Impact of Antivascular Endothelial Growth Factor Therapy in Chronic Subdural Hematoma Patients: A Propensity-Matched Multi-Institutional Cohort Study

Abstract

BACKGROUND:

To evaluate possible associations between anti-VEGF (vascular endothelial growth factor) therapy and cSDH (chronic subdural hematoma) outcomes.

METHODS:

We conducted a cohort study using the TriNetX Research Network, comparing patients with cSDH taking anti-VEGF agents to controls through propensity score matching. Outcomes measured were assessed at 6 months and 1 year follow-up and included cSDH rebleeding, endovascular or surgical cSDH treatment, mortality, headaches, stroke, arterial hypertension, proteinuria, and major bleeding (noncranial).

RESULTS:

After propensity matching, 737 patients were included in both anti-VEGF and control cohorts at 6 months, and 722 patients in each cohort at 1 year. Baseline characteristics were well balanced. At 6-months, the anti-VEGF group had significantly lower odds of rebleeding (odds ratio [OR], 0.204 [95% CI, 0.159–0.26]; P<0.001), craniotomy (OR, 0.340 [95% CI, 0.155–0.680]; P=0.002), and mortality (OR, 0.778 [95% CI, 0.615–0.990]; P=0.037). At 1-year, reduced odds persisted for rebleeding (OR, 0.158 [95% CI, 0.122–0.200]; P<0.001), craniotomy (OR, 0.250 [95% CI, 0.116–0.490]; P<0.001), embolization (OR, 0.380 [95% CI, 0.172–0.770]; P=0.007), and mortality (OR, 0.677 [95% CI, 0.520–0.880]; P=0.003). Arterial hypertension was higher in the anti-VEGF group at 6 months (OR, 1.240 [95% CI, 1.000–1.530]; P=0.048), but not 1 year (OR, 1.110 [95% CI, 0.904–1.350]; P=0.330). No significant differences were observed in headache, stroke, proteinuria, or major bleeding at either time point.

CONCLUSIONS:

Anti-VEGF therapy is associated with significantly reduced rebleeding, reintervention rates, and mortality in patients with cSDH at both 6 months and 1 year. A transient increased incidence of arterial hypertension was noted at 6 months, but other major adverse events were not significantly different. Further randomized, prospective studies are warranted to confirm these results and optimize treatment strategies.

CLINICAL PERSPECTIVE.

What Is New?

• Anti-VEGF (vascular endothelial growth factor) therapy was associated with significantly lower rates of rebleeding, craniotomy, embolization, and all-cause mortality at both 6-month and 1-year follow-up in patients with chronic subdural hematoma, without increases in major bleeding, stroke, or proteinuria—suggesting it may be a safe and effective adjunct to standard management.

• Kaplan–Meier survival analysis demonstrated sustained reductions in cumulative risk of rebleeding, reoperation, and mortality, underscoring the potential long-term benefit of anti-VEGF therapy in reducing recurrence and improving survival in high-risk chronic subdural hematoma populations.

What Are the Clinical Implications?

• Although arterial hypertension was slightly more common in the anti-VEGF group at 6 months, the overall adverse-event profile was comparable to controls. With appropriate monitoring, anti-VEGF therapy may represent a viable nonsurgical treatment strategy that warrants further prospective evaluation.

Chronic subdural hematoma (cSDH) is a prevalent clinical condition, especially among the elderly, with an incidence between 1.7 and 20.6 per 100 000 people annually.^1,2 With an aging population, the incidence of cSDH is expected to rise, creating significant challenges for health care systems.^3–5 This trend emphasizes the need for improved diagnostic and therapeutic strategies.^3,5

