Keywords
retina - vitreous - retinal detachment - epiretinal membrane
Schlüsselwörter
Retina - Netzhautablösung - epiretinale Membran - Vitrektomie
Introduction
Pars plana vitrectomy (ppV) currently represents the most popular and successful surgical
approach to rhegmatogenous retinal detachment (RRD) [1]. While its use and technical aspects are standardized internationally, a multitude
of procedural subtleties based on the individual surgeonʼs discretion have been described.
These include, for example, the use of 23-, 25-, or 27-gauge trocars [2], the use of air or gas tamponades [3], or the addition of macular surgery.
Concerning the latter, peeling of the internal limiting membrane (ILM) represents
the most frequently performed addendum in retinal detachment (RD) repair. Nevertheless,
its potential risks and benefits are debated controversially. A large benefit of ILM
peeling in RRD repair includes the prevention of postoperative fibrotic proliferation
on the macular surface. The formation of epiretinal membranes (ERMs) is one of the
most common postoperative complications after vitrectomy for RD. In 7 – 15% of vitrectomies
for RD, an ERM forms postoperatively [4], [5], [6]. The resulting membrane can exert high tensile forces on the underlying retinal
layer and alter its anatomy [7]. As a result, visual acuity (VA) gain after surgery is limited, metamorphopsia can
occur, and in some cases, reoperation with peeling of the ERMs is required [8]. In the
development of ERMs, vitreous cortex remnants (VCRs) play a major role by serving
as a scaffold on the ILM for fibrocellular proliferation [9], [10], [11], [12]. Peeling of the ILM results in removal of the VCRs and thus a reduced proliferative
stimulus [13], [14], [15], [16], [17], [18].
On the other hand, the ILM is a physiologic component of the retinal architecture.
Thus, ILM removal also carries some risks. Among these are modified postoperative
retinal anatomy with foveal shifting, focal retinal hemorrhage, and dissociated optic
nerve fiber layer (DONFL) [19], [20], [21], [22], [23]. Functionally, paracentral scotomas and reduced central retinal sensitivity may
occur [20], [24]. In addition, removal of the ILM in detached retina is surgically challenging and
much more complex compared to an attached retina [5].
The aim of the present study was to compare the outcomes of ILM peeling in proliferative
vitreoretinopathy (PVR) B RRD with uncomplicated RRD without ILM peeling concerning
VA and retinal reattachment rates.
Methods
Patient selection
For this retrospective cohort study, the Smart Eye Database (SmEyeDat) of the University
Hospital was screened for patients with primary RRD repair between January 2020 and
May 2023 by a single surgeon. The study adhered to the declaration of Helsinki and
local ethics committee approval was obtained. Provided written informed consent for
treatment was obtained and analysis was performed after anonymization. Screening included
all patients with at least a follow-up of 6 months. We included macular on and off
patients. Eyes with a history of trauma, uveitis, proliferative vitreoretinopathy
C, or high myopia > − 6 D were excluded. ILM peeling was performed based on the surgeonʼs
discretion using Membrane Blue as dye (DORC, Zuidland, Netherlands). Patients were
retrospectively divided into two groups. All patients with ILM peeling during RD surgery
and PVR B reaction were included in group 1, and patients who did not undergo ILM
peeling with no PVR were included in group 2.
PVR B reaction was categorized intraoperative by the surgeon. Intraoperative optical
coherence tomography (OCT) was performed in every patient and used for grading.
Intraoperative OCT was used to rule out a concomitant macular hole. Before and after
surgery, each patient underwent a standardized examination including refraction-based
VA testing, air-puff noncontact tonometry, slit lamp examination, and fundoscopy.
Postoperative OCT was performed using the Heidelberg Spectralis (Heidelberg Engineeringʼs
Spectralis HRA + OCT, Heidelberg, Germany) OCT. Clinical data was collected, including
age, sex, date of RD surgery, best-corrected visual acuity (BCVA) at each follow-up
visit, postoperative ERM formation, retinal redetachment, macular edema, and OCT parameters
such as central subfield thickness (CST). Decimal VA was converted to logarithm of
the minimum angle resolution (logMAR) units for analysis. Hand motion and finger counting
was converted to 2.3 and 1.9 as previously described [25].
