Clinical characteristics and surgical outcomes of full-thickness macular hole secondary to retinal vein occlusion

Abstract

Purpose

To describe the clinical characteristics and visual and anatomical surgical prognosis for full-thickness macular hole (FTMH) secondary to retinal vein occlusion (RVO).

Method

This retrospective study reviewed 13 eyes from 13 consecutive patients with FTMH after RVO who underwent pars plana vitrectomy (PPV) combined internal limiting membrane (ILM) peeling or ILM flap insertion. Clinical demographics and spectral-domain optical coherence tomography (SD-OCT) features were evaluated. Multiple linear regression was used to analyze the associated factors for the final best-corrected visual acuity (BCVA).

Results

Of the 13 eyes, 11 had a previous branch RVO (BRVO) and 2 had a central RVO (CRVO). The mean minimum linear diameter (MLD) of these secondary macular holes was 469.31 ± 197.96 μm. Eight subjects (61.5%) showed a FTMH secondary to a long-standing RVO (RVO episode longer than 6 months), and the rest 5 subjects (38.5%) developed a FTMH after a recent onset RVO within 6 months. The two groups shared similar baseline BCVA (P = 0.414), MLD (P = 0.218), and BD (P = 0.453). The mean follow-up period was 10.0 months. All cases showed anatomical closure at the last follow-up, with BCVA significantly improved from 0.90 ± 0.28 logMAR at baseline to 0.45 ± 0.34 logMAR after surgery (P<0.001). The final BCVA of the eyes after recent onset RVO was better than the ones after long-standing RVO (P = 0.012). Regression analysis showed that the final BCVA was significantly associated with the baseline BCVA, RVO duration, and the ellipsoid zone (EZ) integrity (P<0.001).

Conclusion

FTMH secondary to RVO could achieve favorable anatomical and visual acuity success with PPV. The cases secondary to a recent RVO might have a better visual prognosis. The initial BCVA and postoperative EZ recovery were associated with the final BCVA.

Introduction

Retinal vein occlusion (RVO), as one of the most common retinal vascular diseases, could be divided into central and branch RVO (CRVO, BRVO), according to the location of the occlusion [1]. RVO can lead to visual loss by the disease pathology itself or by other complications such as central macular edema (CME), neovascular glaucoma, and vitreous hemorrhage [2].

The development of full-thickness macular hole (FTMH), defined as full-thickness dehiscence of the neurosensory retina at the fovea [3], is rare after RVO. Its formation is currently believed to be related to long-term complications of RVO, such as CME deroofing, epiretinal membrane (ERM) traction, and vitreo-macular interface abnormalities [4]. The formation after intravitreal injection was also documented [5–7]. The clinical characteristics of the FTMHs after RVO need to be elucidated clearly. Furthermore, although most secondary FTMHs after RVO could achieve anatomical closure by pars plana vitrectomy (PPV) combined with internal limiting membrane (ILM) peeling, the visual outcomes might be suboptimal [4–7]. It is necessary to investigate the potential predictive factors for the post-surgical visual recovery in FTMHs secondary to RVO.

The optical coherence tomography (OCT) technology could help us observe the foveal microstructure for macular holes preoperatively and postoperatively [8]. Studies have demonstrated that the restoration of the outer retinal structures, including external limiting membrane (ELM) and ellipsoid zone (EZ), is correlated with the post-surgical visual outcomes in idiopathic FTMHs [9, 10]. However, the relationship between these microstructures and the visual prognosis in FTMHs after RVO remains unclear.

Therefore, this study aimed to describe the baseline clinical features and evaluate the associated factors for the visual outcomes after surgery for FTMH secondary to RVO.

Method

Study population

This retrospective study reviewed the clinical records of consecutive patients with FTMH developed after RVO who underwent vitrectomy at Beijing Tongren Hospital from October 2013 to August 2021. The diagnosis of the FTMH secondary to RVO was confirmed by fundus photographs and fluorescein angiography (FA). Eyes with idiopathic macular holes, high myopia (axial length > 26.0 mm or spherical equivalent < − 6.0 diopters), ocular trauma, previous vitreous surgery, or other significant ocular abnormalities that could impair visual condition were excluded. This study was conducted in accordance with the Declaration of Helsinki and its later amendments.