A major concern in cSDH management is the high rate of rebleeding, reoperation, and associated mortality. Studies indicate that up to 25.1% of patients undergoing burr-hole craniotomy experience postoperative rebleeding,⁶ with the use of anticoagulants as a well-established risk factor and odds ratios (ORs) as high as 2.7.⁷ Reoperation rates vary depending on the surgical technique and patient factors, and range from 9.2% to 29%.^6,8 For instance, burr-hole craniotomy has a reoperation rate of 7.6%, while higher rates are observed with minicraniotomy (13.1%) and twist drill craniotomy (19.5%).⁹ Some studies report an operative mortality of 4%.¹⁰ More importantly, 1-year and 5-year mortality can reach 32% and 18%, respectively.^11–13

Recent advancements have explored nonsurgical options for cSDH management, including the use of Bevacizumab, an anti-VEGF (vascular endothelial growth factor) monoclonal antibody. Bevacizumab works by inhibiting VEGF, a pathway involved in angiogenesis and vessel permeability that contributes to hematoma formation and progression. Evidence supports Bevacizumab’s efficacy in reducing cSDH volume, both through systemic administration¹⁴ and intra-arterial delivery via the middle meningeal artery (MMA).^15,16 This emerging therapy suggests that anti-VEGF treatment may help mitigate rebleeding and recurrence in patients with cSDH, offering an additional nonsurgical approach to the cSDH armamentarium. Other anti-VEGF monoclonal antibodies include Ramucirumab, as well as VEGF receptor-targeting tyrosine kinase inhibitors like sunitinib, sorafenib, axitinib, cabozantinib, pazopanib, regorafenib, and vandetanib.^17–19

This study is the first large-scale database analysis to evaluate the impact of concurrent anti-VEGF therapy on mortality, recurrence, and reoperations in patients with cSDH. Should anti-VEGF therapy prove effective, it could represent a valuable addition to the cSDH treatment options, particularly for patients with high-risk conditions who may not be ideal candidates for endovascular or surgical interventions.

Methods

Data Source

We performed a retrospective study utilizing the TriNetX COVID-19 Research Network (TriNetX, Cambridge, MA; https://live.trinetx.com), a global platform integrating deidentified electronic health records from 94 health care organizations and over 130 million patients. Data are contributed primarily by health care organizations in the United States and encompass demographic details, diagnoses, procedures, medications, and laboratory results. Patients with prior occurrences of these outcomes before the index date were excluded from the analysis. This approach was particularly important for prevalent symptoms. By excluding those with documented diagnoses before the index event, we aimed to capture the incidence of new or worsening outcomes during follow-up periods, allowing us to assess differences in clinical progression. To maintain compliance with ethical and legal standards preventing reidentification, the identities of individual health care organizations are not disclosed. Consequently, this study did not require institutional review board approval. The TriNetX platform provides statistical summaries and aggregate data without access to protected health information or personal identifiers. Data are refreshed daily, and patient characteristics are categorized using the International Classification of Diseases, Tenth Revision (ICD-10) codes. Data and code used to generate analyses/figures will be made available by the corresponding author on reasonable request.

Study Design

Patients with cSDH were identified using the validated ICD-10 code I62.03 and I62.00,²⁰ covering the period from November 3, 2004, to May 3, 2024, to ensure a minimum of 6 months of follow-up. Individuals under 18 years of age were excluded, along with those diagnosed with brain arteriovenous malformations (ICD-10: Q28.2), brain aneurysms or cerebral arteriovenous fistulas (ICD-10: I67.1), malignant neoplasms of the eye, brain, and central nervous system (ICD-10: C69–C72), and malignant neoplasms of the head, face, and neck (ICD-10: C76.0), to minimize potential confounding of anti-VEGF effects.²⁰ These conditions independently increase the risk of intracranial hemorrhage and mortality, introduce alternative vascular pathologies unrelated to cSDH, and often involve prior anti-VEGF exposure, potentially biasing treatment effects.^21–24

Patients were stratified into treatment and control cohorts. The treatment cohort included patients prescribed anti-VEGF agents, including axitinib (RxNorm 1242999), bevacizumab (RxNorm 253337), cabozantinib (RxNorm 1363268), mesylate (RxNorm 1426388), pazopanib (RxNorm 714438), ramucirumab (RxNorm 1535922), regorafenib (RxNorm 1312397), sorafenib (RxNorm 495881), sunitinib (RxNorm 357977), and vandetanib (RxNorm 109413). Because of the database’s inherent constraints, we were unable to determine the specific clinical indications for anti-VEGF therapy. The anti-VEGF mechanism of each of these agents is presented in Figure 1.

