Perioperative Discontinuation of Antiplatelet and Anticoagulation Therapy in Chronic Subdural Hematoma: Impact on in-Hospital Mortality, Functional Outcome, and Recurrence

Abstract

The purpose of this study was to evaluate the effect of perioperative discontinuation of antiplatelet and anticoagulation therapy (AAT) on in-hospital mortality, functional outcome, and recurrence in patients with chronic subdural hematoma (CSDH) with prior AAT use, compared to patients without AAT. We retrospectively analyzed data from patients with symptomatic CSDH who underwent surgery in our department between June 2012 and December 2023. Prior AAT was discontinued at hospital admission and reinitiated 3 weeks postoperatively if CSDH resolution was complete. Clinical and neurological status at admission and discharge were assessed using the modified Rankin Scale, with scores >3 defined as a poor outcome. Among 852 surgically treated CSDH patients (mean age 75.2 years; female-to-male ratio 1:2.1), univariate analysis revealed that prior AAT use was significantly associated with higher in-hospital mortality (p = 0.043) and poor functional outcome (p < 0.001), but not with recurrence (p = 0.434). However, multivariate analysis demonstrated that prior AAT use was not an independent predictor of mortality (p = 0.099) or poor outcome (p = 0.956). Notably, patients with prior (acetylsalicylic acid) aspirin use exhibited significantly lower recurrence rates compared to patients without AAT (p = 0.012) and to the general cohort (p = 0.003). The multivariate analysis showed that prior aspirin use was an independent predictor of reduced recurrence rates (p = 0.015, adjusted odds ratio = 0.504; 95% confidence interval 0.290–0.877). This large-scale study supports the safety of perioperative AAT discontinuation in CSDH patients, showing no significant differences in in-hospital mortality, functional outcome, or recurrence compared to patients without AAT. Intriguingly, prior aspirin therapy was associated with reduced recurrence, suggesting a potential protective role.

Introduction

Chronic subdural hematoma (CSDH) is a frequent neurosurgical condition, particularly in the elderly, characterized by blood accumulation in the subdural space. Its incidence is rising worldwide, largely due to population aging and the increasing use of antiplatelet and anticoagulation therapy (AAT) for the prevention of cardiovascular and cerebrovascular diseases.^1–3 The proportion of CSDH patients receiving AAT varies across studies, ranging from 30% to 60%.^4,5 Perioperative management in this group is challenging. While discontinuation of AAT is routinely undertaken to reduce the risk of postoperative bleeding, it may simultaneously elevate the risk of thromboembolic events.^6,7

Evidence regarding the impact of AAT on surgical outcomes in CSDH is conflicting. Some studies associate AAT with higher rates of postoperative recurrence and mortality,^8–10 whereas others report no significant influence on outcomes,^11,12 leaving the safety and optimal timing of surgery in this population uncertain.

The present study aims to clarify whether surgical treatment of CSDH after perioperative AAT discontinuation is safe. Specifically, we investigate whether stopping AAT before surgery influences in-hospital mortality, functional outcome, or recurrence risk compared with patients not receiving AAT. Leveraging a large patient cohort, this study seeks to provide robust evidence to guide perioperative decision-making in CSDH, balancing the risks of hemorrhagic and thromboembolic complications.

Methods

Study population

From June 2012 to December 2023, we evaluated 938 consecutive patients aged ≥18 years with symptomatic CSDH admitted to our institution for eligibility in this study. Patients with a history of brain surgery (intracranial tumor, acute subdural hematoma, aneurysm clipping, CSDH) or ventriculoperitoneal shunt and concomitant acute lesions (intracranial hemorrhage, traumatic brain injury) were excluded. The study flowchart presents the main exclusion criteria in detail (Fig. 1). This retrospective study was performed in accordance with the ethical standards of the institutional research committee and the Code of Ethics of the World Medical Association (Declaration of Helsinki). The Institutional Review Board approved the study protocol (Medical Faculty, University of Duisburg-Essen, Registration number: 22-10763-BO). All supporting data are available within this article. Anonymized data will be shared upon request from any qualified investigator.

