FUNCTIONAL AND STRUCTURAL

MEASUREMENTS FOR THE ASSESSMENT

OF INTERNAL LIMITING MEMBRANE
PEELING IN IDIOPATHIC MACULAR
PUCKER
SAMIR R. TARI, MD,* ORIT VIDNE-HAY, MD,*
VIVIENNE C. GREENSTEIN, PHD,* GAETANO R. BARILE, MD,*
DONALD C. HOOD, PHD,† STANLEY CHANG, MD*

Objective: To investigate the role of structural and functional measurements in the

assessment of internal limiting membrane (ILM) peeling for the treatment of eyes with
macular pucker.
Methods: Ten patients with macular pucker who underwent pars plana vitrectomy with
ILM peeling were studied prospectively. Visual acuity measurement, standard automated
achromatic perimetry, multifocal electroretinography (mfERG), and optical coherence to-
mography (OCT) were performed before and 3 months after surgery. Four surgical samples
obtained from similar patients were analyzed with electron microscopy.
Results: Three months after surgery, mean visual acuity ⫾ SD was significantly
improved from 0.4 ⫾ 0.11 logMAR to 0.19 ⫾ 0.13 logMAR (P ⱕ 0.002), and mean central
retinal thickness ⫾ SD was significantly decreased 428 ⫾ 73 ␮m to 326 ⫾ 34 ␮m (P ⱕ
0.002). The mfERG response amplitudes were slightly decreased in eight patients, and five
of these patients also had asymptomatic decreases in visual field sensitivity. The electron
micrographs revealed segments of Müller cell footplates on the retinal side of the ILM in all
four specimens.
Conclusion: In this study, the use of mfERG, OCT, and standard automated achromatic
perimetry showed changes in macular function and structure postoperatively. These
measures of visual function and structure allow for better evaluation of the surgical
outcome and understanding of the changes that may occur after ILM peeling.
RETINA 27:567–572, 2007

M acular pucker is a disorder of the vitreomacular

interface in which a fibrocellular growth in the
form of an epiretinal membrane (ERM) induces tan-
From the Departments of *Ophthalmology and †Psychology,
Columbia University, New York, New York.
Supported by a grant from the National Eye Institute (EY02115
and EY09076), an unrestricted grant from Research to Prevent
Blindness, Inc. (New York, NY), and a grant from The Starr
Foundation.
Reprint requests: Vivienne C. Greenstein, PhD, Department of
Ophthalmology, 635 West 165th Street, New York, NY 10032;
e-mail: vcg1@nyu.edu
gential tractional force on the retina, leading to defor-
mation of the retinal architecture. This can result in
impairment of visual acuity and complaints of meta-
morphopsia. The standard treatment for macular
pucker is surgical removal of the ERM. However, it
has been reported that most surgical samples derived
from this surgery have shown the presence of internal
limiting membrane (ILM) along with ERM.1–3
The ILM is an acellular structure that lies in the
vitreoretinal junction. It is formed mainly of collagen
type IV, is bound by the vitreous humor on the inner