Figure 1.

Mechanism of VEGF (vascular endothelial growth factor) in chronic subdural hematoma (cSDH) formation and anti-VEGF inhibitory pathways. Illustrates VEGF’s role in cSDH formation, highlighting key pathways influenced by VEGF signaling. Anti-VEGF agents’ inhibitory effects are shown along the pathway, demonstrating their role in disrupting the VEGF-mediated processes contributing to cSDH development. Created with biorender.com.

The control cohort excluded any anti-VEGF usage throughout the study period. For the treatment group, the first anti-VEGF prescription had to be on or after the initial cSDH diagnosis. Patients were also required to have at least a 6-month follow-up. The index date was defined as the date of the first anti-VEGF prescription for the treatment group or the date of the initial cSDH diagnosis for the control group. Outcomes were evaluated at 6-month and 1-year intervals, covering headache (migraine: ICD-10 G43, excluding G43.6 and G43.A; tension headache: ICD-10 G44.2)^25–27 served as an indicator of cSDH symptom progression, while rebleeding (nontraumatic subdural hemorrhage: ICD-10 I62.0) measured recurrence risk. The need for endovascular (embolization: current procedural terminology 61626/61624) or craniotomy (craniectomy or craniotomy: TriNetX procedure concept 1009097) was tracked to identify management escalation. In addition, potential side effects of anti-VEGF therapy, including arterial hypertension (ICD-10 I10), stroke (ischemic: ICD-10 I63/G45; hemorrhagic: ICD-10 I60/I61), proteinuria (ICD-10 R80), and major bleeding (ICD-10 R57.1) were analyzed. Finally, all-cause mortality was assessed between cohorts. The TriNetX platform reports any patient count of 1 to 10 as ≤10 to ensure patient confidentiality, and this is accordingly reflected in our reported results. The study design is presented in Figure 2.

Figure 2.

Inclusion and exclusion criteria flowchart. cSDH indicates chronic subdural hematoma; FU, follow-up; and VEGF, vascular endothelial growth factor.

Propensity Score Matching

Cohorts were matched 1:1 using a greedy nearest-neighbor algorithm for both 6-month and 1-year follow-up periods. Matching criteria included the index date, demographics (sex, race, ethnicity), and comorbidities, such as coagulation defects and hemorrhagic conditions. Arterial hypertension, hyperlipidemia, type 2 diabetes, overweight/obesity, chronic kidney disease, hypercholesterolemia, atrial fibrillation/flutter, asthma, and any neoplasm diagnosis.^20,28 Neoplasm diagnoses were identified through TrinetX ICD-10 code groupings, and encompassed benign neoplasms (D10-D36), neoplasms of uncertain behavior, polycythemia vera, and myelodysplastic syndromes (D37-D48), melanoma and malignant skin neoplasms (C43-C44), unspecified neoplasms (D49), ill-defined, secondary, and unspecified malignancies (C76-C80), breast cancer (C50), in situ neoplasms (D00-D09), respiratory and intrathoracic malignancies (C30-C39), digestive organ cancers (C15-C26), lymphoid, hematopoietic, and related tissue cancers (C81-C96), and urinary tract malignancies (C64-C68). In addition, cohorts were matched for the medications aspirin (RxNorm 1191), clopidogrel (RxNorm 32968), ticagrelor (RxNorm 1116632), beta blockers (VA code CV100), diuretics (VA code CV700), heparin (RxNorm 5224/235473), warfarin (RxNorm 11289), calcium channel blockers (VA code CV200), ACE (angiotensin-converting enzyme) inhibitors (VA code CV800), hydralazine (RxNorm 5470), enoxaparin (RxNorm 67108), and alpha blockers (VA code CV150).^29–31

