Jackson, Timothy L. PhD, FRCOPHTH*; Nicod, Elena MSc†; Angelis, Aris MSc†; Grimaccia, Federico MD†; Prevost, A. Toby PhD‡; Simpson, Andrew R. H. MRCOphth*,§; Kanavos, Panos PhD†
The onset of posterior vitreous detachment (PVD) is variable and is dependent on several factors including age, refractive error, and sex.1–4 The mean age of onset of PVD is ∼60 years,2,3 and with each decade, the incidence increases so that by the ninth decade, 46% to 86% of eyes have a complete PVD.4–6 Typically, PVD has no discernible consequence other than minimally intrusive visual floaters, but it is possible that PVD alters ocular physiology in a subclinical manner. For example, studies have shown that pharmacologic vitreolysis increases vitreous diffusion coefficients, including oxygen,7,8 and it is hypothesized that PVD leads to similar changes.9 This occurs because the speed of diffusion is inversely related to viscosity and the retrogel aqueous is substantially less viscous than the vitreous.
Neovascular (wet) age-related macular degeneration (AMD), diabetic macular edema (DME), and retinal vein occlusion (RVO) are all characterized by elevated levels of vascular endothelial growth factor, which is known to adversely affect the clinical course of each disease. Posterior vitreous detachment might be expected to enhance the diffusion of endogenous vascular endothelial growth factor away from the macula, and if PVD increases oxygenation, this would also tend to lower macular vascular endothelial growth factor levels. In addition, vitreomacular adhesion (VMA) may be associated with vitreomacular traction (VMT) that can in itself cause macular edema, a hallmark of all three diseases and a major cause of vision loss. It is therefore possible that VMA and an attached vitreous adversely affect the clinical course of DME, AMD, and RVO.
Considering each disease in turn, clinical studies of DME suggest that abnormal VMA may be a contributory factor.10 It has also been reported that patients with DME were less likely to have a PVD and that PVD was associated with a significant reduction in central macular thickness in patients treated with intravitreal injections of triamcinolone.11 Lewis et al12 were the first group to perform pars plana vitrectomy (PPV) on eyes with DME associated with what was described as a “thickened and taut premacular posterior hyaloid,” with positive visual outcomes.
The subject of PPV for DME was reviewed by Laidlaw.13 He concluded that the majority of studies show a reduction in DME and improvement in vision after PPV, but conversely, there are also a number of well-designed studies that do not show such a benefit, and therefore, more evidence is required. Even if the therapeutic benefit of PPV can be questioned, VMT can certainly coexist with DME, and it seems likely that this would aggravate macular edema.
There are several studies indicating that vitreous attachment,14–16 VMA,5,15–17 and VMT18 are all more common in eyes with AMD compared with control eyes. Furthermore, there are some small studies suggesting that surgery to relieve VMT in wet AMD may produce functional and anatomical improvement.18 The link between VMA and AMD was recently reviewed by Simpson et al,19 who concluded that the evidence points to an association between VMA and AMD, even if it could not necessarily be established that the link was causal.
Vitreomacular adhesion or VMT may worsen the clinical course of RVO.20 Reports suggest that a complete PVD reduces the risk of retinal or optic disk neovascularization after central retinal vein occlusion (CRVO).21 It may also reduce the risk of macular edema in both CRVO and branch retinal vein occlusion (BRVO).22 Murakami et al23 induced a PVD with intravitreal tissue plasminogen activator in 36 patients with macular edema secondary to CRVO. Visual acuity and macular thickness improved significantly after treatment. A number of studies also suggested that PPV can reduce the macular edema associated with RVO.24–26
Taken together, these articles suggest that the vitreous may play a role in each disease. The clinical relevance is obvious, in that the relief of VMA could potentially influence disease progression. However, before any causal link or therapeutic benefit can be established, it is important to first establish if there is an association. To this end, we undertook a synthesis of the literature to test the hypothesis that DME, AMD, and RVO are all associated with a higher than normal prevalence of vitreous attachment.