567
568 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
side, and interdigitates with Müller cell footplates on
the outer side. The effects of removal of ILM have
been reported in the literature.1,4,5 In a pilot study by
Park et al,4 peeling of ILM along with ERM produced
better outcomes in both visual acuity and recurrence
rate than ERM peeling alone. Similarly, Bovey et al1
found that ILM removal was associated with better
final visual acuity and a lower incidence of reoccur-
rence of epimacular membrane. However, Sivalingam
et al5 reported that eyes treated with ILM removal
together with ERM removal had a less favorable out-
come with regard to visual acuity than eyes that had
only ERM removal. The outcome measures used in
these studies were visual acuity and recurrence rate of
ERM. The drawback of an outcome measure such as
visual acuity is that it does not reflect the subtle
structural and functional changes that might be occur-
ring after this type of surgery.
In this study, we assessed the outcome of this
surgery using a combination of functional (multifocal
electroretinography [mfERG], standard automated
achromatic perimetry, and visual acuity measurement)
and structural (optical coherence tomography [OCT]
and electron microscopy) techniques to detect changes
that may occur.
Methods
Ten patients with idiopathic macular pucker under-
going pars plana vitrectomy and membrane peeling
were prospectively recruited for this study. The mean
age ⫾ SD of the 4 females and 6 males was 63.9 ⫾ 9.5
years (range, 40 –75 years). Exclusion criteria were
the presence of significant media opacities and/or
other ocular diseases. All subjects gave written in-
formed consent before participating in the study. The
protocol for this study was approved by the committee
of the Columbia University Medical Center Institu-
tional Review Board, and procedures followed the
tenets of the Declaration of Helsinki.
The following tests were performed before surgery
and 3 months after surgery: visual acuity measure-
ment, standard automated achromatic perimetry with
the Humphrey Field Analyzer II (Carl Zeiss Meditech,
Inc., Dublin, CA) utilizing the SITA-standard 10-2
program, OCT, fundus photography, and mfERG.
Cone-mediated mfERG responses were recorded
using VERIS 4.9 software (Electro-Diagnostic Imag-
ing, San Mateo, CA) after pupil dilation (1% tropic-
amide and 2.5% phenylephrine hydrochloride). The
mfERG procedure used in this study has been de-
scribed in detail elsewhere.6,7 Briefly, the stimulus
consisted of an array of 103 hexagons scaled with
eccentricity. The stimulus array subtended 48° hori-
● 2007 ● VOLUME 27 ● NUMBER 5
zontally and 40° vertically and was presented via a
VERIS camera/display/refractor unit. The luminance
of the white hexagons was 200 cd/m2, and the lumi-
nance of the black hexagons was 1 cd/m2; the sur-
round was 100 cd/m2. The mfERG responses were
recorded with a Burian-Allen bipolar contact lens
electrode. The nontested eye was patched. The con-
tinuous record was amplified with the high and low
frequency cutoffs set at 10 Hz and 300 Hz (Grass
PreAmplifier P511J, Quincy, MA), respectively. The
recording was ⬇7 minutes in duration, divided into 32
segments for patient comfort. Any segment contami-
nated by loss of fixation, small eye movements, or
significant artifacts was rejected and rerecorded. Fix-
ation stability was continuously monitored.
A single iteration of artifact rejection was applied to
the raw data, and no spatial averaging was performed.
First-order responses were analyzed by computer soft-
ware (VERIS). The peak response amplitude was
measured from the N1 trough to the P1 peak and
expressed in units of response density (nV/deg2).
Two methods of analyses were performed: area
analysis and individual hexagon analysis. For area
analysis, the array of 103 responses was divided into two
regions: the peeled area or area pl, which was identified
by overlaying the mfERG response array on a retinal
fundus photograph on which the peeled area was
mapped by the performing surgeon after surgery; and a
nonpeeled area or area npl, which corresponded to re-
sponses associated with the remaining stimulus area.
Responses in the two areas were averaged, and the mean
response amplitude values were used for quantitative
analysis. Area npl served as an endogenous control to
correct for interexperimental variability and changes in
media opacity due to surgery. The following correction
formula was used: corrected Ampl plpost ⫽ Ampl plpost/
(Ampl nplpost/Ampl nplpre).
For individual hexagon analysis, the P1 response
amplitude associated with each hexagon was com-
pared before and 3 months after surgery. The postsur-
gical measurements were corrected for interexperi-
mental variability (e.g., changes in media opacity due
to surgery) using the correction formula described
above.
OCT was performed using Zeiss OCT3 (Zeiss
Humphrey Instruments, Dublin, CA). Seven-millime-
ter line scanning and fast macular thickness analyses
were performed. Measurements of the average thick-
ness of the central 1 mm and the macular volume of
the central 6 mm of the retina before and after surgery
were compared.
All surgeries were performed by the same surgeon
as standard three-port pars plana vitrectomy. In 9 eyes,
both ERM and ILM were peeled and identified clini-
 ASSESSMENT OF ILM PEELING ● TARI ET AL 569

cally, and in 1 eye (Patient 10), ERM was peeled

6.88 ⫾ 0.67
Fellow Eye
without identification of ILM. The location and extent

7.52
6.73
6.56
8.50
6.49
7.12
6.40
6.41
6.49
6.59
of ERM and ILM peeled in each eye were drawn on an
enlarged fundus photograph immediately after sur-
Macular Volume (mm3)

gery. Indocyanine green was not used to assist ILM

7.49 ⫾ 0.77
peeling in any of the surgeries. The eyes underwent
Post

8.09
8.99
6.71
7.50
6.75
7.35
6.73
6.90
7.62
8.30
air–fluid exchange before closure.
Four membrane samples were also obtained from
other eyes with idiopathic macular pucker during sur-
gery. The peeled membranes were immediately fixed