Optical coherence tomography
Imaging was performed as previously described elsewhere [26]. Spectral-domain OCT was performed using the Spectralis HRA + OCT (Heidelberg Engineering)
system, including a volume scan (49 B-scans). Auto segmentation was proofed by a physician.
Heidelberg Spectralis SD-OCT (Heidelberg Engineering) was used to evaluate the postoperative
presence of ERMs, foveal profile, CST, presence of macular edema, and the integrity
of the ellipsoid zone (EZ).
Data analysis and statistics
Data management was performed with Microsoft Excel Version 16.72 (Microsoft Corp.,
Redmond, WA, USA) for Mac. For statistical analyses, IBM SPSS Statistics 28 (IBM Germany
GmbH, Ehningen, Deutschland) was used. As the data was not normally distributed, nonparametric
tests were applied. The differences between the groups were assessed using the Mann-Whitney
U test for nominal scaled data and non-normally distributed variables. Continuous
variables are expressed as the mean ± standard deviation and categorial variables
as frequencies. The significance level was set at p < 0.05.
Results
Baseline demographics
In total, 26 patients with 26 eyes were included in our study. All patients with ILM
peeling during RD surgery and PVR B reaction were included in group 1 (13 patients),
and patients who did not undergo ILM peeling with no PVR were included in group 2
(13 patients). The mean age was similar in group 1 (62.38 ± 8.61 years) and group
2 (63.77 ± 7.98) (p = 0.65). Characteristics of groups 1 and 2, including sex, age,
baseline VA, preoperative lens status, and foveal involvement, did not differ significantly
within the two groups and are summarized in [Table 1]. Preoperative PVR B features are also listed in [Table 1].
Table 1 Baseline clinical characteristics of groups 1 and 2. Bold: statistically significant
differences between the two groups. aMann-Whitney U test.
|
Group 1
|
Group 2
|
p value
|
|
n
|
13
|
13
|
|
|
Age (years)
|
62.38 ± 8.61
|
63.77 ± 7.98
|
0.65a
|
|
Baseline: visual acuity
|
1.03 ± 0.96
|
0.84 ± 0.97
|
0.39a
|
|
Pseudophakic
|
5
|
3
|
0.84a
|
|
Foveal involvement Macular on/off
|
4/9
|
7/6
|
0.34a
|
|
PVR B: ERMs preoperative
|
7
|
0
|
< 0.02a
|
|
PVR B: Surface wrinkling
|
13
|
0
|
< 0.01a
|
|
PVR B: Retinal stiffness
|
2
|
0
|
0.51a
|
|
PVR B: Decreased mobility of vitreous
|
3
|
0
|
0.34 a
|
Surgical outcomes
All eyes underwent RD surgery with vitrectomy including intraocular tamponade. Primary
success rates were 92.3% in groups 1 and 2 (12/13) (p = 1.0). The distribution of
intraocular tamponades did not differ significantly between groups and is listed in
[Table 2] (p = 0.96).
Table 2 Distribution of intraocular tamponade during vitrectomy in groups 1 and 2.
|
Tamponade
|
Group 1
|
Group 2
|
|
C2F6
|
9
|
9
|
|
6F6
|
0
|
1
|
|
Silicone oil
|
2
|
1
|
|
Air
|
2
|
2
|
In group 1, five patients had a combined phakovitrectomy surgery, and in group 2,
three patients (p = 0.84). All phakic patients received combined phakovitrectomy,
except for one 51-year old patient in group 2 who only underwent vitrectomy. Postoperatively,
this patient had a slight posterior subcapsular cataract, which, however, was visually
insignificant, with a VA of 0.1 logMAR postoperatively. There were no severe intraoperative
complications. Mean follow-up period was 555.08 ± 376.62 days in group 1 and 808.31 ± 402.12
days in group 2 in total and did not significantly vary in the two groups (p = 0.13).
OCT outcomes
Postoperative ERM formation based on OCT was significantly reduced by ILM peeling
(p = 0.04). Postoperative ERM formation was found in one patient in group 1 and six
patients in group 2. Two patients in group 2 had the need for a second ERM peeling
surgery. [Fig. 1] shows exemplary postoperative OCT scans. The mean CST in group 1 was 339.08 ± 50.14 µm
and in group 2 315.62 ± 33.53 µm; there was no significant difference (p = 0.06).