Materials and Methods
A PubMed MEDLINE database search from 1946 to beginning 2011 was undertaken using Boolean operators with the following keywords and MESH headings: vitreomacular traction; vitreomacular adhesion; taut posterior hyaloid; pars plana vitrectomy + vitreomacular traction; pars plana vitrectomy + macular (o)edema; pars plana vitrectomy + maculopathy; posterior vitreous detachment + age-related macular degeneration; posterior vitreous detachment + macular (o)edema + diabetes; posterior vitreous detachment + maculopathy; retinal vein occlusion + vitreomacular adhesion; retinal vein occlusion + posterior vitreous detachment; retinal vein occlusion + vitreomacular traction. The searches for RVO were repeated using CRVO and BRVO. Further articles were identified from the reference lists of retrieved articles and from relevant textbooks.
A total of 1,025 abstracts were retrieved. The abstracts were then reviewed by a senior retinal expert (T.L.J.) who selected articles for full review, based on whether they might provide data on an association with VMA, VMT, or PVD. Only English- and French-language articles in peer-reviewed journals were eligible. A total of 232 eligible articles were reviewed in full. A further selection was undertaken to retrieve articles that provided data of sufficient quality to undergo pooled statistical analysis. To be included, articles describing DME, AMD, and RVO had to provide information of sufficient clarity such that the reviewer could match the state of the vitreoretinal interface to one the following definitions:
1. Vitreomacular adhesion: adhesion of the vitreous face within the macular region.
2. Vitreomacular traction: VMA causing focal, tractional, distortion of the macula.
3. Complete PVD: separation of the posterior hyaloid face from both the macula and optic disk.
4. Partial PVD: separation of the posterior hyaloid face from either the macula or optic disk, but not both.
Where representative images were provided, these were compared with the standard photograph of VMT of Simpson et al.19 Studies that diagnosed VMT or VMA without optical coherence tomography were excluded.
Of 232 articles, 16 were used for the analysis. For some of these articles, only a certain subset of data could be included: Most commonly, this occurred where the authors provided well-defined data in relation to one vitreous state, such as PVD, but other states, such as VMA, were not described with sufficient clarity, or in a way that could be matched to the study definitions. Because of a lack of large population-based studies, case series and cross-sectional studies were eligible for inclusion and these formed the bulk of the descriptive results. For the statistical metaanalysis, only studies that provided a control group were included.
Three independent reviewers (E.N., A.A., and F.G.) entered data from each article into an electronic data capture form. The data collected for each article included the following: 1) general information about the study (aim, summary, and key findings); 2) methodological details (study design, study population, entry criteria, methods, and study period); 3) primary and secondary outcomes; 4) data on the incidence/prevalence of VMA, PVD, or VMT; and 5) presenting and final visual acuity for interventional studies. To analyze visual outcome across studies, the mean VA data were converted to logarithm of the minimum angle of resolution units (LogMAR).27–30 Counting fingers vision was assigned a LogMAR acuity of 1.6, hand movements 1.9, light perception 2.2, and no light perception 2.5.27–30
Mean prevalence data were weighted based on the number of eyes in each study so that large studies had a greater effect than small studies on the final aggregated results. The data extraction and analysis were performed by two reviewers independently, validated by a third, and finally reviewed by a senior retinal specialist.
Metaanalysis was used to pool within-study comparisons of the relative proportion of VMA and the relative proportion with complete PVD between wet AMD eyes and dry AMD eyes and between each of these AMD groups and control, using RevMan 5 Software (Cochrane Collaboration, Oxford, United Kingdom). The fixed effect method with inverse variance weighting was used or random effects (Der Simonian and Laird) method when there was significant heterogeneity assessed using Cochran Q statistic. All tests were 2 tailed and were assessed at the 5% level of significance. The 95% confidence interval (CI) of pooled effects was used to assist in distinguishing lack of evidence from a negative result. The data from the studies included in the analysis were not reported in sufficient detail to allow adjustment for correlation between eyes in the metaanalysis; hence, the width of the CIs from the metaanalysis may be slightly underestimated.