Examinations

Clinical data, including age, gender, RVO duration (since initial vision loss because of RVO till vitrectomy for FTMHs), and previous treatment for RVO, were recorded. A recent onset RVO was considered as an RVO duration within 6 months [4]. All cases underwent complete ophthalmological examinations, including best corrected visual acuity (BCVA), intraocular pressure (IOP, by Full Auto Tonometer TX-F; Canon Canada, Quebec), axial length (AXL, from IOL Master Biometry, Carl Zeiss Meditec, Jena, Germany), slit-lamp biomicroscopy, binocular indirect ophthalmoscopy, fundus photography (fundus camera, TRC-50; Topcon, Tokyo, Japan), and spectral-domain OCT (SD-OCT, Cirrus high-definition OCT; Carl Zeiss, Dublin, CA).

SD-OCT was used to evaluate the preoperative FTMH status and postoperative foveal microstructures, using Macular Cube 512*128 protocol and the HD 5-line raster. Minimal linear diameter (MLD) was measured as the narrowest distance between the hole edges at the mid retina, by drawing a line roughly parallel to the retinal pigment epithelium (RPE) on a horizontal scan. The basal diameter (BD) was the length of the detached photoreceptors at the level of RPE. The anatomical hole closure was defined as the continuous neurosensory retina at the fovea without RPE exposure [11]. Central foveal thickness (CFT) was measured as the vertical distance between the inner boundaries and RPE. All these quantitative measurements were measured on the same section, using by SD-OCT software’s calipers. The ELM and EZ were considered recovered if the corresponding hyper-reflective lines appeared completely continuous on both vertical and horizontal SD-OCT scans. The main post-surgical outcomes included the primary closure rate, final BCVA, CFT, and the integrity of the ELM and EZ at the last visit.

Surgical procedure

All patients underwent a standard 23G, 3-port PPV by one experienced surgeon (W.L.). Phacoemulsification and intraocular lens implantation were performed first when necessary. After core vitrectomy, the posterior vitreous detachment was induced if it was absent. If present, the ERM was peeled off. For the large FTMHs (MLD > 400 μm) with retinal thinning, the conventional ILM peeling or inverted ILM flap insertion was performed based on individual intraoperative findings. In eyes with conventional ILM peeling, the ILM was removed about 2 disc diameters around the FTMH with or without 0.25% indocyanine green staining. In eyes with inverted ILM flap insertion, the ILM flap was created by partially peeling the ILM with a radius of 0.5-1hole diameter around the macular hole, leaving it adherent to the hole edge. Then the surgeon covered the hole with the inverted ILM flap from all sides using forceps. Laser photocoagulation was applied in ischemic areas found on FA and any retinal degenerative lesions and tears observed during surgery. Finally, fluid-air exchange was performed, then these eyes were filled with gas (14%C2F6 or 12%C3F8) or sterile air. Every patient maintained a face-down position for a week. Further consultation was scheduled 1, 4, and 10 months after surgery, then once a year afterwards.

Statistical analysis

All statistics were performed by SPSS Statistics v26.0 (IBM, Chicago, IL). BCVA was transformed into the logarithm of the minimal angle of resolution (LogMAR) scale for statistical analysis. If continuous variables conformed to a normal distribution, they were presented as mean ± standard deviation (SD) and the t-test was applied; otherwise, they were presented as median and percent intervals, and the Mann-Whitney test was employed. Categorical variables were expressed as frequencies and percentages, and Fisher’s exact test was used to compare. Multivariate linear regression analysis with backward elimination (selection level 5%) was used for analyzing the associated factors with final BCVA, and bootstrapping was performed to validate the findings. Additionally, internal validation was carried out via leave-one-out cross-validation (LOOCV). A P < 0.05 was considered statistically significant.

Result

Baseline characteristics

Totally, 13 eyes of 13 patients (5 men and 8 women) were included, with a mean age of 62.0 ± 2.2 years (range: 44 to 70). 10 cases were combined with a history of BRVO, and 2 cases were after CRVO. One patient came to our clinic with vision loss in the left eye in the last two months. We found a FTMH secondary to a long-standing BRVO (ten years) after examination, who had a previous branch retinal artery occlusion (BRAO) history without remaining severe visual impairment (Fig. 1). 2 cases were combined with vitreomacular traction (VMT). 1 subject had received intravitreal injection (Conbercept), 4 subjects had received laser photocoagulation, and 1 subject had received both intravitreal injection (Ranibizumab) and laser photocoagulation for CME after RVO.

