A macular hole (MH) is a full-thickness retinal defect in the macula. The hole is not a loss of retinal tissue but results from traction by the posterior vitreous cortex, creating a small tear that enlarges into a hole. Spontaneous closure is rare, and over time the hole enlarges with degeneration of retinal pigment epithelial cells.
In 1988, Gass first described the four-stage progression of idiopathic macular holes. In 1991, Kelly and Wendel reported the effectiveness of vitrectomy, and in 1995, Brooks reported improved closure rates with concurrent internal limiting membrane peeling. Currently, vitrectomy with ILM peeling and gas tamponade is the standard surgical procedure.
Macular holes are classified into the following four types based on etiology.
| Type | Predilection layer | Main cause |
|---|---|---|
| Idiopathic | Women in their 60s to 70s | Age-related vitreomacular traction |
| Myopic | Women with high myopia | Posterior staphyloma, macular retinoschisis |
| Traumatic | Young men (20s to 30s) | Blunt ocular trauma |
| Secondary | Various | Retinal detachment, inflammation, drugs (tamoxifen, etc.)4) |
Lamellar macular hole (LMH) is a condition that does not involve a full-thickness defect but rather a partial defect in the inner retinal layers, while the outer retinal layers are preserved. The prevalence is 1.1–3.6%, and it commonly occurs in individuals aged 50–70 years 12). It is important as a related lesion to be differentiated.
A pseudomacular hole is a condition in which the central fovea appears depressed relative to the surrounding epiretinal membrane, and it differs from idiopathic macular hole in that it is not a full-thickness hole.
The incidence of idiopathic macular hole is reported to be 3.14–7.8 per 100,000 population per year 2). In a US population-based study, the incidence was 7.8 per 100,000 per year (8.7 eyes), with a female-to-male ratio of 3.3:1 3). The peak age of onset is in the 60s to 70s, especially the 60s. 72% of patients are female, more than 50% develop the condition between ages 65 and 74, and only 3% are under 55 3).
It is usually unilateral, but the frequency of involvement in the fellow eye is 10–20%. The 5-year risk of fellow eye involvement is 10–15%, reaching up to 28% in the absence of complete posterior vitreous detachment (PVD) 2). Asians have been reported to have a 177% increased risk of iFTMH compared to Caucasians 2). Traumatic cases are more common in young males, in contrast to idiopathic cases.
Macular hole retinal detachment is more common in women with high myopia, accounting for about 5% of retinal detachment cases in Japan, which is higher than the 0.5–2.0% in Western countries.
The risk of fellow eye involvement after unilateral onset is moderate. In patients without posterior vitreous detachment in the fellow eye, the risk reaches up to 28% over 5 years 2). If the fellow eye has complete posterior vitreous detachment, the risk is low. The timing of onset is not necessarily simultaneous, but about 10–20% of cases develop in both eyes.
The main subjective symptoms of idiopathic macular hole are as follows:
Symptoms progress slowly over weeks to months, but are often noticed relatively suddenly 3).
Untreated full-thickness macular holes often have a poor prognosis. About 40–50% of Stage 1 holes progress to full-thickness holes, and about 50% resolve spontaneously with posterior vitreous detachment 3). About 75% of Stage 2 holes progress to Stage 3 or 4 3). In untreated full-thickness macular holes, only 5% maintain visual acuity of 20/50 or better, 55% have 20/100 or worse, and 40% have 20/200 or worse 3). The spontaneous closure rate with observation alone is 2.5–27% (depending on size and duration), and no spontaneous closure occurs in large holes (>400 μm) 2).
The staging system described by Gass in 1995 is still widely used as the standard. Advances in OCT have provided a more detailed understanding of the pathology at each stage.
Stage 1 (Impending/ Occult)
Stage 1A (Impending hole): The fovea is deformed due to posterior vitreous detachment. There is loss of the foveal depression and a yellow spot (100–200 μm) 3). A foveal cystoid space or foveal retinal detachment forms.
Stage 1B (Occult hole): A yellow ring (200–350 μm) is present 3). There is separation of the photoreceptor layer but not yet a full-thickness hole (occult hole).
Natural course: About 50% of Stage 1 holes improve spontaneously.