A postoperative cystoid macular edema was found in five patients in group 1 and in
two patients in group 2 (p = 0.34). Treatment was needed in most patients and included
Nevanac eye drops (Novartis AG, Basel, Switzerland), one patient additionally needed
peribulbar triamcinolone, and one patient received a dexamethasone implant (Ozurdex,
Abbvie, North Chicago, IL, USA). To evaluate the outer retina, the integrity of the
EZ was analyzed. In group 1, two patients and in group 2, three patients showed an
EZ defect in the
OCT scan; there was no significant difference (p = 0.76). OCT outcome parameters are
shown in [Table 3].
Fig. 1 a Group 1 patient (with intraoperative internal limiting membrane peeling) with asymptomatic
mild postoperative epiretinal membrane that required no treatment at that stage. b Group 2 patient (without intraoperative internal limiting membrane peeling) with
symptomatic postoperative epiretinal membrane that required further treatment.
Table 3 Optical coherence tomography parameters during the follow-up period. Bold: statistically
significant differences between the two groups. aMann-Whitney U test. ERM = epiretinal membrane; CME = cystoid macular edema; DONFL = dissociated
optic nerve fiber layer.
|
Group 1
|
Group 2
|
p value
|
|
n
|
13
|
13
|
|
|
ERM formation
|
1
|
6
|
0.04a
|
|
Redetachment
|
1
|
1
|
1.0a
|
|
Central retinal thickness
|
339.08 ± 50.14 µm
|
315.62 ± 33.53 µm
|
0.06a
|
|
CME
|
5
|
2
|
0.34a
|
|
Treatment für CME
|
4
|
2
|
0.52a
|
|
Ellipsoid zone defect
|
2
|
3
|
0.76a
|
|
DONFL
|
5
|
0
|
< 0.01a
|
Visual acuity
VA was tested at every visit. The baseline VA was 1.03 ± 0.96 in group 1 and 0.84 ± 0.97 logMAR
in group 2; there was no significant difference (p = 0.39). We analyzed VA postoperatively,
as well as the VA gain between baseline and postoperative visit ([Table 4], [Fig. 2], and [Fig. 3]). Postoperative VA was 0.26 ± 0.29 in group 1 and 0.15 ± 0.17 in group 2 (p = 0.125).
The VA gain between the baseline visit and postoperative follow-up visit in group
1 was − 0.77 ± 0.97 compared to − 0.7 ± 0.89 in group 2 (p = 0.920).
Table 4 Visual acuity follow-up in groups 1 and 2. Bold: statistically significant differences
between the two groups. aMann-Whitney U test. VA = visual acuity; logMAR = logarithm of the minimum angle of
resolution
|
Group 1
|
Group 2
|
p value
|
|
logMAR VA
|
|
|
|
|
Postoperative follow-up
|
0.26 ± 0.29
|
0.15 ± 0.17
|
0.125a
|
|
logMAR VA Change
|
|
|
|
|
Difference baseline – follow-up visit
|
− 0.77 ± 0.97
|
− 0.7 ± 0.89
|
0.92a
|
Fig. 2 Representation of the preoperative visual acuity (logMAR) in both groups using a
boxplot. ILM = internal limiting membrane; VA = visual acuity.
Fig. 3 Representation of the postoperative visual acuity (logMAR) in both groups using a
boxplot. ILM = internal limiting membrane; VA = visual acuity.
Discussion
Our study shows that, in skilled hands, additional ILM peeling in RRD with PVR B does
not negatively influence VA outcomes or anatomical success, while it dramatically
reduces the postoperative formation of ERMs, which, in the long term, reduces the
burden of reoperation and VA decline in patients with successful RD repair. Moreover,
no negative impact on ILM peeling on OCT biomarkers demonstrating a healthy retinal
architecture was found, except DONFL, which was seen in 38.5% of patients in group
1; however, VA did not significantly differ between our groups.
