·Clinical
Research·
Epiretinal
membrane following pars plana vitrectomy for rhegmatogenous retinal detachment
repair
Ruti Sella1,2, Amir
Sternfeld1,2, Ivan
Budnik3, Ruth Axer-Siegel1,2, Rita Ehrlich1,2
1Department of Ophthalmology, Rabin Medical
Center-Beilinson Hospital, Petach Tikva 4941492, Israel
2Sackler Faculty of Medicine, Tel
Aviv University, Tel Aviv 6997801, Israel
3Department of Pathophysiology,
Sechenov First Moscow State Medical University, Moscow 119146, Russia
Co-first authors: Ruti Sella and Amir Sternfeld
Correspondence to: Ruti Sella. Department of
Ophthalmology, Rabin Medical Center-Beilinson Hospital, Petach Tikva 4941492,
Israel. rutibd@gmail.com
Received:
Abstract
Aim: To determine the rate and
possible contributors for post-pars plana vitrectomy (PPV) epiretinal membrane
(ERM) in patients treated for rhegmatogenous retinal detachment (RRD).
Methods: This prospective, nonrandomized study comprised 47 consecutive patients
(47 eyes) with acute RRD treated with
Results: ERM developed postoperatively
in 23 eyes (48.9%), none necessitated surgical removal. There was a
statistically significant difference between patients with and without ERM
postoperatively in preoperative best corrected visual acuity (median logMAR 1.9
vs 0.3, respectively; P=0.003)
rate of macula-off (69.6% vs 37.5%, respectively, p=0.028), and rate of ≥5 cryo-applications
(55.6% and 18.8%, respectively, p=0.039).
ERM developed mainly between the 1st and 3rd months of
follow-up. Macula-off status increased the risk of ERM, with the odds ratio of
3.81 (P=0.031).
Conclusion: ERM is a frequent post RRD finding, and its development is associated
with macula-off RRD.
KeyWords: epiretinal membrane; pars plana
vitrectomy; rhegmatogenous retinal detachment; cryotherapy; macula-off
DOI:10.18240/ijo.2019.12.09
Citation: Sella
R, Sternfeld A, Budnik I, Axer-Siegel R, Ehrlich R. Epiretinal membrane
following pars plana vitrectomy for rhegmatogenous retinal detachment repair. Int
J Ophthalmol 2019; 12(12):1872-1877
Introduction
Primary rhegmatogenous retinal detachment
(RRD) requires early surgical intervention to prevent loss of vision. Although
the functional and anatomical success rates of pars plana vitrectomy (PPV) for
the treatment of RRD are considered high[1-2],
late postoperative complications may occur, including the formation of an
epiretinal membrane (ERM) and the development of proliferative
vitreoretinopathy (PVR)[3-4]. ERM,
also known as cellophane membrane or macular pucker, principally consists of
retinal pigment epithelial cells, hyalocytes, and retinal glia. It may be
either idiopathic or secondary to a wide variety of conditions, including
intraocular surgery[5-9]. ERM
formation following surgery for RRD may be asymptomatic, warranting follow-up
alone. However, it may also distort the retina causing severe metamorphopsia or
a clinically significant decrease in visual acuity, warranting further surgical
intervention for membrane peel. PVR is characterized by the development of
periretinal membranes which later contract to create retinal traction[10]. Several authors have suggested that ERM following
PPV[11-13] may be considered
an early stage of PVR[10].
ERM has been reported to occur in
4%-8.5% of patients following successful scleral buckling procedure for primary
RRD[14-17]. Rates of 6%-13%[5,16,18-19],
more recently rising to as high as 21%[20-21], have been reported following successful
PPV. There is currently a paucity of large-scale prospective studies assessing
possible contributors to the development of ERM after
SUBJECTS AND Methods
Ethical Approval The study was conducted in
accordance with the Declaration of Helsinki and was approved by the research
Ethics Committee of Rabin Medical Center. All patients had been fully informed
of the purpose and methods of the present study and provided written informed
consent from themselves or their guardians.
A prospective, nonrandomized study
design was used. The cohort included consecutive patients with acute primary
RRD who underwent 23-gauge PPV in a single tertiary center. Exclusion criteria
were PVR stage 3 and higher, diagnosis of ERM prior to surgery, giant retinal tear,
and traumatic or non-RRD.