Fig. 1.

The spectral-domain optical coherence tomography (SD-OCT) images of a full-thickness macular hole (FTMH) after branch retinal vein occlusion (BRVO) in a 53-year-old woman who had a prior history of branch retinal artery occlusion (BRAO). (A). The edge of the FTMH was flattened, the minimal linear diameter of the hole was 557 μm, basal diameter was 1305 μm. The preoperative best-corrected visual acuity (BCVA) was 20/100. (B). The FTMH showed anatomical closure without external limiting membrane (ELM) and ellipsoid zone (EZ) recovery. The final BCVA was 20/67

The mean MLD of these eyes was 469.3 ± 198.0 μm (range:203 to 877), and the mean BD was 1336.2 ± 691.2 (range: 384 to 2843). 8 eyes with MLD larger than 400 μm were considered as large FTMHs after RVO, according to previous classification [12]. Epiretinal membranes were found in 5 eyes (38.5%). There were 3 eyes (23.1%) that had flattened edges of FTMH, and the rest 10 eyes (76.9%) had edematous edges of FTMH. 10 eyes (76.9%) underwent phacoemulsification and intraocular lens implantation at the same time. 8 eyes were treated with laser during surgery. 11 eyes were treated with ILM peeling, and the other 2 eyes were treated with ILM flap insertion. 8 eyes (61.5%) were used sterile air as intraocular tamponade, and others were used expansile gas (1 subject used C3F8, 4 subjects used C2F6). The clinical characteristics of each case were listed in Table 1. The median follow-up was 10.0 months (range: 3 to 24).

Table 1.

Clinical presentation and surgical outcomes of full-thickness macular holes after retinal vein occlusion

No.	Age/Sex	Type of RVO	Duration of RVO (months)	Previous RVO treatment	Concurrent ERM	MLD (µm)		BD (µm)	Baseline BCVA (Snellen)		Surgical intervention			ELM integrity at last visit	EZ band integrity at last visit	Final CFT (µm)	Final BCVA (Snellen)
											Catract surgery	Surgical techniques	Intraocular tamponade
1	74/F	BRVO	24	laser photocoagulation	+	563	2383			20/200	+	ILMP	C2F6	incomplete	incomplete	-	20/100
2	61/M	BRVO	2	none	+	272	808			20/286	+	ILMP	C2F6	incomplete	incomplete	283	20/67
3	64/M	BRVO	6	none	-	750	1887			20/100	+	ILMP	C2F6	complete	incomplete	150	20/33
4	63/M	BRVO	5	none	+	203	1235			20/100	+	ILMP	sterile air	complete	incomplete	368	20/25
5	44/F	BRVO	2	IVC	-	474	1185			20/200	-	ILMP	sterile air	complete	complete	197	20/20
6	63/F	CRVO	12	laser photocoagulation	-	384	384			20/200	+	ILMP	C2F6	incomplete	incomplete	152	20/100
7	66/F	BRVO	12	none	+	489	1509			20/500	+	ILMP	C3F8	complete	incomplete	389	20/200
8	65/M	BRVO	12	none	-	531	1298			20/200	+	ILMP	sterile air	complete	complete	251	20/67
9	54/F	BRVO	12	laser photocoagulation	+	461	941			20/67	-	ILMP	sterile air	complete	complete	125	20/40
10	63 M	BRVO	2	laser photocoagulation	-	207	606			20/67	+	ILMP	sterile air	complete	complete	270	20/20
11	66/F	CRVO	24	laser photocoagulation + IVR	-	877	2843			20/400	+	ILM insertion	sterile air	incomplete	incomplete	156	20/200
12	53/F	BRVO (BRAO history)	120	none	-	557	1305			20/100	-	ILM insertion	sterile air	incomplete	incomplete	73	20/67
13	70/F	BRVO	24	none	-	333	986			20/100	+	ILMP	sterile air	complete	complete	192	20/33

RVO, retinal vein occlusion; ERM, epiretinal membrane; BCVA, best corrected visual acuity; BRVO, branch retinal vein occlusion; CRVO, central retinal vein occlusion; BRAO, branch retinal artery occlusion; IVC, intravitreal injection of Conbercept; IVR, intravitreal injection of Ranibizumab; MLD minimal linear diameter; BD, basal diameter; ILMP, internal limiting membrane peeling; ELM, external limiting membrane; EZ ellipsoid zone

