Posterior Vitreous Detachment Determines the Clinical Impact of Fibrovascular Membrane Fibrosis After Anti-VEGF Treatment in Proliferative Diabetic Retinopathy: A Real-World Study

Abstract

Purpose

To investigate the relationship between posterior vitreous detachment (PVD) status and fibrovascular membrane (FVM) angio-fibrotic transformation following anti-VEGF therapy in advanced proliferative diabetic retinopathy (PDR), and their combined impact on disease progression and visual outcomes.

Methods

This prospective real-world study enrolled 165 treatment-naïve PDR patients with vitreous hemorrhage (VH) and FVM who received intravitreal anti-VEGF injection followed by panretinal photocoagulation(PRP) when possible. PVD status and FVM characteristics were monitored during a median follow-up of 13.0 months. Cox proportional hazards and Kaplan-Meier analyses evaluated factors influencing visual prognosis and timing of surgical intervention.

Results

FVM fibrosis occurred in 43% of patients, with distinct PVD-dependent patterns: predominantly in complete PVD cases, moderately in partial PVD, and rarely without PVD. Multivariate analysis identified macular involvement as the strongest risk factor for poor visual outcomes (HR=3.65), while PVD (HR=0.37) and FVM fibrosis (HR=0.023) were strongly protective. Patients with both PVD and FVM fibrosis maintained stable vision longest (median 28.5 months) and demonstrated significantly longer surgery-free survival (28.3 months) compared to those without either condition (3.2 months, p<0.0001), who frequently presented with vitreoschisis and recurrent VH disproportionate to visible FVM grade.

Conclusion

PVD status critically determines the clinical manifestation and prognosis of post-anti-VEGF FVM fibrosis in PDR. This PVD-dependent pattern challenges the conventional view that FVM fibrosis inevitably leads to “Crunch syndrome” and creates a risk stratification framework that can guide individualized surgical timing decisions, potentially improving functional outcomes in advanced PDR.

Plain Language Summary

Proliferative diabetic retinopathy (PDR) is a serious complication of diabetes and a leading cause of blindness. It often involves the growth of abnormal blood vessels and membranes inside the eye, which can cause severe vision problems. Doctors frequently treat this condition by injecting medications called anti-VEGF agents, which slow the growth of these harmful vessels. However, sometimes these treatments cause the abnormal vessels and membranes to become fibrotic—meaning they harden and contract, potentially causing further issues.

This study aimed to understand how a condition called posterior vitreous detachment (PVD), where the gel-like substance in the eye separates from the retina, influences the outcomes after anti-VEGF treatment. Researchers observed 165 patients with severe PDR who received anti-VEGF injections and monitored how their conditions changed over approximately one year.

The study found that nearly half the patients experienced membrane fibrosis after treatment. Importantly, patients with a complete vitreous detachment showed better results, including less tractional damage to the retina and longer periods without needing surgery. Conversely, patients without vitreous detachment faced more severe complications, needed surgery sooner, and had worse vision outcomes.

These results indicate that the presence of a vitreous detachment significantly impacts whether fibrosis will be beneficial or harmful after treatment. Understanding this relationship helps doctors predict outcomes more accurately and plan treatments better. Ultimately, this could lead to improved vision preservation and fewer surgical interventions for people with severe diabetic retinopathy.

Introduction

Proliferative diabetic retinopathy (PDR) remains a leading cause of severe visual impairment worldwide, particularly in China,¹ where many patients present late with advanced disease characterized by severe vitreous hemorrhage (VH) and extensive fibrovascular membranes (FVM).^2,3 Following intravitreal anti-VEGF therapy, FVM frequently undergoes angio-fibrotic transformation-a transition from vascular proliferation to fibrosis.^4,5 While preliminary evidence suggests potential surgical benefits such as reduced intraoperative bleeding⁶ and easier membrane segmentation and delamination,⁷ the overall impact on the natural course of PDR remains unclear.

