Reinitiating Anticoagulation in Chronic Subdural Hematomas: Does Adding Middle Meningeal Artery Embolization add value? A Multi-Institutional, Multinational Database Study

Abstract

Background:

The safety of early anticoagulation in chronic subdural hematoma (cSDH) patients treated with surgery, with or without middle meningeal artery embolization (MMAE), remains unclear. This study evaluates 6-month outcomes among cSDH patients who received early anticoagulation following surgery with adjunct MMAE and compares outcomes between anticoagulated patients treated with surgery alone versus surgery plus MMAE.

Methods:

Patients aged ≥18 years with cSDH who underwent surgery alone or surgery plus MMAE and received anticoagulation were identified using ICD-10 and RXNORM codes on TriNetX platform. Two analyses were performed: Surgery plus MMAE cohort, stratified by anticoagulation use within 1-month, and anticoagulated patients, stratified by surgery alone versus surgery plus MMAE. Primary outcomes included repeat surgery and 6-month mortality; secondary outcomes included repeat intracranial hemorrhage and cerebral infarction.

Results:

Among 801 patients treated with surgery plus MMAE, 143 received anticoagulation and 658 did not. In this cohort, no significant differences were observed in rates of repeat surgery (OR:1.00;95%CI:0.403–2.483) or 6-month mortality (OR:0.557;95%CI:0.245–1.263) between those who received anticoagulation and those who did not. When comparing 2,301 patients who underwent surgery alone and received anticoagulation to the 143 patients who received anticoagulation following surgery plus MMAE, the latter group had significantly lower mortality (7.9%vs.19.4%; OR:0.356;95%CI:0.169–0.751,p=0.005) and reduced rates of repeat intracranial hemorrhage (54.0%vs.66.9%; OR:0.580;95%CI:0.357–0.942,p=0.027).

Conclusion:

Early anticoagulation following surgery with adjunct MMAE appears safe and is associated with reduced mortality and hemorrhage risk compared to surgery alone in anticoagulated patients.

INTRODUCTION:

In patients with chronic subdural hematomas (cSDH), anticoagulation therapy is often withheld due to concerns about hemorrhage expansion. However, there is also a competing risk of acute ischemic stroke, acute myocardial infarction, and venous thromboembolism. The decision to resume anticoagulation therapy remains uncertain and currently there is insufficient evidence to guide the reinitiation of anticoagulation in high-risk patients^1-4. The practice of resuming anticoagulation following cSDH evacuation is highly variable, with limited literature offering concrete guidance⁵. Nevertheless, some studies have shown no significant differences in hemorrhage expansion or the need for repeat surgery following surgical evacuation in this patient population^6-9. Middle meningeal artery embolization (MMAE), by preventing repetitive microbleeds, has been shown to effectively reduce recurrent hemorrhage and hemorrhage expansion in patients with cSDH compared to surgical evacuation alone. Therefore, patients who undergo surgery with adjunctive MMAE are at a lower risk of developing repeat or worsening hemorrhage, suggesting that it may be safer to initiate anticoagulation in this cohort. However, the impact of reinitiating anticoagulation in this group has not yet been reported.

In light of the lack of supportive evidence on whether anticoagulation should be initiated in high-risk cases following surgery and MMAE in cSDH patients, we present a multi-institutional and multinational retrospective propensity score-matched analysis. The primary objective of this study is to assess the impact of anticoagulation initiation within 1 month of surgery in patients who received a combination of surgery and MMAE. The secondary objective is to assess the impact of anticoagulation initiation within 1 month of surgery in patients who received surgery and MMAE versus those who received surgery alone.

We hypothesize that, in patients with cSDH treated with surgery and MMAE, the initiation of anticoagulation within one month of treatment will have similar odds of repeat surgery, and mortality compared to those who were not initiated on anticoagulation. Similarly, anticoagulated cSDH patients treated with surgery and MMAE may have better outcomes compared to those treated with surgery alone.