There were too few suitable reports with which to determine the incidence of VMA and PVD in eyes with DME. Seven studies of DME provided data on the prevalence of VMT,31–37 with two of the studies labeling the condition as taut posterior hyaloid. In cases where it was specified, the prevalence of VMT or taut posterior was 18% (n = 514). Details are shown in Table 1. All except two of these case series were surgical.33,37 Vitreomacular traction was present in 28.7% of 188 eyes undergoing PPV and in 12.0% of 326 nonsurgical cases.
Age-Related Macular Degeneration
Six studies of AMD detailed the prevalence of VMA.15–18,38,39 These studies were specifically studying the interaction between the vitreomacular interface and AMD and usually included a control group, such as the unaffected fellow eye or eyes with dry AMD. The prevalence of VMA in patients with wet AMD was 22.6% (n = 654), 9.5% (n = 327) in dry AMD, and 7.7% (n = 261) in unaffected control eyes. Details are provided in Table 2. Four studies of wet AMD provided unaffected control eyes.15,17,18,39 The rate of VMA was higher than controls in each of these reports, and in one case, this difference was statistically significant. Combined results are presented in Figure 1A. Eyes with wet AMD were 2.15 times more likely to have VMA than controls (95% CI, 1.34–3.48; p = 0.002). There were two controlled studies of VMA and dry AMD: one showed an increased prevalence of VMA18 the other a reduced prevalence.15 Analysis of both studies combined showed that the likelihood of having VMA was 1.23 times that of controls, but this difference was not statistically significant (95% CI, 0.74–2.36; p = 0.53). Three controlled studies compared the prevalence of VMA in eyes with wet and dry AMD.15,16,18 Eyes with wet AMD were 2.54 times more likely to have VMA, but this difference was not significant (95% CI, 0.88–7.36; p = 0.09), as shown in Figure 1B.
Four studies reported the prevalence of PVD in wet AMD.15,16,18,39 The prevalence of complete PVD was 40.6% (n = 251) in wet AMD, 51.0% (n = 145) in dry AMD, and 55.0% (n = 180) in controls. The prevalence of partial PVD was 30.4% (n = 79), 15.1% (n = 86), and 5.4% (n = 56), respectively. Details are provided in Table 3. Three of these studies provided control eyes.15,18,39 Metaanalysis showed that eyes with wet AMD were 23% less likely to have a complete PVD than control eyes (relative risk 0.77; 95% CI, 0.64–0.93; p = 0.007), as shown in Figure 1C.
There were two controlled studies of PVD and dry AMD: one showed an increased prevalence of PVD15 the other showed a reduced prevalence.18 Combined analysis showed that the likelihood of having PVD was 1.22 times that of controls, but this difference was not statistically significant (95% CI, 0.94–1.58; p = 0.14).
Three controlled studies compared wet AMD and dry AMD eyes.15,16,18 In two studies,15,16 eyes with wet AMD had a significantly lower prevalence of complete PVD than eyes with dry AMD, but the third study did not show a statistically significant difference.18 Metaanalysis showed that eyes with wet AMD had a 44% lower prevalence of complete PVD than eyes with dry AMD, but this difference was not significant (relative risk 0.56; 95% CI, 0.27–1.14, p = 0.11), as shown in Figure 1D.
Two of the analyses (Figure 1, A and D) showed significant heterogeneity. There were no obvious systematic differences between the studies to explain this. The Q-statistics had P values of 0.02 and 0.01 associated with them, which might question the use of a fixed effects model. However, because there were a small number of studies and because the interval estimates of the studies in each metaanalysis showed negligible overlap, it was concluded that there was no strong evidence to overturn the use of random effects metaanalyses. The random effects model does not ignore the inconsistency (as a fixed effects approach would) but incorporates this extra variation into providing a wider 95% CI for the pooled effect.
Retinal Vein Occlusion
Three studies on RVO provided data on the prevalence of PVD.22,24,40 The prevalence of complete PVD was 30.4% (n = 56) in CRVO, 31.0% (n = 71) in BRVO, and 25% (n = 64) in fellow eye controls. The prevalence of partial PVD was 8.9% (n = 56) in CRVO and 19.0% (n = 58) in BRVO, but controls were not provided. Data are summarized in Table 4. No studies provided an independent control. One study used a paired fellow eye control.40 Therefore, there were too few controlled studies to enable metaanalysis. There was insufficient evidence with which to determine the prevalence of VMA in RVO.