Current literature presents conflicting views on FVM fibrotic transformation. Some studies describe “Crunch syndrome” with rapid tractional retinal detachment (TRD) progression and poor visual outcomes,^8–11 while others report spontaneous TRD resolution following anti-VEGF therapy.^12,13 Most randomized controlled trials involve patients previously treated with panretinal photocoagulation (PRP),^11,14 limiting applicability to treatment-naïve advanced cases common in China. Additionally, studies conducting vitrectomy shortly (3–7 days) after anti-VEGF injection^15,16 provide insufficient observation time to assess the natural evolution of fibrotic transformation.

Anatomically, the posterior vitreous cortex, FVM, and retina form a closely connected complex.^17,18 However, no systematic studies have thoroughly explored whether FVM fibrosis, together with posterior vitreous detachment (PVD) status, collaboratively influences the progression and clinical outcomes of PDR. The potential relationship between PVD status and the clinical manifestations of FVM fibrosis represents a critical knowledge gap in our understanding of PDR pathophysiology and management.

Therefore, this prospective real-world study investigates FVM angio-fibrotic transformation following anti-VEGF therapy and its association with PVD status in treatment-naïve advanced PDR patients. By evaluating how these factors influence visual prognosis and surgical timing, we aim to provide insights for clinical decision-making in populations where advanced PDR presentation is common.

Materials and Methods

This prospective cohort study enrolled patients with PDR complicated by VH who consecutively presented to Beijing Tongren Eye center’s retinal clinic between August 2022 and February 2024. This study was approved by the Beijing Tongren Hospital ethics committee and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Inclusion and Exclusion Criteria

Inclusion Criteria

(1) age above 18 years; (2) VH persisting for more than 2 months or recurring more than twice; (3) clinical diagnosis of PDR, or in cases where direct fundus examination was precluded by VH, PDR diagnosis supported by B-ultrasound and contralateral eye examination and subsequently confirmed after PPV or blood absorption; (4) ability to give signed informed consent.

Exclusion Criteria

(1) the presence of severe systemic disease precluding treatment tolerance; (2) an intravitreal injection administered within the past month; (3) neovascular or ghost cell glaucoma refractory to medical treatment; (4) a tractional or rhegmatogenous retinal detachment involving the macula that requires immediate surgery6,19; (5) complete PVD with both adequate PRP treatment and stable FVM for over 6 months duration; (6) previous vitrectomy; (7) corrected VA endophthalmitis, severe complications, or withdrawal from treatment.

Baseline Examinations

All patients underwent comprehensive baseline examination including corrected visual acuity (corrected VA) measured using Snellen decimal visual acuity chart, lens status, intraocular pressure, slit-lamp biomicroscopy, and dilated fundus examination. When fundus visualization was limited by VH, B-scan ultrasonography was performed to assess retinal status and rule out retinal detachment. Spectral-domain optical coherence tomography (OCT) was performed when media clarity permitted.

Treatment Protocol

All patients received intravitreal anti-VEGF injections using ranibizumab (0.5 mg/0.05 mL, Lucentis, Novartis). All patients initially received a single intravitreal anti-VEGF injection and were followed up at weeks 2, 4, 8, 12, 16, 20, and 24. Supplementary PRP or repeated anti-VEGF injections were administered as necessary during follow-up. FVM angio-fibrotic transformation and PVD occurrence were monitored, and the need for PPV (PPV) was assessed based on disease progression. Standard 23-gauge microincision vitrectomy was performed when indicated, with intraoperative documentation of PVD status, FVM grading, retinal detachment, and macular involvement.

Key Definitions

Macular Involvement

TRD within 500μm of the foveal center or FVM covering the macular region.¹⁹

PVD

PVD was defined as meeting any of the following criteria^18,20,21: (1) clearly visible Weiss ring on indirect ophthalmoscopy; (2) evident PVD around organized membranes with clear separation from the retina; (3) PVD confirmed by triamcinolone acetonide staining during surgery; (4) PVD band identified on B-ultrasound.

Fibrovascular Membrane (FVM) Grading^20,22

Grade 0, no adhesions; Grade 1, multiple point adhesions and/or one broad adhesion (≥3 focal adhesions); Grade 2, 1–3 broad posterior to equator adhesions; Grade 3, >3 broad posterior to equator adhesions or ≤2 quadrants of broad anterior to equator adhesions; Grade 4, multiple broad anterior to equator adhesions.