Statistical Analysis

For baseline characteristics, continuous data are presented as means and standard deviations and compared using independent t tests, while categorical data are presented as percentages compared using χ² tests. Measures of association analyses were performed for each outcome of interest; in particular, we evaluated the OR with CIs and P values. Kaplan-Meier survival curve analysis was performed for all outcomes at 1 year.³² A log-rank test was used to identify if the Kaplan-Meier survival curve differed between the 2 cohorts.³² Statistical analyses were performed using both the TriNetX online platform and RStudio. Figures were created in RStudio using the ggplot2 R package and data exported from TriNetX.

Results

Baseline Demographics and Comorbidities

From the health care organizations queried, 73 providers in the control cohort, and 52 in the treatment cohort, responded with patient data. The median time from the initial cSDH diagnosis to anti-VEGF administration was 6 weeks for patients in both follow-up time intervals. In Table 1, baseline characteristics of the anti-VEGF and control cohorts were compared after propensity score matching for patients with at least 6 months of follow-up (737 anti-VEGF versus 737 control) and 1 year of follow-up (722 anti-VEGF versus 722 control). The cohorts were well-balanced overall, with age at index similar between anti-VEGF and control groups (72.26 versus 73.06 years at 6 months and 72.18 versus 72.42 years at 1 year). Female patients comprised a comparable proportion of each cohort (38.54% versus 38.67% at 6 months and 38.37% versus 40.72% at 1 year). Most patients were identified as not Hispanic or Latino (72.19% versus 76.12% at 6 months and 72.58% versus 71.33% at 1 year), and White (73.00% versus 74.63% at 6 months and 73.13% versus 72.16% at 1 year).

Table 1.

Anti-VEGF and Control Cohorts After Matching, Showing Propensity Score Matched Characteristics, Comparing Baseline Demographic and Comorbidity Variables for Patients With at Least 6-Month and 1-Year Follow-Up

Most common comorbidities included essential hypertension (83.58% in the anti-VEGF group versus 87.52% in the control group at 6 months, and 83.52% versus 87.67% at 1 year), hyperlipidemia, unspecified (63.37% versus 68.11% at 6 months and 62.33% versus 63.71% at 1 year), type 2 diabetes (48.71% versus 53.73% at 6 months and 47.51% versus 49.86% at 1 year), and coagulation defects and hemorrhagic conditions (30.12% versus 32.29% at 6 months and 30.89% versus 28.39% at 1 year) with standard mean differences indicating balanced matching across these variables. Neoplasms were balanced across both groups, with similar rates of benign neoplasms (43.28% in anti-VEGF versus 41.79% in control at 6 months; 43.35% versus 41.27% at 1 year), malignant neoplasms of unknown origin (30.39% versus 27.54% at 6 months; 30.47% versus 29.64% at 1 year), and neoplasms of uncertain behavior (24.97% versus 23.88% at 6 months; 25.49% versus 24.10% at 1 year). Medication use was also comparable, with anti-VEGF and control groups showing similar proportions of patients on clopidogrel (15.33% versus 19.81% at 6 months; 15.10% versus 16.07% at 1 year), ticagrelor (2.04% versus 1.49% at 6 months; 1.80% versus 1.39% at 1 year), aspirin (59.16% versus 64.45% at 6 months; 58.73% versus 59.56% at 1 year), and heparin (60.24% versus 64.32% at 6 months; 60.94% versus 62.05% at 1 year; Table 2).

Table 2.

