VMT is defined as vitreo-macular adhesion with structural changes in the underlying macula [2]. At the International Vitreomacular Traction Study (IVTS) group conference in 2012, a new classification system of vitreo-macular traction was developed, based solely upon anatomic changes as found with OCT imaging. To diagnose VMT, the following changes should be observed: (1) evidence of perifoveal vitreous cortex detachment from the retinal surface, (2) macular attachment of the posterior vitreous cortex within a 3-mm radius of the fovea, and (3) distortion of the foveal surface, intraretinal structural changes, elevation of the fovea above the RPE, or a combination of these anatomical changes [2]. VMT is classified by the size of vitreous attachment: focal, if less than 1500 μm, or broad, if greater than 1500 μm.

In the classic form of VMT, the posterior vitreous cortex is separated from the retina in the peripheral fundus but remains attached posteriorly, resulting in antero-posterior traction on a broad, often dumbbell-shaped region encompassing the macular area and optic nerve [49–52] (Fig. 6). The zone of vitreo-macular attachment can be small as seen in Fig. 6 or measure several disk areas in size [51, 52]. Associated findings include a premacular membrane and varying degrees of macular edema with fluorescein leakage. If VMT is severe, traction macular detachment can result, as seen in Fig. 6. Symptoms tend to progress with time, but a full-thickness macular hole rarely develops [53]. Vitreo-retinal surgery to relieve VMT is often associated with anatomic and visual improvement [50, 51, 54–57]. It has been suggested that traction on the fovea arising from APVD is a contributing pathogenic factor in some cases of CME following cataract surgery [58].

A full-thickness macular hole (FTMH) is defined as a foveal lesion with interruption of all retinal layers from the ILM to the RPE, usually diagnosed by OCT imaging. The exact pathophysiologic mechanism is unknown; however, many theories have been proposed, such as vitreous traction, trauma, foveal degeneration, high myopia, and involutional thinning with PVD. The IVTS differentiated FTMHs based on size (small ≤250 μm, medium >250 to <400 μm, large ≥400 μm), vitreous status (with or without VMA), and associated conditions (primary vs. secondary) [41]. This classification has been shown to be useful as it predicts the anatomic and functional success after treatment with pharmacologic vitreolysis or surgical intervention [2].

There is typically no macular edema in MH because the cysts surrounding the hole are tractional rather than exudative [59–62]. The tangential traction theory [63] suggests that shrinkage of the perifoveal vitreous cortex induces an FTMH in four stages [64]. The possible mechanisms that can cause tangential vitreous traction are: fluid vitreous movement and countercurrents, cellular remodeling of the cortical vitreous, and contraction of a cellular membrane on the tapered cortical vitreous after vitreoschisis [65–67]. Indeed, vitreoschisis has been found in half of the eyes with MH [67]. In addition, vitreo-papillary adhesion (VPA) may play a crucial role in the development of MH, since VPA is present in 88.2 % of MH eyes [68]. VPA is also more prevalent in eyes with intraretinal cystoid spaces in both lamellar macular holes and macular pucker [68, 69]. Thus, it has been suggested that VPA influences the vector of tangential forces on the macula causing outward (centrifugal) traction opening a central dehiscence in an FTMH, foveal splitting in lamellar MH, and the formation of intraretinal cysts. In contrast, macular puckers have a very low incidence of VPA, and the vector of tangential forces is believed to be inward (centripetal).

When anomalous PVD is associated with vitreoschisis, a vitreous membrane remains on the macular surface. The term epiretinal membrane (ERM) is often used in literature to refer to this membrane. However, this term is inaccurate for two reasons: The term “epi-” indicates a location next to or beside a structure, in this case, the retina. Thus, the term “epiretinal” could refer to a subretinal location, which is clearly not the case. In addition, all of the related conditions are maculopathies, not retinopathies. Therefore, the accurate term is “premacular membrane” (PMM), which will be used throughout this chapter. In addition, the term “idiopathic” ERM is no longer appropriate, as we know that the cause of this condition is vitreous.

