Pathophysiology of Secondary Macular Hole in Rhegmatogenous Retinal Detachment

Abstract

Purpose: To describe the pathophysiology of secondary macular hole (MH) in rhegmatogenous retinal detachment (RRD).

Methods: A retrospective cohort of 360 consecutive primary fovea-off RRDs presenting to St. Michael's Hospital, Toronto, from January 2012 to September 2022 were included. Preoperative OCT was assessed for bacillary layer detachment (BALAD) abnormalities. Histological sections of normal eyes were assessed to inform OCT interpretations. Primary outcome measure was the progression of BALAD to full-thickness MH (FTMH).

Results: Of the 360 patients, 22.5% (n = 81) had BALAD abnormalities at presentation. Eight percent (29/360) had associated MH, of which 79.3% (23/29) were a BALAD-lamellar hole and 20.7% (6/29) were FTMH. After reattachment, 62% of MHs persisted (18/29), of which 83% (15/18) had BALAD-lamellar holes that subsequently progressed to FTMH in a mean of 8.1 ± 3.2 days. BALAD-lamellar holes had significantly worse postoperative visual acuity (P < 0.001) when compared with other BALAD abnormalities (58/81) or with the rest of the cohort (279/360). OCT spectrum from BALAD to FTMH includes (1) cleavage planes extending from Henle fiber layer into the BALAD; (2) central outer nuclear layer thinning; (3) Müller cell cone loss with tissue remnants at the foveal walls; (4) retinal tissue operculum close to BALAD-MH; and (5) progressive thinning or degradation of the posterior band of BALAD–lamellar hole leading to FTMH. Histological specimens identified foveal regions of low mechanical stability.

Conclusions: BALAD plays a crucial role in the pathophysiology of MH in RRDs, which forms owing to sequential changes in four critical areas: RPE–photoreceptor interface, myoid zone, Henle fiber layer, and Müller cell cone with surrounding tissue. Timely management of fovea-off RRD with BALAD may be prudent to avoid the progression to BALAD-lamellar hole, subsequent FTMH, and worse functional outcomes.

Macular holes (MHs) associated with rhegmatogenous retinal detachment (RRD) occur in two typical scenarios. First, in cases of pathological myopia, the MH is the causative break leading to the detachment.^1,2 The second scenario occurs when a patient has an RRD attributable to a peripheral break with a secondary MH, which can either occur before or as a result of the detachment process itself. Several theories have been proposed to explain the formation of secondary MH after RRD^3–7; however, the pathophysiology remains unknown. 

Although the prevalence of secondary MHs in RRDs varies from approximately 1% to 4%,^7–9 prior studies did not use preoperative OCT. Novel retinal abnormalities are now being recognized in RRD with increasing use of OCT.¹⁰ Bacillary layer detachment (BALAD) and associated abnormalities in a prospective cohort of fovea involving RRDs was recently described. BALAD is an intra-photoreceptor splitting at the inner segment myoid, visualized on OCT as an intraretinal hyporeflective space delineated anteriorly by a hyper-reflective granular band hypothesized to be the external limiting membrane (ELM) and myoid remnants, and posteriorly by another hyper-reflective band of inner segment (myoid and ellipsoid), outer segment, and interphotoreceptor matrix remnants. In some cases, the anterior wall of the BALAD cavity can be lost, resulting in a BALAD–lamellar hole.¹¹ 

This study aims to explore the relationship between BALAD-associated abnormalities and secondary MHs and elaborate on its pathophysiology using high-resolution preoperative OCT and histological specimens from normal eyes. 

Methods

This single-center retrospective cohort study included consecutive patients with primary fovea-off RRD who presented to the tertiary vitreoretinal surgery clinic at St. Michael's Hospital, Toronto, between January 2012 and September 2022. Charts were assessed for gradable baseline SD or swept-source OCT scans. This study was approved by the Research Ethics Board and conducted in accordance with the Declaration of Helsinki. 