Anti-VEGF and Control Cohorts After Matching, Showing Propensity Score Matched Characteristics, Comparing Neoplasm Subtypes and Medication Variables for Patients With at Least 6-Month and 1-Year Follow-Up

Comparison of Clinical Outcomes

At the 6-month follow-up, the anti-VEGF cohort had significantly lower ratios of rebleeding (OR, 0.204 [95% CI, 0.159–0.26]; P<0.001), craniotomy (OR, 0.340 [95% CI, 0.155–0.68]; P=0.002), and mortality (OR, 0.778 [95% CI, 0.615–0.9]; P=0.037). Arterial hypertension was higher in the anti-VEGF group (OR, 1.240 [95% CI, 1.000–1.53]; P=0.048). There were no significant differences in rates of headache (OR, 0.855 [95% CI, 0.445–1.62]; P=0.631), major bleeding (OR, 1.000 [95% CI, 0.403–2.48]; P=0.999), proteinuria (OR, 1.010 [95% CI, 0.407–2.5]; P=0.984), or stroke (OR, 0.576 [95% CI, 0.291–1.1]; P=0.097; Figure 3A; Table 3).

Figure 3.

Forest plot depicting the risk of clinical outcomes, including headache, rebleeding, need for surgery (both open and endovascular), mortality, hypertension (arterial), major bleeding, proteinuria, and stroke. Risk of clinical outcomes is shown at 6-month (A) and 1-year (B) follow-up. *Green indicates significant (P<0.05) odds ratio (OR). Arrows indicate odds ratios (OR) favoring anti-VEGF (vascular endothelial growth factor) therapy when OR <1 and favoring the control group when OR >1. C, Kaplan-Meier survival curves for headache, need for surgery (both open and endovascular), stroke, rebleed, proteinuria, hypertension (arterial), major bleeding, and mortality at 1-year follow-up. The corresponding log-rank test results and associated P value are reported, comparing the treatment cohort (anti-VEGF) in blue to the control cohort in red over the 1-year follow-up period.

Table 3.

Clinical Outcomes in Anti-VEGF and Control Cohorts at 6-Month and 1-Year Follow-Up, Including Progression of cSDH and Major Complications Associated With Anti-VEGF Use

At the 1-year follow-up, the anti-VEGF cohort continued to show significantly lower ratios of rebleeding (OR, 0.158 [95% CI, 0.122–0.2]; P<0.001), craniotomy (OR, 0.250 [95% CI, 0.116–0.49]; P<0.001), embolization (OR, 0.380 [95% CI, 0.172–0.77]; P=0.007), and mortality (OR, 0.677 [95% CI, 0.520–0.88]; P=0.003). No significant differences were found for headache (OR, 1.000 [95% CI, 0.466–2.14]; P=0.999), major bleeding (OR, 1.000 [95% CI, 0.404–2.48]; P=0.999), proteinuria (OR, 1.050 [95% CI, 0.424–2.61]; P=0.913), stroke (OR, 0.506 [95% CI, 0.245–1.00]; P=0.050), or arterial hypertension (OR, 1.110 [95% CI, 0.904–1.35]; P=0.330; Figure 3B; Table 3).

Kaplan-Meier survival curves assessing the cumulative event risk over 1 year, shown in Figure 3C, demonstrated a significantly lower rate of rebleeding (P<0.001), need for craniotomy (P<0.001), embolization rates (P<0.001), stroke (P=0.036), and mortality (P=0.002). Incidence of headache (P=0.541), proteinuria (P=0.436), arterial hypertension (P=0.802), and major bleeding (P=0.100) were nonsignificant.

Discussion

Anti-VEGF drugs, commonly used to treat conditions like age-related macular degeneration, diabetic retinopathy, and cancer, work by inhibiting the VEGF pathway, which has also been implicated in the progression of cSDH.^33–35 Elevated levels of VEGF have been consistently observed in cSDH fluid, suggesting its significant involvement in the disease’s pathogenesis.^36–38 Additionally, the outer membrane of cSDH is particularly rich in VEGF, which promotes the formation of fragile, leaky blood vessels, contributing to the hematoma’s expansion and recurrence. This neovascularization is driven by HIF-1α (hypoxia-inducible factor 1-alpha), which upregulates VEGF expression under hypoxic conditions commonly found in cSDH.³⁷ The high VEGF levels correlate with increased microvessel density and hyperpermeability, leading to the continuous enlargement of the hematoma.^37,38