Pars Plana Vitrectomy Procedure PPV was performed by one of two
surgeons using the Stellaris PC vitreoretinal surgical system (Bausch &
Lomb Inc., Irvine, CA, USA) and the Resight fundus viewing system (Zeiss
Meditech, Dublin, CA, USA). The method of retinopexy and type of tamponade were
left to the surgeon’s discretion. In all cases, either the presence of a
posterior vitreous detachment was confirmed or, if it was not present, active
aspiration was used to induce a separation.
Scleral depression was performed to
identify all breaks. Particular care was taken to ensure all traction was
relieved; this usually warranted excision of the anterior flap with the
vitrector. Retinopexy was performed with cryotherapy or endolaser photocoagulation;
the cryotherapy was performed until whitening covered the edges of the tear in
smaller tears, and in larger tears two cryotherapy balls were applied to cover
all the edges. Cryotherapy was not repeated twice on the same spot. The
cryotherapy was performed under air or heavy liquid on an attached
tear. For tamponade, nonexpansile mixtures of either perfluoropropane (C
Data Collection At completion of surgery, the
surgeon filled out a detailed questionnaire covering the clinical features of
the detachment, including documentation of macular involvement, number and
position of retinal tears, and extent (in clock hours) of retinal detachment,
as well as the course of surgery, including combination with
phacoemulsification for cataract extraction, the use of perfluorocarbon liquid
to flatten the retina, the type of tamponade chosen, and the use and number of
applications of endolaser or cryo-coagulation. Surgical complications were
documented. Demographic and background data on the patients were collected from
the files.
Follow-up Examinations Patients were followed 1d, 1wk, and
1, 3, and 6mo after PPV. At each follow-up visit, best corrected visual acuity
(BCVA) was assessed using Snellen charts followed by a detailed slit-lamp
examination including dilated fundus examination. The diagnosis was confirmed
with spectral-domain optical coherence tomography (SD-OCT; Spectralis,
Heidelberg Engineering, Heidelberg, Germany) at the 1-, 3-, and 6-month visits.
The SD-OCT examinations were performed by a retina specialist blinded to the
procedures performed during surgery. ERM formation was defined as the
appearance of a hyper-reflective line internal to the inner limiting membrane
on the SD-OCT scan. During the follow up time, all patients were treated with
prednisolone 1% and ofloxacin eye drops applied four times daily to the
operated eye starting the day after surgery and continued through the first
post-operative month. The antibiotic treatment was then stopped, and the
steroid drops were gradually tapered by a drop every week over the course of
the second month, as commonly practiced.
Statistical Analysis Statistical analysis was performed
using Statistica 10.0 (Statsoft, Tulsa, OK, USA). Snellen BCVA values were
converted to the logarithm of the minimum angle of resolution (logMAR).
Continuous variables were tested for normality using the Shapiro-Wilk test and
presented as mean±standard deviation (SD) when normally distributed or as
median (interquartile range) when non-normally distributed. Categorical variables
were presented as counts and proportions. Patients with and without ERM were
compared for continuous variables using Student’s t-test or Mann-Whitney
U test, and for proportions using Pearson’s Chi-squared test or Fisher
exact test. Backward stepwise multiple logistic regression analysis was
performed to investigate the relationship between the method of retinopexy
(cryocoagulation, laser photocoagulation, or both), macular status (on or off),
tamponade agent used (silicon or gas) and the number of retinal tears (0-3 or
4+) with the formation of ERM. Two-tailed p
values of less than 0.05 were considered statistically significant.
Results
Of the 59 patients recruited for the
study, 12 were excluded because they were lost to follow-up (n=9), died during
the follow-up period (n=2), or had recurrent RRD due to PVR after the
first surgery (n=1). The remaining 47 patients (47 eyes) were included
in the statistical analysis.
ERM, either diagnosed or confirmed
by SD-OCT, developed postoperatively in 23 eyes (48.9%). In no case was the ERM
significant enough to necessitate surgical removal. The demographic and
baseline features of the patients with and without ERM are detailed in Table 1.
