ORIGINAL ARTICLE
Vitrectomy for diabetic macular edema: a systematic review and meta-analysis Matthew P. Simunovic, MB BChir, PhD,*,† Alex P. Hunyor, MB, BS, FRANZCO,†,‡,§ I-Van Ho, MB, BS, MPH, PhD, FRANZCO†,‡,§ ABSTRACT ● RÉSUMÉ Objective: To systematically review, and perform meta-analysis on, the available data regarding the efficacy of vitrectomy for diabetic macular edema. Design: Systematic review and meta-analysis of published randomized controlled trial data. Methods: We searched PubMed and the Cochrane database for randomized, controlled trials investigating vitrectomy for diabetic macular edema. Structural (foveal thickness) and functional (visual acuity) outcomes were used as the primary outcome measures. Results: Eleven studies met the criteria for inclusion in this review: these studies were heterogenous in their experimental and control interventions, follow-up period, and eligibility criteria. Seven studies compared vitrectomy with the natural history of diabetic maculopathy, with laser, or with intravitreal corticosteroid injection. Four studies compared vitrectomy with internal limiting membrane peeling to vitrectomy alone. One of the latter 4 studies was the only to investigate vitrectomy in patients with vitreomacular traction. Meta-analysis suggests a structural, and possibly functional, superiority of vitrectomy over observation at 6 months. Vitrectomy also appears superior to laser in terms of structural, but not functional, outcomes at 6 months. At 12 months, vitrectomy offers no structural benefit and a trend toward inferior functional outcomes when compared with laser. Conclusions: There is little evidence to support vitrectomy as an intervention for diabetic macular edema in the absence of epiretinal membrane or vitreomacular traction. Although vitrectomy appears to be superior to laser in its effects on retinal structure at 6 months, no such benefit has been proved at 12 months. Furthermore, there is no evidence to suggest a superiority of vitrectomy over laser in terms of functional outcomes. Objet : Revue systématique et méta-analyse des données disponibles concernant l'efficacité de la vitrectomie pour l'œdème maculaire diabétique. Méthodes : Nous avons cherché dans PubMed et la base de données Cochrane les essais aléatoires contrôlés d'investigation concernant la vitrectomie pour traiter l'œdème maculaire diabétique. Les résultats structurels (épaisseur de la fovéa) et fonctionnels (acuité visuelle) ont servi de mesures des premiers résultats. Résultats : Onze études ont satisfait aux critères d'inclusion dans cette revue : ces études étaient hétérogènes dans leurs interventions d'expérience et de contrôle, leur période de suivi et leurs critères d'éligibilité. Sept études ont comparé la vitrectomie à l'histoire naturelle de la maculopathie diabétique, au laser ou à l'injection de corticostéroïde intravitréenne. Quatre études ont comparé la vitrectomie au pelage de la membrane limitante interne avec une vitrectomie seule. Une des ces quatre études était la seule à examiner la vitrectomie chez les patients ayant une traction vitréo-maculaire. La méta-analyse suggère une supériorité structurelle - et possiblement fonctionnelle - de la vitrectomie sur l'observation après six mois. La vitrectomie semble aussi supérieure au laser quant aux résultats structurels, mais non fonctionnels, à 6 mois. À 12 mois, la vitrectomie n'offre aucun bénéfice structurel et démontre une tendance vers les résultats fonctionnels inférieurs, comparativement au laser. Conclusions : Peu d'évidence soutien la vitrectomie comme intervention pour l'œdème maculaire diabétique en l'absence de membrane épirétinienne ou de traction vitréo-maculaire. Bien que la vitrectomie semble être supérieure au laser dans ses effets sur la structure rétinienne à six mois, aucun de ces avantages n'a été prouvé après 12 mois. En outre, rien ne permet de suggérer une supériorité de la vitrectomie sur le laser quant aux résultats fonctionnels.
Diabetic retinopathy is the leading cause of vision loss in those of working age, and diabetic maculopathy is the commonest mechanism via which visual impairment occurs.1 Currently, a number of ophthalmic treatment paradigms may be used to treat diabetic macular edema (DME), including focal and grid laser,2 intravitreal
injection of either corticosteroid3 or anti–vascular endothelial growth factor (anti-VEGF) agents,4 and vitrectomy with, or without, peeling of the internal limiting membrane (ILM).1,5 Trans pars plana vitrectomy (TPPV) was first postulated as a treatment for DME more than 2 decades ago.6 The
From the *Department of Ophthalmology and Visual Science, University of British Columbia, Eye Care Centre, Vancouver General Hospital, Vancouver, B.C.; †Save Sight Institute, Sydney Medical School, University of Sydney; ‡Sydney Eye Hospital, Sydney and §Australian School of Advanced Medicine, Macquarie University, Macquarie Park, New South Wales, Australia
Can J Ophthalmol 2014;49:188–195 0008-4182/14/$-see front matter & 2014 Canadian Ophthalmological Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcjo.2013.11.012
Originally received Jul. 29, 2013. Final revision Oct. 6, 2013. Accepted Nov. 14, 2013 Address correspondence to Matthew P. Simunovic, MB BChir, PhD, Department of Ophthalmology and Visual Science, University of British Columbia, Eye Care Centre, Vancouver General Hospital, 2550 Willow Street, Vancouver BC V5Z 3N9; [email protected]
188
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
Vitrectomy for diabetic macular edema—Simunovic et al. seminal report by Lewis and colleagues6 examined the effects of vitrectomy in patients with diabetes with a clinical appearance that they described as a “thickened taut glistening posterior hyaloid.” Nine of the 10 patients they treated had improved vision after vitrectomy. The subsequent development of optical coherence tomography (OCT) revealed that the pathologic correlate of a thickened taut glistening posterior hyaloid is partial posterior vitreous detachment (PVD) with focal attachment at the fovea,7,8 which may sometimes be associated with a shallow macular detachment.8 Observational and natural history studies have similarly suggested a role for the separation of the vitreous from the retina in the amelioration of DME. For example, Nasrallah and colleagues9 observed that patients with PVD have approximately half the incidence of DME when compared to those with an intact posterior vitreous, whereas Hikichi et al.10 found that spontaneous resolution of diabetic macula edema over a 6-month period was approximately twice as likely in patients with PVD when compared to those without PVD. The means via which vitrectomy might improve DME are unknown.5 Postulated mechanisms include a reduction in tangential and anterior-posterior traction at the macula,6,11–13 increased oxygenation of the vitreous cavity,5 and improved diffusion of vasogenic growth factors.5 Clearly, characterizing the status of the presurgical vitreous is an important factor in any study examining the effects of vitrectomy in DME.5,9,10 Other important factors that may modulate the response to surgery include lens status, previous macular laser, and the presence of macular ischemia. Surgical intervention for DME is generally given consideration when patients have not responded successfully to more conservative interventions such as laser and intravitreal therapy. Typically, such patients suffer from diffuse DME with, or without, concomitant vitreomacular traction (VMT). Although the uncontrolled literature hints at a promising role for vitrectomy in treating diabetic maculopathy,5 its efficacy has not been proved in a single
large prospective, randomized, controlled trial. Several small prospective, randomized, controlled trials, however, have been performed to address this issue. Our aim was to systematically review the randomized, controlled trial literature on pars plana vitrectomy for the treatment of DME and, where possible, to perform meta-analysis of structural and functional outcomes.
