Rapid choroidal neovascularization after non-flap internal limiting membrane peeling for macular hole in a patient with AMD: a case report and literature review

Abstract

Background

Choroidal neovascularization (CNV) following macular hole (MH) surgery is a rare complication. While the exact pathogenesis remains unclear, previous reports in the literature suggest that it can occur in patients with pre-existing age-related macular degeneration (AMD) or macular drusen. In this report, we present a case that highlights the potential biomechanical triggers of this complication in an eye with AMD.

Case presentation

We report the case of a 70-year-old female patient with AMD who presented with a full-thickness macular hole (FTMH). She underwent an uncomplicated 25-gauge pars plana vitrectomy with brilliant blue G-assisted internal limiting membrane (ILM) peeling and perfluoropropane gas tamponade. Notably, the ILM flap technique was not utilized; only standard, non-flap peeling without intraoperative complications was performed. Just 6 weeks postoperatively, the patient experienced a sudden decline in visual acuity due to the rapid development of CNV at the border of the ILM peeled-off area. Intravitreal aflibercept treatment was subsequently planned.

Conclusions

This case suggests that, in eyes with underlying AMD, CNV may develop rapidly after standard ILM peeling even without an ILM flap technique. Biomechanical stress related to ILM removal, potentially involving macular displacement and mechanical stress on Müller cells, may be a possible contributing factor, but causality cannot be established from a single case. Physicians should maintain high vigilance for this vision-threatening complication during the early postoperative period after MH surgery.

Background

A macular hole (MH) is classified as a disorder of the vitreomacular interface, characterized by a complex mechanical origin. The development of MH is influenced by the unique shape of abnormal anteroposterior and tangential vitreous traction resulting from an incomplete posterior vitreous detachment [1]. The intentional removal of the macular internal limiting membrane (ILM) has led to a notable increase in the anatomical success rate in the surgical management of MH. Over the past decade, the peeling of the ILM in the course of surgery for a MH has become a standard procedure, commonly performed by most surgeons. While the intentional removal of the ILM effectively relieves tangential traction, modern optical coherence tomography (OCT) has revealed that this procedure can also induce structural alterations and mechanical stress on the inner retinal surface following ILM peeling, indicating potential progressive retinal damage [2–4].

The development of choroidal neovascularization (CNV) following MH surgery is a rare complication. Previous reports indicate that this complication frequently occurs in patients with pre-existing age-related macular degeneration (AMD) or macular drusen [5–9]. The interval from surgery to CNV formation has been reported to range from 1 month to 30 months. While the exact pathogenesis is not fully understood, researchers have attributed this complication to age-related degenerative changes in the retinal pigment epithelium and Bruch’s membrane, mechanical trauma during ILM peeling, or the toxicity of vital dyes used during surgery [5–9].

In this report, we present a case of rapid CNV development occurring within just 6 weeks following standard, non-flap ILM peeling without intraoperative complications for a full-thickness macular hole (FTMH) in a patient with underlying AMD. A similarly short interval has been described in isolated prior reports [5, 8]; the novelty of this case lies not in the timing alone, but in the specific combination of pre-existing AMD, brilliant blue G-assisted staining, standard non-flap ILM peeling, and CNV localized precisely at the border of the ILM-peeled area, a combination not previously reported together. By highlighting this case, we aim to discuss the potential biomechanical triggers associated with mere ILM removal in MH surgery; we present this as a hypothesis-generating observation rather than evidence of a definitive causal mechanism.

Case presentation

A 70-year-old female patient reported a gradual decline in vision in her left eye (LE) over the past 6 months, with a sudden worsening occurring in the last 4 weeks. Both eyes experienced uncomplicated cataract surgery 2 years ago.

Upon examination, her best corrected visual acuity (BCVA) (Snellen) was measured at 20/20 in the right eye (RE) and 20/200 in the LE. Intraocular pressures were 14 mmHg for both eyes. A well-positioned, centered intraocular lens is present in both eyes. RE’s fundus examination and optical coherence tomography (OCT) revealed small calcified drusen (Fig. 1A). During the retinal examination for LE, drusen and FTMH was noted. Preoperative OCT imaging revealed FTMH as stage 4 with a minimum hole diameter measuring 426 μm and minor changes (small calcified drusen), and no CNV was detected (Figs. 1B and 2A).