The duration of RVO ranged from 2 to 120 months with an average of 19.8 months. 8 eyes (61.5%) had FTMH secondary to a long-standing RVO (RVO episode longer than 6 months), and 5 eyes (38.5%) had FTMH after a recent onset RVO within 6 months. In the comparison of these two groups, we found that only the gender had statistically significant differences (more females in the FTMH after onset RVO group and more males in the FTMH after long-standing RVO group, P = 0.032), while the initial BCVA (P = 0.414), MLD (P = 0.218), and BD (P = 0.453) of the macular hole at baseline showed no significant differences. The detailed comparison between these two groups is shown in Table 2.

Table 2.

Comparison of baseline features and postoperative outcomes between the full-thickness macular hole secondary to onset and long-standing retinal vein occlusion

Parameters	FTMH after onset RVO (n = 5)	FTMH after long-standing RVO (n = 8)	P value	Effect Size (95% CI)
Age, years	median 59.00, IQR 52.50–63.50	median 65.50, IQR 56.25-69.00	0.140a	Median Diff: -3.000 (-13.000 to 8.000)
Gender, male/ female	4/1	1/7	0.032b	OR: 28.000 (1.350 to 580.591)
Type of RVO (BRVO/CRVO)	5/0	6/2	0.487b	OR: 1.333 (0.894 to 1.989)
AXL (mm)	median 24.17, IQR 22.91–24.38	median 23.13, IQR 22.64–23.78	0.092a	Median Diff: 0.845 (-0.760 to 1.620)
MLD (µm)	381.20 ± 233.80	524.38 ± 164.12	0.218c	MD: -143.175 (-98.239 to 384.589)
BD (µm)	1144.20 ± 491.01	1456.13 ± 799.02	0.453c	MD: -311.925 (-569.927 to 1193.777
Combined ERM/or not	3/2	5/3	1.000b	OR: 0.900 (0.091 to 8.899)
Initial BCVA at last visit (LogMar)	0.81 ± 0.25	0.95 ± 0.30	0.414c	MD: -0.139 (0.221 to 0.498)
CFT at last visit (µm)	253.60 ± 83.90	191.14 ± 103.14	0.292c	MD: 62.457 (-187.588 to 62.674)
ELM restoration at last visit (yes/no), n. (%)	4/1 (80.0%)	4/4 (50.0%)	0.565b	OR: 4.000 (-0.299 to 53.468)
EZ band restoration at last visit (yes/no), n. (%)	2/3 (40.0%)	3/5 (37.5%)	1.000b	OR: 1.111 (0.112 to 10.986)
Final BCVA (logMAR)	0.17 ± 0.22	0.62 ± 0.29	0.012c	MD: -0.452 (0.120 to 0.784)
Follow-up period (months)	14.40 ± 9.10	7.25 ± 3.58	0.157c	MD: 7.150 (-18.254 to 3.954)

RVO, retinal vein occlusion; FTMH, full-thickness macular holes; ERM, epiretinal membrane; BRVO, branch retinal vein occlusion; CRVO, central retinal vein occlusion; IQR. interquartile range; AXL, axial length; MLD, minimal linear diameter; BD, basal diameter; BCVA, best corrected visual acuity; CFT, central foveal thickness; logMAR, logarithm of the minimal angle of resolution; ELM, external limiting membrane; EZ, ellipsoid zone; MD: Mean Difference; OR: Odds Ratio; Median Diff: Hodges-Lehmann Median Difference; CI, confidence interval

^a Mann-Whitney test;

^b Fisher’s exact test;

^c T-test

Anatomical outcomes

All cases showed anatomical closure at 1 month after surgery and remained closed until the last follow-up. One case received intravitreal injection (Ranibizumab) at 4 months after surgery because of the persistent macular edema and choroidal neovascularization (CNV) emerged at 1 month after surgery, which were resolved during later follow-ups (Fig. 2). The mean CFT at the last visit was 217.2 ± 96.9 μm. The number of eyes with ELM restoration gradually increased from 6 at 1 month to 8 at 4 months postoperatively, while the number of eyes with EZ band restoration increased from 2 to 4. At the last visit, there were 8 eyes (61.5%) recovered ELM community and 5 eyes (38.5%) restored both ELM and EZ band integrity. The integrity recovery of the EZ band was always preceded by the restoration of ELM integrity.