Stages 2–4 (Full-Thickness Hole)
Stage 2: Full-thickness hole <400 μm. It may be eccentric with a flap 3). The posterior vitreous cortex is connected to the fovea by a flap.
Stage 3: Full-thickness hole ≥400 μm. It is associated with a free pseudo-operculum. Posterior vitreous detachment is incomplete.
Stage 4: Posterior vitreous detachment is complete. A glial ring is observed on ophthalmoscopy. A Weiss ring is present.
In 2013, the International Vitreomacular Traction Study Group (IVTS) established a classification based on OCT findings. Full-thickness macular holes are classified based on the minimum diameter 2).
| Category | Minimum Diameter |
|---|---|
| Small (S) | < 250 μm |
| Medium (M) | 250–400 μm |
| Large (L) | > 400 μm |
Generally, Stage 2 or later and medium or larger holes are considered active surgical indications. The CLOSE Study Group proposed further subclassification of large holes into L (>400–≤550 μm), XL (>550–≤800 μm), XXL (>800–≤1000 μm), and Giant (>1000 μm). Reports indicate that closure rates drop below 90% for holes exceeding 500 μm 2).
Traumatic macular holes are classified into the following types based on OCT findings 14).
| Type | OCT Findings | Characteristics |
|---|---|---|
| Type I | Macular edema/cystic changes | No hole formation |
| Type II | Lamellar hole | Partial-thickness defect |
| Type III | Full-thickness hole (small to medium) | IVTS classification S/M |
| Type IV | Full-thickness hole (large) | IVTS classification L |
| Type V | Hole with retinal detachment | Indication for emergency surgery |
Lamellar holes are broadly classified into two types according to Govetto et al.: “degenerative type” and “tractional type” 13). Hubschman et al. further redefined the degenerative type as “LMH” and the tractional type as “ERM foveoschisis (ERMF)” 12).
The degenerative type is characterized by nonlinear cavitation of the inner retina, loss of foveal tissue, and lamellar hole-associated epiretinal proliferation (LHEP). The tractional type is characterized by the presence of a contractile epiretinal membrane and retinal separation at the level of the Henle fiber layer 13). The degenerative type is more likely to involve disruption of the ellipsoid zone (EZ) and tends to have a poorer visual prognosis.
Stage 1 is an impending hole that has not yet progressed to a full-thickness hole, and about 50% improve spontaneously 3). Observation is usually recommended. However, because there is a risk of progression, regular follow-up with OCT is important.
The main cause of idiopathic macular hole is vitreomacular traction during the early stages of age-related posterior vitreous detachment (PVD) 2).
In the human eye, there is a physiological liquefied pocket (pre-posterior vitreous cortex pocket: Kishi pocket) anterior to the macula. The cortex forming the posterior wall of the pocket pulls on the macula with aging and contributes to the development of macular holes.
The process of development is as follows:
The extent of outer retinal defect correlates with the range and intensity of vitreomacular traction; broader adhesion leads to more extensive photoreceptor loss 2). When glial repair attempts fail, glial cells migrate and contract on the ILM at the hole margin, enlarging the hole through tangential traction 2).
When blunt trauma is applied to the eye, anteroposterior compression and equatorial expansion occur simultaneously. This deformation concentrates tangential traction on the macula, tearing the foveal tissue and forming a hole. In young individuals, the vitreous is firmly attached to the retina (no PVD), so external force is easily transmitted directly to the macula via the vitreous 14). Approximately 85% of TMH cases show no posterior vitreous detachment.
High-power lasers, such as accidental exposure to laser pointers, can also be a cause.
OCT is the gold standard for diagnosis and management of macular holes 3). It allows detailed evaluation of the retinal structure at the macular hole. The following information can be obtained:
To exclude a pseudohole, OCT should confirm that there is no full-thickness retinal defect in the depressed area. It is important to perform a volume scan and examine continuously.
As basic examinations at the first visit, in addition to OCT, best-corrected visual acuity (BCVA), intraocular pressure (IOP), and peripheral retinal examination (screening for tears) should be performed 3).