METHODS We performed a literature search using PubMed and the Cochrane database using the keywords “diabetic macular (o)edema,” “macular (o)edema,” and “vitrectomy.” Studies were limited to randomized, controlled trials. Each study was evaluated using the criteria specified in the Delphi list of quality assessment (Table 1).14 Meta-analysis was conducted via the RevMan 5 software package15; in the analysis of data, we performed separate meta-analyses for different treatment and/or control interventions, which were heterogeneous (Table 2). Structural outcomes were determined by computing mean differences for foveal thickness assessed by OCT using inverse variance analysis and a random-effects model with 95% confidence intervals. Functional outcomes were analyzed by computing odds ratios for achieving an improvement or loss of Z2 logMAR lines (0.2 logMAR units) using the Mantel– Haenszel method with random effects and 95% confidence intervals. In the event of missing data for the outcome measures defined earlier, authors were contacted directly. If the data remained unobtainable, standard deviations for change in foveal thickness were reverse calculated from statistical estimates of probability16 or other summary statistics.17 In the event that no such probabilities or summary statistics were given for a comparison or measure,18 standard deviations were imputed (using the most conservative estimates from similar studies). Because no data regarding likelihood of achieving a Z2 line
Table 1—Quality of evidence of included studies (assessment based on the Delphi list for quality assessment of randomized, controlled trials)14
Study Stolba et al.16 Yanyali et al.19 Kumar et al.21 Yanyali et al.20 Patel et al.18 Thomas et al.17 Hoerauf et al.22 Bahadir et al.24 Bardak et al.23 Doi et al.25 Figueroa et al.26
Allocation Randomized Concealed Yes Yes Yes Yes Yes Yes Yes† Yes Yes Yes Yes‡
NA NA NA NA NA NA Yes Yes NK NA NK
Baseline Group Similarity
Eligibility Specified
Blinded Outcome Assessor
Care Provider Blinded
Patient Blinded
Point Estimates and Measure of Variability Presented
No Yes Yes Yes Yes Yes* Yes Yes NK Yes Yes‡
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Yes Yes Yes Yes Yes NK Yes NK NK NK NK
NA NA NA NA NA NA NA NA NA NA NK
NA NA NA NA NA NA NK NK NK NA NK
No Yes Yes Yes No Yes Yes Yes Yes Yes Yes
Intention to Treat Score Yes Yes Yes Yes NK Yes NK NK NK Yes NK
4 6 6 6 4 5 6 5 3 5 4
NK, not known. *Differences in baseline characteristics controlled for in analysis. †Patients with pre-existing posterior vitreous detachment not randomized, but not included in meta-analysis. ‡Patients with glaucoma did not receive intravitreal triamcinolone (IVTA). All other patients were randomized to trans pars plana vitrectomy (TPPV)/IVTA or TPPV/internal limiting membrane peeling/IVTA.
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
189
Vitrectomy for diabetic macular edema—Simunovic et al. Table 2—Summary of randomized, controlled trials investigating vitrectomy for diabetic macular edema No. of Eyes
Study Stolba et al.16 Yanyali et al.19 Kumar et al.21
Yanyali et al.20
Patient Group
56
Diffuse DME without VMT/PVD or ischemia 20 (paired) Diffuse DME without VMT/PVD or ischemia 24 Diffuse DME without VMT/PVD or ischemia and VA (r6/60) 24 (paired) Diffuse DME without VMT/PVD or ischemia
Patel et al.18
15
Thomas et al.17
40
Hoerauf et al.22
40
Bahadir et al.24
58
Bardak et al.23
17
Doi et al.25
40
Figueroa et al.26 42
Intervention TPPV
TPPV/ILM peel with ICG 0.1% TPPV/ILM peeling with trypan blue TPPV/ILM peeling with ICG 0.1% DME resistant to laser TPPV without VMT/PVD or ischemia and VA 6/15 to 6/60 TTPV/ILM Diffuse DME without VMT/PVD or ischemia peeling with ICG 0.05% and VA r6/12 Diffuse DME without TPPV/ILM VMT or ischemia peeling with trypan blue or ICG 0.05% Thickened posterior TPPV/ILM hyaloid with DME peeling with ICG 0.25% Diffuse DME without TPPV/ILM PVD peeling with ICG 0.25% or triamcinolone (paired) Diffuse DME without TPPV VMT/PVD or ischemia Diffuse DME without VMT or ischemia
TPPV/IVTA/ILM peeling with trypan blue
Control
Follow-up
Observed
6 months
Observed
12 months No difference in VA between groups 6 months No difference in VA between groups
Grid laser
Functional Outcomes Improved VA with TPPV
Grid laser
6 months
Grid laser
12 months Improved VA with laser at 12 months Increased perifoveal sensitivity with TPPV 12 months No difference in VA between groups
Reduction in foveal thickness with TPPV (27 μm) and laser (107 μm); no statistical comparison made No significant difference in central macular thickness
TPPV
6 months
Foveal thickness reduced after ILM peeling
TPPV
12 months No difference in VA between groups
Macular thickness not examined
6 months
Macular thickness not examined
Grid laser
TPPV with triamcinolone assistance IVTA
No difference in VA between groups
Structural Outcomes Foveal thickness reduced after TPPV Foveal thickness reduced after TPPV Foveal thickness reduced after TPPV
No difference in VA between groups
No difference in VA between groups
Foveal thickness reduced after TPPV
12 months No difference in VA between groups
TPPV/IVTA
Decreased central macular thickness with TPPV at 12 months Decreased macular 18 months No difference in VA between thickness at 6 months in groups except at month 3 (TPPV/IVTA/ILM peel superior both groups No difference in macular to TPPV/IVTA) thickness from baseline in both groups at 12 months and beyond
DME, diabetic macular edema; VMT, vitreomacular traction; PVD, posterior vitreous detachment; TPPV, trans pars plana vitrectomy; VA, visual acuity; ILM, internal limiting membrane; ICG, indocyanine green; IVTA, intravitreal triamcinolone.
improvement or loss in acuity could be obtained from 2 studies comparing TPPV with laser at 12 months, analysis of change in logMAR acuity was performed. A probability of less than 0.05 was deemed to be statistically significant.
RESULTS A total of 11 studies were identified that met the inclusion criteria for this study and are summarized in Tables 1 and 2. These studies were heterogenous in their inclusion and exclusion criteria, surgical intervention, and follow-up period. Furthermore, these studies also varied in their comparisons: 2 studies compared vitrectomy with the natural history of the disease, 4 compared vitrectomy with laser, 1 study compared vitrectomy with intravitreal steroid injection, 3 compared vitrectomy alone with combined vitrectomy with ILM peeling, and 1 investigated vitrectomy combined with ILM peeling, intravitreal steroid, or both. Outcomes of all studies included visual acuity, and the majority (9/11) also included an assessment of foveal thickness.
190
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
Studies of trans pars plana vitrectomy versus natural history
Yanyali et al.19 compared vitrectomy with the natural history of DME in 20 eyes of 10 patients who had previously undergone (Z6 months preceding the study) grid laser photocoagulation. Their study assigned TPPV randomly to either eye and specifically excluded those with features suggestive of VMT on OCT. They demonstrated that vitrectomy resulted in a significant decrease in foveal thickness at 12 months postprocedure (mean reduction of 165.8 vs 37.8 μm for the intervention and control groups, respectively). The majority of this improvement occurred within a month of surgery and was not paralleled by changes in function: no difference in visual acuity was found between treated and control eyes. Stolba and colleagues16 compared the effects of TPPV in 56 eyes with DME that had previously undergone grid laser (Z4 months preceding the study) and that also had an attached hyaloid. They found that TPPV was associated with a significant reduction in foveal thickness when compared with the natural history of the disease (60.3μm reduction in mean foveal for TPPV at 6 months vs an
Vitrectomy for diabetic macular edema—Simunovic et al.
Fig. 1 — A, Trans pars plana vitrectomy (TPPV) versus natural history at 6 months: improvement in foveal thickness (positive values better). B, TPPV versus natural history at 6 months: likelihood of achieving Z2 lines improvement in acuity. C, TPPV versus natural history at 6 months: likelihood of Z2 lines loss in acuity.
increase of 17.7 μm in the observed group). Although Stolba and colleagues16 found an improvement in visual acuity after vitrectomy, they also elected to perform cataract surgery in selected patients (those 460 years with mild cataract) at the time of surgery, thus introducing an important confounding factor. Meta-analysis of studies16,19 comparing the natural history of DME with TPPV suggests that the latter is associated with structural benefits (Fig. 1A) at 6 months. Meta-analysis also suggests a possible functional benefit of TPPV at 6 months in terms of likelihood of gaining Z2 lines in acuity (Fig. 1B), with the caveat of a confounding factor in the form of simultaneous cataract extraction in 1 of the included studies.16 There was no evidence for a decreased likelihood of visual loss of Z2 lines after TPPV (Fig. 1C). Studies of trans pars plana vitrectomy versus laser
Yanyali and colleagues’20 study compared TPPV with ILM peeling to grid laser in 24 eyes of 12 patients with bilateral diffuse DME who had not previously undergone macular laser. Patients with features of VMT were specifically excluded. They demonstrated that TPPV/ ILM peeling results in superior structural improvements
over laser (foveal thickness reduced by 219.4 vs 28.5 μm, respectively). Mean improvement in logMAR acuity (0.22 vs 0.1 for TPPV and laser, respectively) demonstrated a trend toward superior outcomes after TPPV/ ILM peeling, although this failed to reach statistical significance. Kumar and colleagues21 also compared TPPV/ILM peeling with grid laser in 24 eyes with DME resulting in severe visual impairment (visual acuity r6/60). Patients who had undergone macular laser within the preceding 6 months and those with VMT were specifically excluded. Postprocedural changes in foveal thickness suggested a significant benefit of TPPV/ ILM peeling (300.9-μm decrease at 6 months) over grid laser (106.6-μm decrease at 6 months). No significant difference in visual acuity was found between the 2 interventions. In a small pilot study of 15 patients that was directed at determining the number of subjects required to definitively compare TPPV with laser, Patel et al.18 found a slight, but statistically significant, superiority of laser over TPPV in terms of acuity. Although both TPPV and laser were associated with an improvement in macular thickness, the effect was greater for laser (27 and 107 μm, respectively). Notably, however, they found that patients undergoing TPPV had improved perifoveal CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
191
Vitrectomy for diabetic macular edema—Simunovic et al.