Patients with prior history of vitrectomy or ocular disease that could influence foveal morphology were excluded. Data collected included demographic and clinical characteristics, treatment modality and surgical outcome. Best-corrected visual acuity (BCVA) was obtained at baseline, 3, 6, and 12 months after RRD repair. BCVA was evaluated using the Snellen chart and then converted to logMAR. Patients were treated (by RHM) with either pneumatic retinopexy, pars plana vitrectomy (PPV), scleral buckle, or combined PPV and scleral buckle. 

Patients underwent imaging with the PLEX Elite 9000 SD-OCT and/or Cirrus 5000 SD-OCT (Carl Zeiss, Dublin, CA, USA), using high-definition horizontal foveal scans (×100), and 6 × 6-mm macular cube with 512 × 512 or 512 × 128 scans. The entire volume scan was assessed preoperatively for foveal BALAD and associated structural abnormalities and postoperatively for discontinuity of the outer retinal bands. Descriptive statistics used for normally distributed data were mean with standard deviation. Comparisons of BCVA logMAR between subgroups were conducted using a two-sided t-test, with an alpha of 0.05. Analyses were performed using IBM SPSS version 26 (IBM Corp., Armonk, NY, USA). A multinomial regression analysis was performed to assess predictive factors for the development of BALAD and BALAD–lamellar holes, including covariates that might affect the progression of BALAD-related abnormalities, such as age, gender, duration of fovea-off (regarded as duration of central vision loss), height of foveal detachment, and presence of baseline cystoid edema on OCT. Covariates strongly correlated with other variables in the model, such as the baseline visual acuity and extent of detachment in clock hours, were excluded to avoid multicollinearity. 

An annotated collection of high-resolution histological sections of the normal macula of aged donors were assessed to assist with in vivo OCT interpretations. These histological sections belong to Project MACULA (https://projectmacula.org/normal/), an online image bank made available by author CAC and the Department of Computer Science of the University of Alabama at Birmingham. Details on tissue acquisition, preparation, and microscopy are available.^12,13 Thirty-four foveal histological sections were carefully reviewed, assessing possible areas of low mechanical stability within the foveola, evidenced by artifactual separations along tissue planes. 

Results

Three hundred sixty patients with primary fovea-off RRD and gradable baseline OCT scans were assessed for structural changes of the central bouquet. BALAD-related abnormalities were observed in 22.5% of patients (81/360) at presentation. The mean age of patients with BALAD-related abnormalities was 63.5 ± 8.89 years, 55% (45/81) were male, and the mean baseline visual acuity was 1.64 ± 0.41. Associated MHs were present in 8% of the cohort (29/360), of which 79.3% (23/29) were BALAD-lamellar holes and 20.7% (6/29) were FTMHs. Demographic details and clinical characteristics of patients presenting with associated MHs are summarized in Table. 

Table.

 

Baseline Demographic and Clinical Characteristics of Patients (n = 29 Patients) With a Secondary MH at Presentation

 

The mean time to surgery from the baseline OCT to the primary RRD repair procedure was 0.96 ± 2.34 days for patients with any spectrum of BALAD-related abnormalities (81/360) and 0.94 ± 3.03 days for patients presenting with a BALAD–lamellar hole. When assessing predictive factors for developing BALAD-related abnormalities and progressing to BALAD–lamellar holes, the height of foveal detachment and baseline cystoid edema were the only significant predictors (P < 0.001). Age, gender, and duration of fovea-off were not significant predictors (P = 0.87, P = 0.44, and P = 0.94, respectively). 

Regarding the surgical approach for patients with RRD and secondary MH, 45% of eyes (13/29) were repaired with pneumatic retinopexy, and 55% (16/29) were repaired with PPV, including one combined PPV/scleral buckle. After reattachment, 62% of MHs (18/29) persisted, 31% (9/29) closed with the RRD repair and no internal limiting membrane (ILM) peeling, and 7% (2/29) had additional procedures to close the MH during the initial repair. The hole closure rate of patients who had only one PPV without ILM peeling was 57% (8/14). Of persistent MHs post-RRD repair, 83% (15/18) had a BALAD–lamellar hole with subsequent progression to FTMH in 8.1 ± 3.2 days, and 17% (3/18) already had a FTMH preoperatively. Most patients with post-RRD repair MHs (14/18) required ILM peeling to close the MH (one patient is currently on the list for surgery), with four patients having late spontaneous hole closure within 3 months. Of patients who underwent PPV to close the MH, 20% (3/15) did not achieve closure at the 6-month follow-up. 