Table 1 Demographics and baseline
characteristics of the patients with and without ERM
|
Baseline characteristic |
Groups |
p |
|
|
ERM |
Non-ERM |
||
|
No. of eyes |
23 |
24 |
|
|
Age (y), mean±SD |
67.1±9.6 |
64.0±10.5 |
|
|
BCVA at
diagnosis (logMAR), median (IQR) |
1.9 (0.3-2.2) |
0.3 (0.1-1.5) |
0.003b |
|
Lens status, n (%) |
|
|
|
|
Phakia |
10 (43.5) |
6 (25.0) |
|
|
Pseudophakia |
13 (56.5) |
18 (75.0)d |
|
|
Macula status, n (%) |
|
|
|
|
On |
7 (30.4) |
15 (62.5) |
|
|
Off |
16 (69.6) |
9 (37.5) |
|
|
RRD extent (h), mean±SD |
5.8±2.5 |
5.2±2.1 |
|
|
RRD location (quadrants) |
|
|
|
|
Upper temporal |
17 (73.9) |
17 (70.8) |
|
|
Lower temporal |
13 (56.5) |
11 (45.8) |
|
|
Upper nasal |
11 (47.8) |
10 (41.7) |
|
|
Lower nasal |
6 (26.1) |
7 (29.2) |
|
|
No. of tears, n (%) |
|
|
|
|
0-3 |
14 (60.9) |
20 (83.3) |
|
|
>4 |
9 (39.1) |
4 (16.7) |
|
BCVA: Best corrected visual acuity;
ERM: Epiretinal membrane; IQR: Interquartile range; RRD: Rhegmatogenous retinal
detachment; SD: Standard deviation. aStudent’s t-test; bMann-Whitney
U test; cPearson’s Chi-squared test; dIncluding
one aphakic patient.
On comparison of the two groups, the
ERM group was found to have a significantly lower baseline mean BCVA [logMAR
1.9 (0.3-2.2) vs logMAR 0.3 (0.1-1.5), p=0.003] and a higher rate of macula-off (69.6% vs
37.5%, p=0.028). The ERM
group also had a higher proportion of patients with 4 or more retinal tears
(39.1% vs 16.7%), but the difference from the non-ERM group was not
statistically significant (p=0.085).
Other baseline variables were similar in the two groups, and no preoperative
risk factors for the formation of ERM were identified. Overall, there was a
predominance of upper temporal quadrant involvement in the RRD (73.9% in the
ERM group and 70.8% in the non-ERM group).
The surgical and follow-up data of
the ERM and non-ERM groups are detailed in Table 2. Significantly more patients
in the ERM group received 5 or more cryo-applications (55.6% vs 18.8%, p=0.039). The ERM group also
received more laser applications (1283±657 vs 948±427), but the
between-group difference was not statistically significant (p=0.122). The two groups were
similar for all other intraoperative variables, including the retinopexy method
selected by the surgeon, the tamponade agent, use of heavy liquid, and
performance of cataract surgery in combination with the RRD repair. At the end
of the follow-up period, the BCVA was logMAR 0.2 (0.2-0.70) in the ERM group
and logMAR 0.2 (0.10-0.5) in the non-ERM group (p=0.65). Figure 1 describes the change in BCVA from
baseline to the end of follow-up in both groups. The course of ERM formation
during follow-up is described in Figure 2.
Table 2 Surgical and follow-up data
of the patients with and without ERM
|
Parameters |
Groups |
p |
|
|
ERM |
Non-ERM |
||
|
No. of eyes |
23 |
24 |
|
|
Retinopexy, n (%) |
|
||
|
Laser therapy |
5 (21.7) |
10 (41.7) |
|
|
Cryotherapy |
9 (39.1) |
8 (33.3) |
|
|
Both |
9 (39.1) |
6 (25.0) |
|
|
Laser applications, mean±SD |
1283±657 |
948±427 |
|
|
360° laser applications |
|
||
|
Yes |
8 (34.8) |
9 (37.5) |
|
|
No |
15 (65.2) |
15 (62.5) |
|
|
Cryo applications |
0.039d |
||
|
1-4 |
8 (44.4) |
13 (81.3) |
|
|
+5 |
10 (55.6) |
3 (18.7) |
|
|
Tamponade, n (%) |
|
||
|
Gas |
20 (87.0) |
23 (97) |
|
|
Silicone |
3 (13.0) |
1 (3.0) |
|
|
PFC, n (%) |
|
||
|
Yes |
6 (26.0) |
5 (20.8) |
|
|
No |
17 (74.0) |
19 (79.2) |
|
|
Combined cataract surgery, n
(%) |
1.000d |
||
|
Yes |
2 (8.7) |
1 (4.2) |
|
|
No |
21 (91.3) |
23 (95.8) |
|
|
Final BCVA (logMAR), median (IQR) |
0.2 (0.2-0.7) |
0.2 (0.1-0.5) |
0.65b |
BCVA: Best corrected visual acuity;
ERM: Epiretinal membrane; IQR: Interquartile range; RRD: Rhegmatogenous retinal
detachment; PFC: Perfluorocarbon; SD: Standard deviation. aStudent’s
t-test; bMann-Whitney U test; cPearson’s
Chi-squared test; dFisher exact test.