FIG. 1.

Study flowchart.

Routine clinical care and treatment protocol of CSDH

AAT was discontinued upon hospital admission and re-established 3 weeks after surgery in case of complete resolution of the hematoma. In all patients, hematoma evacuation was performed under general anesthesia using either a burr hole or a minicraniotomy with placement of a subdural drain whenever possible. Prior to surgery, a single dose of antibiotics was administered prophylactically. Postoperatively, patients spent 48 h in the intermediate care unit or the intensive care unit and remained on bed rest. On the second postoperative day, a cranial computer tomography scan was performed, and the subdural drain was removed. Patients were usually discharged around day 7 after surgery.

Patient data collection

The patients’ medical records were retrospectively reviewed to collect data on: age, sex, Charlson comorbidity index (CCI), American Society of Anesthesiologists Physical Status Classification System (ASA status), history of a recent trauma, as well as comorbidities such as epilepsy, dementia, arterial hypertension, diabetes mellitus, chronic kidney disease (CKD), atrial fibrillation (AF), chronic heart disease (CHD), cerebral vascular accident or transient ischemic accident (CVA/TIA), thrombophilia, AAT, seizure 24 h prior to admission, neurological condition, and functional status on admission using the Glasgow coma scale (GCS), the Markwalder grading scale, and the modified Rankin Scale (mRS), respectively. Furthermore, CSDH location (unilateral vs. bilateral), operative technique (burr hole vs. minicraniotomy) and preoperative hemoglobin level, and platelet count were analyzed. During hospitalization, neurological condition on discharge, in-hospital mortality, and postoperative complications—including postoperative bleeding, postoperative seizure, recurrence, myocardial infarction, pulmonary embolism, and postoperative stroke were thoroughly assessed. Recurrence was defined as an episode of radiological hematoma progression, with or without development of symptoms, that required reoperation. Functional status was documented at discharge, with an mRS score >3 considered indicative of a poor outcome. All data were verified by the senior author (M.C.) prior to analysis.

Definition of AAT discontinuation

AAT was discontinued prior to surgery (acetylsalicylic acid [aspirin] and clopidogrel 5–7 days, direct oral anticoagulants [DOAC] 48 h, and vitamin K antagonists until the Quick value exceeded 70%, i.e., International normalised ratio < 1.3). AAT was reinitiated 3 weeks after surgery if the hematoma had completely resolved. In cases of neurological deterioration during the discontinuation period, desmopressin or prothrombin, proconvertin, Stuart factor and antihemophilic factor B was administered to reverse the effects of aspirin and vitamin K antagonists, respectively, to normalize coagulation status.

Statistical analyses

All analyses were performed using SPSS version 29.0.2.0 (IBM, Chicago, IL). Categorical data are presented as frequencies and percentages. Normally distributed continuous variables are expressed as the mean and standard deviation. Nonnormally distributed continuous variables are expressed as the median and interquartile range. The Mann–Whitney U test was used to assess statistical associations between AAT and other continuous variables. Associations between AAT and other dichotomous variables were determined using the chi-square test and the Fisher’s exact test (when the expected cell frequency was <5). The factors in the univariate analysis, which were associated with the study endpoints (in-hospital mortality, functional outcome at discharge and recurrence) with a p value < 0.1, were implemented in the multivariate analysis. Binomial logistic regression (forward likelihood ratio method) was used to determine predictors of the study endpoints, and adjusted odds ratio (aOR) and 95% confidence intervals (95% CI) were calculated. Prediction models based on the results of the multivariate analysis were computed. p-Values < 0.05 in two-sided testing were considered statistically significant. Receiver operating characteristic (ROC) curve analyses were conducted to determine the accuracy of prediction models.

