Proliferative Vitreoretinopathy: Pathophysiology and Therapeutic Approaches

Abstract

Proliferative vitreoretinopathy (PVR) is a complication of retinal detachment (RD) that is characterized by the development of retinal stiffness and contractile membranes on the surface or underside of the retina. It can occur in primary RD and make repair more challenging, or it can occur following initial successful RD repair and lead to re-detachment. Though our understanding of the pathophysiology underlying PVR membrane formation has grown based on cellular and animal models, there remains no currently approved medical therapy for treatment or prevention of PVR. Though some pharmacologic agents remain under active investigation, many have failed to show consistent benefit in human trials despite promising results from preclinical models. Further research is essential not only to enhance our understanding of PVR pathophysiology but also to identify novel therapeutic strategies for treating PVR in human patients.

Introduction

Though most patients with retinal detachment (RD) have excellent outcomes after surgical repair, 5–10% develop proliferative vitreoretinopathy (PVR), a complication characterized by the development of retinal stiffness and contractile membranes on the surface or underside of the retina. Despite initial successful surgical repair of RD, PVR can contribute to retinal re-detachment, which typically requires re-operation and is associated with worse visual outcomes¹. The incidence of PVR is even higher in the context of trauma; some have estimated it to be as high as 50%^2,3. The clinical appearance of PVR is diverse, and it can occur in numerous clinical settings. Some examples are provided in Figure 1. Though it has been described by many names, such as “massive vitreous retraction,” “massive preretinal retraction,” and “massive periretinal proliferation,” based on clinical observations throughout the latter half of the 20^th century, the standardized term proliferative vitreoretinopathy as well as its initial classification system were first described in 1983⁴.

Figure 1.

Proliferative vitreoretinopathy (PVR) in diverse clinical contexts, including (A) in primary retinal detachment (RD) without prior procedures or surgery, (B) in trauma (i.e., perforating globe injury) where PVR developed predominately at entry/exit sites, and (C) after RD repair (in this case, after scleral buckle/pars plana vitrectomy) leading to re-detachment.

The classification system for PVR was updated by Machemer and colleagues in 1991 and is reproduced in Table 1⁵. In brief, Grade A PVR is characterized by vitreous haze, vitreous pigment clumps, or pigment clusters on the retina. Grade B PVR is defined by wrinkling of the inner retinal surface, retinal stiffness, retinal vessel tortuosity, rolled/irregular edges of retinal breaks, or decreased mobility of the vitreous. Finally, Grade C PVR features full-thickness retinal folds that are subdivided based on both their location (i.e., anterior or posterior to the equator) and their contraction type (i.e., diffuse, subretinal, circumferential, or focal). For Grade C PVR, the number of clock-hours of involvement is also designated. This standardized classification system has been essential for comparison of PVR severity among patients in clinical studies.

Table 1.

Classification System for Proliferative Vitreoretinopathy

Grade	Features
A	Vitreous haze; vitreous pigment clumps; pigment clusters on inferior retina
B	Wrinkling of inner retinal surface; retinal stiffness; vessel tortuosity; rolled or irregular edge of retinal break; decreased mobility of vitreous
C, posterior (1–12)	Posterior to equator: focal, diffuse, or circumferential full-thickness folds; subretinal strands (number of clock-hours of involvement)
C, anterior (1–12)	Anterior to equator: focal, diffuse, or circumferential full-thickness folds; subretinal strands; anterior displacement; condensed vitreous with strands (number of clock-hours of involvement)

Adapted from Machemer et al., 1991⁵

Prior studies have examined the clinical risk factors that predispose individuals to developing PVR. The risk factors are numerous and include but are not limited to aphakia, pre-operative presence of PVR, cigarette smoking, presence of choroidal detachment, giant retinal tears, large area of retinal detachment, large area of retinal break, vitreous hemorrhage, and uveitis (Table 2)^6–9. Notably, almost all identified risk factors except for cigarette smoking are not modifiable. Thus, identification of these risk factors may aid in identifying patients at high risk for developing PVR, but they can rarely be changed to improve a patient’s risk profile.

Table 2.

Clinical Risk Factors for Developing Proliferative Vitreoretinopathy (PVR)

Risk Factors
Aphakia
Pre-operative presence of PVR
Cigarette smoking
Choroidal detachment
Giant retinal tears
Large area of retinal detachment
Large area of retinal break
Vitreous hemorrhage
Uveitis

Pathophysiology of Proliferative Vitreoretinopathy

The complex pathophysiology underlying PVR remains under active investigation. It is hypothesized that a retinal break is a prerequisite for developing PVR. A retinal break leads to intravitreal dispersion of retinal pigment epithelial (RPE) cells and glial cells, as well as breakdown of the blood-retinal barrier. Under the right conditions, these normally polarized and highly organized RPE cells undergo epithelial-mesenchymal transformation (EMT) and acquire a phenotype resembling fibroblasts, leading to extracellular matrix deposition and formation of contractile PVR membranes. Breakdown of the blood-retinal barrier allows for the influx of various cytokines and growth factors that create the pathological biological milieu that promotes inflammation, induces cell migration, and causes extracellular matrix deposition, contributing to the development of PVR. A simplified schematic depicting this hypothesized pathophysiology is shown in Figure 2.