Fig. 1.

Preoperative OCT imaging of both eyes. (A) OCT of the right eye demonstrating a preserved foveal contour with sub-RPE elevations consistent with small calcified drusen. BCVA was 20/20. (B) OCT of the left eye reveals a stage 4 FTMH with elevated, cystic edges and a minimum hole diameter of 426 μm. Underlying RPE irregularities consistent with drusen are noted, with no evidence of associated CNV. BCVA was 20/200

Fig. 2.

Sequential OCT imaging of the left eye. (A) Preoperative imaging reveals a stage 4 FTMH with a minimum diameter of 426 μm and underlying small calcified drusen, with no evidence of CNV. (B) Four weeks following pars plana vitrectomy with ILM peeling, OCT confirms the successful anatomical closure of the macular hole. (C) Six weeks postoperatively, corresponding to a rapid decline in visual acuity, OCT demonstrates a fibrovascular PED and hemorrhage superonasal to the closed hole, exactly at the border of the ILM peeled-off area. (D) Two weeks after the first intravitreal aflibercept injection, the scan shows persistent subretinal fluid and a massive fibrovascular PED, corresponding to the dull-white subretinal membrane observed clinically. Note: this section is centered slightly inferior to the fovea due to image quality constraints directly through the foveal center; the lesion itself remains superonasal to the fovea, as confirmed on fundus examination and in Fig. 2C

She underwent 25-gauge pars plana vitrectomy using the Dorc Eva vitrectomy system (DORC, Netherlands Ophthalmology Research Center, a ZEISS company). Following performing core vitrectomy, a posterior vitreous detachment was established and expanded to the boundary of the vitreous base. The ILM was colored with 0.025% brilliant blue G dye (DORC Dutch Ophthalmic Research Center, a ZEISS company) for exactly 1 min, and no re-staining was performed. The ILM edge was engaged and elevated directly with 25-gauge ILM forceps (Grieshaber, Alcon Laboratories, Fort Worth, Texas, USA) using a pinch-and- peeling method, without the use of a secondary initiating instrument. Illumination was provided by a xenon light source (BrightStar, DORC), with direct macular illumination during ILM peeling strictly limited to approximately 2 min. A 12% perfloropropan (C3F8) tamponade was administered, and the patient was instructed to maintain a face-down position for a duration of 3 to 4 days. Intraoperative images of the standard ILM peeling procedure are demonstrated in (Fig. 3A-D). Four weeks following the procedure, the gas was fully absorbed, resulting in an improvement of BCVA to 6/20 and MH was closed (Fig. 2B). Approximately two weeks following the last examination, the patient exhibited a rapid decline in vision lasting for a duration of 2 days. The BCVA had decreased to 20/200. A fundus examination showed drusen and subretinal hemorrhage in the superonasal portion of the MH, at the border of the ILM peeled-off area. OCT revealed fibrovascular pigment epithelial detachment (PED) and hemorrhage (Fig. 2C). A monthly intravitreal aflibercept regimen (2 mg/0.05 mL, Bayer, Germany) was planned. Two weeks after the first injection, the fundus examination revealed subretinal fluid and a dull-white subretinal membrane located superonasal of the fovea. OCT revealed subretinal fluid and large fibrovascular PED (Fig. 2D). The patient did not return for further scheduled visits thereafter and therefore received only a single injection. Written informed consent was obtained from the patient for the publication of this case report and any accompanying images.

Fig. 3.

Intraoperative photographs of the left eye during macular hole surgery. Sequential frames (A-D) demonstrate the standard ILM peeling procedure. After staining with 0.025% brilliant blue G dye, the ILM is carefully removed utilizing a pinch-and-peeling method with 25-gauge ILM forceps. Notably, only standard peeling was performed without the application of an ILM flap technique. Underlying macular drusen are also visible during surgical manipulation

Discussion

This case demonstrates the possibility of rapid CNV development following only standard ILM peeling, without applying the ILM flap technique, for a large FTMH in a patient with underlying AMD. Although previous studies have reported CNV development after MH surgeries that included ILM peeling, our case is particularly noteworthy due to the extremely short interval of just 6 weeks between the uncomplicated surgery and the onset of CNV. Furthermore, it suggests that even without additional surgical manipulations such as the flap technique, ILM removal itself may act as a potential trigger for CNV in eyes with pre-existing age-related degenerative changes.