FVM Angio-Fibrotic Transformation

Transformation of FVM from predominantly vascular proliferation to predominantly fibrous tissue proliferation following intravitreal anti-VEGF injection.^4,5

Crunch Syndrome

FVM fibrosis with worsening TRD following intravitreal anti-VEGF injection or PRP treatment.^8–11,23

Outcome Measures

The primary outcome measure was corrected visual acuity (corrected VA) at the final follow-up. Corrected VA was measured using Snellen decimal visual acuity chart, with measurements performed under consistent lighting conditions by trained ophthalmic technicians. Snellen decimal visual acuity values were converted to logMAR for statistical analysis. Secondary outcome measures included FVM angio-fibrotic transformation and PVD occurrence with timing, the number of anti-VEGF injections, laser treatment and PPV procedures during the treatment course, changes in VH, and ocular complications.

Study Endpoints

(1) follow-up until January 20, 2025, or the final visit; (2) withdrawal due to worsening systemic disease.

Statistical Analysis

Data were collected using EPIDATA and analyzed using R software (Version 4.4.2). Visual acuity (VA) results were analyzed using logMAR. Normally distributed continuous variables were presented as mean ± standard deviation, while non-normally distributed continuous variables were presented as median (IQR (IQR)). Continuous variables were compared using t-tests or Mann–Whitney U-tests, and categorical variables were analyzed using chi-square tests or Fisher’s exact tests.

Poor visual outcome was defined as final VA below 0.05 (≤3/60), and failure to achieve expected visual prognosis was defined as stabilized or improved VA but <0.1, or deterioration of VA compared to baseline for more than 3 lines. Follow-up duration in months was used for survival analysis. Cox proportional hazards models were employed to evaluate factors affecting visual prognosis, and Kaplan-Meier survival curves were constructed to analyze the impact of key variables.

Results

Baseline Characteristics

This study enrolled 165 patients with PDR complicated by VH (Table 1). The cohort had a mean age of 52.5±11.4 years and was predominantly male (60%). Most patients had longstanding diabetes (median duration 10 years) with significant systemic comorbidities, including hypertension (68.5%) and chronic kidney disease (24.8%).

Table 1.

Baseline Characteristics of the Study Population

Characteristic	Value
Demographics
Age (years), mean ± SD	52.5 ± 11.4
Male sex, n (%)	72 (60.0)
Diabetes Characteristics
Diabetes duration (years), median[IQR]	10 [5,17]
Type 2 diabetes, n (%)	115 (95.8)
Systemic Comorbidities
Hypertension, n (%)	113 (68.5)
Cerebrovascular accident (%)	44 (26.7)
Coronary heart disease (%)	32 (19.4)
Chronic kidney disease, n (%)	41 (24.8)
Baseline Ocular Characteristics
Anti-VEGF injections, n (%)	35 (21.2)
PRP not performed, n (%)	106 (64.2)
Pseudophakic, n (%)	19 (11.5)
Initial corrected VA, n (%)
<0.05	95 (57.6)
0.05–0.1	29 (17.6)
>0.1	28 (24.8)
NVI, n (%)	18 (10.9)
Dense vitreous hemorrhage, n (%)	95 (57.6)

Abbreviations: corrected VA, corrected visual acuity; IQR, interquartile range; PRP, panretinal photocoagulation; NVI, neovascularization of iris.

Notably, nearly two-thirds of patients (64.2%) had not received prior photocoagulation, and only 21.2% had received anti-VEGF therapy. Visual impairment was severe at baseline, with 57.6% having corrected VA (Table 1).

Follow-Up Characteristics

The median follow-up duration was 13.0 months [IQR: 6.8–19.6]. Fourteen patients (8.5%) received ≥3 anti-VEGF intravitreal injections. During follow-up, 57 patients (34.5%) underwent partial PRP and 37 (22.4%) completed PRP. Among all patients receiving PRP treatment (n=94), only 5 (5.3%) required repeated anti-VEGF injections, while the remaining patients received PRP alone. The median follow-up was 13.0 months [IQR: 6.6–19.1] in the patients underwent PPV and 13.0 months [IQR: 7.4–21.2] in the patients without PPV.