Detailed fundus examination under mydriasis is fundamental. In full-thickness macular holes, a grayish macular rim (reflecting subretinal fluid accumulation), yellow deposits at the hole base, and RPE changes are observed.
| Test | Main Role | Features |
|---|---|---|
| OCT | Definitive diagnosis, staging, differential diagnosis | Gold standard |
| Slit-lamp microscopy | Clinical diagnosis | Direct observation of the hole under mydriasis |
| Watzke-Allen test | Confirmation of full-thickness hole | Useful for differentiation from pseudohole |
| Amsler chart | Detection of metamorphopsia | Qualitative assessment of subjective symptoms |
The following diseases require differentiation. All are distinguished from macular hole by the absence of full-thickness retinal defect on OCT.
OCT allows easy differentiation. In pseudohole, there is no full-thickness retinal defect, and foveal depression due to epiretinal membrane is observed. The Watzke-Allen test is negative in pseudohole and positive in full-thickness hole. It is important to confirm the absence of full-thickness retinal defect in the depressed area on OCT and to verify continuously with volume scan.
Treatment for macular hole is vitreous surgery. There is no medical therapy. In patients aged 50 years or older, cataract surgery is often performed simultaneously.
A meta-analysis of 5,480 cases (Rahimy 2016) showed a closure rate difference of 98.82% in the ILM peeling group vs. 92.88% in the non-peeling group (P<0.0001), and the reopening rate was significantly reduced (25% → 0%) 3). A meta-analysis of 4 RCTs (317 cases, Stage 2–4) also showed a reduced need for additional surgery in the ILM peeling group, and a larger area of ILM peeling was associated with less metamorphopsia 3). ILM peeling is cost-effective, and a cost-effectiveness analysis based on RCTs showed that ILM peeling was superior to non-peeling over 6 months 3).
The closure rate with the standard procedure is 91–98% 3). In the RCOphth meta-analysis, the surgery group had a visual acuity advantage of 0.16 logMAR (95% CI −0.23 to −0.09) compared to the observation group, and the odds ratio for hole closure was 31.4 (95% CI 14.9–66.3), clearly demonstrating the effectiveness of surgery 2). In UK real-world data (1,483 eyes), the closure rate was 95.7%, VA improved from 0.78 to 0.42 logMAR (approximately 4 lines improvement), and 64.2% improved by 0.3 logMAR or more 2).
The median postoperative visual acuity is approximately 20/40 (0.5), with visual improvement sustained for up to 3 years and maintained for 5–10 years 3). Since BCVA worsens by 0.008 logMAR for each 1-month increase in symptom duration (≈1 ETDRS letter loss every 2 months), early surgical intervention is important 3) (12 RCTs, IPD meta-analysis of 940 eyes). When disease duration exceeds 2–3 years, the closure rate drops to 63% and visual prognosis becomes poor 3). Shorter disease duration and smaller hole size are associated with higher closure rates and better visual prognosis.
A Cochrane review of 8 RCTs including 709 eyes found no significant difference in hole closure rates between prone and non-prone groups. Even for large holes (≥400 μm), the closure rate was 94% in the prone group vs. 84% in the non-prone group; for small holes, it was 100% vs. 96%. A meta-analysis (251 cases, 5 RCTs) suggests that prone positioning is unnecessary for holes <400 μm but beneficial for holes ≥400 μm 3). A UK survey (2018) reported that 82% of surgeons still recommend positioning restrictions, with durations varying: 1 day (19%), 2–4 days (30%), 5–6 days (23%), and ≥1 week (9%) 2).
Large macular holes (>400 μm), long-standing cases, and secondary holes associated with high myopia, trauma, or inflammation are considered refractory.
Inverted ILM flap technique: This technique is applied to large holes ≥1/3 disc diameter or primary non-closure cases. The ILM peeled to the edge of the hole is inverted and placed over or into the hole. In reoperations, a free ILM flap obtained from outside the macula can be placed into the hole to promote closure. A systematic review of 4 small RCTs suggests that this technique may provide better BCVA improvement than standard ILM peeling (low evidence) and superior closure rates (moderate evidence) 3).
ILM flap without gas tamponade (Szeto 2025): This gasless technique uses a temporal ILM flap (in 92.9% of cases) and perfluorocarbon liquid (PFCL) to unfold and stabilize the flap 16).