METHODS:

This study was conducted using the TriNetX platform (https://trinetx.com/) and adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Study Population and Inclusion Criteria

The study population was defined on the TriNetX server using an approach similar to that reported in the literature^10-12.

Patients aged 18 years and older with cSDH were identified from the TriNetX database for the period between May 06, 2020, and May 07, 2025, using ICD-10 codes (Supplemental Table 1). Patients who underwent surgery alone or in combination with MMAE within one-month of cSDH diagnosis were identified using ICD-10 PCS codes (Supplemental Table 1). Within each group, those who were initiated on anticoagulation within one month were identified using RXNORM codes. Patients who received aspirin or clopidogrel during this period were excluded based on RXNORM codes listed in Supplemental Table 1.

Two separate analyses were conducted. In the first analysis, patients who underwent both surgery and adjunct MMAE were divided into two groups: those who received anticoagulation within one month and those who did not (Supplemental Table 2). This analysis addressed the first research question—evaluating the effects of anticoagulation in patients treated with a combination of surgery and MMAE. The second analysis focused exclusively on patients who were initiated on anticoagulation and compared those who underwent surgery with adjunct MMAE to those who received surgery alone (Table 1). This comparison aimed to assess the impact of adjunct MMAE on outcomes in anticoagulated patients.

Table 1: Study Demographics (before and after propensity-score matching) in cSDH patients who received anticoagulation within 1-month.

Anticoagulation initiation	Before PSM				After PSM
	Surgery + MMAE (n=143)	Surgery alone (n=2,301)	Std Diff.	p-value	Surgery + MMAE (n=139)	Surgery alone (n=139)	Std Diff.	p-value
Demographics
Age at Index	72.5 ± 11.0	64.1 ± 17.7	0.566	<0.001	72.4 ± 11.1	73.3 ± 11.5	0.077	0.519
Sex
Female	30 (21.1%)	691 (30.0%)	0.205	0.024	30 (21.6%)	31 (22.3%)	0.017	0.885
Male	107 (75.4%)	1,468 (63.8%)	0.253	0.005	104 (74.8%)	102 (73.4%)	0.033	0.784
Race
Black or African American	14 (9.9%)	296 (12.9%)	0.095	0.296	14 (10.1%)	18 (12.9%)	0.09	0.452
White	103 (72.5%)	1,498 (65.1%)	0.161	0.07	103 (74.1%)	89 (64.0%)	0.219	0.069
Asian	11 (7.7%)	108 (4.7%)	0.127	0.101	10 (7.2%)	15 (10.8%)	0.126	0.295
American Indian or Alaska Native	0 (0%)	11 (0.5%)	0.098	0.409	0 (0%)	0 (0%)	--	--
Unknown Race	10 (7.0%)	248 (10.8%)	0.131	0.16	10 (7.2%)	10 (7.2%)	<0.001	1
Other Race	10 (7.0%)	126 (5.5%)	0.065	0.429	10 (7.2%)	10 (7.2%)	<0.001	1
Native Hawaiian or Other Pacific Islander	10 (7.0%)	14 (0.6%)	0.34	<0.001	0 (0%)	10 (7.2%)	0.394	0.001
Unknown Ethnicity	22 (15.5%)	409 (17.8%)	0.061	0.489	22 (15.8%)	19 (13.7%)	0.061	0.612
Not Hispanic or Latino	111 (78.2%)	1,671 (72.6%)	0.129	0.149	108 (77.7%)	112 (80.6%)	0.071	0.555
Hispanic or Latino	10 (7.0%)	221 (9.6%)	0.093	0.311	10 (7.2%)	10 (7.2%)	<0.001	1
Diagnosis
Transient ischemic attack (TIA)	16 (11.3%)	214 (9.3%)	0.065	0.436	16 (11.5%)	23 (16.5%)	0.145	0.227
Cerebral infarction	13 (9.2%)	193 (8.4%)	0.027	0.749	12 (8.6%)	17 (12.2%)	0.118	0.327
Hypertension	100 (70.4%)	1,500 (65.2%)	0.112	0.203	97 (69.8%)	105 (75.5%)	0.129	0.282
Diabetes	38 (26.8%)	572 (24.9%)	0.043	0.611	38 (27.3%)	47 (33.8%)	0.141	0.241
Alcohol use	10 (7.0%)	236 (10.3%)	0.115	0.217	10 (7.2%)	10 (7.2%)	<0.001	1
Smoking	33 (23.2%)	374 (16.3%)	0.176	0.03	33 (23.7%)	32 (23.0%)	0.017	0.887
Obesity	30 (21.1%)	307 (13.3%)	0.207	0.009	28 (20.1%)	27 (19.4%)	0.018	0.88
Chronic kidney disease (CKD)	19 (13.4%)	245 (10.6%)	0.084	0.309	18 (12.9%)	20 (14.4%)	0.042	0.727
Atrial fibrillation	43 (30.3%)	425 (18.5%)	0.278	0.001	41 (29.5%)	34 (24.5%)	0.114	0.344
Asthma	10 (7.0%)	119 (5.2%)	0.078	0.333	10 (7.2%)	10 (7.2%)	<0.001	1
Ischemic heart diseases	37 (26.1%)	499 (21.7%)	0.103	0.222	37 (26.6%)	45 (32.4%)	0.126	0.293
Pulmonary embolism	10 (7.0%)	104 (4.5%)	0.108	0.167	10 (7.2%)	10 (7.2%)	<0.001	1
Deep venous thrombosis	10 (7.0%)	119 (5.2%)	0.078	0.333	10 (7.2%)	10 (7.2%)	<0.001	1
Medication
Antiplatelets (prior use)	28 (19.7%)	300 (13.0%)	0.181	0.023	28 (20.1%)	31 (22.3%)	0.053	0.66
Anticoagulants (prior use)	121 (85.2%)	983 (42.7%)	0.987	<0.001	118 (84.9%)	120 (86.3%)	0.041	0.733