Following APVD with vitreoschisis, premacular membranes (PMM) can contract and cause metamorphopsia and visual impairment. Light microscopic studies of these membranes have shown the presence of astrocytes and RPE cells. However, there are likely other cells that have a similar morphological appearance, namely, hyalocytes. Zhao et al. examined ILM from 79 patients with macular pucker or VMT and found that hyalocytes are one of the major cell types [70]. These cells can elicit the migration of monocytes from the circulation further increasing the population of mononuclear phagocytes in these membranes. Sebag hypothesized that MP results when VS splits the posterior vitreous cortex anterior to the hyalocytes, leaving a thick and cellular membrane attached to the macula. A recent study found that half of all eyes with MP have more than one site of retinal contraction leading to a higher incidence of intraretinal cysts and significantly more macular thickening [71]. It is not clear whether these cysts are the result of traction as in the case of MH, or exudation and macular edema, which is found in about 40 % of eyes with MP [72].

Premacular membranes (PMM) are associated with macular dysfunction most likely due to an impairment of the neural retinal layers, particularly disruption of the inner and outer segments of photoreceptors [73, 74]. With time, PMM contraction in MP can result in disruption of the blood-retinal barrier in the underlying retina with exudation, identifiable on fluorescein angiography (Fig. 7). The extracellular accumulation of fluid can lead to cystoid changes, particularly in chronic MP. Thus, in addition to the distortions induced by the PMM in MP, ME causes blurred vision subjectively and decreased visual acuity objectively [75].

Vitreous surgery is the definitive cure for MP, with reduction/elimination of distortions [76], resolution of ME, and restoration of normal macular thickness and function. Visual acuity improves in 87 % of cases with restoration of normal macular thickness in 94 % of patients [77, 78]. MP does not presently appear to be treatable by pharmacologic vitreolysis, as the presence of a PMM is a relative contraindication to this form of therapy for vitreo-maculopathies (see below).

It has been shown that full-thickness vitreo-macular adhesion (VMA) and traction (VMT) play a role in exudative age-related macular degeneration (AMD). Anomalous PVD with persistent VMA is believed to promote or at least predispose to choroidal neovascularization, while a total PVD appears to be somewhat protective against wet AMD [79–82]. Clinical studies [79] found a twofold higher prevalence of PVD (diagnosed by ultrasound) in eyes with dry AMD compared to exudative AMD and a fivefold higher prevalence of VMA (diagnosed by OCT) in wet AMD compared to eyes with dry AMD. This was subsequently confirmed in two separate paired eye studies [79, 83] and was recently confirmed by a meta-analysis of 1,025 qualified articles, which found that wet AMD has double the prevalence of VMA and is less likely to have PVD [84]. Thus, although the exact mechanism is yet to be discovered, anomalous PVD seems to be a risk factor for exudative AMD [85–87]. A recent study [88] looked at 30 eyes of 25 patients with VMA surrounded by shallow detachment of the posterior vitreous cortex. The area of adhesion corresponded to a site of retinal angiomatous proliferation (RAP) in 88.2 %. In 73.3 % there was evidence of visible vitreo-retinal traction lines directed toward the choroidal neovascular membrane. Additionally, vitreo-papillary adhesion could be detected on the temporal margin of the optic disk in 83.3 %, about as often as in macular holes [68, 69], and the posterior vitreous cortex seemed to be split in the area of vitreo-papillary attachment, constituting vitreoschisis.