Results

Baseline characteristics

The selection process of the patients eligible for this study is presented in Figure 1. Accordingly, 852 patients (mean age: 75.2 ± 12 years, female/male: 1/2.1, CCI: 4.4 ± 2.1, ASA status: 2.8 ± 0.6) met the inclusion criteria. Of the included patients, 186 had AF (21.8%), 160 had CHD (18.8%), 84 had CKD (10.2%), and 443 received AAT (52%), which was perioperatively discontinued. Prior to admission, 134 patients (15.7%) were diagnosed with dementia, 40 patients (4.7%) with epilepsy, and 650 (76.3%) reported a recent head trauma. At admission, 393 patients (46.1%) had a GCS score < 15, and 400 individuals (46.9%) had a poor mRS score (Table 1).

Table 1.

Baseline Characteristics

Variables	n/%
	Age (mean ± SD)	75.2 ± 11.9
Sex (ratio female : male)	1:2.1
Charlson comorbidity index (mean ± SD)	4.4 ± 2.1
ASA status (mean ± SD)	2.8 ± 0.6
Recent head trauma	650/76.3
History of epilepsy	40/4.7
Dementia	134/15.7
Arterial hypertension	581/68.2
Diabetes mellitus	193/22.7
Chronic kidney disease	87/10.2
Atrial fibrillation	186/21.8
Chronic heart disease	160/18.8
Cerebral vascular/transient ischemic accident	118/13.8
Operative technique (burr hole)	741/87
CSDH site (bilateral)	177/20.8
Prior antiplatelet/anticoagulant therapy	443/52
Thrombophilia	59/6.9
At admission: mRS >3	400/46.9
At discharge: mRS >3	312/36.6
Preoperative seizure <24 h	51/6
Preoperative neurological deficits	528/62
Preoperative altered state of consciousness	393/46.1
Postoperative seizure	57/6.7
Postoperative bleeding	62/7.3
Postoperative pneumonia	27/3.2
Postoperative pulmonary embolism	3/0.4
Postoperative myocardial infarction	3/0.4
Postoperative stroke	3/0.3
New neurological deficit at discharge	33/3.9
Total hospital stay in days	11.2 ± 7
Recurrence	122/14.3
In-hospital mortality	31/3.6

ASA, American Society of Anesthesiology; CSDH, chronic subdural hematoma; SD, standard deviation.

Characteristics and predictors of in-hospital mortality

In-hospital mortality was observed in 31 patients (3.6%). In the univariate analysis, in-hospital mortality was statistically significantly associated with prior AAT use (p = 0.043), age (p = 0.005), CCI (p < 0.001), ASA status (p = 0.001), dementia (p = 0.005), CKD (p = 0.031), thrombophilia (p = 0.016), preoperative mRS (p < 0.001), preoperative GCS score (p < 0.001), preoperative Markwalder grading (p < 0.001), preoperative platelet count (p < 0.001), preoperative hemoglobin level (p = 0.003), minicraniotomy (p = 0.013), and postoperative bleeding (p = 0.005). Furthermore, epilepsy (p = 0.052), arterial hypertension (p = 0.075), and AF (p = 0.075) showed only a non-significant statistical trend. After adjusting for all confounders, binomial logistic regression revealed that prior AAT use does not predict in-hospital mortality (p = 0.099) (Table 2 and Fig. 2A). A ROC curve analysis determined the model’s discriminative capacity and was excellent with an area under the curve (AUC) of 0.87 (95% CI 0.78–0.95) (Fig. 2B).

Table 2.