Angio-Fibrotic Switch in FVM and Its Impact on PDR Progression in Different PVD States

During follow-up, PVD status was documented in all patients, with 33 (20.0%) having complete PVD, 81 (49.1%) having partial PVD, and 46 (27.9%) without PVD at the final assessment. Meanwhile, 71 patients (43.0%) experienced angio-fibrotic switch in FVMs, with 10 patients (6.1%) developing Crunch syndrome. The clinical manifestations and impact on PDR progression varied significantly with PVD status (p<0.001).

In the complete PVD group (n=33), the incidence of FVM angio-fibrotic switch was highest (81.1%), predominantly manifesting as FVM separating from previously attached retina (17 cases) or isolated FVM (13 cases) separating from posterior vitreous cortex, with rare progression of TRD (1 case of tractional tear).

In the partial PVD group (n=81), FVM angio-fibrotic switch occurred in 40.7% of patients, primarily presenting as isolated FVM with vitreous separation (5 cases) and loosened FVM-retinal adhesion without TRD progression (23 cases). Crunch syndrome was observed in 5 cases: 4 with expanding FVM accompanied by TRD progression and 1 with tractional tear development.

In the group without PVD (n=46), the incidence of FVM angio-fibrotic switch was lowest (17.4%). Among the 8 cases with FVM angio-fibrotic switch, 4 exhibited FVM fibrosis contraction without TRD progression, while 4 developed Crunch syndrome with expanded FVM and TRD progression. The remaining 38 patients (82.6%) who presented with a clinical paradox: all had multiple sites of mild grade 1 FVM (100%, 38/38), yet most experienced recurrent vitreous hemorrhage (73.7%, 28/38) and pre-retinal or sub-internal limiting membrane hemorrhage (78.9%, 30/38)—a severity of hemorrhagic manifestations that could not be adequately explained by the relatively minimal FVM observed. During follow-up, their condition deteriorated with FVM progression to grade 2–3 (76.3%, 29/38) and continued recurrent hemorrhages (86.8%, 33/38).

Following initial anti-VEGF intravitreal injection, the median time to development of FVM angio-fibrotic switch was 12 weeks (IQR: 8–16), to Crunch syndrome (10 cases) was 12 weeks (IQR: 4–14), and to complete separation of fibrotic FVM from retina (35 cases) was 13 months (IQR: 11–14).

Factors Influencing Final Visual Acuity

Multivariate Cox analysis demonstrated that macular involvement (HR=3.65,95% CI:1,39–9.55, p=0.008) was an independent risk factor for poor visual prognosis. Protective factors included PVD development (HR=0.37, 95% CI 0.16–0.83, p=0.016), FVM angio-fibrotic switch (HR=0.023,95% CI 0.002–0.30, p=0.004), and PPV (HR=0.024,95% CI: 0.002–0.29, p=0.003) (Concordance=0.82, p<0.01) (Figure 1).

Figure 1.

Forest plot of hazard ratios for poor visual outcome in PDR patients with vitreous hemorrhage. The horizontal lines represent 95% confidence intervals. Asterisks indicate statistical significance: *p<0.05, **p<0.01.

Impact of FVM Angio-Fibrotic Switch and PVD on Visual Prognosis

Kaplan-Meier analysis revealed that patients with concurrent FVM angio-fibrotic switch and PVD maintained the best visual outcomes, with median duration of corrected VA >0.05 reaching 28.5 months. Conversely, patients without PVD or FVM angio-fibrotic switch exhibited the poorest visual prognosis, with median duration of corrected VA >0.05 limited to 6.3 months (p=0.0011) (Figure 2).

Figure 2.

Kaplan-Meier analysis of survival time for maintaining corrected VA >0.05 considering PVD and FVM fibrosis.