Szeto et al. (2025) reported a gasless technique for large holes (including >30% with high myopia, MLD ≥500 μm) 16). The closure rate was comparable between the gasless group (94.1%) and conventional group (95.2%) (P=0.812). Early postoperative visual acuity was significantly better in the gasless group (at 1 week and 1 month). The rate of foveal gliosis was lower in the gasless group (4.9% vs. 20.0%, P=0.043). OCT evaluation was possible from the day after surgery, and there was no patient burden from positioning restrictions 16).
Surgical techniques using other biomaterials (Romano 2025)1):
Song et al. (2024) reported surgical outcomes for 8 eyes with macular hole associated with retinitis pigmentosa6). Hole closure was achieved in all eyes that underwent ILM peeling; ILM free flap transplantation was used for large holes. Postoperative visual acuity improved in 1 eye and remained stable in 7 eyes.
In the natural course, visual acuity decreases to 0.1 or less in many cases. Surgery achieves hole closure in over 90% of cases and improves visual acuity. However, metamorphopsia often persists. If the hole does not close after initial surgery, reoperation with extended ILM peeling or autologous ILM transplantation is considered.
Because spontaneous closure is possible (10–67%), initial management after injury is a period of observation. Spontaneous closure is thought to occur when glial cells bridge the defect. Younger patients have a relatively higher spontaneous closure rate due to greater glial cell proliferative capacity.
If spontaneous closure does not occur, vitrectomy (PPV) with internal limiting membrane (ILM) peeling is the gold standard14). Closure rates of 82–96% have been reported14). Although the efficacy of ILM peeling is not fully established, the same surgical technique as for idiopathic holes is used. Postoperatively, gas tamponade (SF₆ or C₃F₈) is performed, and prone positioning is required. For large or refractory holes, ILM flap technique or amniotic membrane transplantation are options14).
In a meta-analysis by Zhou et al. (2021), comparing closure rates between PPV and spontaneous closure, the surgical group showed significantly higher closure rates 15). Prolonged waiting may reduce closure rates and, even if closure occurs, visual recovery may be insufficient.
| Situation | Management |
|---|---|
| Immediately after injury / Small | Observation first |
| No spontaneous closure | PPV + ILM peeling |
| Large / Difficult closure | ILM flap / Amniotic membrane transplantation |
Vitrectomy for traumatic macular holes due to blunt trauma has been reported to achieve closure rates of over 90%. Visual improvement depends on hole closure and the presence and location of other injuries besides the macular hole.
In high myopia, a macular hole may progress to retinal detachment. The standard procedure is vitrectomy with internal limiting membrane peeling and gas (or silicone oil) tamponade. Even if posterior vitreous detachment appears to have occurred, vitreous cortex remains attached to the retina over a wide area. Residual vitreous cortex is removed under triamcinolone visualization, and the internal limiting membrane is peeled for 2–3 disc diameters using vital dye. The reattachment rate is about 70%, lower than that for typical rhegmatogenous retinal detachment (over 90%).
Observation: Many LMHs remain stable over time, and asymptomatic idiopathic cases do not require treatment. Regular OCT follow-up is performed to monitor structural changes.
LHEP-sparing surgery: Surgery is considered when there is symptomatic visual decline, progressive metamorphopsia, or worsening foveal profile. Recently, techniques that preserve or embed the LHEP without peeling have gained attention.
In a meta-analysis by Yu et al. (2025) of 8 studies, the postoperative BCVA improvement in the LHEP-sparing surgery group was significant at −0.25 logMAR (95% CI −0.30 to −0.21, P<0.00001)11). Compared with conventional peeling, the difference in BCVA improvement was −0.19 logMAR (P<0.0001), favoring the sparing group. The proportion of patients with postoperative EZ restoration was also significantly higher in the sparing group (OR 2.55; 95% CI 1.48 to 4.38)11). No postoperative FTMH formation was reported in the LHEP-sparing group11).
If visual acuity is good and there are no subjective symptoms, surgery is generally not performed and observation is often chosen. If subjective symptoms such as decreased vision or severe metamorphopsia are present, vitrectomy is considered, similar to the treatment of epiretinal membrane. In vitrectomy, the epiretinal membrane and internal limiting membrane are peeled.