Fig. 2.

The spectral-domain optical coherence tomography (SD-OCT) images of a 66-year-old woman with full-thickness macular hole (FTMH) secondary to branch retinal vein occlusion (BRVO). (A). The FTMH with an edematous edge, had a minimal linear diameter of 489 μm, and a basal diameter of 1509 μm. The preoperative best-corrected visual acuity (BCVA) was 20/500. (B). At 1 month postoperatively, the FTMH closed, and a choroidal neovascularization (CNV) was found at the nasal of the foveola. The visual acuity improved to 20/400. (C). At 4 months postoperatively, the visual acuity was kept at 20/400. The patient received an intravitreal injection (Ranibizumab). (D). At the last visit, the CNV subsided, and cystoid spaces could be found at the fovea. The final BCVA was 20/200

Visual outcomes and associated factors

The mean BCVA significantly improved from 0.9 ± 0.3 logMAR at baseline to 0.5 ± 0.3 logMAR (P<0.001). Postoperatively, the median visual acuity improved from 20/200 to 20/67. 6 subjects (46.2%) achieved a final BCVA of 20/40 or above. The eyes with long-standing RVO showed worse visual acuity than those after recent onset RVO (0.6 ± 0.3 vs. 0.2 ± 0.2 logMAR BCVA, P = 0.012). A comparison of the postoperative characteristics of the two groups revealed that the final CFT (P = 0.292) and the recovery rates for the ELM (P = 0.565) and EZ bands (P = 1.000) were not significantly different.

We compared the surgical outcomes between the eyes using sterile air and long-acting gas as intraocular tamponade. There were no significant statistical differences in final BCVA (0.3 ± 0.3 vs. 0.6 ± 0.3 logMAR BCVA, P = 0.134), final CFT (204.0 ± 92.1 μm vs. 243.5 ± 115.3 μm, P = 0.619), and the recovery rates of ELM (75.0% vs. 40.0%, P = 0.293) and EZ band (62.5% vs. 0.0%, P = 0.075) at the last visit.

Multivariate linear regression analysis showed that the related factors to the final BCVA included the baseline BCVA, RVO duration, and the EZ band integrity at the last visit (F = 47.182, P<0.001, R² = 0.920) (Table 3). The bootstrap estimates, presented in Table 4, showed consistent coefficients and comparable standard errors, thereby validating the initial model’s reliability. The LOOCV yielded a mean absolute error (MAE) of 0.10, a mean-square error (MSE) of 0.02, and an R² of 0.925, confirming the model’s robustness and the absence of overfitting.

Table 3.

Factors associated with final BCVA in full-thickness macular hole secondary to retinal vein occlusion using multivariable linear regression analysis

Variables	β	Coefficient β	SE	95% CI	P Value
Baseline BCVA	0.593	0.491	0.114	0.336 to 0.849	0.001
RVO duration	0.364	0.540	0.057	0.235 to 0.493	<0.001
EZ integrity at the last visit	-0.232	-0.344	0.062	-0.371 to -0.093	0.004

RVO, retinal vein occlusion; BCVA, best corrected visual acuity; EZ, ellipsoid zone; SE, standard error; CI, confidence interval

Table 4.

Bootstrapping of the regression coefficients

Variables	β0	SE	95% CI
Baseline BCVA	0.593	0.130	0.341 to 0.850
RVO duration	0.364	0.066	0.220 to 0.481
EZ integrity at the last visit	-0.232	0.074	-0.363 to -0.062

SE, standard error; CI, confidence interval

Discussion

In this study, we demonstrated the clinical characteristics of FTMH formed secondary to RVO. These atypical macular holes achieved favorable anatomical and functional success after operation.

Previous studies have suggested several possible pathogeneses of FTMHs after RVO. Inflammation caused by RVO could lead to retinal ischemia and hemorrhage, which increased the expression of vascular endothelial growth factor (VEGF), transforming growth factor-β, and other inflammatory factors [13]. These factors disrupted the blood-retinal barrier, causing the onset and progression of macular edema [14]. The cystoid degeneration and the circulatory disturbances could furthermore induce the thinning of the inner wall at the foveal cystoid spaces. When the central foveal cysts deroofed, the FTMH formed [15, 16].