Impact of FVM Angio-Fibrotic Switch and PVD on Surgical Timing

A significant interaction between PVD and FVM angio-fibrotic switch was associated with a delayed need for surgical intervention (Log rank test, p<0.0001). Patients with both PVD and FVM angio-fibrotic switch (n=63) demonstrated the longest median PPV-free survival period of 28.3 months (95% CI: 23.5–33.1). Within this group, 52 patients (82.5%) received partial or complete PRP, 46 (73.0%) experienced VH resolution, and PDR stabilization was achieved.

In contrast, patients with PVD but without FVM angio-fibrotic switch (n=56) and those with FVM angio-fibrotic switch but without PVD (n=8) had shorter surgery-free survival periods of 6.7 months (95% CI: 5.1–8.3) and 7.9 months (95% CI: 5.6–10.2), respectively. Patients lacking both PVD and FVM angio-fibrotic switch (n=38) had the shortest median PPV-free survival of only 3.2 months (95% CI: 2.3–4.1). Patients lacking both PVD and FVM angio-fibrotic switch (n=38) had the shortest median PPV-free survival of only 3.2 months (95% CI: 2.3–4.1) (Figure 3).

Figure 3.

Kaplan-Meier analysis of surgery-free survival time (PPV-free survival) considering PVD and FVM fibrosis.

Follow-Up Outcomes

At final follow-up, corrected VA was <0.05 in 26 patients (15.8%), between 0.05–0.1 in 20 patients (12.1%), between 0.1–0.5 in 95 patients (57.6%), and >0.5 in 24 patients (14.5%). Vision improved by ≥3 lines in 96 patients (58.2%), while 136 patients (82.4%) achieved stable visual acuity of ≥0.1 or showed improvement.

PPV was performed in 123 patients (74.5%). Complications included sustained intraocular pressure elevation requiring either two or more intraocular pressure-lowering medications for over 2 weeks (4 cases) or selective laser trabeculoplasty (2 cases) (6 cases total), postoperative VH (10 cases), diabetic macular edema (12 cases), and outer retinal atrophy (28 cases).

Discussion

This prospective study demonstrates how posterior vitreous detachment (PVD) status fundamentally determines the clinical trajectory of fibrovascular membrane (FVM) fibrosis following anti-VEGF therapy in treatment-naïve PDR patients. Rather than uniformly causing tractional complications as traditionally assumed, FVM fibrosis manifests differently across PVD states: protective separation in complete PVD versus continued progression without PVD. This PVD-dependent pattern provides a new framework for risk stratification and surgical decision-making in advanced PDR.

Study Design and Characteristics

This investigation employed a Real-World Study design focusing on treatment-naïve PDR patients with VH and FVM who had not received prior anti-VEGF therapy or PRP. This population represents the predominant presentation pattern in Chinese clinical practice, differing significantly from PDR cohorts in developed countries where early intervention is more common.^13,24,25 This distinction enhances the external validity of the findings for regions with similar healthcare access patterns.

To strengthen internal validity while maintaining clinical relevance, carefully considered exclusion criteria were implemented: (1) patients requiring immediate vitrectomy due to macular-involving retinal detachment, allowing sufficient observation time for FVM and PVD changes; (2) patients with already stabilized disease, to prevent confounding from previous treatment effects; and (3) patients with absolute contraindications to vitrectomy, who would require atypical management strategies and present high dropout risk.

This methodological approach ensured an adequate observation window (median follow-up: 13.0 months [IQR: 6.8–19.6]) to capture the complete evolution of post-anti-VEGF combined with PRP changes, including FVM contraction (median time: 3 months [IQR: 2–4]), development of PVD, and long-term outcomes.