Fig. 2 — A, Trans pars plana vitrectomy (TPPV) versus laser. Structural (reduction in foveal thickness) outcomes at 6 months (positive values better). B, TPPV versus laser. Functional (likelihood of achieving Z2 lines improvement in acuity) outcomes at 6 months.
microperimetric sensitivities postprocedure, whereas laser patients had decreased sensitivities. One patient in the TPPV group suffered a loss of vision of 0.3 logMAR units at 12 months. Thomas and colleagues’17 feasibility study of TPPV/ILM peeling versus laser recruited 40 patients with DME and at least 1 previous episode of grid laser (Z3 months preceding recruitment). They found a mean deterioration in acuity of 0.05 logMAR at 12 months after TPPV/ILM peeling, whereas repeat laser was associated with an improvement of 0.03 logMAR (no significant difference). Although TPPV/ILM peeling was associated with improved structural outcomes over laser at 12 months (73-μm decrease vs 29-μm decrease in central macular thickness), this also failed to reach statistical significance.
Meta-analysis of studies comparing TPPV with laser suggests that TPPV affords a structural, but not a functional, benefit over laser (see Fig. 2) at 6 months after surgery. At 12 months, there was no evidence for a structural benefit of TPPV over laser and a trend toward inferior functional outcomes was observed, although this failed to reach statistical significance (Fig. 3). Only 1 patient undergoing TPPV was reported with a loss of vision of Z0.2 logMAR unit (no such losses were reported after laser; meta-analysis was not performed).18 Trans pars plana vitrectomy versus trans pars plana vitrectomy with internal limiting membrane peeling
Three studies have compared TPPV with ILM peeling with TPPV alone. Hoerauf and colleagues’22 study of 40
Fig. 3 — A, Trans pars plana vitrectomy (TPPV) versus laser. Structural (reduction in foveal thickness) outcomes at 12 months (positive values better). B, TPPV versus laser. Functional (baseline – final logMAR acuity) outcomes at 12 months.
192
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
Vitrectomy for diabetic macular edema—Simunovic et al.
Fig. 4 — Trans pars plana vitrectomy (TPPV) with internal limiting membrane peeling versus TPPV. Likelihood of an improvement in acuity by Z2 lines.
eyes with diffuse DME did not suggest a superiority of 1 surgical intervention over the other in terms of visual acuity, although they found that TPPV/ILM peeling resulted in significantly greater resolution in macular edema at 3 and 6 months. Bardak and colleagues’23 small study also found no evidence for a functional superiority of TPPV/ ILM peeling over TPPV alone; they did not compare structural outcomes. Bahadir and colleagues24 compared TPPV/ILM peeling with TPPV alone in patients with a “thickened posterior hyaloid” on clinical and retinal thickness analyzer examination. They similarly found no evidence for improved visual outcomes with the addition of ILM peeling to TPPV (structural outcomes not assessed). Only 2 of 3 studies were included in meta-analysis because they both excluded patients with VMT, whereas the third specifically investigated such patients. There was no evidence for a difference in visual acuity outcomes between TPPV/ILM peeling and TPPV alone (Fig. 4). Because only 1 study comparing TPPV/ILM peeling with TPPV alone included foveal thickness data, meta-analysis of this outcome could not be performed. Trans pars plana vitrectomy with internal limiting membrane peeling versus intravitreal steroid
One study compared vitrectomy with intravitreal corticosteroid therapy: Doi and colleagues25 evaluated TPPV/ ILM peeling against intravitreal triamcinolone (IVTA) over a 12-month period. Although their study protocol allowed re-treatment after 6 months and permitted cataract extraction within the study period, no eyes underwent either of these additional interventions. In terms of reducing foveal thickness, their results suggest an early superiority of IVTA (at 1 month) and a later superiority of TPPV/ILM peeling (at 12 months). No significant difference was found between the 2 interventions in terms of visual acuity. Trans pars plana vitrectomy/intravitreal triamcinolone versus trans pars plana vitrectomy/internal limiting membrane peeling versus trans pars plana vitrectomy/internal limiting membrane peeling/ intravitreal triamcinolone
Figueroa et al.26 investigated the effects of TPPV combined with ILM peeling, IVTA, or both. Only patients in the latter 2 arms (n ¼ 31) were randomized, as those with a prior history of glaucoma were exclusively
allocated to the first group (n ¼ 11). Nevertheless, they found no significant difference between the 2 randomized interventions in terms of visual acuity, except at 3 months (combined TPPV/ILM peeling/IVTA superior to TPPV/ IVTA); furthermore, they did not find a statistically significant improvement in visual acuity from baseline for either intervention. All 3 groups of patients demonstrated an average improvement in macular thickness at 6 months, which subsequently regressed to baseline by 12 months.
DISCUSSION There is concordance between 2 prospective, randomized, controlled studies comparing vitrectomy with the natural history of DME that TPPV is associated with improvements in macular structure. Meta-analysis also supports the assertion that TPPV is associated with an increased likelihood of achieving a Z2 line improvement in acuity, although these results may be confounded by the methodology used in 1 of the studies, which permitted simultaneous cataract extraction as part of the primary procedure in some of the experimental subjects.16 Studies comparing vitrectomy with laser have produced mixed results. Meta-analysis of studies including 6-month follow-up suggests a superiority of TPPV over laser for improvement of foveal thickness, whereas meta-analysis of studies including a 12-month follow-up suggests no such benefit at this later time point. Meta-analysis of visual acuity outcomes does not support a superiority of TPPV over laser at 6 months and suggested a trend toward inferior outcomes after TPPV at 12 months. In terms of surgical interventions, the evidence suggests that the addition of ILM peeling may afford improved structural, but not functional, outcomes over vitrectomy alone.22 It is also important to note that several of the studies included in this systematic review used indocyanine green (ICG) to aid ILM peeling. This agent has subsequently been implicated in retinal toxicity and could have theoretically had a negative effect on treatment outcomes when used,27–31 especially at concentrations of Z0.1%. However, it should also be noted that studies that included the use of ICG staining generally found in favour of TTPV/ILM peeling (see Table 2). CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
193
Vitrectomy for diabetic macular edema—Simunovic et al. The management of patients with isolated DME unresponsive to laser or intravitreal therapy remains a clinical dilemma. At this point in time, there is little evidence to support vitrectomy as an alternative treatment. Although some studies suggest a possible benefit of TPPV with, or without, ILM peeling over either the natural history of the disease or over grid laser, others fail to reveal a benefit. Part of the problem in assessing the studies included in this systematic review is their heterogeneity in control interventions and follow-up periods, which necessitated separate meta-analyses and also prevented formal assessment of publication bias. These issues are compounded by the fact that all studies to date have included relatively small sample sizes. By way of comparison, intravitreal anti-VEGF has recently been shown to be superior to laser in diffuse DME; however, not one of the studies included in this review would be powered to detect the difference in the proportion of patients achieving an improvement in acuity of Z2 lines demonstrated in the RESTORE study (N ¼ 126, assuming a power of 0.80 and an α ¼ 0.05).32 Nevertheless, meta-analysis supports the assertion that vitrectomy results in superior early (i.e., 6-month) structural outcomes when compared with observation and grid laser. However, conclusive evidence for a functional benefit is lacking. Although meta-analysis of 2 randomized, controlled trials would suggest a possible functional improvement over the natural history of the disease, 1 of these studies introduced an important confounding factor of simultaneous cataract extraction in a subgroup of TPPV patients. Although a trend was observed toward improved functional outcomes over grid laser at 6 months, this failed to reach statistical significance; furthermore, there was a trend toward inferior functional outcomes compared with laser at 12 months. One particular difficulty in using conventional acuity measures in the assessment of post-TPPV retinal function in phakic eyes is that vitrectomy is cataractogenic. Thus, it could be argued that improvements in retinal function after TPPV, as assessed by conventional measures of acuity, may have been obscured by postoperative lenticular changes observed in some of the included studies (see Table 2). One of the possible mechanisms by which vitrectomy may improve retinal structure and function in patients with DME is via relief of tangential and/or anterior-posterior macular traction. Some of the nonrandomized control trial literature suggests that those with VMT, including the socalled thickened taut glistening posterior hyaloid, may derive a particular benefit from surgical intervention.17 However, such patients have been included in only 1 randomized, controlled trial to date (which compared vitrectomy alone with vitrectomy with ILM peeling).24 A further issue is that since the publication of the clinical trials included in this systematic review, a new gold standard treatment for diffuse DME has emerged: intravitreal antiVEGF therapy.32 This fact, combined with observation that
194
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
vitrectomy does not appear to afford improved functional outcomes over laser in patients with DME, suggests that there is no compelling argument for vitrectomy as a treatment for DME in the absence of VMT. However, given the suggestion that those with various forms of concomitant VMT may derive a particular benefit from vitrectomy and ILM peeling,33 this patient group may make a particularly suitable population to study prospectively in a randomized, controlled trial comparing anti-VEGF agents with vitrectomy.