In addition, some researchers thought that the secondary ERM caused by persistent ischemia after RVO could contract and exert extra tangential traction on the damaged macular retina, which might cause FTMH formation [5, 17, 18]. In this study, we found 38.5% of the eyes combined with ERM, which is slightly lower than the incidence of 50.0% reported in the study by Mishra et al. [4].

Peripheral vitreous traction might be exerted by vitreous incarceration at the site of the injection, leading to the posterior vitreous detachment [19–21]. When macular thickness was dramatic and showed rapid shrinkage after injection or laser therapy, an FTMH might form [22]. Meanwhile, the transient increase in the intraocular pressure could perturb the blood-retinal barrier, inducing tangential tractional forces at the surface of the retina [23, 24]. Some studies considered that the vitreous adherence to the fovea also represented an etiopathogenetic factor in the development of FTMH after RVO [25, 26]. The small number of our cases for each prior treatment modality precluded any meaningful statistical subgroup analysis to determine their individual effects on baseline and postoperative features.

Some secondary macular holes might be difficult to close and have worse structural and functional outcomes because of the primary disease [27]. Several studies have reported that most FTMHs secondary to RVO could achieve anatomical closure after surgery and had improved visual acuity [4–7]. All the eyes in our study provided satisfactory anatomic outcomes, with a similar visual prognosis to previous studies. Previous studies had observed that the atypical macular holes with edematous edges usually had more satisfactory surgical outcomes than those with atrophic edges [28]. The included eyes with both edematous edges and flattened edges showed anatomical success after vitrectomy. In our research, 76.9% of the macular holes showed edematous edges, in contrast to Mishra’s study, 75.0% of their study eyes with rounded and raised edges [4].

Furthermore, we found that the large FTMHs after RVO achieved anatomic closure, with ILM peeling (6/8, 75.0%) and ILM flap insertion (2/8, 25.0%). According to other studies, the intraocular tamponade for FTMH after RVO was usually long-acting gas, such as C3F8 and SF6 [4–7]. Some studies have suggested using long-acting gas in FTMHs after RVO, because of the lower surgical success rates, inferior to idiopathic macular holes [27]. Though we used sterile air for most of the eyes (61.5%), the FTMHs were all closed with vision improved. It seems that the use of sterile air appeared feasible to achieve anatomical closure for FTMH after RVO, as Zhang K et al. have reported in the research of highly myopic macular holes [29].

All the eyes of our study showed improvement in visual function after vitrectomy, 46.2% of the eyes achieved a final BCVA of 20/40 or above. We found the final BCVA was associated with the initial BCVA, RVO duration, and postoperative EZ recovery. The size of the holes depicted by MLD and BD exhibited no correlations to the final BCVA, as well as the postoperative CFT, similar to the conclusion in the previous study on traumatic macular holes [30]. The eyes secondary to recent onset RVO had better final BCVA than those after long-standing RVO. We hypothesized that this might be due to retinal dysfunction resulting from prolonged retinal ischemia and hypoxia because of the occlusion [2]. The vascular cells would lose progressively. The persistent ischemia could conduce to epithelium degeneration and irreversible damage to macular ganglion cells. The visual acuity might tend to worsen over time [31]. These facts suggested that complete assessments at the early stage and regular follow-ups of RVO patients are crucial.

Mishra et al. [4] reported that the FTMHs with a RVO history of 6 months or more had a better visual prognosis after surgery than the ones after a recent onset RVO. They explained this paradoxical result as the ones with recent onset RVO combined with persistent intra- or sub-retinal edema, which led to lesser visual acuity gains than the cases with RVO longer than 6 months. Because there is little research on FTMH secondary to RVO at present, the exact reason for this difference is not clear. We speculated that the major reason for this discrepancy might be attributed to their shorter follow-up period of 6 months. The central photoreceptor regeneration might proceed longer than 6 months. The restoration of the outer retina by photoreceptor migration and realignment was the basis for higher postoperative visual acuity [31, 32]. Therefore, the visual acuity of the eyes in Mishra’s study would likely improve after the edema regressed and the outer retinal layers recovered in further follow-ups. But the observed discrepancies may not solely reflect the true difference but could also be influenced by variations in cohort characteristics, limited by small sample sizes.