Characteristics of FVM Fibrosis in Chinese PDR Patients Following Anti-VEGF Therapy

The study identified an overall incidence of 43.0% for FVM angio-fibrotic switch following anti-VEGF therapy. The median occurrence time was 12 weeks (IQR: 8–12 weeks) for FVM fibrosis without TRD progression (61/71) and 12 weeks (IQR: 4–14 weeks) for Crunch syndrome with TRD progression (10/71)—both later than the approximately 13 days reported in previous studies.^11,14,26 FVM contraction leading to separation from retina (35/71) required an extended timeframe (median 13.0 months, IQR: 6.8–19.6 months). Crunch syndrome constituted 6.1% of all FVM fibrosis cases, consistent with previously reported rates (1.5%-18.4%).^26–28

Four factors likely contributed to this delayed FVM fibrotic response: (1) High prevalence of severe VH masked early fibrotic changes despite initiating observations at two weeks post-injection; (2) More extensive FVM with stronger retinal adhesions diminished single anti-VEGF injection efficacy, with circular or large-area membranes limiting TRD progression; (3) Combined PRP therapy (56.9% of patients) potentially counteracted the acute FVM fibrotic response caused by anti-VEGF treatment. Unlike anti-VEGF agents which rapidly suppress neovascularization (with studies like Protocol T²³ and CLARITY²⁹ showing significant NVE regression within 4 weeks of injection), PRP exerts a mild and gradual inhibitory effect on NVE, as demonstrated by Zhou et al³⁰ who reported only 60% regression of NVE at 3 months post-PRP and the meantime interval for complete NVD regression was 15.2±3.5 weeks,³¹ and by Gross et al²³ who observed complete NVE regression requiring 6–12 months in most cases with PRP. Additionally, Bressler et al¹² showed that 54% of eyes in the PRP group exhibited some level of NV regression at 2 years after PRP, while Gross et al²⁴ documented 70% regression at five years after PRP. Similarly, Bressler et al²⁵ reported that TRD following PRP primarily occurs as a delayed complication (2.6% within 60 days), with a cumulative probability of 6.8% at one year. (4) Lower anti-VEGF injection frequency (median: one injection; 8.5% receiving ≥3 injections) reduced Crunch syndrome risk. Tan et al’s systematic review⁸ identified multiple anti-VEGF injections as a significant risk factor for Crunch syndrome, with higher cumulative doses accelerating fibrotic contraction.

These findings suggest that current follow-up protocols for PDR patients should be extended. While monitoring during the previously reported 2–4 week post-injection period when Crunch syndrome typically occurs in other studies remains important, the data show that PDR patients need follow-up for at least 3–6 months, with special attention to weeks 8–12 when delayed fibrotic changes typically appear in this cohort. The 13-month median time to FVM separating from retina further indicates that assessment of complete treatment outcomes requires much longer observation periods than previously recognized.

Impact of Post-Anti-VEGF Fibrotic FVM-Posterior Vitreous Cortex Complex Changes on PDR

The results reveal PVD status as the decisive factor in determining how the fibrotic FVM-posterior vitreous cortex complex influences PDR progression. Significantly higher rates of FVM angio-fibrotic switch were observed in patients with complete or partial PVD (81.1% and 40.7%, respectively). Notably, as PVD occurred at FVM attachment areas during treatment, fibrotic FVM predominantly manifested in two distinct patterns: separation from the retinal surface (45.9%), where the fibrotic FVM detached from its retinal attachments while maintaining connection to the posterior vitreous cortex, or as isolated fibrous tissue that remained adherent to the retinal surface but detached from the posterior hyaloid interface (35.1%). Both manifestations were rarely accompanied by TRD progression, suggesting a benign evolution pattern. These observations challenge the conventional assumption that FVM fibrosis inevitably leads to TRD progression, indicating instead that PVD status critically determines whether fibrotic transformation will result in disease improvement in PDR. The benign presentation of fibrotic FVM in patients with PVD contrasts markedly with the classical Crunch syndrome characterized primarily by TRD progression.^8,23,32,33 While previous studies have observed that fibrotic FVM may separate from the retina, potentially reducing TRD,¹³ and that post-anti-VEGF reduction in FVM-retinal adhesion can facilitate surgical membrane peeling,²⁸ they did not identify PVD status as a key determinant in this process. The current investigation identifies PVD as essential for facilitating the favorable separation of fibrotic FVM from retinal or posterior hyaloid surfaces, providing evidence from non-PPV cohorts for a relationship previously only suggested in post-surgical analyses.