Disclosure: The authors have no proprietary or commercial interest in any materials discussed in this article.
REFERENCES 1. O’Doherty M, Dooley I, Hickey-Dwyer M. Interventions for diabetic macular oedema: a systematic review of the literature. Br J Ophthalmol. 2008;92:1581-90. 2. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study report number 1. Early Treatment Diabetic Retinopathy Study research group. Arch Ophthalmol. 1985;103:1796-806. 3. Lam DS, Chan CK, Mohamed S, et al. Intravitreal triamcinolone plus sequential grid laser versus triamcinolone or laser alone for treating diabetic macular edema: six-month outcomes. Ophthalmology. 2007;114:2162-7. 4. Scott IU, Edwards AR, Beck RW, et al. A phase II randomized clinical trial of intravitreal bevacizumab for diabetic macular edema. Ophthalmology. 2007;114:1860-7. 5. Laidlaw DA. Vitrectomy for diabetic macular oedema. Eye (Lond). 2008;22:1337-41. 6. Lewis H, Abrams GW, Blumenkranz MS, Campo RV. Vitrectomy for diabetic macular traction and edema associated with posterior hyaloidal traction. Ophthalmology. 1992;99:753-9. 7. Massin P, Duguid G, Erginay A, Haouchine B, Gaudric A. Optical coherence tomography for evaluating diabetic macular edema before and after vitrectomy. Am J Ophthalmol. 2003;135:169-77. 8. Kaiser PK, Riemann CD, Sears JE, Lewis H. Macular traction detachment and diabetic macular edema associated with posterior hyaloidal traction. Am J Ophthalmol. 2001;131:44-9. 9. Nasrallah FP, Jalkh AE, Van Coppenolle F, et al. The role of the vitreous in diabetic macular edema. Ophthalmology. 1988;95: 1335-9. 10. Hikichi T, Fujio N, Akiba J, Azuma Y, Takahashi M, Yoshida A. Association between the short-term natural history of diabetic macular edema and the vitreomacular relationship in type II diabetes mellitus. Ophthalmology. 1997;104:473-8. 11. Harbour JW, Smiddy WE, Flynn HW Jr, Rubsamen PE. Vitrectomy for diabetic macular edema associated with a thickened and taut posterior hyaloid membrane. Am J Ophthalmol. 1996;121: 405-13. 12. Pendergast SD, Hassan TS, Williams GA, et al. Vitrectomy for diffuse diabetic macular edema associated with a taut premacular posterior hyaloid. Am J Ophthalmol. 2000;130:178-86. 13. Recchia FM, Ruby AJ, Carvalho Recchia CA. Pars plana vitrectomy with removal of the internal limiting membrane in the treatment of persistent diabetic macular edema. Am J Ophthalmol. 2005;139: 447-54. 14. Verhagen AP, de Vet HC, de Bie RA, et al. The Delphi list: a criteria list for quality assessment of randomized clinical trials for conducting systematic reviews developed by Delphi consensus. J Clin Epidemiol. 1998;51:1235-41. 15. Rev Man. Version 5.1 [computer program]. Copenhagen: The Cochrane Collaboration; 2011. 16. Stolba U, Binder S, Gruber D, Krebs I, Aggermann T, Neumaier B. Vitrectomy for persistent diffuse diabetic macular edema. Am J Ophthalmol. 2005;140:295-301.
Vitrectomy for diabetic macular edema—Simunovic et al. 17. Thomas D, Bunce C, Moorman C, Laidlaw DA. A randomised controlled feasibility trial of vitrectomy versus laser for diabetic macular oedema. Br J Ophthalmol. 2005;89:81-6. 18. Patel JI, Hykin PG, Schadt M, et al. Diabetic macular oedema: pilot randomised trial of pars plana vitrectomy vs macular argon photocoagulation. Eye (Lond). 2006;20:873-81. 19. Yanyali A, Horozoglu F, Celik E, Ercalik Y, Nohutcu AF. Pars plana vitrectomy and removal of the internal limiting membrane in diabetic macular edema unresponsive to grid laser photocoagulation. Eur J Ophthalmol. 2006;16:573-81. 20. Yanyali A, Nohutcu AF, Horozoglu F, Celik E. Modified grid laser photocoagulation versus pars plana vitrectomy with internal limiting membrane removal in diabetic macular edema. Am J Ophthalmol. 2005;139:795-801. 21. Kumar A, Sinha S, Azad R, Sharma YR, Vohra R. Comparative evaluation of vitrectomy and dye-enhanced ILM peel with grid laser in diffuse diabetic macular edema. Graefes Arch Clin Exp Ophthalmol. 2007;245:360-8. 22. Hoerauf H, Bruggemann A, Muecke M, et al. Pars plana vitrectomy for diabetic macular edema. Internal limiting membrane delamination vs posterior hyaloid removal. A prospective randomized trial. Graefes Arch Clin Exp Ophthalmol. 2011;249:997-1008. 23. Bardak Y, Cekic O, Tig SU. Comparison of ICG-assisted ILM peeling and triamcinolone-assisted posterior vitreous removal in diffuse diabetic macular oedema. Eye (Lond). 2006;20:1357-9. 24. Bahadir M, Ertan A, Mertoglu O. Visual acuity comparison of vitrectomy with and without internal limiting membrane removal in the treatment of diabetic macular edema. Int Ophthalmol. 2005;26(1-2):3-8.
25. Doi N, Sakamoto T, Sonoda Y, et al. Comparative study of vitrectomy versus intravitreous triamcinolone for diabetic macular edema on randomized paired-eyes. Graefes Arch Clin Exp Ophthalmol. 2012;250:71-8. 26. Figueroa MS, Contreras I, Noval S. Surgical and anatomical outcomes of pars plana vitrectomy for diffuse nontractional diabetic macular edema. Retina. 2008;28:420-6. 27. Gale JS, Proulx AA, Gonder JR, Mao AJ, Hutnik CM. Comparison of the in vitro toxicity of indocyanine green to that of trypan blue in human retinal pigment epithelium cell cultures. Am J Ophthalmol. 2004;138:64-9. 28. Goldstein M, Zemel E, Loewenstein A, Perlman I. Retinal toxicity of indocyanine green in albino rabbits. Invest Ophthalmol Vis Sci. 2006;47:2100-7. 29. Narayanan R, Kenney MC, Kamjoo S, et al. Toxicity of indocyanine green (ICG) in combination with light on retinal pigment epithelial cells and neurosensory retinal cells. Curr Eye Res. 2005;30:471-8. 30. Querques G, Prascina F, Iaculli C, Noci ND. Retinal toxicity of indocyanine green. Int Ophthalmol. 2008;28:115-8. 31. Tokuda K, Zorumski CF, Izumi Y. Involvement of illumination in indocyanine green toxicity after its washout in the ex vivo rat retina. Retina. 2009;29:371-9. 32. Mitchell P, Bandello F, Schmidt-Erfurth U, et al. The RESTORE study: ranibizumab monotherapy or combined with laser versus laser monotherapy for diabetic macular edema. Ophthalmology. 2011;118: 615-25. 33. Haller JA, Qin H, Apte RS, et al. Vitrectomy outcomes in eyes with diabetic macular edema and vitreomacular traction. Ophthalmology. 2010;117:1087-93.e1083.