In both analyses, patients who received antiplatelet therapy within the one-month window were excluded, as the sample size was insufficient to evaluate the effects of antiplatelets on outcomes.

Data source

The TriNetX research network (Cambridge, MA, USA) consists of over 275 million patients from 120 healthcare organizations across 30 countries. It offers realtime access to de-identified healthcare records and integrates new patients, observations, and results daily, all harmonized to standard terminologies such as ICD-10 and LOINC. Clinical variables are extracted through natural language processing of clinical documents, and robust quality assurance measures ensure data reliability before inclusion in the database. To protect patient privacy, the platform provides only aggregate counts and statistical summaries.

Study outcomes

The primary outcomes of the study included the the need for repeat surgery, and mortality at six-months. Secondary outcome measures included repeat intracranial hemorrhage and cerebral infarction following surgery. The outcome measures were define using ICD-10 codes described in the Supplemental Table 1.

Statistical analysis

All analyses were performed using the TriNetX platform. Univariate analyses were conducted using the chi-square test and t-test.

Propensity score matching was performed in a 1:1 ratio using the covariates listed in Tables 2 and Supplemental Table 2. Matching was performed using greedy nearest neighbor algorithms with a caliper width of 0.1 pooled standard deviations. Covariate balance was assessed using standardized mean differences, with values greater than 0.1 considered indicative of residual imbalance. Statistical significance was defined as a two-sided alpha of less than 0.05. Odds ratios (OR) with 95% confidence intervals (CI) were calculated for all analyses. The TriNetX platform uses logistic regression analysis to generate propensity scores for individual subjects and randomizes row order to eliminate bias associated with nearest neighbor matching.

Table 2: Study outcomes following anticoagulation initiation in patients who underwent surgery and adjunct MMAE versus surgery alone.