There are a limited number of studies in the literature reporting the relationship between MH surgery and the development of CNV and examining case characteristics and etiopathogenesis mechanisms (Table 1). In earlier reports, Banker et al. [5] noted the development of CNV 6 weeks after macular hole surgery; however, their surgical technique did not involve ILM peeling, but rather the removal of epiretinal membranes if present. Similarly, Berinstein et al. [6] noted CNV development after macular hole surgery in an eye with significant macular drusen, and their surgical technique also involved only epiretinal membrane removal without ILM peeling. Tabandeh et al. [7] reported that in a series of 9 patients aged between 51 and 76 years, CNV development occurred between 6 weeks and 30 months after surgery, with 78% of patients having pre-existing drusen and 56% having undergone ILM peeling during surgery; they suggested that the primary mechanism was the acceleration of the age-related degenerative process in the RPE and Bruch’s membrane by surgery. Similarly, Oh et al. [8] reported the development of CNV in a 64-year-old patient with drusen in both eyes at 1 month following ILM peeling, attributing this to age-related changes as well as mechanical trauma during ILM peeling or indocyanine green (ICG) dye toxicity. Lee and colleagues [9] reported detecting CNV seven months after surgery in a 73-year-old patient with drusen; they argued that the surgical components provided an angiogenic stimulus by causing mechanical damage or multinuclear giant cell activation in the degenerated Bruch membrane. This complication can also be seen in cases without underlying AMD; indeed, Natarajan and his team directly attributed CNV that developed 25 weeks after surgery in a 70-year-old diabetic patient to mechanical manipulation during surgery, light toxicity, or RPE damage caused by trypan blue dye [10].

Table 1.

Clinical characteristics, surgical techniques, and onset time of choroidal neovascularization after macular hole surgery in the literature

Source	Age/Gender	Preop AMD/drusen	ILM Technique	Time to CNV
Banker 1997	Not specified	Widespread abnormality of the RPE	Standard vitrectomy, ERM peeling (if present), gas tamponade	6 weeks
Berinstein 2000	Not specified (mean age 68.9 in study)	Yes (AREDS Category 2 or 3 drusen)	Standard vitrectomy, ERM peeling (if present), gas tamponade	Not specified
Tabandeh 2004 (series, 9 eyes)	mean 68	Drusen in 7/9 eyes (78%)	5/9 ILM peeling	6 weeks-30 months
Oh 2012	64 F	Yes, bilateral	ILM peeling (ICG)	1 month
Lee JH 2015	73 F	Yes (+ neovascular AMD in fellow eye)	ILM peeling	7 months
Natarajan 2006	70 F	None (diabetic)	ILM peeling (trypan blue)	25 weeks

AMD, age-related macular degeneration; CNV, choroidal neovascularization; F, female; ICG, indocyanine green; ILM, internal limiting membrane