Univariate and Multivariate Analysis of Predictors of in-Hospital Mortality

Predictors	Univariate analysis		Multivariate analysis
	Test statistic	p Value*	aOR (95% CI)	p Value*
Age	2.501 b	0.005	—	0.160
Sex	0.616 a	0.447	—	—
Charlson Comorbidity Index	17.946 b	<0.001	1.374 (1.130–1.670)	0.001
ASA status	11.465 b	0.001	—	0.180
Recent head trauma	0.504 a	0.518	—	—
Epilepsy	4.844 a	0.052	—	0.167
Dementia	9.474 a	0.005	2.771 (1.107–6.936)	0.029
Hypertension	3.646 a	0.075	—	0.171
Diabetes mellitus	3.023 a	0.122	—	—
Chronic kidney disease	5.368 a	0.031	—	0.659
Atrial fibrillation	3.514 a	0.075	—	0.260
Chronic heart disease	1.042 a	0.346	—	—
CVA/TIA	0.469 a	0.607	—	—
Thrombophilia	7.712 a	0.016	—	0.113
Preoperative seizure <24 h	0.012 a	1.000	—	—
CSDH location	0.039 a	1.000	—	—
Prior AAT use	4.639 a	0.043	—	0.099
Preoperative mRS score	11.993 b	<0.001	—	0.991
Preoperative Markwalder grading score	17.211 b	<0.001	—	0.275
Preoperative hemoglobin	10.351 b	0.003	—	0.253
Preoperative platelet count	7.426 b	<0.001	0.991 (0.985–0.997)	0.006
Preoperative GCS score	38.204 b	<0.001	0.73 (0.631–0.845)	<0.005
Preoperative neurological deficits	0.208 b	0.707	—	—
Operative technique	7.272 a	0.013	—	0.160
Postoperative seizure	1.989 a	0.259	—	—
Postoperative bleeding	11.166 a	0.005	4.886 (1.640–14.556)	0.004
Postoperative myocardial infarction	7.572 a	0.105	—	—
Postoperative pulmonary embolism	0.038 a	1.000	—	—
Postoperative stroke	7.572 a	0.105	—	—

—Stepwise exclusion from model.

^aChi-Square test used.

^bMann–Whitney-U test used.

*p Value of <0.05 was deemed significant. Predictors with a p value of <0.1 were included in the model.

AAT, antiplatelet and anticoagulant therapy; ASA, American Society of Anesthesiology; CI, confidence interval; CSDH, chronic subdural hematoma; CVA, cardiovascular accident; GCS, Glasgow Coma Scale; mRS, Modified Ranking Scale; OR, odds ratio; TIA, transient ischemic attack.

FIG. 2.

Independent predictors of mortality do not include prior AAT use. (A) Independent predictors of mortality are shown in a forest plot after stepwise exclusion of relevant predictors. These include the Charlson Comorbidity Index (p = 0.001), dementia (p = 0.029), preoperative platelet count (p = 0.006), postoperative bleeding (0 = 0.004) and preoperative GCS (p < 0.001). (B) ROC curve analysis determined a discriminative capacity of the model with an AUC of 0.87. GCS, Glasgow Coma Scale; AAT, antiplatelet and anticoagulant therapy; ROC, receiver operating characteristic; AUC, area under the curve.

Characteristics and predictors of functional outcome at discharge

At discharge, 312 individuals (36.6%) had a poor outcome. In the univariate analysis, functional outcome at discharge was statistically significantly associated with prior AAT use (p < 0.001), age (p < 0.001), CCI (p < 0.001), ASA status (p = 0.001), recent head trauma (p = 0.009), epilepsy (p < 0.001), dementia (p < 0.001), arterial hypertension (p = 0.015), diabetes mellitus (p < 0.001), CKD (p < 0.001), AF (p < 0.001), CHD (p = 0.002), CVA/TIA (p = 0.01), preoperative neurological deficits (p < 0.001), preoperative GCS-score (p < 0.001), preoperative Markwalder grading scale (p < 0.001), preoperative mRS (p < 0.001), preoperative hemoglobin level (p < 0.001), operative technique (p = 0.02), postoperative bleeding (p = 0.006), postoperative seizure (p < 0.001), postoperative myocardial infarction (p = 0.049), and postoperative stroke (p = 0.049). After adjusting for all confounders, binomial logistic regression revealed that prior AAT use does not predict functional outcome at discharge (p = 0.956) (Table 3 and Fig. 3A). A ROC curve analysis determined the model’s discriminative capacity and was outstanding with an AUC of 0.92 (95% CI 0.90–0.94) (Fig. 3B).

Table 3.