Figure 1 Box-and-whiskers plots of
changes in BCVA during follow-up The boxes
span the 25th to the 75th percentile; the whiskers span
the lowest to the highest observations; and the line inside each box denotes
the median. The figure shows the median BCVAs of the ERM and non-ERM groups
prior to surgery and at every follow-up visit. BCVA was significantly worse in
the ERM group prior to surgery. At each time point, the groups were compared
using the Mann-Whitney U test.
Figure 2 Proportion of patients with
ERM during follow-up.
In the majority of patients (n=12,
48.9% of the ERM group, 25.6% of the whole cohort), ERM developed between the
first and third postoperative months; in only 5 patients (21.7%, 10.6%) was ERM
evident already at one month.
To further investigate the
association between method of retinopexy, macular status, tamponade agent used
and number of retinal tears with the formation of ERM, we performed a backward
stepwise multiple logistic regression analysis. As shown in Table 3, only
macular status was found significant in the final model.
Table 3 Factors associated with risk
of ERM
|
Independent variable |
B |
SE of B |
p |
Odds ratio |
95%CI for odds ratio |
|
Macular status, off |
1.338 |
0.619 |
0.031 |
3.810 |
1.132-12.816 |
|
Constant |
-0.762 |
0.458 |
0.096 |
0.467 |
N/A |
B: Unstandardized coefficient; N/A:
Not applicable; SE: Standard error.
χ2(1)=4.942, P<0.026; Nagelkerke R2=0.133;
based on the final model of backward stepwise multiple logistic regression
analysis. Independent variables: Method of retinopexy, cryocoagulation=0, laser
photocoagulation=1, both=1; Macular status, on=0, off=1; Tamponade agent,
silicon=0, gas=1; Number of retinal tears, 0-3 tears=0, 4+ tears=1.
Interpreting these data, the odds of
ERM formation increases by a factor of 3.81 (95%CI 1.13-12.82, p=0.031) if the macula is off. The
other variables are not associated with ERM formation.
Discussion
Our study prospectively evaluated
the incidence and potential contributors to the formation of ERM, as evaluated
by SD-OCT, following
The incidence of ERM following PPV
for RRD was 48.9%. Nevertheless, none of the patients with ERM required an
internal limiting membrane (ILM) peel procedure during follow-up. The ERM had
no adverse effect on BCVA. While we describe a statistically significant
difference between patients with and without ERM postoperatively in
preoperative BCVA, this may have been impacted by the lens status at
presentation, however, we found no significant difference in the lens status
between the ERM and non-ERM groups as stated in Table 1, and the lens status
was not found to be a factor associated with ERM in multivariate analysis.
SD-OCT is the current gold standard
method for ERM diagnosis. Its use has made it possible to identify cases of ERM
that would otherwise be overlooked. We speculate that in the pre-SD-OCT era,
the technological limitations of clinical examination or low-resolution time
domain OCT may have led to an underdiagnosis of postoperative ERM formation[5,12,16]. This is
supported by two recent studies in which SD-OCT was used to evaluate the
incidence of ERM formation following PPV for RRD with and without ILM peeling during
the initial surgery[21-22].
Both found a similar postoperative ERM incidence of 21%[21-22], twice the rate cited in the earlier literature.
However, both studies used a retrospective design and were therefore prone to
recruitment bias. The even higher incidence in the present study might be
attributable to the strict follow-up regimen. The performance of OCT in all
patients at each check point, regardless of BCVA or visual symptoms, could have
yielded a high rate of diagnosis of ERM, including many cases that were
clinically nonsignificant. However, a direct comparison with previous studies
is impractical owing to the major differences in study design, OCT equipment
used, and exclusion criteria.