Figure 2.

Simplified schematic of hypothesized pathophysiology underlying development of proliferative vitreoretinopathy (PVR): (A-B) retinal break causing retinal detachment leads to (C) dispersion of retinal pigment epithelium (RPE) cells into the vitreous space where they undergo epithelial-mesenchymal transformation (EMT) in the presence of various molecular mediators and (D) lead to formation of contractile membranes. Figure created with BioRender.com.

Numerous molecular mediators have been implicated as possible contributors to the development of PVR, as they have been identified at higher concentrations in the intraocular fluids of patients with PVR. A list of selected molecular mediators is provided in Table 3, which is reproduced from a recent review on this topic¹⁰. These mediators include inflammatory cytokines, growth factors, immune cell chemokines, and molecules involved in wound healing. Given the extensive number of molecules that have been identified, the mechanisms underlying PVR are complex and likely involve dysregulation of numerous immunologic and cellular pathways. To date, no single causative molecule has been identified.

Table 3.

Molecules Implicated in the Pathophysiology of Proliferative Vitreoretinopathy

Molecular Mediators
Interleukin 1 beta (IL-1β), IL-4, IL-5, IL-6, IL-8, IL-15, IL-16
Tumor necrosis factor alpha (TNF-α)
Transforming growth factor beta (TFG-β)
Vascular endothelial growth factor (VEGF)
Basic fibroblast growth factor (bFGF)
CC/CXC chemokines (CXCL8, CCL2, CXCL8)
Matrix metalloproteinases (MMPs)
Monocyte chemoattractant protein 1 (MCP-1)
Interferon gamma (IFN-γ)
Interferon gamma-induced protein 10 (IP-10)
Lipocalin 2 (LCN-2)

Reproduced from Ananikas et al., 2022¹⁰

There may also be a genetic predisposition for developing PVR. A study in 2010 identified a single nucleotide polymorphism (SNP) in the lymphotoxin alpha (LTA) gene within the tumor necrosis factor (TNF) locus that was significantly associated with PVR¹¹. Subsequently, the Genetics on PVR Study Group identified polymorphisms in four additional genes significantly associated with PVR: tumor protein 53 (P53), murine double minute 2 (MDM2), B-cell lymphoma 2 (BCL2), and Bcl2-associated X protein (BAX)^12–14. While the functional consequences of these polymorphisms remain uncertain, these associations may provide insight into the underlying pathophysiology that contributes to PVR development.

Pharmacologic Treatments for Proliferative Vitreoretinopathy

To date, there is no approved medical therapy for treating or preventing PVR. Here, we discuss the numerous treatments that have been tested. Though many have failed to show a consistent benefit, some remain under active investigation. A summary is provided in Table 4.

Table 4.

Summary of Possible Pharmacologic Treatments for Proliferative Vitreoretinopathy

Drug Class	Drug	Summary of Findings
Corticosteroids	Intravitreal triamcinolone acetonide	No benefit found
	Dexamethasone implant (Ozurdex; Allergan, an AbbVie company)	No benefit found
Anti-vascular endothelial growth factor Agents	Intravitreal bevacizumab	No benefit found
Antimetabolites or Chemotherapeutics	Oral colchicine	No benefit found
	Intravitreal daunorubicin	Contradictory results; not in widespread use
	Intravitreal 5-fluorouracil (5-FU) and low molecular weight heparin	Contradictory results; not in widespread use
	Intravitreal methotrexate	Still under active investigation
Vitamin A Derivatives	Oral 13-cis-retinoic acid	Possible benefit but small study
	Oral isotretinoin	Possible benefit; not in widespread use
Ribozymes	VIT100 (Immusol)	No benefit found

Corticosteroids

Given their widespread use in reducing intraocular and systemic inflammation, steroids have been evaluated as a possible preventive therapy for PVR. In a randomized clinical trial (RCT), 75 eyes with RD and Grade C PVR undergoing vitrectomy with silicone oil (SiO) tamponade were randomized to adjunctive 4 mg triamcinolone acetonide (TA) injected into the silicone oil-filled vitreous cavity versus no TA injection. There was no statistically significant difference in the retinal reattachment rate at 6 months (84.2% in the TA group; 78.4% in the control group; P = 0.50), nor any difference in the secondary outcome measures, such as visual acuity, rate of recurrent PVR, reoperation rate, or rise of intraocular pressure¹⁵.