Although burr-hole craniotomy remains the primary surgical intervention for cSDH, it addresses only the immediate need for hematoma evacuation. Its inability to mitigate underlying pathophysiologic processes, such as inflammation and membrane formation^39–41 contributes to the 10% to 20% recurrence rate,^2,42,43 which leads to additional surgeries, increased complications, and added health care costs. Anti-VEGF therapy, by directly targeting VEGF-driven angiogenesis and hyperpermeability, may offer an adjunct or alternative with long-term benefits.

Anti-VEGF agents such as bevacizumab bind VEGF-A and prevent activation of endothelial receptors. This blocks angiogenic signaling, limits endothelial proliferation, and lowers vascular permeability. In chronic subdural hematoma, inhibition of VEGF likely stabilizes the fragile neovasculature of the outer membrane and reduces microvascular leak and small recurrent hemorrhages.⁴⁴ Dexamethasone lowers VEGF levels in postoperative cSDH samples and is linked to lower recurrence, which supports a causal role for VEGF in rebleeding.⁴⁵ Macrophage-derived VEGF has been identified as a driver of ongoing microbleeds in cSDH, so VEGF blockade would counter this process.⁴⁶ Matrix metalloproteinases such as MMP-9 (matrix metalloproteinase-9), which are associated with leaky capillaries and recurrence in cSDH, are regulated downstream of VEGF and may decline when VEGF signaling is suppressed.⁴⁶

Our findings support this hypothesis, with anti-VEGF-treated patients showing significant reductions in rebleeding and reoperation rates at 6 months and 1 year, suggesting that anti-VEGF therapy could enhance outcomes by preventing hematoma re-accumulation and reducing reliance on repeated surgeries. Notably, our study also demonstrated a statistically significant reduction in mortality among the anti-VEGF cohort at both 6-month and 1-year follow-up. This finding suggests a potential survival advantage for patients with cSDH receiving anti-VEGF therapy, which may be attributed to its effect in reducing recurrence, surgical intervention, and the complications associated with these. These positive effects align in general with promising results shown in previous studies. In a previously published case report, complete clearance of cSDH was demonstrated in a patient with a concurrent glioblastoma after starting bevacizumab, and was efficacious in reducing cSDH volume.¹⁴ Khalife et al¹⁵ reported the first case of intraarterial administration of bevacizumab through the MMA, with near total resolution of the cSDH on repeat imaging at 6 months. Following this preliminary experience, the same group did the aforementioned treatment in 12 hemispheres and 8 patients. The study reported no treatment-related complications, with all treated hemispheres achieving at least a 50% reduction in hematoma size by 3 months and no recurrences requiring further intervention during that period.¹⁶

Beyond anti-VEGF agents, other pharmacological treatments have demonstrated efficacy in modulating cSDH’s vascular and inflammatory pathways. Atorvastatin, for example, supports vascular stabilization by modulating VEGF levels, promoting vascular maturation within the hematoma wall.⁴⁷ In combination with dexamethasone, atorvastatin has shown an additive effect in reducing VEGF expression and endothelial inflammation,⁴⁸ leading to reduced hematoma volume and improved neurological outcomes,^49,50 though studies report mixed results on long-term efficacy.^51,52 Additionally, statins paired with surgical evacuation have shown enhanced postoperative hematoma reduction.⁵³

Steroids, particularly dexamethasone, further reinforce the potential of pharmacological approaches in cSDH. Dexamethasone has been shown to lower VEGF levels postoperatively³⁶ and decrease the need for reoperation by up to 62% compared with placebo.⁴⁹ However, steroids carry a significant risk profile, with potential adverse effects including elevated mortality rates, underscoring the need for cautious application.^49,54 Antifibrinolytics, such as tranexamic acid, have also been explored, demonstrating reductions in hematoma volume in initial studies.⁴⁹ However, the limited data on safety and long-term outcomes suggest that antifibrinolytics should be used conservatively until more comprehensive studies are available.