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
195
Vitrectomy for diabetic macular edema: a systematic review and meta-analysis Matthew P. Simunovic, MB BChir, PhD,*,† Alex P. Hunyor, MB, BS, FRANZCO,†,‡,§ I-Van Ho, MB, BS, MPH, PhD, FRANZCO†,‡,§ ABSTRACT ● RÉSUMÉ Objective: To systematically review, and perform meta-analysis on, the available data regarding the efficacy of vitrectomy for diabetic macular edema. Design: Systematic review and meta-analysis of published randomized controlled trial data. Methods: We searched PubMed and the Cochrane database for randomized, controlled trials investigating vitrectomy for diabetic macular edema. Structural (foveal thickness) and functional (visual acuity) outcomes were used as the primary outcome measures. Results: Eleven studies met the criteria for inclusion in this review: these studies were heterogenous in their experimental and control interventions, follow-up period, and eligibility criteria. Seven studies compared vitrectomy with the natural history of diabetic maculopathy, with laser, or with intravitreal corticosteroid injection. Four studies compared vitrectomy with internal limiting membrane peeling to vitrectomy alone. One of the latter 4 studies was the only to investigate vitrectomy in patients with vitreomacular traction. Meta-analysis suggests a structural, and possibly functional, superiority of vitrectomy over observation at 6 months. Vitrectomy also appears superior to laser in terms of structural, but not functional, outcomes at 6 months. At 12 months, vitrectomy offers no structural benefit and a trend toward inferior functional outcomes when compared with laser. Conclusions: There is little evidence to support vitrectomy as an intervention for diabetic macular edema in the absence of epiretinal membrane or vitreomacular traction. Although vitrectomy appears to be superior to laser in its effects on retinal structure at 6 months, no such benefit has been proved at 12 months. Furthermore, there is no evidence to suggest a superiority of vitrectomy over laser in terms of functional outcomes. Objet : Revue systématique et méta-analyse des données disponibles concernant l'efficacité de la vitrectomie pour l'œdème maculaire diabétique. Méthodes : Nous avons cherché dans PubMed et la base de données Cochrane les essais aléatoires contrôlés d'investigation concernant la vitrectomie pour traiter l'œdème maculaire diabétique. Les résultats structurels (épaisseur de la fovéa) et fonctionnels (acuité visuelle) ont servi de mesures des premiers résultats. Résultats : Onze études ont satisfait aux critères d'inclusion dans cette revue : ces études étaient hétérogènes dans leurs interventions d'expérience et de contrôle, leur période de suivi et leurs critères d'éligibilité. Sept études ont comparé la vitrectomie à l'histoire naturelle de la maculopathie diabétique, au laser ou à l'injection de corticostéroïde intravitréenne. Quatre études ont comparé la vitrectomie au pelage de la membrane limitante interne avec une vitrectomie seule. Une des ces quatre études était la seule à examiner la vitrectomie chez les patients ayant une traction vitréo-maculaire. La méta-analyse suggère une supériorité structurelle - et possiblement fonctionnelle - de la vitrectomie sur l'observation après six mois. La vitrectomie semble aussi supérieure au laser quant aux résultats structurels, mais non fonctionnels, à 6 mois. À 12 mois, la vitrectomie n'offre aucun bénéfice structurel et démontre une tendance vers les résultats fonctionnels inférieurs, comparativement au laser. Conclusions : Peu d'évidence soutien la vitrectomie comme intervention pour l'œdème maculaire diabétique en l'absence de membrane épirétinienne ou de traction vitréo-maculaire. Bien que la vitrectomie semble être supérieure au laser dans ses effets sur la structure rétinienne à six mois, aucun de ces avantages n'a été prouvé après 12 mois. En outre, rien ne permet de suggérer une supériorité de la vitrectomie sur le laser quant aux résultats fonctionnels.
Diabetic retinopathy is the leading cause of vision loss in those of working age, and diabetic maculopathy is the commonest mechanism via which visual impairment occurs.1 Currently, a number of ophthalmic treatment paradigms may be used to treat diabetic macular edema (DME), including focal and grid laser,2 intravitreal
injection of either corticosteroid3 or anti–vascular endothelial growth factor (anti-VEGF) agents,4 and vitrectomy with, or without, peeling of the internal limiting membrane (ILM).1,5 Trans pars plana vitrectomy (TPPV) was first postulated as a treatment for DME more than 2 decades ago.6 The
From the *Department of Ophthalmology and Visual Science, University of British Columbia, Eye Care Centre, Vancouver General Hospital, Vancouver, B.C.; †Save Sight Institute, Sydney Medical School, University of Sydney; ‡Sydney Eye Hospital, Sydney and §Australian School of Advanced Medicine, Macquarie University, Macquarie Park, New South Wales, Australia
Can J Ophthalmol 2014;49:188–195 0008-4182/14/$-see front matter & 2014 Canadian Ophthalmological Society. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jcjo.2013.11.012
Originally received Jul. 29, 2013. Final revision Oct. 6, 2013. Accepted Nov. 14, 2013 Address correspondence to Matthew P. Simunovic, MB BChir, PhD, Department of Ophthalmology and Visual Science, University of British Columbia, Eye Care Centre, Vancouver General Hospital, 2550 Willow Street, Vancouver BC V5Z 3N9; [email protected]
188
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
Vitrectomy for diabetic macular edema—Simunovic et al. seminal report by Lewis and colleagues6 examined the effects of vitrectomy in patients with diabetes with a clinical appearance that they described as a “thickened taut glistening posterior hyaloid.” Nine of the 10 patients they treated had improved vision after vitrectomy. The subsequent development of optical coherence tomography (OCT) revealed that the pathologic correlate of a thickened taut glistening posterior hyaloid is partial posterior vitreous detachment (PVD) with focal attachment at the fovea,7,8 which may sometimes be associated with a shallow macular detachment.8 Observational and natural history studies have similarly suggested a role for the separation of the vitreous from the retina in the amelioration of DME. For example, Nasrallah and colleagues9 observed that patients with PVD have approximately half the incidence of DME when compared to those with an intact posterior vitreous, whereas Hikichi et al.10 found that spontaneous resolution of diabetic macula edema over a 6-month period was approximately twice as likely in patients with PVD when compared to those without PVD. The means via which vitrectomy might improve DME are unknown.5 Postulated mechanisms include a reduction in tangential and anterior-posterior traction at the macula,6,11–13 increased oxygenation of the vitreous cavity,5 and improved diffusion of vasogenic growth factors.5 Clearly, characterizing the status of the presurgical vitreous is an important factor in any study examining the effects of vitrectomy in DME.5,9,10 Other important factors that may modulate the response to surgery include lens status, previous macular laser, and the presence of macular ischemia. Surgical intervention for DME is generally given consideration when patients have not responded successfully to more conservative interventions such as laser and intravitreal therapy. Typically, such patients suffer from diffuse DME with, or without, concomitant vitreomacular traction (VMT). Although the uncontrolled literature hints at a promising role for vitrectomy in treating diabetic maculopathy,5 its efficacy has not been proved in a single
large prospective, randomized, controlled trial. Several small prospective, randomized, controlled trials, however, have been performed to address this issue. Our aim was to systematically review the randomized, controlled trial literature on pars plana vitrectomy for the treatment of DME and, where possible, to perform meta-analysis of structural and functional outcomes.