This study investigated whether eyes with DME, wet AMD, or RVO were more likely to have an attached vitreous than control eyes. About DME, it was not possible to conclude on the relative prevalence of VMA or PVD, because of a lack of appropriate controls. There was insufficient evidence to determine if RVO is associated with a higher or lower than normal prevalence of VMA. In the subset of patients undergoing PPV for treatment of DME, VMT was present in 28.7% of cases, more than seen in the nonsurgical series (12.0%), but the latter analysis comprised only 2 series and needs to be interpreted with caution. It should be noted that the degree of retinal distortion needed to diagnose VMT is hard to quantify: some clinicians may label an eye as having VMA, whereas others may diagnose VMT. At the time of publication, none of the authors had access to a photographic standard19 that could be used to define VMT.
By contrast, there were controlled studies of AMD of sufficient quality that, taken together, suggest that eyes with wet AMD are twice as likely to have VMA and 23% less likely to have a PVD than control eyes. The question that then follows is whether release of VMA would be beneficial. While some studies of PPV suggest that it might,18,41–45 there is still no sufficient evidence to establish that the presence of VMA contributes to disease activity in AMD or to present PPV as a proven therapy. A review by Simpson et al19 drew a similar conclusion. The authors noted that the apparent association between AMD and VMA does not imply that VMA contributes to disease activity, even though it might. Indeed, it is conceivable that AMD causes VMA. This uncertainty is increased by the fact that the studies were mostly cross-sectional studies of prevalence and they did not study the incidence of either VMA or PVD in relation to wet AMD. It should also be noted that the present pooled analysis showing an association between VMA and wet AMD came from only 3 studies of 188 eyes.
At present, there are no reports of large randomized controlled trials of any intervention designed to release VMA in eyes with AMD. However, Microplasmin for Intravitreous Injection-5 (MIVI-5) trial is a randomized controlled trial of 100 participants with AMD that is designed to assess intravitreal ocriplasmin, an agent designed to chemically induce PVD and release VMA. The trial is likely to substantially add to the evidence base and is expected to report in 2013 (clinicaltrial.gov identifier: NCT00913744).
The literature on RVO suggested that, unlike AMD, PVD was more frequent than control eyes. The literature was however extremely limited, and only one study of CRVO provided a fellow eye control. As such, even with pooled analysis, the poor quality of the evidence means it is not possible to draw a conclusion.
The main strength of the present analysis is that it undertook a large survey of the literature and pooled data across studies that met defined criteria. The main weakness stems from the variable quality of the literature and the fact that some of the metaanalyses combined studies with significant heterogeneity. Some studies used a clustered design in which outcomes from both eyes of participants were included in the analysis. However, coefficients for the within-person correlation between eyes were not universally reported with the consequence that 95% CIs for the estimated effects may not be accurate. As the mean cluster size across all studies is at most two, and as intracluster correlation is bounded between zero and one, we anticipate that the impact could be only a small underestimation of the width of CIs. None of the studies provided Level I evidence, but a small number of controlled studies of wet AMD did provide Level II evidence that was used for metaanalysis.46
In conclusion, this synthesis of the literature supports the hypothesis that AMD is associated with an attached vitreous. There is insufficient evidence to conclude on the prevalence of vitreous attachment in DME or RVO. Further controlled studies of RVO and DME are needed.
1. Hayreh S, Jonas J. Posterior vitreous detachment: clinical correlations. Ophthalmologica 2004;218:333–343.
2. Yonemoto J, Ideta H, Sasaki K, et al.. The age of onset of posterior vitreous detachment. Graefes Arch Clin Exp Ophthalmol 1994;232:67–70.
3. Chuo JY, Lee TY, Hollands H, et al.. Risk factors for posterior vitreous detachment: a case-control study. Am J Ophthalmol 2006;142:931–937.
4. Akiba J. Prevalence of posterior vitreous detachment in high myopia. Ophthalmology 1993;100:1384–1388.
5. Weber-Krause B, Eckardt U. Incidence of posterior vitreous detachment in eyes with and without age-related macular degeneration. An ultrasonic study [German]. Ophthalmologe 1996;93:660–665.