Concerning clinical decision-making, our study adds evidence supporting ILM peeling
in cases with primary RD with PVR B reaction. It seems to dramatically reduce the
redetachment rate. While redetachment in PVR B cases has been described to be up to
34% [13], we only found a redetachment in 7.7% of our ILM peeled PVR B cases. We thereby
could achieve a reduction of the redetachment rate to the levels of uncomplicated
RDs. As our data suggest, this benefit is not outweighed by potential anatomical damage,
for example, foveal shifting, focal retinal hemorrhage, and DONFL [19], [20], [21], [22], [23]. Concerning foveal shifting, the retina itself is already displaced due to the RD,
which probably renders this effect secondary to ILM peeling less important. Many retinal
surgeons even advocate performing ILM peeling
in patients with foveal splitting, a special anatomical configuration where the demarcation
line of the RD goes directly through the fovea and can cause severe postoperative
metamorphopsia [27]. Besides, potential negative effects of ILM peeling, such as paracentral scotomas
and reduced central retinal sensitivity, may also be caused by the RD itself [20], [24].
The efficacy of simultaneous ILM peeling with removal of the ERM to prevent recurrence
has been demonstrated in a number of studies in recent years. However, most of these
studies refer to idiopathic and not secondary ERMs [28]. Currently, there is controversy regarding the potential beneficial effects of ILM
peeling during RD surgery other than ERM recurrence reduction [24]. Based on recent literature, many authors recommend ILM peeling only for patients
with complicated RD [17], [29].
As a limitation, our study provides primary outcomes at follow-ups of 6 months.
Although our follow-up is relatively short, most ERM formations are detected within
the first 3 months postoperatively [4], [15], [18], [30]. So with a mean follow-up of 681.69 ± 389.94 days, we should have detected most
of the ERM formations. Metamorphopsia was, unfortunately, not recorded in all patients,
so it was omitted from the analysis. However, a large percentage of ERMs remain asymptomatic
and OCT imaging provides a sensitive diagnosis [8], [31].
As a caveat, longer-term retinal changes secondary to any intervention involving the
ILM might only be obvious after many months or years. For example, Klaas et al. recently
suggested that structural changes, e.g., foveal asymmetry secondary to posterior vitreous
detachment, can cause a domino effect that can cause lamellar or full-thickness macular
holes after a long period of 6 – 27 months [32]. This can also apply to ILM removal, which may cause substantial histologic damage
to the Mueller cells within the ILM. The mechanical stress of the basal lamina caused
by peeling can be transmitted to deeper retinal layers and lead to modifications of
retinal structures [33]. Lamas-Francis et al. performed a meta-analysis in 2023 examining studies with ILM
peeling in RD surgery, finding no significant difference in terms of VA outcome depending
on the peeling condition. However, in terms of redetachment rate, a significant
reduction was found in patients with ILM peeling [4]. Several other studies also demonstrated a reduction in ERM formation with ILM peeling
but found no significant difference in postoperative VA [14], [15], [27]. Obata et al. even observed poorer VA in patients with ILM peeling and macula-OFF
RD [16]. Eissa et al. found a worse postoperative VA and macular sensitivity in patients
with ILM removal. Additionally, they described retinal pits in all patients with ILM
peeling, which they explained by more mechanical trauma during surgery. Nevertheless,
there was no correlation of dimples with the VA or macular sensitivity [17]. However, there are further studies that demonstrated a better visual outcome after
ILM peeling [5], [18]. Nam and Kim described a better
postoperative VA in macula-ON patients when ILM peeling was performed. They attribute
this to the prevention of ERM formation, which could reduce VA postoperatively [18]. Although ILM peeling might not lead to a better postoperative VA, it can lead to
better anatomical success and macular compliance by decreasing the contraction forces
in PVR retinal redetachment [13]. Consistent with the studies mentioned above, postoperative VA was not significantly
worse in patients who underwent ILM peeling than in patients without peeling in our
study. Regarding the redetachment rate, we could not find any difference between the
two groups.
Additionally, our study is limited by the retrospective design, the relatively small
sample size, and the surgeonʼs bias of selecting patients for ILM Peeling.
In summary, we found that ILM peeling in primary rhegmatogenous ablatio with PVR B
can significantly reduce the risk of ERMs in high-risk patients and improve their
prognosis comparable to uncomplicated RRD. No negative effects on functional outcome
were detected. Further studies with larger cohorts, longer follow-up, and additional
functional testing (e.g., microperimetry) are needed to investigate the long-term
impact of ILM peeling on visual function and retinal redetachment.
Conclusion
Already known:
Newly described:
-
This study suggests that intraoperative removal of the ILM in PVR B RRD can improve
functional and morphological outcomes to levels obtained in uncomplicated RRD without
PVR.
-
ILM peeling does not appear to negatively affect postoperative VA.
