The vitreoretinal interface plays an important role in the pathogenesis of many retinal disorders. It is a common belief that traction at this interface contributes to the retinal pathology observed in proliferative diabetic retinopathy, proliferative vitreoretinopathy, macular pucker, diabetic macular edema, vitreomacular traction, macular holes, and retinal detachments. Traction is believed to be mediated by the vitreous cortex and by fibrocellular proliferation, and therefore complete removal of the cortical hyaloid is often a principal goal of vitreoretinal surgery (1). The surgical management of these disorders often targets this interface by separating the posterior hyaloid from the internal limiting membrane (ILM), thereby creating a posterior vitreous detachment (PVD). Mechanical separation of the vitreoretinal interface with vitrectomy may not remove all of the vitreous or completely separate the vitreoretinal junction (2), as cortical vitreous fibrils are left behind on the ILM (3). In addition, incomplete removal of the vitreous may result in surgical failure (4).

Vitrectomy with peeling of the ILM necessitates direct mechanical manipulation of the macula, and although it is generally safe, it can potentially result in trauma to the retina (5,6). Mechanical separation of the posterior hyaloid from the retina in adult primates demonstrates incomplete vitreoretinal separation and frequently causes damage to the macula and optic disc, including partial or full thickness foveal tears, damage to the optic nerve, separation of the ILM from the retina, and avulsion of the ganglion cells and other retina layers (2). Therefore, the development of a pharmacologic agent that can aid in vitreous liquefaction and PVD creation may facilitate surgical separation of the posterior hyaloid, thus reducing intraoperative time and potential complications. Pharmacologic vitreous liquefaction may help to reduce vitreous viscosity, thereby facilitating removal during vitrectomy and reducing surgical time, especially when using smaller gauge instruments such as 25-gauge and 23-gauge vitrectors.

Intraocular enzymatic agents have been discussed for many years, one of the earliest being alpha-chymotrypsin and its effect on zonular proteins during intracapsular cataract extraction. The search for an appropriate agent to manipulate the vitreous has included several enzymes, most of which are autologous enzymes that activate other endogenous enzymes. Unfortunately, many of the earliest studied agents were fraught with failure due to lack of efficacy, retinal toxicity, or difficult preparation. Microplasmin is a recombinant human enzyme that appears to be a promising agent for pharmacologic manipulation of the vitreous. It has recently been shown to cause vitreous liquefaction and cleavage of the vitreoretinal interface with a single intravitreal injection (7, 8, 9 and 10). Pharmacologic alteration of the vitreous with microplasmin and other agents is an evolving modality that will likely be used more frequently for therapy and preventative measures (11).

In the past, the vitreous was considered unimportant in the development of vitreoretinal pathology (7). With recent advancements in biochemistry and vitreous anatomy, we are now able to appreciate the complex arrangement of the vitreous gel and its influence on retinal diseases. The vitreous is a clear, semisolid gel containing hyaluronic acid interspersed in a framework of parallel collagen fibrils coursing in an anteroposterior direction (7). Posterior to the pars plana, the concentration of collagen and hyaluronic acid is greatest in the vitreous cortex, which lies along the inner retinal surface (12). The collagen fibrils, which are condensed to form an outer layer of the vitreous cortex, are adherent to the ILM of the retina (12). Glycoproteins such as laminin and fibronectin are located at the vitreoretinal junction (13) and are believed to contribute to the adhesion of the posterior vitreous cortex to the ILM.

As the human vitreous ages, syneresis and liquefaction of the gel occur with the development of pools of fluid usually in the premacular region or in the central part of the vitreous cavity. Foos and Wheeler (14) found a strong correlation between increasing amounts of vitreous syneresis and the prevalence of PVD. This suggests that the human vitreous gel can tolerate only a certain amount of liquefaction and instability before PVD occurs. Age-related PVD usually occurs as an acute event. A tear in the posterior cortical vitreous allows fluid from the central part of the liquefied vitreous gel to then pass through the break in the vitreous cortex and separate the surrounding cortical vitreous from the retina.

Similar to the development of an endogenous age-related PVD, the success of pharmacologic vitreolysis depends on simultaneous vitreous liquefaction and separation of the vitreoretinal interface. Sebag (7,15) has used the term “anomalous PVD” to describe the situation in which these two processes are uncoupled, whereby the extent of vitreous liquefaction exceeds the degree of vitreoretinal interface weakening. Anomalous PVD may lead to traction at the vitreoretinal interface and subsequent pathologic conditions including vitreomacular traction syndrome, macular holes, and retinal tears (15).

Prior to the advent of microplasmin, multiple other pharmacologic agents have been developed for enzymatic manipulation of the vitreous. Enzymes such as dispase, chondroitinase, hyaluronidase, tissue plasminogen activator (tPA), and plasmin have had variable success (7,16). Dispase was initially believed to be a good candidate for pharmacologic vitreolysis due to its ability to hydrolyze several proteins including type IV collagen and fibronectin. In fact, dispase can lead to the creation of a PVD (17,18); however, it causes anterior chamber and vitreous inflammation, epiretinal membranes, preretinal and intraretinal hemorrhages, cataract, electroretinogram (ERG) amplitude reductions, and ultrastructural damage to
the retina (17, 18 and 19). This intraocular toxicity has limited its clinical utility. Chondroitinase lyses the proteoglycan chondroitin sulfate, which is associated with the vitreoretinal interface (20). A recent masked, placebo-controlled, in vivo study concluded that chondroitinase failed to produce a PVD (21).

Hyaluronidase (Vitrase) is a highly purified ovine enzyme that primarily digests the proteoglycan hyaluronan, which constitutes a large component of the vitreous body. Hyaluronidase has been suggested as an agent to liquefy the central vitreous with the assumption that it may also lead to a PVD after an extended period of time. Hyaluronidase was originally targeted toward patients with dense vitreous hemorrhages in the hope of causing vitreous liquefaction (and settling of blood) to allow for laser photocoagulation in cases of proliferative diabetic retinopathy. A recent prospective, double-masked phase III trial concluded that hyaluronidase is more effective than placebo injections in clearing vitreous hemorrhages (22). Although it has been shown to decrease vitreous macromolecule size suggesting a role for vitreous liquefaction (7), other reports have concluded that intravitreal injection of hyaluronidase cannot induce a PVD in animal models (10,23 and 24). As mentioned earlier, vitreous liquefaction without simultaneous separation of the vitreoretinal interface may induce untoward effects, including vitreomacular traction syndrome, macular holes, and retinal tears (7,15). Nevertheless, a recent phase III trial concluded that there were no serious safety issues and the incidence of retinal detachment was not statistically different between hyaluronidase and controls (25).