Finally, MMA embolization represents another innovative nonsurgical option, with recurrence rates as low as 2.4% for recurrent cases and 4.1% for primary cases.^55,56 This way, MMA embolization showed potential to reduce reoperations to 4.2%, which is significant when compared to standard surgical techniques.⁵⁷ Lower open surgery rates among anti-VEGF users may reflect clinician preference or perceived contraindication (black-box warnings) rather than reduced disease activity, which is why including embolization provides a complementary end point less influenced by surgical deferral.

Despite the potential benefits highlighted in this work, anti-VEGF agents require careful monitoring due to their possible adverse effects,^34,35,58 including arterial hypertension, proteinuria, thromboembolic events, and potential endocrine dysfunctions from vascular regression in healthy tissues.^34,58–61 Our data did not show significant differences in major adverse events, including proteinuria and bleeding, between the anti-VEGF and control groups, providing an encouraging safety signal. Nevertheless, the observed increase in arterial hypertension at the 6-month follow-up among anti-VEGF users highlights the need for careful patient selection and risk management. The 1-year Kaplan-Meier survival curves and log-rank test revealed no significant difference in the cumulative risk of arterial hypertension, suggesting no change in the overall Anti-VEGF cohort risk profile over time.³²

Anti-VEGF agents can raise blood pressure through on-target vascular effects. VEGF promotes endothelial nitric oxide and prostacyclin release and supports microvascular growth. When VEGF signaling is inhibited, vasodilatory pathways such as eNOS (endothelial nitric oxide synthase)-mediated nitric oxide are reduced, capillary rarefaction develops, and endothelin-1-related vasoconstriction increases.⁶² Hypertension is therefore a frequent toxicity across this drug class. Bevacizumab is associated with any grade hypertension in about 25% of patients and with grade 3 or higher events in about 80%.⁶³ Sunitinib shows any grade events in about 22% and grade 3 or higher events in about 70%.⁶⁴ Comparative analyses indicate that all approved VEGF pathway inhibitors increase hypertension risk several-fold compared with control.^65,66 These observations have practical implications for patients with chronic subdural hematoma, who are often older and already hypertensive. Close blood pressure surveillance during early treatment cycles is appropriate, with timely initiation or intensification of antihypertensive therapy and temporary dose interruption for severe elevations in line with expert guidance.^64,67 In our study, hypertension was higher at 6 months in the anti-VEGF cohort, but this difference was not present at 1 year, and other major adverse events were not different between groups. This pattern supports cautious use with active monitoring in this population.

Overall, the multi-center signal shows a consistent association between anti-VEGF exposure and lower cSDH recurrence, intervention rates, and mortality, achieved without a disproportionate safety cost. This pattern across timepoints and institutions positions anti-VEGF therapy as a high-priority candidate for prospective evaluation alongside other emerging pharmacological and endovascular approaches.

Limitations

This study has several limitations due to its retrospective, observational nature and reliance on data from the TriNetX database. First, while propensity score matching was used to balance key demographic and clinical variables between the anti-VEGF and control groups, the observational design limits the ability to establish causality. Because this is a retrospective observational study, residual confounding is likely, and the findings should be interpreted as associations rather than evidence of causation. Furthermore, despite the matching process, unmeasured confounders may remain, such as the severity of cSDH and detailed clinical indications for anti-VEGF therapy. In particular, several anti-VEGF agents carry surgical warnings that may lead clinicians to defer elective open procedures, introducing selection bias in the open-surgery end point. The use of de-identified, aggregated data poses additional challenges. TriNetX data restricts access to granular information such as specific dosage, administration route, and treatment duration for anti-VEGF agents. As such, patients were included in the anti-VEGF cohort based solely on prescription records. This method has been used in prior validated TriNetX studies,^68–71 but it prevents confirmation of whether patients received regular dosing at indicated regimens. Consequently, residual uncertainty remains regarding the intensity and duration of anti-VEGF exposure. It also limits our understanding of adherence, which could have influenced patient outcomes, particularly given that anti-VEGF agents are typically prescribed for indications outside of cSDH. Consequently, the majority of patients included in the anti-VEGF cohort most likely received the medication for reasons unrelated to their cSDH, which could have introduced unknown confounders. A potential for surveillance bias also remains because patients on anti-VEGF therapy might have more frequent medical encounters, which could increase the likelihood of detecting outcomes compared with the control group. Additionally, we grouped diverse anti-VEGF agents into 1 exposure, which introduces heterogeneity and may mask drug-specific effects. Future studies should stratify by agent, indication, dose, and route.