Our case, while perfectly consistent with the high-risk patient profile in the literature in terms of the history of AMD before surgery, is noteworthy for the development of CNV only 6 weeks after surgery, despite performing only standard, non-flap ILM peeling (with brilliant blue G) without intraoperative complications. It is well established that advanced patient age and the presence of pre-existing AMD significantly lower the structural and metabolic threshold of the outer retina. In such highly vulnerable, drusen-stressed RPE layers, the potential synergistic phototoxic profile of vital dyes combined with endoillumination light must be rigorously addressed to rule out iatrogenic acceleration of the pathology. Unlike the cases reported by Oh et al. [8] and Natarajan et al. [10], where potentially toxic dyes like trypan blue or ICG were utilized, our use of the safer brilliant blue G dye significantly minimizes the likelihood of a chemically driven CNV. Furthermore, to mitigate the risk of synergistic phototoxicity, our surgical parameters were strictly controlled: the BBG exposure was limited to just 1 min without any re-staining, and the direct macular exposure to the xenon endoillumination light source during the ILM peeling was restricted to approximately 2 min. These durations remain well below the exposure thresholds associated with measurable RPE phototoxicity in experimental studies of BBG and xenon/LED endoillumination [11–13]. While a mild phototoxic or chemical contribution to the already vulnerable RPE cannot be entirely excluded, the strict limitation of these iatrogenic factors emphasizes the significance of mechanical forces. This short 6-week period parallels the earliest cases in Tabandeh’s series, Banker’s early report and Oh’s 1-month-old case [5, 7, 8]. Unlike earlier reports by Banker [5] and Berinstein [6] where ILM peeling was not performed, the localization of the CNV exactly at the border of the ILM peeled-off area in our patient strongly highlights the specific mechanical impact of ILM removal. Notably, the CNV localized to the border of the peeled area rather than the central site of forceps engagement, arguing against direct focal instrumentation trauma and supporting diffuse tangential traction at the margin of ILM removal as the more likely trigger. To explain this rapid angiogenic response, structural and vascular alterations following vitrectomy and ILM peeling must be considered. Studies have shown that MH surgery induces changes in choriocapillaris, indicating a sudden increment in reperfusion [14, 15]. Additionally, following vitrectomy and ILM peeling, the choroid can exhibit a significant displacement towards the optic disc in a centripetal manner [16]. These choroidal movements and sudden reperfusion changes could serve as triggers for CNV in eyes with pre-existing AMD undergoing uncomplicated non-flap MH surgery. Furthermore, because the retinal vasculature is predominantly situated in the inner retina, any macular displacement reflects an inner retinal shift. There are multiple biomechanical factors that can lead to this displacement of the inner retina after vitrectomy accompanied by ILM peeling [17–19]. This sudden mechanical shift exerts significant stress on the retinal cytoarchitecture, particularly on the Müller cells, which have the ILM as their basal membrane. Müller cells play critical roles in regulating the blood-retina barrier and trigger vascular growth by producing vascular endothelial growth factor (VEGF) under hypoxic and mechanical stress conditions [20].

We hypothesize that the biomechanical displacement of the inner retina following even a standard, non-flap ILM peeling may generate sufficient mechanical stress on Müller cells to contribute to an acute release of angiogenic factors. We propose that this combination of altered choroidal microcirculation, macular displacement, and Müller cell-mediated VEGF release following ILM removal may have contributed to the rapid CNV development observed in this case, although this mechanism remains speculative and cannot be established as causal from a single observation. Alternative explanations cannot be fully excluded, including spontaneous conversion from non-neovascular to neovascular AMD independent of surgery, pre-existing subclinical CNV not detected by structural OCT alone, postoperative inflammatory changes, and surgery- or light-related stress.

The patient did not return for further examination after the 2-week follow-up visit following her first intravitreal aflibercept injection, and therefore did not receive the remaining planned injections; consequently, longer-term imaging to determine whether secondary atrophic changes developed in this eye was not available, which we note as a limitation of this report. Additionally, both the preoperative exclusion of CNV and the subsequent diagnosis of CNV relied on clinical examination and structural OCT findings alone, without confirmatory fluorescein angiography, indocyanine green angiography or OCT angiography, which were not available at our institution; therefore, subclinical occult CNV at baseline cannot be entirely excluded. This represents another limitation of this report.

This case supports the need for careful early postoperative monitoring for CNV in AMD eyes undergoing MH surgery. Although the localization of the lesion at the border of the ILM-peeled area suggests that surgical biomechanical factors may have contributed, spontaneous AMD-related neovascular conversion or pre-existing subclinical CNV cannot be fully excluded. Physicians should remain highly vigilant for this vision-threatening complication during the early postoperative follow-up period after MH surgery.

Author contributions

BK and BYT analyzed the patient data. BK, BYT, EK and SÖ interpreted findings. BK and BYT were major contributor in writing the manuscript. All authors read and approved the final manuscript.

Funding

Not applicable.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

The authors certify that they have obtained all appropriate patient consent forms. On the form, the patient has given her consent for her images and other clinical information to be reported in the journal. The patient understands that their names and initials will not be published and that all due efforts will be made to conceal her identity, but anonymity cannot be guaranteed.

Consent for publication

Written informed consent was obtained from the patient for the publication of this case report and any accompanying images.

Competing interests

The authors declare no competing interests.