In contrast, patients with partial or absent PVD demonstrated a greater tendency for fibrotic FVM-posterior vitreous cortex complexes to manifest as Crunch syndrome. Although the overall incidence was relatively low (6.2% and 8.7%, respectively), these cases typically presented with more severe clinical manifestations, frequently accompanied by significant TRD progression or retinal tear formation. Previous research has established the association between incomplete PVD and both PDR progression and TRD exacerbation.^17,30 The data further suggest that the combination of fibrotic FVM contraction with incomplete PVD may significantly intensify tractional forces on the retina, accelerating TRD development or even inducing retinal tears, thereby leading to rapid disease deterioration.

Additionally, a distinct clinical pattern emerged among patients without PVD—a condition consistent with vitreoschisis frequently observed in PDR patients.^18,34 In this group (n=46), the incidence of FVM angio-fibrotic switch was markedly lower (17.4%) compared to patients with complete or partial PVD. Among the 8 cases with FVM angio-fibrotic switch, 4 exhibited FVM fibrosis without TRD progression, while 4 developed Crunch syndrome with expanded FVM and TRD progression. Of particular interest were the remaining 38 patients (82.6%) who presented with a clinical paradox: all had multiple sites of mild grade 1 FVM (100%, 38/38), yet most experienced recurrent vitreous hemorrhage (73.7%, 28/38) and pre-retinal or sub-internal limiting membrane hemorrhage (78.9%, 30/38)—a severity of hemorrhagic manifestations that could not be adequately explained by the relatively minimal FVM observed. This discordance between mild visible FVM and severe hemorrhagic presentations suggests an underlying vitreoschisis, where neovascularization remains concealed within split posterior vitreous layers.³⁵ During follow-up, their condition deteriorated with FVM progression to grade 2–3 (76.3%, 29/38) and continued recurrent hemorrhages (86.8%, 33/38), supporting the hypothesis that undetected neovascularization remained tightly adherent to the posterior vitreous interface. Kaplan-Meier analysis revealed these patients had both the poorest visual outcomes (median duration of corrected VA >0.05 was only 6.3 months; p=0.0011).

These findings collectively suggest that comprehensive assessment of both PVD status and FVM characteristics should be integrated into the clinical evaluation and management planning for PDR patients receiving anti-VEGF therapy, potentially allowing for more personalized treatment approaches and improved visual outcomes.

Factors Influencing Visual Prognosis and Special Populations Requiring Attention

Multivariate Cox analysis revealed that both FVM angio-fibrotic switch (HR=0.023, 95% CI: 0.002–0.30, p=0.004) and PVD (HR=0.37, 95% CI: 0.16–0.83, p=0.016) serve as independent protective factors for visual function, significantly reducing the risk of visual deterioration. Conversely, macular involvement (HR=3.65, 95% CI: 1.39–9.55, p=0.008) emerged as a strong independent risk factor for poor visual outcomes. These findings establish a clear hierarchy of prognostic factors in PDR patients receiving anti-VEGF therapy, with macular status, PVD development, and FVM fibrotic transformation collectively determining visual prognosis.

The protective effect of PVD and FVM fibrosis was further substantiated by Kaplan-Meier survival analysis, which demonstrated striking differences in visual preservation between patient subgroups. Patients with concurrent FVM fibrosis and PVD maintained vision the longest (median survival period for vision >0.05 was 28.5 months, 95% CI: 24.1–32.9), while those without either fibrotic FVM or PVD experienced the most rapid visual decline (median visual survival period of only 6.3 months, 95% CI: 4.8–7.9; p=0.0011). This marked disparity in outcomes suggests that the combination of PVD and FVM fibrosis creates particularly favorable vitreoretinal interface dynamics.