9.19 ⫾ 1.06
with 2.5% glutaraldehyde in 0.1 mol/L Sorenson
11.10

10.43
9.06
7.09
9.52
8.95
8.98
8.94
8.95

8.89
Pre

buffer (pH 7.2) for at least 12 hours. Samples were

then postfixed with 1% OsO4 also in Sorenson buffer
33.10 ⫾ 2.73 427.70 ⫾ 73.39 325.70 ⫾ 34.25 for 1 hour. En block staining was performed using 1%
tannic acid. After dehydration and embedding, the
Central Foveal Thickness

tissue was cut in thin sections (60 nm). The sections

Post

372
320
272
339
348
361
328
345
291
281

were stained with uranyl acetate and lead citrate and

examined under a JEOL JEM-1200 EXII electron
Table 1. Preoperative and Postoperative Measurements

(␮m)

microscope (JEOL, Peabody, MA).

505
378
286
451
535
400
466
477
402
377
Pre

Results
All 10 surgeries were successful in removing the
pucker and improving the retinal morphology as evi-
denced clinically and by OCT.
Foveal Threshold (dB)

Post

Preoperative and postoperative visual acuities, fo-

33
35
34
38
34
30
32
33
28
34

veal threshold obtained with standard automated ach-

romatic perimetry, central retinal thickness, and mac-
ular volume for the 10 patients are listed in Table 1.
31.20 ⫾ 3.08

All these measures showed improvement after sur-

Pre

gery. Examples of preoperative and postoperative

33
31
34
38
28
31
30
30
28
29

OCT thickness maps (Patients 1, 4, and 8) are shown

in Figure 1; the postoperative maps of three patients
showed a decrease in retinal thickness. Visual acuity
0.19 ⫾ 0.13

and retinal thickness measurements are shown in Fig-

Visual Acuity (logMAR)

Post

0.10
0.10
0.00
0.18
0.18
0.30
0.10
0.40
0.40
0.18

ure 2. All 10 patients had improvement in visual

acuity after surgery. The mean value ⫾ SD changed
from 0.40 ⫾ 0.11 logMAR to 0.19 ⫾ 0.13 logMAR
(P ⱕ 0.002). Central retinal thickness and macular
0.40 ⫾ 0.11

volume were significantly decreased after surgery.

0.18
0.30
0.40
0.40
0.48
0.40
0.40
0.48
0.60
0.40
Pre

Again, all 10 patients had a decrease in both central

retinal thickness and macular volume. Mean central
retinal thickness ⫾ SD changed from 428 ⫾ 73 ␮m to
326 ⫾ 34 ␮m (P ⱕ 0.002). Mean macular volume ⫾
Sex

M
M
M

M
M

M
F

F
F
F

SD decreased from 9.21 ⫾ 1.04 mm3 to 7.58 ⫾ 0.77

mm3 (P ⱕ 0.001). Although there was a decrease in
63.90 ⫾ 9.49

volume, the values were greater than those for the

Age (y)

58
40
66
68
66
69
67
62
75
68

fellow eyes (6.70 ⫾ 0.38 mm3; P ⱕ 0.05) (one patient

[Patient 5] was excluded due to the presence of mac-
ular pucker in the fellow eye).
Mean foveal threshold ⫾ SD as measured by the
Mean ⫾ SD
Patient