Outcome Measures at 6-months	Anticoagulation initiation in 1 month
	Surgery + MMAE (n=139)	Surgery alone (n=139)	Odds Ratio (95% CI), p-value
Needed Repeat Surgery	10 (7.2%)	10 7.2%)	1.0 (0.403–2.484), p = 1.000
Mortality	11 (7.9%)	27 (19.4%)	0.356 (0.169–0.751), p = 0.005
Intracranial Hemorrhage	75 (54.0%)	93 (66.9%)	0.580 (0.357–0.942), p = 0.027
Cerebral Infarction	10 (7.2%)	17 (12.2%)	0.556 (0.245–1.262), p = 0.156

RESULTS:

A total of 2,301 patients who underwent surgery alone and 143 patients who underwent surgery with adjunct MMAE were identified as having received anticoagulation within one month of surgery, while 658 patients who underwent surgery with adjunct MMAE did not receive anticoagulation within the same timeframe (Table 1 and Supplemental Table 2).

Surgery + MMAE: Anticoagulation Initiation vs No Anticoagulation

Following propensity score matching, no significant differences in key outcomes were observed between patients who received anticoagulation and those who did not, within the cohort treated with surgery and adjunct MMAE (Supplemental Table 2). Patients who initiated anticoagulation had similar odds of requiring repeat surgical intervention (OR: 1.00, 95% CI: 0.403–2.483, p = 1.000) and mortality at six months post-surgery (OR: 0.557, 95% CI: 0.245–1.263, p = 0.156).

Anticoagulation Initiation: Surgery with Adjunct MMAE vs Surgery Alone

In propensity score–matched analyses, rates of repeat surgery were comparable between the two cohorts (OR: 1.00, 95% CI: 0.403–2.484, p = 1.000). However, mortality was significantly lower in the adjunct MMAE group compared to the surgery alone group (7.9% vs 19.4%; OR: 0.356, 95% CI: 0.169–0.751, p = 0.005).

Further exploratory analysis revealed that the rate of repeat intracranial hemorrhage was significantly lower in the adjunct MMAE group (54.0% vs 66.9%; OR: 0.580, 95% CI: 0.357–0.942, p = 0.027), potentially contributing to the reduced mortality. In contrast, the odds of cerebral infarction were similar between the two groups (7.2% vs 12.2%; OR: 0.556, 95% CI: 0.245–1.262, p = 0.156) (Table 2 and Figure 1).

Figure 1:

Bar chart illustrating proportion of patients needing repeat surgery, mortality, repeat hemorrhage and infarction at 6-months across two cohorts.

DISCUSSION:

Our study demonstrates that initiating anticoagulation following surgery and MMAE does not result in differences in the need for repeat surgery, or mortality at six months. Additionally, initiating anticoagulation after surgery and MMAE was associated with lower odds of mortality and repeat intracranial hemorrhage compared to those who underwent surgery alone.

There is considerable variation in current practice patterns and perceived risks of embolic and hemorrhagic complications associated with anticoagulation reinitiation after cSDH evacuation. A recent survey of Canadian neurosurgeons (n=55) and stroke neurologists (n=26) on the reinitiation of anticoagulation after surgical evacuation of SDH revealed notable differences in practice patterns³. The perceived risk of stroke and SDH reaccumulation varied among participants and was influenced by the timing of anticoagulation reinitiation. This uncertainty stems from the limited data available on the effects of anticoagulation on hemorrhage expansion, the need for repeat surgery, and mortality. Existing literature on this subject has largely focused on patients who underwent surgery alone. Recent meta-analyses have shown no increased risk of repeat hemorrhage in patients who underwent surgery, although evidence remains limited.