Another RCT tested the possible role of an adjunctive slow-release dexamethasone implant (Ozurdex; Allergan, an AbbVie company, Chicago, IL) placed at the time of pars plana vitrectomy (PPV) with SiO tamponade for RD with Grade C PVR and at the time of SiO removal¹⁶. The dexamethasone implant did not lead to a statistically significant improvement in anatomic success of surgical repair (49.3% in the dexamethasone group; 46.3% in the control group; P = 0.73), though it was associated with a lower rate of cystoid macular edema.

Anti-Vascular Endothelial Growth Factor Agents

Agents that target vascular endothelial growth factor (VEGF) have revolutionized the therapeutic landscape for patients with exudative age-related macular degeneration and diabetic retinopathy. VEGF has also been implicated in PVR pathogenesis in cell and animal models¹⁷. Therefore, anti-VEGF strategies have emerged as a possible strategy for PVR.

One prospective study randomized 38 eyes with RD and Grade C PVR undergoing PPV/SiO to injection of 1.25 mg bevacizumab into the SiO at the end of surgery versus no injection¹⁸. There was no statistically significant difference in the rate of retinal re-detachment with PVR between the groups (47.3% in bevacizumab group; 36.8% in control group; P = 0.5). Another small study found that adjunctive intravitreal bevacizumab following PPV for eyes with RD with up to Grade B PVR did not lead to improved anatomic success or improved best-corrected visual acuity at either 3 or 6 months¹⁹. Finally, one study explored whether serial intra-SiO bevacizumab injections given intraoperatively and at post-operative months 1, 2, and 3 had benefit for patients undergoing PVR-related surgery²⁰. The study found no improvement in retinal reattachment rate or visual acuity in the bevacizumab group versus control. A meta-analysis of the prospective studies that had been published to date in 2018 reached a similar conclusion of no clear benefit of bevacizumab for treating PVR²¹. Despite promising preclinical data, anti-VEGF strategies do not appear to have benefit for PVR in human studies.

Colchicine

Colchicine is an anti-inflammatory agent that has numerous effects. Among these effects is the ability to inhibit microtubule assembly, interfere with cell proliferation, and prevent cellular migration. Though commonly used for treatment and prevention of gout, it has also been shown in animal models to decrease incidence and severity of tractional retinal detachments²². Oral colchicine was thus explored as a possible therapy for treating PVR associated with proliferative diabetic retinopathy, sickle retinopathy, trauma, and venous occlusive disease, but no significant benefit was found²³. Recently, Ahmadieh and colleagues performed a RCT in which they randomized 184 eyes with RRD undergoing scleral buckle to oral colchicine (1 mg twice daily) versus placebo²⁴. They found no significant decrease in retinal re-detachment rate between the groups (11.8% in colchicine group; 15.3% in the placebo group; P = 0.39).

Daunorubicin

Daunorubicin is a chemotherapeutic agent that inhibits cellular proliferation and migration. The Daunorubicin Study Group explored the utility of adjunctive daunorubicin given during PPV. In an RCT, 286 eyes with Stage C2 or worse PVR undergoing planned PPV/SiO were randomized to standardized surgery with adjunctive daunorubicin perfusion versus surgery alone²⁵. The retinal reattachment rate was not significantly different between the groups (62.7% in the daunorubicin group; 54.1% in the control group; P = 0.07, one-tailed), though the daunorubicin group had significantly less need for reoperations during the one-year post-operative period (65.5% no reoperations in the daunorubicin group; 53.9% in the control group; P = 0.005, one-tailed). A similar study was performed by Kumar et al. in which 30 eyes with RRD and advanced PVR were randomized to intravitreal daunorubicin at the end of PPV versus no injection. Though the study is difficult to interpret due to small sample size and short follow-up period, patients who received daunorubicin had a higher rate of complete retinal attachment at 3 months (86.6% in the treatment group versus 66.6% in the control group).

5-Fluorouracil and Heparin

5-fluorouracil (5-FU) is a cytotoxic chemotherapeutic agent that has also been explored as a possible therapy for preventing and treating PVR. In 2001, a prospective RCT was performed in which 174 patients undergoing primary vitrectomy for RD who were deemed high risk for developing PVR were randomized to 5-FU and low molecular weight heparin (LMWH) versus placebo²⁶. The incidence of postoperative PVR was significantly lower in the 5-FU/LMWH group (12.6% in 5-FU/LMWH group; 26.4% in control group; P = 0.02). However, the PRIVENT trial, a prospective RCT of 325 patients with primary RD deemed at high risk for PVR, did not replicate this result²⁷.