EZ band integrity has been reported as an independent prognostic factor for visual acuity outcomes after treatment for both FTMH and RVO [32–34]. After the recovery of the ELM defect mediated by Müller cells, the photoreceptor cells underwent centripetal displacement. The disrupted EZ band would be restored over a longer period than that of ELM [32, 35]. In our research, 46.2% of the eyes showed complete restoration of ELM integrity at 1 month postoperatively, but only 15.4% of the eyes achieved EZ band recovery. 61.5% of the eyes had intact ELM, and 38.5% of the eyes restored both ELM and EZ band at the last visit. In previous studies focused on the idiopathic and traumatic FTMHs, the postoperative visual acuity was more highly correlated with the EZ defect rather than the ELM defect [30, 36]. Hence, the integrity of the EZ band might be a dominant predictor of the visual acuity prognosis after the surgical repairment of FTMH after RVO. And it might take a longer time to get a complete recovery than our follow-up period. Probably, more intact EZ bands would be observed if we could take further consultation.

There was one patient who had a hyper-reflective lesion that broke through the RPE layer 1 month after vitrectomy and subsided after being treated with intravitreal injection of Ranibizumab. We hypothesized that the patient developed CNV [37]. CNV following RVO could be triggered by VEGF upregulation from chronic ischemia and exacerbated by macular edema-induced damage to the RPE and Bruch’s membrane [38]. Furthermore, RVO elevated the hydrostatic pressure in venous collaterals, facilitating anastomoses between fibrovascular complexes and the retina. When Bruch’s membrane was breached, these anastomoses might shunt retinal venous flow into the choroid, which nourished the neovascularization [39, 40]. Additionally, vitrectomy could alter the retinal microenvironment, activate microglia, and promote the secretion of inflammatory and pro-angiogenic cytokines, thereby fostering CNV development [41].

This study has several limitations. Its retrospective design precluded the systematic collection of patient-reported outcomes, thereby omitting the patient’s perspective on functional improvement. The small sample size, derived from a single surgeon and center, and the inclusion of only surgically managed patients, introduced the risks of selection bias and unmeasured confounding. The variations in concurrent cataract surgery and follow-up duration may influence the interpretation of visual outcomes and long-term stability. The 3 phakic eyes combined cataract progression after the operation, and might cause little vision loss, one with a final BCVA of 20/40, one with 20/67, and the other reached 20/20 at the final follow-up examination. There was no significant difference in the final BCVA between the eyes that underwent cataract surgery or not (0.5 ± 0.4 vs. 0.3 ± 0.3 logMAR, P = 0.339). Therefore, the patients who did not undergo cataract surgery might have no significant impact on the results of this present study. For patients who participated in follow-up for six months or less (n = 5) and 10 months at least (n = 8) after the operation, there were significant differences in only the preoperative BCVA (0.7 ± 0.2 vs. 1.0 ± 0.3 logMAR, P = 0.026), whereas the other preoperative and postoperative anatomical parameters and visual acuity did not significantly differ. The small sample size precluded the use of multivariate models to adjust for all potential baseline confounders, such as differences in gender distribution between groups. Consequently, our findings should be considered preliminary and hypothesis-generating, warranting validation through future prospective, multi-center studies with larger cohorts and standardized protocols.

In conclusion, this small retrospective series provides preliminary evidence that the PPV combined ILM peeling or ILM flap insertion could achieve favorable anatomical and visual acuity improvement in FTMHs after RVO. Eyes with FTMHs secondary to a recent RVO might have a better prognosis after surgery. The ones with better initial BCVA and postoperative EZ recovery usually achieved better final BCVA. In addition, early assessments and regular follow-ups of RVO patients are needed to avoid the formation of secondary FTMHs.

Author contributions

Xijin Wu: substantial contributions to the conception and design of the work; acquisition, analysis, and interpretation of data for the work; drafting and revising the work; final approval of the version to be published; and agreement to be accountable for all aspects of the work. Xiaohan Yang, Biying Qi, Yanping Yu, Xinbo Wang, Qinlang Jia, Xiao Feng, and Wu Liu: revising the work; final approval of the work, and agreement to be accountable for all aspects of the work.

Funding

No funding was received for this research.

Data availability

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the article.

Declarations

Ethical approval

All procedures performed in studies involving human participants were adhered to the Helsinki declaration and its lateral amendments. Approval was granted by the ethical review committee of Beijing Tongren Hospital, Capital Medical University.

Consent to participate

Informed consent was obtained from all individual participants included in this study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.