In patients presenting with BALAD and no inner lamellar hole (58/81), the mean logMAR BCVA was 1.57 ± 0.44, 0.71 ± 0.46, 0.58 ± 0.42, and 0.39 ± 0.32 at baseline and postoperative months 3, 6, and 12, respectively. In contrast, in BALAD-lamellar holes (23/81), the mean logMAR BCVA was 1.80 ± 0.30, 0.89 ± 0.50, 0.82 ± 0.49, and 0.77 ± 0.56, at baseline and postoperative months 3, 6, and 12, respectively, despite surgical closure of subsequent FTMH. There was a statistically significant reduced BCVA in patients presenting with BALAD–lamellar holes preoperatively versus those with BALAD alone at all time points (all P < 0.001). Similarly, postoperative 12-month BCVA was significantly worse in patients with BALAD–lamellar hole compared with patients with fovea-off RRD and no BALAD abnormalities (279/360), 0.77 ± 0.56 versus 0.42 ± 0.47, respectively (P < 0.001). When comparing 12-month BCVA in patients with any of the findings within the BALAD spectrum to patients with fovea-off RRD and no BALAD abnormalities, the former had significantly worse VA (0.51 ± 0.44 vs. 0.42 ± 0.47; P < 0.001). 

The minimum linear width of secondary MHs at baseline and post-RRD repair was 376 ± 158.2 µm and 393.7 ± 141.4 µm, respectively (P = 0.71). At 3 months postoperatively, 10 patients still had a FTMH. At 6 months postoperatively, among patients with MH closure and gradable OCT (17/25), the assessment of the integrity of outer retinal bands revealed that 100% (17/17) had discontinuity or loss of the ellipsoid zone and 47% (8/17) had discontinuity or loss of the ELM. Four patients still had an open hole at 6 months postoperatively and were not included in this analysis. The 1-year OCT data were not reported because only 17 patients had a gradable OCT, which in the vast majority of cases (16/17) remained unchanged from the 6-month follow-up. 

An in-depth analysis of OCT scans was conducted to assess specific morphologic features that appeared to represent a progression from BALAD to BALAD–lamellar hole. These observations and interpretations are presented in Figures 1 through 6. Fifty-eight cases of foveal BALAD had an intact anterior wall. Of those, 20% (12/58) presented with oblique hyporeflective cleavage planes extending from the Henle fiber layer (HFL)–outer plexiform layer, through the ELM, into the BALAD cavity (Figs. 1A–D). In a large proportion of cases, these cleavage planes were found to be at the junction of the Müller cell cone (MCC) and the foveal walls (Figs. 1A and 1C). Among these same 58 cases, we also observed a significant thinning of the central ONL above the BALAD in 19% of patients (11/58) (Figs. 2A–D). In some of these cases, there was no ONL visible, with only a thin hyper-reflective tissue connecting the foveal walls (Figs. 2A and B). 

Figure 1.

 

HFL cystic degeneration in RRD with associated foveal BALAD. (A–D) OCT foveal scans in cases of fovea-off RRD demonstrating foveal BALAD (arrows) with oblique hyporeflective cleavage planes (arrowheads) extending from the HFL–outer plexiform layer (OPL) through the ELM, into the intraretinal cavity formed by the BALAD. In some patients (1A and 1C), these cystic changes were observed close to the location of the junction between the MCC and HFL at the foveal walls (arrowheads).

Figure 2.

 

Central outer nuclear layer thinning in RRD with associated foveal BALAD. (A–D) OCT foveal scans in fovea-off RRDs demonstrating foveal BALAD with significant thinning of the central outer nuclear layer (ONL) immediately anterior to the BALAD cavity (star), where the central bouquet of cones is located. In some patients (1A and 1B), the ONL seemed to be completely absent, with only a very thin anterior hyper-reflective bridge of tissue connecting the foveal walls. In one patient (2D), a hyporeflective cyst (arrowhead) was also located at the junction of the MCC and HFL at the foveal walls.

