyes were mainly the results of complex traction. And the shape of LMH may depend on the direction and chronicity of the tractional force [23].
The present study had certain limitations because it was retrospective in nature. For instance, the actual time of FD or FTMH development could not be determined precisely. The strengths of the study were that we performed OCT regularly at intervals of 3-6 months, which increased the detection rate and that the follow-up period was longer than in previous studies.
In the present study, LMHs in highly myopic eyes had an unstable clinical course and were at risk of developing to FD or FTMH. Specifically, nearly 40% of these eyes developed FD or FTMH and required vitrectomy during a median follow-up of 67 months. Eyes with extensive retinoschisis and V-shaped LMH had more concentrated and larger traction directed towards the floor, and this was associated with structural worsening. Elevated tissue within the LMH, which may have a reparative or cushioning function, was a strong protective factor. While eyes that had LMH with retinoschisis were at higher risk of FD or FTMH than those with MTM in general, LMH without retinoschisis may progress into FTMH without obvious warning signs. Therefore, frequent follow-up in highly myopic eyes with LMH is warranted.