The clinical significance of these findings becomes apparent when considering their implications for surgical decision-making. Integrating subgroup analyses revealed three distinct patient populations with different management needs:

1. Patients lacking both PVD and FVM fibrosis demonstrate the highest risk profile as previously discussed. Given their short surgery-free interval (median 3.2 months) and poor visual prognosis (median duration of vision >0.05 only 6.3 months), early aggressive vitrectomy is warranted for this group to prevent irreversible visual damage—a management strategy supported by similar findings in vitreoschisis-predominant PDR cases reported by Schwartz et al.¹⁷

2. Patients with both PVD and FVM fibrosis represent the opposite end of the spectrum, exhibiting a benign clinical course characterized by separation of FVM from the retina or posterior vitreous cortex. These patients maintain stable vision for extended periods (median 28.5 months), with most successfully receiving PRP treatment (82.5%) and experiencing significant VH absorption (73.0%). This favorable natural history aligns with Ono et al’s observations¹⁸ regarding the protective effect of PVD against PDR progression, supporting a more conservative approach with delayed vitrectomy and close follow-up.

3. Patients with either PVD or FVM fibrosis (but not both) occupy an intermediate risk position. While their clinical course necessitates eventual vitrectomy (at 6.7 and 7.9 months respectively), the timing allows for proper preparation and optimization of surgical conditions. The moderate delay may prove beneficial as partial PVD simplifies posterior vitreous cortex dissection during surgery, while FVM fibrosis often reduces membrane-retinal adhesion strength, potentially facilitating safer and more effective surgical outcomes.

These patterns highlight the importance of moving beyond traditional macular-focused assessment in PDR management. Comprehensive evaluation of both PVD status and FVM fibrotic characteristics provides critical information for clinical decision-making, enabling more personalized timing of surgical intervention and potentially improving functional outcomes. For patients with favorable prognostic factors, watchful waiting with appropriate PRP may be justified, while those with poor prognostic indicators may benefit from earlier surgical referral, even before conventional thresholds for intervention are reached.

Study Limitations

This study has several important limitations. First, the assessment of key variables including FVM fibrosis, PVD status, and vitreoretinal interface characteristics relied on subjective clinical judgment despite standardized criteria. This subjective evaluation was further complicated by dense VH at baseline, preventing direct OCT visualization of the posterior hyaloid-retinal relationship. Second, the single-center design with predominantly advanced PDR cases may limit generalizability to broader populations or earlier disease stages. Third, uneven distribution across subgroups, particularly the small sample of patients with FVM fibrosis without PVD, reduced statistical power for certain comparisons. Finally, the 13-month median follow-up period was insufficient to establish optimal vitrectomy timing after FVM fibrosis development, especially since this duration coincided with the median time for fibrotic FVM separation from the retina (13 months, IQR: 11–14). Longer follow-up periods are needed to better evaluate long-term outcomes across various combinations of PVD status and FVM fibrotic patterns.

Conclusions

This prospective study demonstrates that in treatment-naïve PDR patients, the angio-fibrotic transformation following anti-VEGF therapy is significantly influenced by PVD status. Both factors independently protect against visual deterioration, while macular involvement remains the strongest predictor of poor outcomes.

This creates a clear risk stratification framework: patients with both PVD and FVM fibrosis demonstrate stable disease with excellent visual preservation, supporting conservative management. Conversely, patients lacking both factors—particularly those with suspected vitreoschisis—experience rapid progression requiring early surgical intervention.

By identifying this critical interplay between vitreoretinal interface characteristics and FVM transformation, this study enables ophthalmologists to move beyond macular-focused assessment toward comprehensive evaluation of the FVM-posterior vitreous complex when determining management strategies, facilitating personalized treatment decisions that optimize functional outcomes in PDR.

Funding Statement

No funding was received for this study.

Abbreviation:

FVM, Fibrovascular membrane; PDR, Proliferative diabetic retinopathy; VEGF, Vascular endothelial growth factor; PVD, Posterior vitreous detachment; VH, Vitreous hemorrhage; TRD, Tractional retinal detachment; PRP, Panretinal photocoagulation; PPV, Pars plana vitrectomy; RWS, Real-world studies; NVE, Neovascularization elsewhere; VA, visual acuity; OCT, Optical coherence tomography; logMAR, Logarithm of the minimum angle of resolution; IQR, Interquartile range.

Data Sharing Statement

Data supporting this study are openly available in Mendeley Data at https://data.mendeley.com/datasets/gdkpjnjv45/1.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

This study was approved by the Beijing Tongren Hospital ethics committee and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

Disclosure

The authors report no competing interests.