METHODS We performed a literature search using PubMed and the Cochrane database using the keywords “diabetic macular (o)edema,” “macular (o)edema,” and “vitrectomy.” Studies were limited to randomized, controlled trials. Each study was evaluated using the criteria specified in the Delphi list of quality assessment (Table 1).14 Meta-analysis was conducted via the RevMan 5 software package15; in the analysis of data, we performed separate meta-analyses for different treatment and/or control interventions, which were heterogeneous (Table 2). Structural outcomes were determined by computing mean differences for foveal thickness assessed by OCT using inverse variance analysis and a random-effects model with 95% confidence intervals. Functional outcomes were analyzed by computing odds ratios for achieving an improvement or loss of Z2 logMAR lines (0.2 logMAR units) using the Mantel– Haenszel method with random effects and 95% confidence intervals. In the event of missing data for the outcome measures defined earlier, authors were contacted directly. If the data remained unobtainable, standard deviations for change in foveal thickness were reverse calculated from statistical estimates of probability16 or other summary statistics.17 In the event that no such probabilities or summary statistics were given for a comparison or measure,18 standard deviations were imputed (using the most conservative estimates from similar studies). Because no data regarding likelihood of achieving a Z2 line
Table 1—Quality of evidence of included studies (assessment based on the Delphi list for quality assessment of randomized, controlled trials)14
Study Stolba et al.16 Yanyali et al.19 Kumar et al.21 Yanyali et al.20 Patel et al.18 Thomas et al.17 Hoerauf et al.22 Bahadir et al.24 Bardak et al.23 Doi et al.25 Figueroa et al.26
Allocation Randomized Concealed Yes Yes Yes Yes Yes Yes Yes† Yes Yes Yes Yes‡
NA NA NA NA NA NA Yes Yes NK NA NK
Baseline Group Similarity
Eligibility Specified
Blinded Outcome Assessor
Care Provider Blinded
Patient Blinded
Point Estimates and Measure of Variability Presented
No Yes Yes Yes Yes Yes* Yes Yes NK Yes Yes‡
Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
Yes Yes Yes Yes Yes NK Yes NK NK NK NK
NA NA NA NA NA NA NA NA NA NA NK
NA NA NA NA NA NA NK NK NK NA NK
No Yes Yes Yes No Yes Yes Yes Yes Yes Yes
Intention to Treat Score Yes Yes Yes Yes NK Yes NK NK NK Yes NK
4 6 6 6 4 5 6 5 3 5 4
NK, not known. *Differences in baseline characteristics controlled for in analysis. †Patients with pre-existing posterior vitreous detachment not randomized, but not included in meta-analysis. ‡Patients with glaucoma did not receive intravitreal triamcinolone (IVTA). All other patients were randomized to trans pars plana vitrectomy (TPPV)/IVTA or TPPV/internal limiting membrane peeling/IVTA.
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
189
Vitrectomy for diabetic macular edema—Simunovic et al. Table 2—Summary of randomized, controlled trials investigating vitrectomy for diabetic macular edema No. of Eyes
Study Stolba et al.16 Yanyali et al.19 Kumar et al.21
Yanyali et al.20
Patient Group
56
Diffuse DME without VMT/PVD or ischemia 20 (paired) Diffuse DME without VMT/PVD or ischemia 24 Diffuse DME without VMT/PVD or ischemia and VA (r6/60) 24 (paired) Diffuse DME without VMT/PVD or ischemia
Patel et al.18
15
Thomas et al.17
40
Hoerauf et al.22
40
Bahadir et al.24
58
Bardak et al.23
17
Doi et al.25
40
Figueroa et al.26 42
Intervention TPPV
TPPV/ILM peel with ICG 0.1% TPPV/ILM peeling with trypan blue TPPV/ILM peeling with ICG 0.1% DME resistant to laser TPPV without VMT/PVD or ischemia and VA 6/15 to 6/60 TTPV/ILM Diffuse DME without VMT/PVD or ischemia peeling with ICG 0.05% and VA r6/12 Diffuse DME without TPPV/ILM VMT or ischemia peeling with trypan blue or ICG 0.05% Thickened posterior TPPV/ILM hyaloid with DME peeling with ICG 0.25% Diffuse DME without TPPV/ILM PVD peeling with ICG 0.25% or triamcinolone (paired) Diffuse DME without TPPV VMT/PVD or ischemia Diffuse DME without VMT or ischemia
TPPV/IVTA/ILM peeling with trypan blue
Control
Follow-up
Observed
6 months
Observed
12 months No difference in VA between groups 6 months No difference in VA between groups
Grid laser
Functional Outcomes Improved VA with TPPV
Grid laser
6 months
Grid laser
12 months Improved VA with laser at 12 months Increased perifoveal sensitivity with TPPV 12 months No difference in VA between groups
Reduction in foveal thickness with TPPV (27 μm) and laser (107 μm); no statistical comparison made No significant difference in central macular thickness
TPPV
6 months
Foveal thickness reduced after ILM peeling
TPPV
12 months No difference in VA between groups
Macular thickness not examined
6 months
Macular thickness not examined
Grid laser
TPPV with triamcinolone assistance IVTA
No difference in VA between groups
Structural Outcomes Foveal thickness reduced after TPPV Foveal thickness reduced after TPPV Foveal thickness reduced after TPPV
No difference in VA between groups
No difference in VA between groups
Foveal thickness reduced after TPPV
12 months No difference in VA between groups
TPPV/IVTA
Decreased central macular thickness with TPPV at 12 months Decreased macular 18 months No difference in VA between thickness at 6 months in groups except at month 3 (TPPV/IVTA/ILM peel superior both groups No difference in macular to TPPV/IVTA) thickness from baseline in both groups at 12 months and beyond
DME, diabetic macular edema; VMT, vitreomacular traction; PVD, posterior vitreous detachment; TPPV, trans pars plana vitrectomy; VA, visual acuity; ILM, internal limiting membrane; ICG, indocyanine green; IVTA, intravitreal triamcinolone.
improvement or loss in acuity could be obtained from 2 studies comparing TPPV with laser at 12 months, analysis of change in logMAR acuity was performed. A probability of less than 0.05 was deemed to be statistically significant.
RESULTS A total of 11 studies were identified that met the inclusion criteria for this study and are summarized in Tables 1 and 2. These studies were heterogenous in their inclusion and exclusion criteria, surgical intervention, and follow-up period. Furthermore, these studies also varied in their comparisons: 2 studies compared vitrectomy with the natural history of the disease, 4 compared vitrectomy with laser, 1 study compared vitrectomy with intravitreal steroid injection, 3 compared vitrectomy alone with combined vitrectomy with ILM peeling, and 1 investigated vitrectomy combined with ILM peeling, intravitreal steroid, or both. Outcomes of all studies included visual acuity, and the majority (9/11) also included an assessment of foveal thickness.
190
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
Studies of trans pars plana vitrectomy versus natural history
Yanyali et al.19 compared vitrectomy with the natural history of DME in 20 eyes of 10 patients who had previously undergone (Z6 months preceding the study) grid laser photocoagulation. Their study assigned TPPV randomly to either eye and specifically excluded those with features suggestive of VMT on OCT. They demonstrated that vitrectomy resulted in a significant decrease in foveal thickness at 12 months postprocedure (mean reduction of 165.8 vs 37.8 μm for the intervention and control groups, respectively). The majority of this improvement occurred within a month of surgery and was not paralleled by changes in function: no difference in visual acuity was found between treated and control eyes. Stolba and colleagues16 compared the effects of TPPV in 56 eyes with DME that had previously undergone grid laser (Z4 months preceding the study) and that also had an attached hyaloid. They found that TPPV was associated with a significant reduction in foveal thickness when compared with the natural history of the disease (60.3μm reduction in mean foveal for TPPV at 6 months vs an
Vitrectomy for diabetic macular edema—Simunovic et al.
Fig. 1 — A, Trans pars plana vitrectomy (TPPV) versus natural history at 6 months: improvement in foveal thickness (positive values better). B, TPPV versus natural history at 6 months: likelihood of achieving Z2 lines improvement in acuity. C, TPPV versus natural history at 6 months: likelihood of Z2 lines loss in acuity.
increase of 17.7 μm in the observed group). Although Stolba and colleagues16 found an improvement in visual acuity after vitrectomy, they also elected to perform cataract surgery in selected patients (those 460 years with mild cataract) at the time of surgery, thus introducing an important confounding factor. Meta-analysis of studies16,19 comparing the natural history of DME with TPPV suggests that the latter is associated with structural benefits (Fig. 1A) at 6 months. Meta-analysis also suggests a possible functional benefit of TPPV at 6 months in terms of likelihood of gaining Z2 lines in acuity (Fig. 1B), with the caveat of a confounding factor in the form of simultaneous cataract extraction in 1 of the included studies.16 There was no evidence for a decreased likelihood of visual loss of Z2 lines after TPPV (Fig. 1C). Studies of trans pars plana vitrectomy versus laser
Yanyali and colleagues’20 study compared TPPV with ILM peeling to grid laser in 24 eyes of 12 patients with bilateral diffuse DME who had not previously undergone macular laser. Patients with features of VMT were specifically excluded. They demonstrated that TPPV/ ILM peeling results in superior structural improvements
over laser (foveal thickness reduced by 219.4 vs 28.5 μm, respectively). Mean improvement in logMAR acuity (0.22 vs 0.1 for TPPV and laser, respectively) demonstrated a trend toward superior outcomes after TPPV/ ILM peeling, although this failed to reach statistical significance. Kumar and colleagues21 also compared TPPV/ILM peeling with grid laser in 24 eyes with DME resulting in severe visual impairment (visual acuity r6/60). Patients who had undergone macular laser within the preceding 6 months and those with VMT were specifically excluded. Postprocedural changes in foveal thickness suggested a significant benefit of TPPV/ ILM peeling (300.9-μm decrease at 6 months) over grid laser (106.6-μm decrease at 6 months). No significant difference in visual acuity was found between the 2 interventions. In a small pilot study of 15 patients that was directed at determining the number of subjects required to definitively compare TPPV with laser, Patel et al.18 found a slight, but statistically significant, superiority of laser over TPPV in terms of acuity. Although both TPPV and laser were associated with an improvement in macular thickness, the effect was greater for laser (27 and 107 μm, respectively). Notably, however, they found that patients undergoing TPPV had improved perifoveal CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
191
Vitrectomy for diabetic macular edema—Simunovic et al.