Complications common to vitrectomy may occur.
Preoperative visual acuity, hole size, and duration of symptoms affect visual prognosis. The median postoperative visual acuity is approximately 20/40 (0.5), with improvement sustained up to 3 years and maintained for 5–10 years 3). Metamorphopsia often persists. Longer duration of symptoms leads to worse prognosis; beyond 2–3 years, the closure rate drops to 63% 3). In traumatic cases, associated injuries besides the macular hole also affect visual prognosis.
The development of idiopathic macular holes is closely related to the early stages of posterior vitreous detachment (PVD) 2). In normal aging, PVD begins around the fovea and progresses in four stages. In Stage 1, the vitreous detaches perifoveally but remains attached at the fovea, and finally in Stage 4, detachment from the optic disc completes the process 2).
Macular holes develop as a pathological condition of PVD Stage 1. Abnormal vitreous attachment persists at the fovea, and dynamic traction from eye movements causes separation between the Müller cell cone and photoreceptors 2). On OCT, this process is observed as intraretinal cysts.
If anteroposterior traction persists, the ILM and ELM break, leading to retinal dehiscence and a full-thickness hole 2).
The exact mechanism by which ILM peeling and gas tamponade during vitrectomy close the hole is not fully understood. Early after ILM peeling, the macula shifts slightly nasally and toward the fovea along nerve fiber bundles. This retinal movement is thought to contribute to approximation of the hole edges.
Recent OCT studies consistently show the involvement of traction in LMH formation. In a study of 50 highly myopic eyes by Hsia et al. (2023), traction was identified in all LMH formation processes 9). Four traction-related formation processes were identified.
Spontaneous closure of some LMHs has been reported. Catania et al. (2024) reported spontaneous closure in 11 of 187 degenerative cases (5.9%) and 10 of 200 mixed cases (5.0%) 10). The median time to closure was 4 years 10).
In the closure group, the frequency of hyperreflective inner border (HIB) at the cavity margin and linear hyperreflective outer plexiform layer (LHOP) was significantly higher than in the stable group 10). These findings may reflect coordinated activation of microglia and Müller cells 10).
Ocriplasmin is a 27 kDa serine protease that enzymatically cleaves the adhesion between the vitreous and retina 1). The dose is a single intravitreal injection of 0.125 mcg/0.1 mL 2). Indications are limited to cases with “persistent VMT + iFTMH <400 μm + no ERM” 2).
In the MIVI-TRUST trials, vitreomacular adhesion release was achieved in 26.5% of the ocriplasmin group (vs. 10.1% placebo) at 28 days post-injection, and the hole closure rate was 40.6% in the ocriplasmin group (vs. 10.6% placebo). In the RCOphth IPD meta-analysis (1,067 cases), iFTMH closure rate was 36.8% (control 9.3%, OR 6.1), VA improvement was +5.97 letters (control +3.33 letters, difference +2.32 letters) 2).
Reported side effects include ERG changes (amplitude reduction 40%, 81.3% recovered over time), dyschromatopsia (4.5% vs. control 0.6%), and lens subluxation 2). In real-world clinical practice, efficacy tends to be lower than in RCTs, and adoption rates are declining 2).
In the DRCR Retina Network Protocol AH trial, 0.3 mL of 100% C₃F₈ was injected intravitreally for small MH (median 79 μm). The closure rate was 29% (95% CI 16–45%), but retinal tears/detachment occurred in 12% (7/59 eyes, 95% CI 6–23%), leading to trial termination 2). It is contraindicated in patients with peripheral retinal fragility 2).
New biomaterials are being explored for refractory or recurrent macular holes 1).
LHEP-sparing surgery has been rapidly developing in recent years. A technique of embedding a double flap of peeled ILM and LHEP into the macular hole has also been reported 11). Future prospective randomized controlled trials are needed 11).
HIB and LHOP detected by image processing may serve as predictive markers for spontaneous closure of LMH 10). Future prospective studies are expected to validate these markers.
Retinal regeneration therapy using stem cells is at an experimental stage 1). It is expected to be applied for photoreceptor regeneration and functional recovery after macular hole closure.