5-FU/LMWH has also been tested in patients with established PVR and found to have no beneficial effect²⁸. Finally, 5-FU/LMWH has been tested as an adjunctive treatment for all patients undergoing primary vitrectomy for RD, not just those at high risk for PVR²⁹. In this study of 641 patients, adjunctive 5-FU/LMWH was not associated with improvement in anatomic or visual success rate or with reduction in failure due to development of PVR. In fact, among patients with macula-on RD, 5-FU/LMWH was associated with worse visual acuity. Given this potential for harm and unclear efficacy, 5-FU and LMWH have not become widely used for preventing or treating PVR.

Methotrexate

Methotrexate is an anti-folate antimetabolite that, among its other uses, is used in ophthalmology for treating intraocular lymphoma. Numerous small studies have described a possible benefit of methotrexate in preventing or treating PVR. Methotrexate has been shown to be capable of reducing growth and inducing cell death of PVR cells cultured in vitro from PVR membranes from human patients³⁰. A recent meta-analysis of 240 eyes with PVR or with high risk of PVR from 8 studies found no significant difference in retinal reattachment rates in eyes that received intravitreal methotrexate. However, this analysis was limited by lack of randomization, variable case selection, and variable dosing regimens within these studies³¹. Aldeyra Therapeutics (Lexington, MA) recently announced in a press release that Part 1 of their Phase 3 GUARD trial achieved its primary endpoint by showing a benefit of intravitreal methotrexate for preventing PVR. In Part 1 of this study, 68 patients received serial injections of intravitreal methotrexate and, relative to historical controls, had statistically significant reduction in retinal re-detachment over the subsequent 6 months (P = 0.024). Of note, though there was “numerical superiority,” there was no statistically significant difference in outcomes between patients randomized to intravitreal methotrexate compared to those that were randomized to routine surgical care in this study. These results have not yet been published in a peer-reviewed journal.

Retinoic Acid

Retinoic acid has been shown to inhibit growth of retinal pigment epithelium (RPE) cells in vitro and thus has been hypothesized to have a potential role in preventing PVR. A prospective RCT tested this possibility by randomizing 35 eyes with primary RD and PVR to oral 13-cis-retinoic acid (RA) at 10 mg twice daily versus no treatment³². Patients who received RA had a higher rate of retinal attachment (93.8% versus 63.2%; P = 0.047) compared to the control group.

The DELIVER trial explored the role of low-dose isotretinoin on PVR. This prospective, open-label study enrolled 51 eyes with recurrent PVR-related RD and 58 eyes with primary RD and high-risk features for developing PVR and assigned them to 20 mg isotretinoin daily for 12 weeks following surgery, comparing their anatomic success rate to historical controls³³. The results suggested that isotretinoin may help reduce the risk of developing PVR, though it has no effect on established PVR. Notably, nearly all patients (97.2%) who received isotretinoin reported the side effect of dry skin and dry mucous membranes.

Ribozymes

VIT100 (Immusol, San Diego, CA) is a ribozyme that targets the gene proliferating cell nuclear antigen (PCNA). Because PCNA is essential for DNA replication, inhibition/degradation of PCNA should theoretically reduce cell proliferation and was thought to have a possible role in preventing PVR. Though this strategy showed promise by preventing and treating PVR in rabbits, it failed to prevent PVR recurrence in patients with Grade C PVR or worse³⁴.

Emerging Approaches

Given the lack of a clear therapeutic strategy for preventing and treating PVR, there are still multiple ongoing efforts to uncover novel therapeutic approaches based on cellular and animal models. Given the hypothesis that EMT of RPE cells may be the pathologic cause of PVR, some efforts have focused on approaches that may prevent this process. As an example, we recently identified that runt-related transcription factor 1 (RUNX1) is highly expressed on surgically removed PVR specimens from humans and may contribute to EMT of RPE cells based on in vitro studies³⁵. We developed a small-molecule inhibitor of RUNX1, Ro5-3335, which we showed could inhibit the development of experimental PVR when delivered topically in a rabbit model³⁵. In rodent models, microRNA 194 (miR-194) and H89, an inhibitor of protein kinase A, have also been shown to attenuate experimental PVR^36,37. Human studies are necessary to determine whether these candidate molecules show similar benefit for patients with PVR.

Conclusions

Surgical innovation has greatly improved outcomes for patients who develop RD. Nonetheless, the development of PVR is a major complication that is often associated with need for re-operation and worse visual outcomes. Despite significant research efforts, we have an incomplete understanding of the pathophysiology underlying PVR. Though many agents have been tested as possible therapies to prevent or treat PVR, none have achieved success in a Phase 3 trial. Continued research efforts in this field are essential.