Several potential predisposing
factors have been reported for the development of ERM in patients after surgery
for RRD. These include preoperative macular detachment, vitreous hemorrhage
(VH), low BCVA, and numerous or large equatorial retinal breaks[5,15-17,22];
intraoperative placement of a high number of cryocoagulation spots; and lack of
postoperative use of systemic steroids[12]. In
our study, none of the patients presented with VH. In concordance with some
previous studies[12,22] but
not others[5], the ERM group was characterized by
significantly more cryocoagulation spots intraoperatively than the non-ERM
group, which may possibly be attributable to an ensuing inflammatory reaction.
The assumption that ERM is an early stage of PVR and that the development of
PVR can be influenced by modifying the surgery-induced inflammatory reaction
and disruption to the blood-retinal barrier was previously reported by Koerner et
al[12], based on a positive effect of
postoperative steroids in reducing the early stages of PVR after RRD surgery.
Interestingly, the number of cryocoagulation spots in their study was
correlated to the percentage of eyes with ERM. This finding is in
line with previous evidence of more frequent postoperative autoimmune reactions
against retinal antigens in eyes after excessive cryocoagulation therapy[23]. Some surgeons choose to use cryotherapy during RRD
repair; it is reassuring that small amount of cryotherapy applications did not
increase the risk of ERM formation. A possible explanation might be that the
performance of cryotherapy under air prevents the dispersion of retinal pigment
epithelium cells. The larger numbers of cryotherapy might have increased the
risk for ERM formation due to increased inflammatory reaction or greater number
of retinal tears that require more cryotherapy applications and more laser
applications that were found to be also related to increased risk of ERM. In
our study, the multiple logistic regression analysis, did not point at an
association between cryotherapy, as the go-to method of retinopexy, and the
development of ERM post-surgery. This, however, does not preclude the possible
contribution of a high number of cryotherapy applications, to the development of
ERM, as shown in Table 2.
Other possible contributors to the
formation of ERM following PPV for RRD are the number, size, and location of
retinal tears. This may be explained by the dispersion of retinal pigment
epithelium cells through larger, peripheral, or more numerous retinal breaks,
which later serve as a scaffold for ERM formation[5].
In our study, the ERM group contained a larger proportion of patients with 4 or
more retinal tears than the non-ERM group, in agreement with the results of
Katira et al[5] but not those of Nam and
Kim[20], although the difference between our
groups did not reach statistical significance. Moreover, there were cases of ERM
among our patients with RRD and macular involvement, as suggested by Rezar et
al[24]. We found no statistically significant
differences in demographic parameters between the ERM and non-ERM groups, or in
lens status at the time of surgery, retinopexy method selected by the surgeon,
or tamponade agent. Intraoperative considerations, such as performance of
combined PPV and phacoemulsification or use of heavy liquid to flatten the
retina, had no effect on the risk of ERM formation.
ILM peeling is not routinely done at
our institution. Although ILM peeling has been suggested as a means to diminish
ERM formation following PPV[20-21],
we found that it was not required in any of our patients during the 6-month
follow-up after PPV owing to good visual acuity. This may suggest that routine
ILM peeling has little clinical benefit in terms of avoiding complications of
PPV for RRD.
The main limitations of our study
are the relatively small sample size which prevented us from evaluating each
possible contributor as an independent risk factor for the formation of ERM,
and the short follow-up time which may have masked some late postoperative
cases of ERM. Moreover, the prospective design of our study wherein patients
with acute RRD were recruited at presentation did not allow for early,
presurgery, clinical screening for ERM formation in the macula-off patients,
which may have been one of the reasons for ERM overdiagnosis. It is, however,
reassuring that the vast majority of the ERM group (78.7%) was diagnosed 1-6mo
postoperatively, and in only 21.7% was ERM evident already at the one-month
follow-up, indicating that most patients did not have ERM prior to surgery.
In summary, in the present
single-center study, ERM formation occurred in nearly half the patients treated
with
ACKNOWLEDGEMENTS
Conflicts of Interest: Sella R, None; Sternfeld A, None; Budnik
I, None; Axer-Siegel R, None; Ehrlich R, None.
References