The reliance of the TriNetX platform on ICD-10 coding, despite using previously validated ICD-10 codes for cSDH, also introduces potential inaccuracies in identifying diagnoses, outcomes, and interventions, as coding errors or variations in clinical practice may affect data accuracy. Additionally, the lack of imaging data prevents verification of cSDH characteristics, such as hematoma size or membrane status, which are relevant factors in determining treatment and prognosis. The lack of procedural details, such as intraoperative complications and cause of death, also limits the interpretation of outcomes, as reported mortality represents all-cause mortality rather than cSDH-specific mortality.

Future Directions

To better define the role of anti-VEGF in CSDH, prospective, randomized controlled trials are essential. Such studies should investigate the optimal dosage, administration route (systemic versus intra-arterial via the middle meningeal artery), and timing to maximize therapeutic benefit while minimizing adverse effects. Future research should also assess the long-term safety profile of anti-VEGF agents, particularly monitoring for known side effects such as arterial hypertension, proteinuria, and thromboembolic events. In addition, less frequent but clinically relevant anti-VEGF toxicities such as impaired wound healing and endocrine effects related to vascular regression in normal tissues should be monitored. This is critical, as these therapies have a black box warning against their use in elective surgery or serious hemorrhage due to increased bleeding risk.⁷² Additionally, investigating combination therapies, such as pairing anti-VEGF agents with pharmacological treatments such as statins or dexamethasone, may provide additive or synergistic effects, helping to stabilize vascular structures within the hematoma and reduce recurrence risk. Research should also focus on identifying patient subgroups that may benefit most from anti-VEGF therapy, including those with high recurrence risk or those who are not candidates for surgery.

Beyond timing of initiation, future studies should explicitly test the duration of anti-VEGF therapy, such as fixed short-course versus maintenance strategies, and prospectively evaluate outcomes after discontinuation to detect potential rebound or recurrence. Designs that incorporate time-updated exposure, such as extended Cox models or target-trial emulation, and prespecified stop rules guided by imaging and symptoms will be critical to define minimum effective duration and whether maintenance confers durable benefit in cSDH.

Conclusions

Anti-VEGF therapy may offer a potential approach to cSDH management by addressing underlying vascular mechanisms that drive hematoma recurrence. When combined with other adjunctive therapies such as atorvastatin, dexamethasone, and potentially MMA embolization, anti-VEGF agents may reduce the need for invasive procedures and improve long-term outcomes in patients with cSDH. Further research, ideally through randomized controlled trials, is needed to confirm these associative findings and fully characterize the long-term safety of this possible therapeutic avenue.

ARTICLE INFORMATION

Acknowledgments

Our study exclusively utilized deidentified patient records from TriNetX (ISO 27001:2013) and refrained from using individually identifiable data. As such, our study was exempt from requiring institutional review board approval.

Sources of Funding

This research was supported by the University of Texas Medical Branch (UTMB) Institute for Translational Sciences, supported in part by a Clinical and Translational Science Award (UL1 TR001439) from the National Center for Advancing Translational Sciences at the National Institutes of Health (NIH). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Disclosures

None.