Univariate and Multivariate Analysis of Predictors of Poor Functional Outcome

Predictors	Univariate analysis		Multivariate analysis
	Test statistic	p Value*	aOR (95% CI)	p Value*
Age	118732.5 b	<0.001	1.037 (1.015–1.059)	0.001
Sex	0.427 a	0.543	—	—
Charlson Comorbidity Index	120148.5 b	<0.001	—	0.455
ASA status	107458 b	<0.001	1.815 (1.239–2.658)	0.002
Recent head trauma	7.132 a	0.009	—	0.794
Epilepsy	12.112 a	0.001	3.979 (1.385–11.429)	0.010
Dementia	80.487 a	<0.001	2.587 (1.477–4.531)	<0.001
Hypertension	6.148 a	0.015	—	0.343
Diabetes mellitus	26.540 a	<0.001	—	0.242
Chronic kidney disease	20.207 a	<0.001	—	0.375
Atrial fibrillation	12.929 a	<0.001	—	0.747
Chronic heart disease	10.048 a	0.002	—	0.386
CVA/TIA	6.932 a	0.01	—	0.585
Thrombophilia	0.533 a	0.488	—	—
Preoperative seizure <24 h	2.547 a	0.133	—	—
CSDH location	0.448 a	0.540	—	—
Prior AAT use	24.500 a	<0.002	—	0.956
Preoperative mRS score	177.075 a	<0.001	32.793 (19.612–54.832)	<0.001
Preoperative Markwalder grading score	109672 b	<0.001	—	0.932
Preoperative hemoglobin level	52019 b	<0.001	0.810 (0.723–0.909)	<0.001
Preoperative platelet count	79848.5 b	0.880	—	—
Preoperative GCS score	45885.5 b	<0.001	—	0.760
Preoperative neurological deficits	16.403 a	<0.001	—	0.509
Operative technique	5.752 a	0.02	—	0.346
Postoperative seizure	26.618 a	<0.001	4.534 (1.836–11.196)	0.001
Postoperative bleeding	7.945 a	0.006	3.310 (1.423–7.702)	0.005
Postoperative myocardial infarction	5.211 a	0.049	—	0.347
Postoperative pulmonary embolism	1.733 a	0.366	—	—
Postoperative stroke	5.211 a	0.049	—	0.288

—Stepwise exclusion from model.

^aChi-Square test used.

^bMann–Whitney-U test used.

*p Value of <0.05 was deemed significant. Predictors with a p value of <0.1 were included in the model.

AAT, antiplatelet and anticoagulant therapy; ASA, American Society of Anesthesiology; CI, confidence interval; CSDH, chronic subdural hematoma; CVA, cardiovascular accident; GCS, Glasgow Coma Scale; mRS, Modified Ranking Scale; OR, odds ratio; TIA, transient ischemic attack.

FIG. 3.

Independent predictors of poor functional outcome do not include prior AAT use. (A) Independent predictors of poor functional outcome (mRS > 3) are shown in a forest plot after stepwise exclusion of relevant predictors. These include age (p = 0.001), epilepsy (p = 0.01), dementia (p = 0.001), postoperative bleeding (p = 0.005), preoperative mRS (0 < 0.001), postoperative seizure (p = 0.001), ASA status (p = 0.002) and preoperative hemoglobin level (p < 0.001). (B) ROC curve analysis determined a discriminative capacity of the model with an AUC of 0.92. mRS, Modified Ranking Scale; ASA, American Society of Anesthesiologists; AAT, antiplatelet and anticoagulant therapy; ROC, receiver operating characteristic; AUC, area under the curve.

Characteristics and predictors of recurrence

Recurrence occurred in 122 patients (14.3%). In the univariate analysis, recurrence was not statistically significantly associated with prior AAT use (p = 0.434). We then analyzed the association of recurrence between patients without AAT (n = 409) and patients with prior aspirin therapy (n = 217, p = 0.012), vitamin K antagonists (n = 116, p = 0.666), and DOAC (n = 57, p = 1.000). Interestingly, we found that patients with prior aspirin therapy had statistically significantly lower recurrence rates than patients without prior AAT use (p = 0.012). A further analysis comparing patients with prior aspirin therapy (n = 217) with all other patients (without prior AAT use and with prior AAT use other than aspirin) showed a statistically significant association of prior aspirin use with lower recurrence rates (p = 0.003). Furthermore, addressing patients with prior aspirin therapy and patients without prior AAT, recurrence was associated with preoperative platelet count (p = 0.030), whereas female sex (p = 0.097) showed a nonsignificant statistical trend. After adjusting for all confounders (sex and preoperative platelet count), binomial logistic regression revealed that patients with prior aspirin therapy had significantly lower recurrence rates (p = 0.015, aOR = 0.504; 95% CI 0.290–0.877) (Table 4). The other variables did not reach statistical significance. Nevertheless, the model’s discriminative capacity was determined by a ROC curve analysis and was poor, with an AUC of 0.57 (95% CI 0.51–0.64) (Fig. 4).