In patients presenting with BALAD–lamellar hole (23/81), 26% (6/23) had tissue remnants of medium reflectivity at the inner edges of the foveal walls, contiguous with the adjacent ONL (Figs. 3A–D). In the remaining cases (17/23), the edges of the hole were at the foveal slopes, where all retinal layers were present, with a complete absence of the MCC and central ONL (Figs. 4A–C). In one case, an operculum of retinal tissue attached to thin vitreous strands was observed near the BALAD–lamellar hole, which was thought to represent remnants of the MCC (Figs. 5A–C). The hyper-reflective posterior band of the BALAD was found to undergo progressive thinning or degradation, consistent with the concurrent morphologic stage of RRD observed in the parafoveal region¹⁴ (Figs. 4A–C). Finally, after reattachment, all holes had a flat appearance for a mean of 9.3 ± 4.2 days, after which cystic changes would occur, and the MH edges would evert (Figs. 6A–C). In all cases of persistent BALAD–lamellar hole post-RRD repair, the posterior aspect of the BALAD degenerated, progressing to FTMH (Figs. 6B and C). 

Figure 3.

 

Morphology of the inner edges of the BALAD lamellar hole in RRD. (A–D) OCT foveal scans in cases of fovea-off RRD demonstrating BALAD lamellar holes with a residual posterior hyper-reflective band of presumed photoreceptor inner and outer remnants and interphotoreceptor matrix. These cases had residual tissue remnants of medium reflectivity at the inner edges of the foveal walls (arrowheads), contiguous with the adjacent outer nuclear layer, likely representing remnants of the MCC.

Figure 4.

 

Degradation of the posterior band of BALAD lamellar hole in RRD. (A–C) OCT foveal scans in cases of fovea-off RRD demonstrating BALAD-lamellar holes with a residual posterior hyper-reflective band of photoreceptor inner segment and outer segment remnants. The edges of these holes had all retinal layers present, with a complete absence of the MCC and central outer nuclear layer. From A–C, we observe varying features in the parafoveal outer retina consistent with the morphologic stage of detachment: (A) stage 3b, high-frequency and high-amplitude outer retinal corrugations (ORCs); (B) stage 4, loss of definitions of the ORCs with multiple hyper-reflective dots; and (C) stage 5, loss of outer retinal tissue, observed as moth-eaten appearance of the photoreceptor inner segment and outer segment, with a residual posterior granular hyper-reflective band. The residual posterior band of the BALAD-lamellar hole becomes thinner and more translucent from 4A to 4C (arrows).

Figure 5.

 

Retinal operculum attached to vitreous strands close to BALAD lamellar hole in RRD. (A and B) OCT foveal scan in fovea-off RRD demonstrating a lamellar hole with a residual posterior hyper-reflective band, consisting of the remaining posterior border of the BALAD (arrow), in the same patient. Contrast and brightness were progressively increased from A to C to allow better visualization of the structures. In this scan, an operculum of retinal tissue (star) attached to thin vitreous strands (arrowheads) was observed near the BALAD-lamellar hole. (C) Manual reconstruction of the vitreous strands was performed to enhance the visualization.

Figure 6.

 

Degradation of the posterior band of BALAD lamellar hole after retinal reattachment. (A–C) Time-lapse OCT foveal scans of fovea-off RRD demonstrating the evolution of a BALAD lamellar hole with a residual posterior bridge of photoreceptor inner and outer segment remnants (A) (arrow). After retinal reattachment, a faint bridge of photoreceptor remnants (arrow) can still be seen on day 7 (B), which will subsequently degenerate on day 12 (C), progressing to a FTMH (star). From B-C, the edges of the hole evert with increased cystic changes.

From Project MACULA, the most common areas of artifactual tissue disruption observed in the histological sections and their cellular and anatomical relationships within the fovea were described in Figure 7A through D. We observed in some cases (1) a separation of the neurosensory retina from the RPE, (2) BALAD, (3) disruption within the HFL, and (4) cystic changes within the central foveola, with a separation of the MCC from the surrounding structures. 