6. Snead MP, Snead DR, Mahmood AS, Scott JD. Vitreous detachment and the posterior hyaloid membrane: a clinicopathological study. Eye (Lond) 1994;8(pt 2):204–209.
7. Sebag J, Ansari RR, Suh KI. Pharmacologic vitreolysis with microplasmin increases vitreous diffusion coefficients. Graefes Arch Clin Exp Ophthalmol 2007;245:576–580.
8. Quiram PA, Leverenz VR, Baker RM, et al.. Microplasmin-induced posterior vitreous detachment affects vitreous oxygen levels. Retina 2007;27:1090–1096.
9. Stefansson E. Physiology of vitreous surgery. Graefes Arch Clin Exp Ophthalmol 2009;247:147–163.
10. Nasrallah FP, Jalkh AE, Van Coppenolle F, et al.. The role of the vitreous in diabetic macular edema. Ophthalmology 1988;95:1335–1339.
11. Sivaprasad S, Ockrim Z, Massaoutis P, et al.. Posterior hyaloid changes following intravitreal triamcinolone and macular laser for diffuse diabetic macular edema. Retina 2008;28:1435–1442.
12. Lewis H, Abrams GW, Blumenkranz MS, Campo RV. Vitrectomy for diabetic macular traction and edema associated with posterior hyaloidal traction. Ophthalmology 1992;99:753–759.
13. Laidlaw DA. Vitrectomy for diabetic macular oedema. Eye (Lond) 2008;22:1337–1341.
14. Ondes F, Yilmaz G, Acar MA, et al.. Role of the vitreous in age-related macular degeneration. Jpn J Ophthalmol 2000;44:91–93.
15. Krebs I, Brannath W, Glittenberg C, et al.. Posterior vitreomacular adhesion: a potential risk factor for exudative age-related macular degeneration? Am J Ophthalmol 2007;144:741–746.
16. Robison CD, Krebs I, Binder S, et al.. Vitreomacular adhesion in active and end-stage age-related macular degeneration. Am J Ophthalmol 2009;148:79–82.e2.
17. Lee SJ, Lee CS, Koh HJ. Posterior vitreomacular adhesion and risk of exudative age-related macular degeneration: paired eye study. Am J Ophthalmol 2009;147:621–626.e1.
18. Mojana F, Cheng L, Bartsch DU, et al.. The role of abnormal vitreomacular adhesion in age-related macular degeneration: spectral optical coherence tomography and surgical results. Am J Ophthalmol 2008;146:218–227.
19. Simpson A, Petrarca R, Jackson TL. Vitreomacular adhesion and neovascular age-related macular degeneration—a review. Surv Opthalmol 2012;57:498–509.
20. Ophir A, Trevino A, Martinez MR. Extrafoveal vitreous traction associated with branch retinal vein occlusion. Eur J Ophthalmol 2010;20:733–739.
21. Hikichi T, Konno S, Trempe CL. Role of the vitreous in central retinal vein occlusion. Retina 1995;15:29–33.
22. Takahashi MK, Hikichi T, Akiba J, et al.. Role of the vitreous and macular edema in branch retinal vein occlusion. Ophthalmic Surg Lasers 1997;28:294–299.
23. Murakami T, Takagi H, Ohashi H, et al.. Role of posterior vitreous detachment induced by intravitreal tissue plasminogen activator in macular edema with central retinal vein occlusion. Retina 2007;27:1031–1037.
24. Charbonnel J, Glacet-Bernard A, Korobelnik JF, et al.. Management of branch retinal vein occlusion with vitrectomy and arteriovenous adventitial sheathotomy, the possible role of surgical posterior vitreous detachment. Graefes Arch Clin Exp Ophthalmol 2004;242:223–228.
25. Hvarfner C, Larsson J. Vitrectomy for non-ischaemic macular oedema in retinal vein occlusion. Acta Ophthalmol Scand 2006;84:812–814.
26. Kumagai K, Furukawa M, Ogino N, et al.. Long-term outcomes of vitrectomy with or without arteriovenous sheathotomy in branch retinal vein occlusion. Retina 2007;27:49–54.
