1719
Literature Search
We searched the MEDLINE, EMBASE, and Cochrane Central Register of Controlled Trials databases systematically with no language restrictions
from inception until May 2014. Two independent
reviewers (T.U. and T.N.) performed the internetbased searches. The key terms used for the systematic
search were macular edema [MeSH], internal
limiting membrane, inner limiting membrane,
diabetic retinopathy [MeSH], and vitrectomy
[MeSH]. For each key term, we searched using commands of /exp and adj6 to include as many relevant articles as possible. We collected reference lists
of original studies manually, and review articles were
1721
29
Design
PC
15
15
41
17
308
178
66
62
19
20
Bahadir et al30
PC
Kumagai et al31,32
RO
RO
Hoerauf et al35
RCT
13
12
NR
11
10
NR
446
NR
NR
29
28
19
20
17
19
PPV alone
PPV + ILM peeling
PPV alone
PPV + ILM peeling
PPV alone
PPV + ILM peeling
PPV alone
PPV + ILM peeling
PPV alone
PPV + ILM peeling
10
10
12
12
12
12
12
12
6
6
NR, not reported; PC, prospective comparative; PPV, pars plana vitrectomy; RO, retrospective observational.
Sensitivity Analysis
Applying the leave-one-out method did not change
the statistical signicance of the meta-analysis,
except for the following: when one retrospective
study33,34 was removed, the mean difference in
BCVA improvement before and after surgery became
statistically signicant in the favor of ILM peeling
(0.06 [95% CI: 0.010.11] by logMAR, equivalent
to 2 ETDRS letters, P = 0.03).
Discussion
The role of ILM peeling in the treatment of DME is
controversial, and sufcient data to clarify its role have
not yet been available. The results of the present metaanalysis showed no difference in visual acuity outcomes and reduction of macular edema by optical
Fig. 2. Comparison of baseline visual acuity (VA) and CMT between patients who underwent vitrectomy with ILM peeling (ILM peeling group) and
without ILM peeling (ILM nonpeeling group). Mean and SD values were presented by logMAR and micrometers for VA and CMT, respectively. Metaanalysis was performed for the mean difference between the two groups using inverse variance method in random effects.
1723
Fig. 3. Comparison of postoperative visual acuity (VA) and its change after surgery between patients who underwent vitrectomy with ILM peeling
(ILM peeling group) and without ILM peeling (ILM nonpeeling group). Mean and SD values were presented by logMAR. Meta-analysis was performed
for the mean difference between the two groups using inverse variance method in random effects.
Fig. 4. Comparison of CMT and its change after surgery between patients who underwent vitrectomy with ILM peeling (ILM peeling group) and
without ILM peeling (ILM nonpeeling group). Mean and SD values were presented by micrometers. Meta-analysis was performed for the mean
difference between the two groups using inverse variance method in random effects.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
1725