Figure 7.

 

Artifactual histological tissue disruptions in probable areas of low mechanical stability. Histological sections (A–D) of normal eyes (aged 90, 83, 62, and 71 years, respectively) demonstrate possible areas of low mechanical stability contributing to BALAD in RRD and its progression to BALAD-lamellar hole and FTMH. All specimens exhibited artifactual separation of the neurosensory retina from the RPE (star in A). Scale bars in D and D1 apply to A to D and A1 to D1, respectively. Histological layers are labeled in D1. (A) Photoreceptor outer segments are shortened by detachment (black arrowheads). Outer segment tips and apical processes of RPE cells (black arrows) are visible on the outer aspect of the detachment. (B) Artifactual BALAD delineated anteriorly by the ELM and myoid remnants (left, green arrowheads) and posteriorly by a band of photoreceptor inner segment/outer segment remnants (right, green arrowheads). Cystic spaces within the HFL, surrounding the MCC, and within the central outer nuclear layer (ONL) are also seen (black arrowheads). (C) Multiple cystic spaces within the MCC and central ONL (black arrowheads) and at the HFL–outer plexiform (OPL) interface at the fovea walls (black arrowheads). (D) MCC separating from the adjacent HFL at the foveal walls and from the central outer processes and ONL (black arrowheads). Ch, Choroid; GCL, ganglion cell layer; ILM, inner limiting membrane; INL, inner nuclear layer; IPL, inner plexiform layer; inner segment/outer segment, inner segment and outer segment of photoreceptors; ONL, Outer nuclear layer; OPL, outer plexiform layer. Figure prepared by Jeffrey D. Messinger, DC.

Discussion

In this large retrospective study, 22.5% of patients (81/360) had evidence of BALAD-related abnormalities in the fovea, confirming that this is a frequent feature in fovea-involving RRDs.¹¹ Additionally, the prevalence of secondary MHs (8%) was greater than that reported in the literature (1%–4%),^7–9 which can be explained by the fact that none of the previous studies used preoperative OCT. Most interestingly, 79.3% of all RRD-associated MHs were BALAD–lamellar holes with an intact posterior band of photoreceptor and interphotoreceptor matrix remnants. 

Although BALAD–lamellar holes were previously imaged after RRD repair,^5,15 they have not been recognized as related to BALAD and only recently have we been able to suggest a relationship between these two entities.¹¹ Our OCT findings suggest that BALAD is associated with MH formation in RRD and that it likely plays a crucial role in its pathophysiology. We also used a cohort of histological specimens of normal eyes to assess the anatomy of the fovea. The imaging and histological data suggest that the progression from BALAD to BALAD–lamellar hole is initiated by mechanical and degenerative changes in the fovea occurring sequentially at regions of low mechanical stability. 

Understanding Regions of Low Mechanical Stability in the Fovea

RPE–Photoreceptor Interface

The adhesion of photoreceptors with the underlying RPE occurs secondary to several anatomical and metabolic factors, including RPE apical processes, interphotoreceptor matrix, intraocular and choroidal oncotic and hydrostatic pressures, oxygenation, and temperature.¹⁶ However, it is important to note that the subretinal space is a potential space, in which junctional proteins do exist, but their function is still not well-understood.¹⁷ The multifactorial nature of this adhesiveness is consistent with the observation of postmortem retinal detachment in many histological sections,¹⁸ suggesting that the RPE and neurosensory retina interface is indeed an area of low mechanical stability (Fig. 7A). 