Fig. 2 — A, Trans pars plana vitrectomy (TPPV) versus laser. Structural (reduction in foveal thickness) outcomes at 6 months (positive values better). B, TPPV versus laser. Functional (likelihood of achieving Z2 lines improvement in acuity) outcomes at 6 months.
microperimetric sensitivities postprocedure, whereas laser patients had decreased sensitivities. One patient in the TPPV group suffered a loss of vision of 0.3 logMAR units at 12 months. Thomas and colleagues’17 feasibility study of TPPV/ILM peeling versus laser recruited 40 patients with DME and at least 1 previous episode of grid laser (Z3 months preceding recruitment). They found a mean deterioration in acuity of 0.05 logMAR at 12 months after TPPV/ILM peeling, whereas repeat laser was associated with an improvement of 0.03 logMAR (no significant difference). Although TPPV/ILM peeling was associated with improved structural outcomes over laser at 12 months (73-μm decrease vs 29-μm decrease in central macular thickness), this also failed to reach statistical significance.
Meta-analysis of studies comparing TPPV with laser suggests that TPPV affords a structural, but not a functional, benefit over laser (see Fig. 2) at 6 months after surgery. At 12 months, there was no evidence for a structural benefit of TPPV over laser and a trend toward inferior functional outcomes was observed, although this failed to reach statistical significance (Fig. 3). Only 1 patient undergoing TPPV was reported with a loss of vision of Z0.2 logMAR unit (no such losses were reported after laser; meta-analysis was not performed).18 Trans pars plana vitrectomy versus trans pars plana vitrectomy with internal limiting membrane peeling
Three studies have compared TPPV with ILM peeling with TPPV alone. Hoerauf and colleagues’22 study of 40
Fig. 3 — A, Trans pars plana vitrectomy (TPPV) versus laser. Structural (reduction in foveal thickness) outcomes at 12 months (positive values better). B, TPPV versus laser. Functional (baseline – final logMAR acuity) outcomes at 12 months.
192
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
Vitrectomy for diabetic macular edema—Simunovic et al.
Fig. 4 — Trans pars plana vitrectomy (TPPV) with internal limiting membrane peeling versus TPPV. Likelihood of an improvement in acuity by Z2 lines.
eyes with diffuse DME did not suggest a superiority of 1 surgical intervention over the other in terms of visual acuity, although they found that TPPV/ILM peeling resulted in significantly greater resolution in macular edema at 3 and 6 months. Bardak and colleagues’23 small study also found no evidence for a functional superiority of TPPV/ ILM peeling over TPPV alone; they did not compare structural outcomes. Bahadir and colleagues24 compared TPPV/ILM peeling with TPPV alone in patients with a “thickened posterior hyaloid” on clinical and retinal thickness analyzer examination. They similarly found no evidence for improved visual outcomes with the addition of ILM peeling to TPPV (structural outcomes not assessed). Only 2 of 3 studies were included in meta-analysis because they both excluded patients with VMT, whereas the third specifically investigated such patients. There was no evidence for a difference in visual acuity outcomes between TPPV/ILM peeling and TPPV alone (Fig. 4). Because only 1 study comparing TPPV/ILM peeling with TPPV alone included foveal thickness data, meta-analysis of this outcome could not be performed. Trans pars plana vitrectomy with internal limiting membrane peeling versus intravitreal steroid
One study compared vitrectomy with intravitreal corticosteroid therapy: Doi and colleagues25 evaluated TPPV/ ILM peeling against intravitreal triamcinolone (IVTA) over a 12-month period. Although their study protocol allowed re-treatment after 6 months and permitted cataract extraction within the study period, no eyes underwent either of these additional interventions. In terms of reducing foveal thickness, their results suggest an early superiority of IVTA (at 1 month) and a later superiority of TPPV/ILM peeling (at 12 months). No significant difference was found between the 2 interventions in terms of visual acuity. Trans pars plana vitrectomy/intravitreal triamcinolone versus trans pars plana vitrectomy/internal limiting membrane peeling versus trans pars plana vitrectomy/internal limiting membrane peeling/ intravitreal triamcinolone
Figueroa et al.26 investigated the effects of TPPV combined with ILM peeling, IVTA, or both. Only patients in the latter 2 arms (n ¼ 31) were randomized, as those with a prior history of glaucoma were exclusively
allocated to the first group (n ¼ 11). Nevertheless, they found no significant difference between the 2 randomized interventions in terms of visual acuity, except at 3 months (combined TPPV/ILM peeling/IVTA superior to TPPV/ IVTA); furthermore, they did not find a statistically significant improvement in visual acuity from baseline for either intervention. All 3 groups of patients demonstrated an average improvement in macular thickness at 6 months, which subsequently regressed to baseline by 12 months.
DISCUSSION There is concordance between 2 prospective, randomized, controlled studies comparing vitrectomy with the natural history of DME that TPPV is associated with improvements in macular structure. Meta-analysis also supports the assertion that TPPV is associated with an increased likelihood of achieving a Z2 line improvement in acuity, although these results may be confounded by the methodology used in 1 of the studies, which permitted simultaneous cataract extraction as part of the primary procedure in some of the experimental subjects.16 Studies comparing vitrectomy with laser have produced mixed results. Meta-analysis of studies including 6-month follow-up suggests a superiority of TPPV over laser for improvement of foveal thickness, whereas meta-analysis of studies including a 12-month follow-up suggests no such benefit at this later time point. Meta-analysis of visual acuity outcomes does not support a superiority of TPPV over laser at 6 months and suggested a trend toward inferior outcomes after TPPV at 12 months. In terms of surgical interventions, the evidence suggests that the addition of ILM peeling may afford improved structural, but not functional, outcomes over vitrectomy alone.22 It is also important to note that several of the studies included in this systematic review used indocyanine green (ICG) to aid ILM peeling. This agent has subsequently been implicated in retinal toxicity and could have theoretically had a negative effect on treatment outcomes when used,27–31 especially at concentrations of Z0.1%. However, it should also be noted that studies that included the use of ICG staining generally found in favour of TTPV/ILM peeling (see Table 2). CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
193
Vitrectomy for diabetic macular edema—Simunovic et al. The management of patients with isolated DME unresponsive to laser or intravitreal therapy remains a clinical dilemma. At this point in time, there is little evidence to support vitrectomy as an alternative treatment. Although some studies suggest a possible benefit of TPPV with, or without, ILM peeling over either the natural history of the disease or over grid laser, others fail to reveal a benefit. Part of the problem in assessing the studies included in this systematic review is their heterogeneity in control interventions and follow-up periods, which necessitated separate meta-analyses and also prevented formal assessment of publication bias. These issues are compounded by the fact that all studies to date have included relatively small sample sizes. By way of comparison, intravitreal anti-VEGF has recently been shown to be superior to laser in diffuse DME; however, not one of the studies included in this review would be powered to detect the difference in the proportion of patients achieving an improvement in acuity of Z2 lines demonstrated in the RESTORE study (N ¼ 126, assuming a power of 0.80 and an α ¼ 0.05).32 Nevertheless, meta-analysis supports the assertion that vitrectomy results in superior early (i.e., 6-month) structural outcomes when compared with observation and grid laser. However, conclusive evidence for a functional benefit is lacking. Although meta-analysis of 2 randomized, controlled trials would suggest a possible functional improvement over the natural history of the disease, 1 of these studies introduced an important confounding factor of simultaneous cataract extraction in a subgroup of TPPV patients. Although a trend was observed toward improved functional outcomes over grid laser at 6 months, this failed to reach statistical significance; furthermore, there was a trend toward inferior functional outcomes compared with laser at 12 months. One particular difficulty in using conventional acuity measures in the assessment of post-TPPV retinal function in phakic eyes is that vitrectomy is cataractogenic. Thus, it could be argued that improvements in retinal function after TPPV, as assessed by conventional measures of acuity, may have been obscured by postoperative lenticular changes observed in some of the included studies (see Table 2). One of the possible mechanisms by which vitrectomy may improve retinal structure and function in patients with DME is via relief of tangential and/or anterior-posterior macular traction. Some of the nonrandomized control trial literature suggests that those with VMT, including the socalled thickened taut glistening posterior hyaloid, may derive a particular benefit from surgical intervention.17 However, such patients have been included in only 1 randomized, controlled trial to date (which compared vitrectomy alone with vitrectomy with ILM peeling).24 A further issue is that since the publication of the clinical trials included in this systematic review, a new gold standard treatment for diffuse DME has emerged: intravitreal antiVEGF therapy.32 This fact, combined with observation that
194
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
vitrectomy does not appear to afford improved functional outcomes over laser in patients with DME, suggests that there is no compelling argument for vitrectomy as a treatment for DME in the absence of VMT. However, given the suggestion that those with various forms of concomitant VMT may derive a particular benefit from vitrectomy and ILM peeling,33 this patient group may make a particularly suitable population to study prospectively in a randomized, controlled trial comparing anti-VEGF agents with vitrectomy.