MMAE reduces microhemorrhages within the hematoma, thereby preventing hematoma expansion, lowering the risk of rebleeding, and reducing the need for repeat surgery. By targeting the underlying pathophysiology of recurrent hemorrhage, adjunctive MMAE leads to improved outcomes compared to standard care^10,11,13-23. Accordingly, we observed significantly lower rates of repeat hemorrhage and mortality following the initiation of anticoagulation in patients treated with surgery plus MMAE compared to those who underwent surgery alone. This suggests that initiating anticoagulation may be safer in patients who receive MMAE, as the primary source of rebleeding is better controlled. In contrast, withholding anticoagulation in high-risk patients—such as those with acute ischemic stroke, myocardial infarction, or venous thromboembolism—can have catastrophic consequences. Therefore, adjunctive MMAE may allow for earlier and safer initiation of anticoagulation and may be considered in high-risk patients for whom anticoagulation is indicated.

This study has several limitations that must be acknowledged. First, the inherent challenges of database research include variability in data definitions and coding. Second, there is limited functional outcome information, as assessments of functional independence (such as the modified Rankin scale) were not available. Third, access to imaging data and DICOM files was not possible, which limits the assessment of hematoma characteristics such as thickness, intrinsic density, presence of septations, midline shift, or intracranial herniation. Fourth, information on the presence of simultaneous cSDH on the contralateral side was not available. To mitigate the unbalanced distribution of cases, we only included the surgical cSDH group, then classified them based on whether they received anticoagulation or not, with the hope of reducing the chances of uneven distribution of cases based on severity. However, it remains possible that patients who underwent surgery alone had less severe disease compared to those who received adjunct MMAE with surgery. Specifically, the hematoma size might have been larger in those who received surgery and MMAE compared to those who received surgery alone. If so, this would favor the surgery-alone group; however, despite this limitation, better outcomes were observed with surgery and MMAE compared to surgery alone. Fifth, it is still possible that patients initiated on anticoagulation had less severe symptoms or smaller hematomas than those who did not receive anticoagulation, which could explain the observed outcomes, where the anticoagulation group did not have worse results. Sixth, information on the dose, duration, and route of anticoagulation was not available, which may have resulted in the inclusion of patients who received anticoagulation for prophylactic rather than therapeutic purposes. Seventh, for the outcome analysis of repeat hemorrhage, the dataset did not allow differentiation between new and residual hemorrhage. As a result, the reported rates may be erroneously elevated; however, this limitation should affect both cohorts equally. Nevertheless, the inclusion of data from 120 healthcare organizations worldwide enhances the generalizability of our findings and represents the largest sample to date.

In conclusion, cSDH patients treated with both MMAE and surgery who were initiated on anticoagulation had similar odds of repeat surgery, and reduced mortality compared to those who were not initiated on anticoagulation. The use of adjunctive MMAE in surgical cSDH cases was associated with reduced mortality and repeat intracranial hemorrhage compared to surgery alone with anticoagulation initiation. Therefore, MMAE may be beneficial for high-risk individuals who require early anticoagulation initiation. Further prospective studies are needed to validate these findings.

KEY MESSAGES.

What is already known on this topic?

Anticoagulation therapy is often withheld in cSDH patients due to concerns about hemorrhage expansion, though there is also the risk of acute ischemic stroke, myocardial infarction, and venous thromboembolism. Limited research in cSDH patients treated with surgery alone has shown no significant differences in clinical outcomes. However, no studies have assessed the impact of anticoagulation initiation in patients receiving adjunct MMAE.

What this study adds

This study provides robust, real-world, multi-institutional, and multinational data on the effects of anticoagulation initiation following surgery and MMAE. It demonstrates that anticoagulation initiation in the MMAE group was associated with lower mortality at 6 months compared to surgery alone. No significant differences in repeat surgery were noted in the surgery + MMAE group between those who received anticoagulation and those who did not.

How this study might affect research, practice or policy.

This study may prompt further investigations into the safety and timing of anticoagulation reinitiation, particularly in patients treated with MMAE and surgical evacuation for cSDH.

Abbreviation:

cSDH

chronic subdural hematoma

MMAE

middle meningeal artery embolization