Photoreceptor IS Myoid Zone

Another area of low stability is the myoid zone of the ISs. The postmortem occurrence of BALAD is a well-known artifactual phenomenon in histological sections (Fig. 7B), suggesting it to be an area that can be disturbed.¹⁹ In exudative processes, one hypothesis is that outwardly directed forces promoting the adhesion of the outer segment and RPE exceed the tensile force of the myoid zone, which will then break down and allow fluid to accumulate.¹⁹ However, the occurrence of BALAD in RRD seems to have a different mechanism. In this case, the continuous influx of hypo-osmolar liquified vitreous through an open break, overwhelms the physiologic adhesive forces between the neurosensory retina and the RPE. As the RRD progresses, the modulus of elasticity of the outer retina changes and corrugations are formed as a compensatory response to compressive forces generated from interphotoreceptor matrix hydration/lateral expansion.^20,21 However, the unique arrangement of the central Müller glia prevents the fovea from corrugating, which likely leads to tractional stretching forces within the central bouquet, resulting in a breakdown of the myoid with subsequent fluid accumulation.¹¹ 

MCC and Adjacent Henle Fiber Layer

The anatomy of the foveola comprises an inner layer of atypical Müller cell nuclei and thin lamelliform processes forming a cystic plait below the ILM and anterior to the bouquet of cones.^22,23 The outer processes of these Müller cells do not form a functional column with the surrounding cones and their cytoplasm display an increasingly watery appearance closer to the ELM. At the interface between the MCC and the fovea walls, these atypical outer cell processes are also surrounded by extracellular fluid-filled cysts.²³ Therefore, these distinct characteristics in addition to the absence of intermediate glial filaments, and of connections with adjacent typical Müller glia at the foveal walls, and the reduced number of zonulae adherens in the central ELM, are hypothesized to impact foveal stability, and its susceptibility to tissue deformation^23,24 Histological sections (Figs. 7A–D) from normal eyes also suggest areas of low mechanical stability within the HFL at the foveal walls (Fig. 7C) and between the MCC and the surrounding tissue (Fig. 7D). These same regions have been documented to be disrupted in other degenerative and tractional abnormalities of the fovea.^23,25–27 

Pathophysiology of MH in RRD

RPE–Photoreceptor Interface

Several theories have been proposed to explain secondary MH formation in RRD. This study provides evidence in support of BALAD representing a critical stage in its pathophysiology, because almost 80% of all RRD-associated MHs are BALAD–lamellar holes. OCT and histological findings suggest that a sequential disruption along natural tissue planes is key in this pathophysiological process. The first area of low mechanical stability to be exposed is the RPE–photoreceptor interface, owing to the continuous influx of liquified vitreous through an open break, leading to an RRD.¹⁶ 

Photoreceptor IS Myoid Zone

In the setting of an acute and extensive detachment, progressive hydration and lateral expansion of the outer retina will lead to compensatory corrugations.²¹ However, the rigid scaffold provided by the MCC likely prevents the fovea from corrugating, which in turn may lead to traction within the central bouquet, leading to BALAD.¹¹ Of note, outer retinal corrugations in RRD should not be confused with those previously described in AMD.²⁸ 

MCC and Adjacent HFL

Once BALAD has occurred, cystic changes can extend from the HFL into the BALAD cavity, disrupting its anterior wall (Figs. 1A–D). Similarly, Shukla et al.³ suggested that the dehiscence of retinal tissue with chronic CME in long-standing RRDs could be related to the pathogenesis of secondary MHs. Additionally, degenerative enlargement of cystic cavities within the HFL–outer plexiform layer also occurs in degenerative lamellar holes.²⁹ Our analysis of histological sections suggests that there are postmortem tissue disruptions between the MCC and foveal walls, corroborating the likelihood of low mechanical stability at this interface (Figs. 7C and D). In our clinical cases, once the ELM is compromised, we hypothesize that the MCC becomes unstable, which may lead to the dehiscence and loss of the MCC if any stresses are exerted on the foveola. 