Disclosure: The authors have no proprietary or commercial interest in any materials discussed in this article.
REFERENCES 1. O’Doherty M, Dooley I, Hickey-Dwyer M. Interventions for diabetic macular oedema: a systematic review of the literature. Br J Ophthalmol. 2008;92:1581-90. 2. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study report number 1. Early Treatment Diabetic Retinopathy Study research group. Arch Ophthalmol. 1985;103:1796-806. 3. Lam DS, Chan CK, Mohamed S, et al. Intravitreal triamcinolone plus sequential grid laser versus triamcinolone or laser alone for treating diabetic macular edema: six-month outcomes. Ophthalmology. 2007;114:2162-7. 4. Scott IU, Edwards AR, Beck RW, et al. A phase II randomized clinical trial of intravitreal bevacizumab for diabetic macular edema. Ophthalmology. 2007;114:1860-7. 5. Laidlaw DA. Vitrectomy for diabetic macular oedema. Eye (Lond). 2008;22:1337-41. 6. Lewis H, Abrams GW, Blumenkranz MS, Campo RV. Vitrectomy for diabetic macular traction and edema associated with posterior hyaloidal traction. Ophthalmology. 1992;99:753-9. 7. Massin P, Duguid G, Erginay A, Haouchine B, Gaudric A. Optical coherence tomography for evaluating diabetic macular edema before and after vitrectomy. Am J Ophthalmol. 2003;135:169-77. 8. Kaiser PK, Riemann CD, Sears JE, Lewis H. Macular traction detachment and diabetic macular edema associated with posterior hyaloidal traction. Am J Ophthalmol. 2001;131:44-9. 9. Nasrallah FP, Jalkh AE, Van Coppenolle F, et al. The role of the vitreous in diabetic macular edema. Ophthalmology. 1988;95: 1335-9. 10. Hikichi T, Fujio N, Akiba J, Azuma Y, Takahashi M, Yoshida A. Association between the short-term natural history of diabetic macular edema and the vitreomacular relationship in type II diabetes mellitus. Ophthalmology. 1997;104:473-8. 11. Harbour JW, Smiddy WE, Flynn HW Jr, Rubsamen PE. Vitrectomy for diabetic macular edema associated with a thickened and taut posterior hyaloid membrane. Am J Ophthalmol. 1996;121: 405-13. 12. Pendergast SD, Hassan TS, Williams GA, et al. Vitrectomy for diffuse diabetic macular edema associated with a taut premacular posterior hyaloid. Am J Ophthalmol. 2000;130:178-86. 13. Recchia FM, Ruby AJ, Carvalho Recchia CA. Pars plana vitrectomy with removal of the internal limiting membrane in the treatment of persistent diabetic macular edema. Am J Ophthalmol. 2005;139: 447-54. 14. Verhagen AP, de Vet HC, de Bie RA, et al. The Delphi list: a criteria list for quality assessment of randomized clinical trials for conducting systematic reviews developed by Delphi consensus. J Clin Epidemiol. 1998;51:1235-41. 15. Rev Man. Version 5.1 [computer program]. Copenhagen: The Cochrane Collaboration; 2011. 16. Stolba U, Binder S, Gruber D, Krebs I, Aggermann T, Neumaier B. Vitrectomy for persistent diffuse diabetic macular edema. Am J Ophthalmol. 2005;140:295-301.
Vitrectomy for diabetic macular edema—Simunovic et al. 17. Thomas D, Bunce C, Moorman C, Laidlaw DA. A randomised controlled feasibility trial of vitrectomy versus laser for diabetic macular oedema. Br J Ophthalmol. 2005;89:81-6. 18. Patel JI, Hykin PG, Schadt M, et al. Diabetic macular oedema: pilot randomised trial of pars plana vitrectomy vs macular argon photocoagulation. Eye (Lond). 2006;20:873-81. 19. Yanyali A, Horozoglu F, Celik E, Ercalik Y, Nohutcu AF. Pars plana vitrectomy and removal of the internal limiting membrane in diabetic macular edema unresponsive to grid laser photocoagulation. Eur J Ophthalmol. 2006;16:573-81. 20. Yanyali A, Nohutcu AF, Horozoglu F, Celik E. Modified grid laser photocoagulation versus pars plana vitrectomy with internal limiting membrane removal in diabetic macular edema. Am J Ophthalmol. 2005;139:795-801. 21. Kumar A, Sinha S, Azad R, Sharma YR, Vohra R. Comparative evaluation of vitrectomy and dye-enhanced ILM peel with grid laser in diffuse diabetic macular edema. Graefes Arch Clin Exp Ophthalmol. 2007;245:360-8. 22. Hoerauf H, Bruggemann A, Muecke M, et al. Pars plana vitrectomy for diabetic macular edema. Internal limiting membrane delamination vs posterior hyaloid removal. A prospective randomized trial. Graefes Arch Clin Exp Ophthalmol. 2011;249:997-1008. 23. Bardak Y, Cekic O, Tig SU. Comparison of ICG-assisted ILM peeling and triamcinolone-assisted posterior vitreous removal in diffuse diabetic macular oedema. Eye (Lond). 2006;20:1357-9. 24. Bahadir M, Ertan A, Mertoglu O. Visual acuity comparison of vitrectomy with and without internal limiting membrane removal in the treatment of diabetic macular edema. Int Ophthalmol. 2005;26(1-2):3-8.
25. Doi N, Sakamoto T, Sonoda Y, et al. Comparative study of vitrectomy versus intravitreous triamcinolone for diabetic macular edema on randomized paired-eyes. Graefes Arch Clin Exp Ophthalmol. 2012;250:71-8. 26. Figueroa MS, Contreras I, Noval S. Surgical and anatomical outcomes of pars plana vitrectomy for diffuse nontractional diabetic macular edema. Retina. 2008;28:420-6. 27. Gale JS, Proulx AA, Gonder JR, Mao AJ, Hutnik CM. Comparison of the in vitro toxicity of indocyanine green to that of trypan blue in human retinal pigment epithelium cell cultures. Am J Ophthalmol. 2004;138:64-9. 28. Goldstein M, Zemel E, Loewenstein A, Perlman I. Retinal toxicity of indocyanine green in albino rabbits. Invest Ophthalmol Vis Sci. 2006;47:2100-7. 29. Narayanan R, Kenney MC, Kamjoo S, et al. Toxicity of indocyanine green (ICG) in combination with light on retinal pigment epithelial cells and neurosensory retinal cells. Curr Eye Res. 2005;30:471-8. 30. Querques G, Prascina F, Iaculli C, Noci ND. Retinal toxicity of indocyanine green. Int Ophthalmol. 2008;28:115-8. 31. Tokuda K, Zorumski CF, Izumi Y. Involvement of illumination in indocyanine green toxicity after its washout in the ex vivo rat retina. Retina. 2009;29:371-9. 32. Mitchell P, Bandello F, Schmidt-Erfurth U, et al. The RESTORE study: ranibizumab monotherapy or combined with laser versus laser monotherapy for diabetic macular edema. Ophthalmology. 2011;118: 615-25. 33. Haller JA, Qin H, Apte RS, et al. Vitrectomy outcomes in eyes with diabetic macular edema and vitreomacular traction. Ophthalmology. 2010;117:1087-93.e1083.
CAN J OPHTHALMOL — VOL. 49, NO. 2, APRIL 2014
195