The biochemistry [30] and structure [89] of vitreous is altered by diabetes resulting in diabetic vitreopathy [48], which plays a role in the pathobiology of proliferative diabetic vitreo-retinopathy. Ultrasonography [90] and histopathology [91] have demonstrated vitreoschisis in proliferative diabetic retinopathy as well as in DME [92], which is the most common cause of vision loss in diabetic patients. In DME, cytokines and chemokines such as VEGF, fibroblast growth factor-2, and protein kinase C induce proliferation of hyalocytes and astrocytes in the posterior vitreous cortex [93]. These cells may strengthen adhesion of the posterior vitreous cortex to the ILM and perpetuate VMA, promoting VMT and contributing to DME [93]. Clinical studies have shown that the surgical removal of the posterior vitreous in DME patients contributes to less ischemia and decreased vasopermeability [94–96]. These findings may not only be due to relief of VMT by vitrectomy but also the removal of various molecules that promote DME, since vitreous appears to be a repository for various molecules that promote DME [97, 98]. Recent studies have shown that diabetic patients have more pro-inflammatory molecules in the vitreous body compared to nondiabetics [99]. In particular, IL-1B, caspase 1, and chemokines are increased significantly in vitreous, suggesting that chronic low-grade inflammation may be a factor in the pathogenesis of DME [99]. TNFα, which is also found in vitreous of patients with DME, is known to increase retinal endothelial permeability [100]. Other key molecules thought to promote DME by breakdown of the blood-retinal barrier are endothelial adhesion molecules, such as ICAM1, VCAM1, platelet-endothelial cell adhesion molecule 1 (PECAM-1), and P-selectin [101, 102].

Globally, an estimated 13.9 million people are affected by branch retinal vein occlusion (BRVO) and 2.5 million by central retinal vein occlusion (CRVO) [103], affecting males and females equally with a mean age of 65 years for CRVO. Five percent of RVO occurs in people over 80 years of age [104, 105]. The primary risk factor for RVO is atherosclerosis [105, 106], although arterial hypertension, cardiovascular disease, hyperlipidemia, diabetes mellitus, hypercoagulation, thrombophilia, and inflammatory disease have also been associated with RVO.

Retinal vein thrombosis causes elevation of intravascular pressure in vessels distal to the occlusion. The increased transmural hydrostatic pressure gradient in retinal capillaries results in transudation of fluid into the extracellular space. Associated hypoxia releases cytokines such as VEGF that break down the blood-retinal barrier, causing further leakage in both central and branch RVOs [107–110]. Although the exact etiology of RVO is not understood, vitreous seems to play a role [111]. Vitreo-papillary adhesion has been documented in patients with ischemic central retinal vein occlusion (CRVO), resulting in a secondary serous neurosensory detachment that usually involves the macula [112]. Thick peripapillary tissue may be due to the growth of fibrous tissue in the area of VPA following hemorrhagic CRVO.

Although the current mainstay in treating ME due to RVO is intravitreal anti-VEGF injection [113], vitrectomy has been shown to be relatively beneficial for non-ischemic CRVO [114–116] and BRVO [116–119]. In addition, pharmacologic vitreolysis with autologous plasmin enzyme in eyes with BRVO has resulted in PVD and significant reduction of ME and improvement of VA in 88 % of eyes [119]. These observations suggest a role for the cortical vitreous and PVD status in BRVO. Future studies should explore whether pharmacologic vitreolysis could provide a useful adjunct to other forms of therapy for ME due to RVO.

Chromodissection is a term that refers to the use of dyes to stain the ILM so as to facilitate membrane peeling [120]. This technique is especially useful in macular hole surgery [121], achieving a primary anatomic closure rate of 94 % using indocyanine green (ICG) chromodissection and 89 % with trypan blue [122].

ILM removal by chromodissection remains a controversial topic in macular pucker (MP) surgery. The rationale for ILM removal relates to the high incidence of vitreoschisis in MP and the risk of leaving a layer behind if ILM removal is not performed. Further, removal of ILM may reduce the incidence of MP recurrence due to removal of a scaffold that would enable cells to re-proliferate.