Humphrey Field Analyzer II improved from 31.2 ⫾

3.1 dB to 33.1 ⫾ 2.7 dB. The trend toward increased
1
2
3
4
5
6
7
8
9
10

foveal sensitivity did not however reach statistical

570 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
Fig. 1. Optical coherence tomography 6-mm thickness map before
(Pre) and 3 months after (Post) surgery showing a decrease in retinal
thickness after surgery. P1, Patient 1; P4, Patient 4; P8, Patient 8.
significance. Although foveal sensitivity was in-
creased for five patients, it remained unchanged for
four, and was minimally decreased for one.
Figure 3 shows the mfERG (P1) amplitudes in the
peeled areas before and after surgery. Seven patients
had a decrease in P1 amplitude 3 months after surgery.
Two patients (Patients 6 and 9) had essentially no
change, and another patient (Patient 10) who had only
the ERM peeled had an increase in amplitude. Al-
though the decreases in amplitudes between preoper-
ative and postoperative values were very small (1.1–
7.2 nV/deg2), when the responses of individual
hexagons were examined within the peeled area, 8
patients had at least 1 region in which the mfERG
responses associated with ⱖ5 contiguous hexagons
● 2007 ● VOLUME 27 ● NUMBER 5
Fig. 3. Multifocal electroretinography P1 response amplitudes in the
peeled areas. Preoperative values, filled triangles (Pre); postoperative
values, unfilled squares (Post). *Patient 10 had only the epiretinal
membrane peeled.
were decreased ⬎20% compared with preoperative
values. Figure 4 shows the area with peeled ILM and
the areas that had decreased mfERG response ampli-
tudes.
Although the standard automated achromatic pe-
rimetry results showed a trend toward improvement in
foveal thresholds after surgery and no significant dif-
ference in mean deviation or pattern ⫾ SD, five of
seven patients with decreased mfERG amplitudes
(Fig. 3) had areas of increased threshold values on
both total and pattern deviation plots. The pattern
deviation plots showed new clusters of contiguous test
points with increased threshold values. These test
points were within the peeled areas.
Ultrastructural analysis of the peeled membranes
showed that all four samples included an ILM with
Müller cell remnants on the retinal side. Three of the
samples also had collagen fibers and fibroblast-like
Fig. 2. Top left, Visual acu-
ity (logMAR) before (Pre)
and 3 months after (Post) sur-
gery. Top right, Central reti-
nal thickness before and 3
months after surgery (n ⫽ 10;
error bars, ⫾1 SD). Bottom,
Macular volume (n ⫽ 10; er-
ror bars, ⫾1 SD). OCT, opti-
cal coherence tomography.
 ASSESSMENT OF ILM PEELING ● TARI ET AL 571

Fig. 4. A, Fundus photo-

graph of Patient 8 showing
the area with peeled internal
limiting membrane (shaded in
gray) and areas in which mul-
tifocal electroretinography
(mfERG) response ampli-
tudes were decreased after
surgery (shaded in green). B,
Fundus photograph of Patient
6 who did not have any
changes according to peeled
vs. nonpeeled analysis but
had areas of decreased
mfERG response amplitudes
when analyzed by individual
hexagons.

cells as seen in Figure 5, while the fourth sample
included only ILM and Müller cell remnants (Fig. 6).
Discussion
In our study of patients undergoing surgery for
idiopathic macular pucker, all 10 patients had im-
provement in visual acuity after surgery. The increase
in foveal sensitivity and the decrease in central foveal
thickness and macular volume shown by OCT are all
findings that are consistent with improvement in vi-
sual acuity and represent the successful outcome of
surgery. The mfERG however showed areas in which
response amplitudes were slightly decreased postop-
eratively. These areas with decreased mfERG re-
sponse amplitudes often extended outside the areas
tested by OCT and the 10-2 Humphrey visual field.
What could be the cause of the decreases in mfERG
response amplitudes? Although the nerve fiber layer
and ganglion cell layer are closest to the ILM and
more likely to be affected by ILM peeling than other
retinal structures, damage to these layers is unlikely to
result in a decrease in mfERG responses. There is
evidence that the standard human mfERG response is
largely shaped by cells of the outer retina, the on and
off bipolar cells and the photoreceptors,8 and that