Contributors of Further Damage

Ah Kiné et al.⁷ hypothesized that MHs secondary to RRDs could be associated with proliferative vitreoretinopathy (PVR), an idea supported by Cunningham et al.,⁸ who showed an association between PVR and simultaneous MH in a retrospective series. However, the presence of PVR may simply be an indicator of the chronicity of the detachment. In animal experiments, the ONL can lose ≤80% of its cell population after 90 days of detachment.^30,31 OCT scans demonstrate a thinning of the ONL and bacillary layer in chronic detachments, and histopathological studies corroborate these findings, revealing a complete loss of inner segments and outer segments and severe ONL atrophy in similar cases.³² Additionally, photoreceptor loss was shown to be associated with a degeneration of the outer processes of Müller cells of the foveal walls,³³ and we hypothesize that BALAD may contribute to photoreceptor loss via retrograde damage to cell bodies. Thus, secondary MH formation may be facilitated by cellular degeneration in long-standing RRDs, rather than the PVR itself. We observed BALAD cases with a noticeably thin central ONL (Figs. 2A–D), in which progressive cystic degeneration or any forces applied to the tissue could cause dehiscence of the MCC. Once a BALAD–lamellar hole occurs, the posterior band of photoreceptor and interphotoreceptor matrix remnants eventually degenerate (Figs. 6A–C). Although a full-thickness defect was observed in 20.7% of MHs at presentation, this finding likely represents prior cases of BALAD–lamellar holes, in which the posterior band degenerated (Figs. 4A–C). 

Management of Secondary MH in RRD

All patients achieved retinal reattachment irrespective of MH closure, consistent with the literature.^3,5,34 Thirty-one percent of patients attained closure after reattachment alone (without ILM peel), and 7% closed with primary combined procedures. Sixty-two percent of MHs persisted, of which most (16/20) required a PPV with ILM peeling, with a final closure rate of 80% (12/15). Although ILM peeling may lead to higher closure rates in RRD–MH cases,^9,34–37 some still favor no ILM peeling in primary RRD repair.^3,7 Data on modified techniques in the management of MHs secondary to RRD are limited. However, ILM peeling with a free-flap insertion technique was shown to produce better hole closure results.^38,39 Irrespective of the closure mechanism, all patients with OCTs at 6 months postoperatively had discontinuity or loss of the ellipsoid zone. 

Further studies will be required to confirm if there is an ideal technique and timing to perform surgery in these cases. Because the pathophysiology of secondary MHs in RRD is likely related to rapid degenerative processes, the visual prognosis may be worse than in idiopathic MHs. Patients with BALAD–lamellar hole had significantly worse postoperative visual acuity at all time points when compared with patients with BALAD-related abnormalities earlier in the spectrum (P < 0.001) or to those with no BALAD (P < 0.001). Therefore, preventing the progression from BALAD to BALAD–lamellar hole is particularly important, which may be achieved by reattaching the fovea in a timely fashion. This process allows the BALAD to quickly resolve, as shown in our previous study,¹¹ suggesting that these cases may be more urgent. 

Limitations

Although this is the largest dataset of BALAD-related abnormalities in fovea-involving detachments in the literature, MHs in RRD are still relatively uncommon. Therefore, larger prospective cohort studies are required to evaluate how variations in surgical management can influence the recovery/progression of these BALAD lesions. 

Conclusions

This study provides significant insight into the formation of secondary MH in RRD, because BALAD is shown to play a crucial role in its pathophysiology. OCT features suggest that hydration, tractional forces, and cystic and cellular degeneration are key processes, occurring sequentially in the following areas of low mechanical stability: RPE–photoreceptor interface, myoid zone, foveal wall HFL and MCC, and surrounding tissue. BALAD–lamellar holes generally do not resolve after RRD repair, with all cases progressing to FTMH at approximately 1 week. Postoperative visual acuity was found to be significantly worse in BALAD–lamellar holes versus BALAD abnormalities earlier in the spectrum (P < 0.001) at all time points, despite surgical closure of subsequent FTMHs. The timely management of fovea-off RRDs with associated BALAD may be most prudent to avoid the progression to BALAD–lamellar hole and worse functional outcomes. 

Acknowledgments

The creation of the Project MACULA website was supported by NIH R01EY06109 (CAC), International Retinal Research Foundation, and the Edward N. and Della L. Thome Memorial Foundation Awards Program in Age-Related Macular Degeneration Research. Institutional support to the Department of Ophthalmology and Visual Sciences (University of Alabama at Birmingham) comes from Research to Prevent Blindness (NYC) and the EyeSight Foundation of Alabama. 

Disclosure: I. Martins Melo, None; A. Jhaveri, None; A. Bansal, None; W.W. Lee, None; P.L. Oquendo, None; C.A. Curcio, None; R.H. Muni, None