Failure of MH surgery results from the incomplete peeling of the posterior vitreous cortex, the presence of subclinical premacular membrane (PMM) causing a residual traction at the hole, or inadequate gliosis [123, 124]. After initial successful closure, there can also be reopening of an MH, sometimes even years later [122, 123, 125–128]. The presence and progression of PMM formation has been correlated with MH recurrence, the mechanism thought to be that the PMM exerts tangential traction causing foveal dehiscence [123, 124].

Macular edema (ME) is associated with a sevenfold increase in the reopening of MH [129], with inflammatory fibrinolysis considered to be the causative process [122, 128]. Post-vitrectomy cataracts occur in 76 % of cases after 2 years [130–134], and cataract surgery after MH surgery has been associated with the reopening of the hole within 6 months of cataract extraction [122, 125, 128]. The proposed hypothesis for this event is the development of ME and/or PMM formation after cataract surgery due to breakdown of the blood-retinal barrier introducing postoperative inflammatory mediators. In order to avoid these complications, some surgeons have elected to combine MH surgery with phacoemulsification [130–132]. Lastly, YAG capsulotomy of posterior capsular opacification has been implicated in MH reformation [133], with the mechanism thought to be a perifoveal vitreous contraction related to the laser pulse, although it is not clear how this could occur.

Reoperation for recurrent/persistent MH and MP is successful if the cause is incomplete removal of the PMM. In this instance, ILM peeling with chromodissection can be performed with a number of stains such as ICG, trypan blue, triamcinolone acetonide, and Brilliant Blue G [67, 135–138]. In cases where adequate PMM and ILM peeling have been performed in the central macula, it has been suggested that more extensive ILM removal toward the vascular arcades may be an option [139]. However, reoperations can at times be associated with profound loss of vision [140]. Patients usually describe the sudden onset of a scotoma that occurs soon after surgery. There is significant loss of visual acuity, an afferent pupillary defect (APD), and altitudinal visual field loss. This presentation has been identified as resulting from vitrectomy with membrane peeling that has injured the retinal nerve fiber layer, resulting in an inner retinal optic neuropathy (IRON) [141]. Unlike AION, the visual field loss in IRON is not altitudinal, and no disk edema is present [141]. Studies [141] have determined that the risk of IRON is markedly diminished if reoperations are undertaken at least 6 months after the initial surgery. This is presumably due to the protection of the inner retina provided by resynthesis of the ILM, a process that takes many months [142].

The current mainstay of treating ME from RVO is VEGF inhibition with ranibizumab, bevacizumab, or aflibercept. Frequent injections are usually needed to show benefit [143, 144], and rebound edema is sometimes worse than pretreatment edema [145]. In addition, the possible neurotoxic effect of chronic pan-VEGF suppression in an ischemic retina has raised concerns [146]. Other therapies have been advocated such as laser-induced chorio-retinal anastomosis, retinal intravascular tPA infusion, and surgery by radial optic neurotomy or arteriovenous sheathotomy [147]. Considering the aforementioned putative role of vitreous in RVO, there may be value in performing only vitrectomy with surgical induction of PVD or pharmacologic vitreolysis to induce PVD as an adjunct to the medical management of ME from RVO.

DME is the leading cause of vision loss in patients with both types 1 and 2 diabetes. Corticosteroids have been used to suppress inflammatory cell proliferation and migration and block the upregulation and/or activity of many pro-inflammatory cytokines involved in leukocyte stasis and breakdown of the blood-retinal barrier [148], but there are complications that follow repeated steroid injections. VEGF inhibition is an FDA-approved pharmacologic approach for the management of DME. Studies using pegaptanib [149], ranibizumab [150, 151], and bevacizumab [152] have demonstrated favorable clinical responses. It has recently been shown that aflibercept is superior to ranibizumab, which is in turn superior to bevacizumab in the treatment of DME [153].

Vitreous seems to play an important role in DME, both via VMA (see above) and VMT [154, 155]. There is also evidence that even subclinical perifoveal vitreous detachment may play a role in DME [156]. Thus, it would be reasonable to consider that the induction of PVD by pharmacologic vitreolysis could provide a useful adjunct to the medical therapy of DME.