Fig. 5. A, Electron microscopy (EM) of peeled sample from an eye

with macular pucker. The sample contains internal limiting membrane
(ILM), Müller cell foot end plate (M), collagen fibers, and fibroblast-
like cells (FLC). B, Higher-magnification EM showing a Müller cell
end plate interdigitating with ILM and transport vesicles (small ar-
rows). C, Higher-magnification EM showing collagen fibers and part of
an FLC nucleus and cytoplasm.
Fig. 6. Electron microscopy (EM) of peeled sample from an eye with
macular pucker. A, The sample included internal limiting membrane
(ILM) and Müller cell end plate (M) on the retinal side. No collagen or
cells were present on the vitreal side. B, Higher-magnification EM of a
Müller cell end plate.
572 RETINA, THE JOURNAL OF RETINAL AND VITREOUS DISEASES
removal of ganglion cell activity has little effect on
amplitude.7,9
According to our electron microscopy findings, the
four samples of peeled ILM included ILM, collagen
fibers, Müller cell footplate processes, as well as oc-
casional astrocytes and macrophage-like cells. These
findings are consistent with the results of other inves-
tigators.10 –13 The Müller cell plays a key role in the
maintenance of retinal homeostasis and functional-
ity,14 –16 and hence dysfunction of this vital cell could
result in modulating the synaptic transmission in the
retinal neural circuitry15 and affect mfERG response
amplitudes.
Terasaki et al17 found that focal electroretinography
amplitudes were decreased after vitrectomy with ILM
peeling rather than vitrectomy without ILM peeling.
They attributed these results to “some dysfunction or
physiologic changes in the Müller cells.” Another
possible explanation is based on our finding that the
retinal thickness measurements, although improved,
remained increased in comparison with those of the
fellow eye, suggesting the presence of residual retinal
anatomical abnormalities. This may be due to a gliotic
reaction resulting from the long-standing ERM or to
residual edema.
In conclusion, the use of a combination of structural
and functional tests showed improvement in visual
acuity, foveal sensitivity, central retinal thickness, and
macular volume. However, the mfERG showed re-
gions of slightly decreased response amplitudes. The
exact cause of these subtle changes is not fully under-
stood, but we suggest that dysfunction of Müller cells
might play a role. We also suggest that assessment of
the outcome of ILM peeling should, in addition to the
measurement of visual acuity and analysis of recur-
rence rates, include measurements of macular function
and structure. These measures not only will provide
better assessment of the surgical outcome but also will
improve our understanding of the changes that may
occur after ILM peeling.
Key words: internal limiting membrane peeling,
macular pucker, Müller cells, multifocal electroreti-
nography, optical coherence tomography, standard au-
tomated achromatic perimetry.
● 2007 ● VOLUME 27 ● NUMBER 5
References
1. Bovey EH, Uffer S, Achache F. Surgery for epimacular
membrane: impact of retinal internal limiting membrane re-
moval on functional outcome. Retina 2004;24:728–735.
2. Gaudric A, Fardeau C, Goberville M, et al. Ablation of the
internal limiting membrane, macular unfolding and visual
outcome in surgery of idiopathic epimacular membranes. J Fr
Ophtalmol 1993;16:571–576.
3. Smiddy WE, Maguire AM, Green WR, et al. Idiopathic
epiretinal membranes. Ultrastructural characteristics and clin-
icopathologic correlation. Ophthalmology 1989;96:811–821.
4. Park DW, Dugel PU, Garda J, et al. Macular pucker removal
with and without internal limiting membrane peeling: pilot
study. Ophthalmology 2003;110:62–64.
5. Sivalingam A, Eagle RC Jr, Duker JS, et al. Visual prognosis
correlated with the presence of internal-limiting membrane in
histopathologic specimens obtained from epiretinal mem-
brane surgery. Ophthalmology 1990;97:1549–1552.
6. Hood DC, Holopigian K, Greenstein V, et al. Assessment of
local retinal function in patients with retinitis pigmentosa
using the multi-focal ERG technique. Vision Res 1998;38:
163–179.
7. Hood DC. Assessing retinal function with the multifocal
technique. Prog Retin Eye Res 2000;19:607–646.
8. Hood DC, Seiple W, Holopigian K, Greenstein V. A com-
parison of the components of the multifocal and full-field
ERGs. Vis Neurosci 1997;14:533–544.
9. Hood DC, Frishman LJ, Saszik S, Viswanathan S. Retinal
origins of the primate multifocal ERG: implications for the
human response. Invest Ophthalmol Vis Sci 2002;43:1673–
1685.
10. Rezende FA, Kapusta MA. Internal limiting membrane: ul-
trastructural relationships, with clinical implications for mac-
ular hole healing. Can J Ophthalmol 2004;39:251–259.
11. Fekrat S, Wendel RT, de la Cruz Z, Green WR. Clinicopath-
ologic correlation of an epiretinal membrane associated with
a recurrent macular hole. Retina 1995;15:53–57.
12. Messmer EM, Heidenkummer HP, Kampik A. Ultrastructure
of epiretinal membranes associated with macular holes.
Graefes Arch Clin Exp Ophthalmol 1998;236:248–254.
13. Wolf S, Schnurbusch U, Wiedemann P, et al. Peeling of the
basal membrane in the human retina: ultrastructural effects.
Ophthalmology 2004;111:238–243.
14. Newman EA, Frambach DA, Odette LL. Control of extracel-
lular potassium levels by retinal glial cell K⫹ siphoning.
Science 1984;225:1174–1175.
15. Newman EA. Glial modulation of synaptic transmission in
the retina. Glia 2004;47:268–274.
16. Newman EA. Acid efflux from retinal glial cells generated by
sodium bicarbonate cotransport. J Neurosci 1996;16:159–
168.
17. Terasaki H, Miyake Y, Nomura R, et al. Focal macular ERGs
in eyes after removal of macular ILM during macular hole
surgery. Invest Ophthalmol Vis Sci 2001;42:229–234.