27. Jackson TL, Larsson J, Tanner V, Williamson TH. Combined phaco-emulsification cataract extraction and pars plana vitrectomy without intra-ocular lens implantation. Ophthalmologica 2001;215:271–275.
28. Leyland MD, Schulenburg WE. Combined phacoemulsification—vitrectomy surgery: technique, indications and outcomes. Eye (Lond) 1999;13(pt 3a):348–352.
29. Westheimer G. Scaling of visual acuity measurements. Arch Ophthalmol 1979;97:327–330.
30. Javitt JC, Brenner MH, Curbow B, et al.. Outcomes of cataract surgery. Improvement in visual acuity and subjective visual function after surgery in the first, second, and both eyes. Arch Ophthalmol 1993;111:686–691.
31. Massin P, Duguid G, Erginay A, et al.. Optical coherence tomography for evaluating diabetic macular edema before and after vitrectomy. Am J Ophthalmol 2003;135:169–177.
32. Gandorfer A, Rohleder M, Grosselfinger S, et al.. Epiretinal pathology of diffuse diabetic macular edema associated with vitreomacular traction. Am J Ophthalmol 2005;139:638–652.
33. Thomas D, Bunce C, Moorman C, Laidlaw AH. Frequency and associations of a taut thickened posterior hyaloid, partial vitreomacular separation, and subretinal fluid in patients with diabetic macular edema. Retina 2005;25:883–888.
34. Rouberol F, Kodjikian L, Fleury J, Grange JD. Diffuse diabetic macular edema surgery: prospective study of seven cases followed up with optical coherence tomography [French]. J Fr Ophtalmol 2005;28:474–479.
35. Shah SP, Patel M, Thomas D, et al.. Factors predicting outcome of vitrectomy for diabetic macular oedema: results of a prospective study. Br J Ophthalmol 2006;90:33–36.
36. Yamamoto T, Takeuchi S, Sato Y, Yamashita H. Long-term follow-up results of pars plana vitrectomy for diabetic macular edema. Jpn J Ophthalmol 2007;51:285–291.
37. Ophir A, Trevino A, Fatum S. Extrafoveal vitreous traction associated with diabetic diffuse macular oedema. Eye 2010;24:347–353.
38. Lee SJ, Koh HJ. Effects of vitreomacular adhesion on anti-vascular endothelial growth factor treatment for exudative age-related macular degeneration. Ophthalmology 2011;118:101–110.
39. Nomura Y, Ueta T, Iriyama A, et al.. Vitreomacular interface in typical exudative age-related macular degeneration and polypoidal choroidal vasculopathy. Ophthalmology 2010;118:853–859.
40. Kado M, Jalkh AE, Yoshida A, et al.. Vitreous changes and macular edema in central retinal vein occlusion. Ophthalmic Surg 1990;21:544–549.
41. Ikeda T, Sawa H, Koizumi K, et al.. Pars plana vitrectomy for regression of choroidal neovascularization with age-related macular degeneration. Acta Ophthalmol Scand 2000;78:460–464.
42. Liu YL, Lin CP, Yang CM. Transient regression of choroidal neovascularization membrane after vitrectomy in age-related macular degeneration with vitreomacular traction syndrome. Acta Ophthalmol 2011;89:e291–e293.
43. Roller AB, Mahajan VB, Boldt HC, et al.. Effects of vitrectomy on age-related macular degeneration. Ophthalmology 2010;117:1381–1386.
44. Sakamoto T, Sheu SJ, Arimura N, et al.. Vitrectomy for exudative age-related macular degeneration with vitreous hemorrhage. Retina 2010;30:856–864.
45. Schulze S, Neugebauer A, Kroll P. Appearance of age-related macular degeneration in vitrectomized and nonvitrectomized eyes: an intraindividual case study. Acta Ophthalmol 2012;90:244–247.
46. U.S. Preventive Services Task Force. Guide to Clinical Preventive Services: Report of the U.S. Preventive Services Task Force. Darby, Pa: DIANE Publishing; 1989:24.
© 2013 by Ophthalmic Communications Society, Inc.