Leakage from choroidal neovascularization (CNV) in neovascular age-related macular degeneration (AMD) results in intraretinal, as well as subretinal, fluid accumulation [157–159]. Studies have found that in exudative AMD, decreased visual acuity correlates with increased macular thickness [159, 160]. Anti-VEGF treatment with intravitreous ranibizumab has resulted in significant reduction of central retinal thickness (which includes both intraretinal and subretinal fluid components) as early as 1 day after the first injection [160]. Clinical trials have shown a significant reduction in central retinal thickness and improvement in visual acuity with anti-VEGF therapy [160]. It is important to note that reduction of macular edema overlying the CNV is not always accompanied by visual gain, most likely due to the atrophy of photoreceptors, loss of RPE, fibrotic scarring, and other irreversible alterations.

The importance of vitreous in the pathophysiology of exudative AMD was described above. Vitreous also appears to play a role in the therapy of exudative AMD [161]. In eyes with VMA, the attached posterior vitreous cortex might create a barrier effect interfering with the influx of nutrients and oxygen to, as well as the efflux of pro-angiogenic cytokines from, the macula [84]. A study in 110 eyes with wet AMD but no VMA found that there was significant improvement in VA after treatment with anti-VEGF injections (P < 0.05) [162], while eyes with VMA did not have a significant improvement. This was confirmed by another study [163] of 255 treatment-naïve subjects, also finding that eyes with PVD improved with fewer injections. Thus, inducing PVD may not only eliminate any contribution of vitreous to the pathophysiology of exudative AMD but may be a way to enhance the pharmacotherapy of AMD as well. This was recently achieved via pharmacologic vitreolysis using ocriplasmin in 100 patients with exudative AMD and VMA. Those patients who achieved VMA resolution required fewer anti-VEGF injections [164]. These findings were recently confirmed by a study of p.r.n. dosing with ranibizumab in 34 treatment-naïve eyes with VMT, compared to 29 treatment-naïve eyes without VMT, as well as to other variable dosing studies (CATT, HARBOR, PrONTO, SUSTAIN). The findings showed that after 1 year of treatment, eyes with VMT had no significant improvement in visual acuity or central macular thickness, while eyes without VMT had significant improvements in both, and the degree of improvement was comparable to those of previous large studies [165].

Given the apparent role of vitreous in both the pathophysiology [166] and therapy [84] of exudative AMD, there may be benefit in eliminating the contribution of vitreous through adjunctive therapy to separate vitreous from the macula. While this can be effectively achieved with vitrectomy, a less invasive approach is preferable, such as with pharmacologic vitreolysis.

Vitrectomy surgery is the mainstay of therapy to eliminate the role of vitreous in various retinal diseases, but it is considerably invasive, expensive, and can lead to complications [141, 167, 168]. Thus, alternative methods for the induction of PVD have been investigated, in particular, pharmacologic vitreolysis [169]. To this end, various agents have been tested over the past 20 years. According to their mechanism of action, these agents can be categorized as “enzymatic” or “nonezymatic.” Enzymatic agents include tissue plasminogen activator (tPA) [170], plasmin [171], microplasmin [172], nattokinase [173], chondroitinase [174], dispase [175], and hyaluronidase [176]. Nonenzymatic agents include urea/Vitreosolve and arginine-glycine-aspartate peptides [177].

Recently, Sebag [177] proposed a classification system in which pharmacologic vitreolysis agents are categorized based on their biologic effect: those that induce vitreous liquefaction (liquefactants) and those that induce dehiscence at the vitreo-retinal interface (interfactants). Table 1 classifies agents based on these biologic activities. Of note is that there have been seven agents tested to date. Five projects have failed or been disbanded, one agent is currently under continuing investigation, and only one agent has been approved for clinical pharmacologic vitreolysis.