Table 4.

Univariate and Multivariate Analysis of Predictors of Recurrence

Predictors	Univariate analysis		Multivariate analysis
	Test statistic	p Value*	aOR (95% CI)	p Value*
Age	44544 b	0.996	—	—
Sex	2.874 a	0.097	—	0.052
Charlson Comorbidity Index	21796 b	0.854	—	—
ASA status	21.556 b	0.689	—	—
Recent head trauma	1.063 a	0.320	—	—
Epilepsy	0.308 a	0.613	—	—
Dementia	0.033 a	0.873	—	—
Hypertension	0.060 a	0.901	—	—
Diabetes mellitus	0.150 a	0.768	—	—
Chronic kidney disease	0.979 a	0.381	—	—
Atrial fibrillation	1.248 a	0.308	—	—
Chronic heart disease	0.669 a	0.495	—	—
CVA/TIA	0.385 a	0.557	—	—
Thrombophilia	1.290 a	0.342	—	—
Preoperative seizure <24 h	0.031 a	1.000	—	—
CSDH location	3.382 a	0.185	—	—
Prior AAT use	0.754 a	0.434	—	—
Prior aspirin use	6.360 a	0.012	0.504 (0.29–0.88)	0.015
Prior VKA use	0.229 a	0.666	—	—
Prior DOAC use	0.006 a	1.000	—	—
Preoperative mRS score	0.002 a	1.000	—	—
Preoperative Markwalder grading score	21327.5 b	0.650	—	—
Preoperative hemoglobin level	21747 b	0.710	—	—
Preoperative platelet count	17935 b	0.030	—	0.066
Preoperative GCS score	22443 b	0.685	—	—
Preoperative neurological deficits	0.589 a	0.460	—	—
Operative technique	0.717 a	0.495	—	—
Postoperative seizure	0.641 a	0.500	—	—
Postoperative bleeding	0.295 a	0.643	—	—
Postoperative myocardial infarction	0.448 a	1.000	—	—
Postoperative pulmonary embolism	0.149 a	1.000	—	—
Postoperative stroke	0.149 a	1.000	—	—

—Stepwise exclusion from model.

^aChi-Square test used.

^bMann–Whitney-U test used.

*p Value of < 0.05 was deemed significant. Predictors with a p value of < 0.1 were included in the model.

AAT, antiplatelet and anticoagulant therapy; ASA, American Society of Anesthesiology; CI, confidence interval; CSDH, chronic subdural hematoma; CVA, cardiovascular accident; DOAC, direct oral anticoagulants; GCS, Glasgow Coma Scale; mRS, Modified Ranking Scale; OR, odds ratio; TIA, transient ischemic attack; VKA, vitamin K antagonists.

FIG. 4.

Independent predictors of recurrence. This includes only prior aspirin therapy, compared to all other variables (p = 0.015, aOR = 0.504; 95% CI 0.290–0.877). ROC curve analysis determined a discriminative capacity of the model with an AUC of 0.57 (95% CI 0.51–0.64). AAT, antiplatelet and anticoagulant therapy; ROC, receiver operating characteristic; AUC, area under the curve.

Discussion

This comparative study includes a large patient cohort and offers robust, contemporary evidence on the perioperative management of CSDH patients with prior AAT use.

In our cohort, discontinuation of AAT did not independently affect in-hospital mortality, compared with patients without prior AAT use. Multivariate analysis indicated that in-hospital mortality among patients with prior AAT use was more likely related to underlying comorbidities, such as higher CCI and presence of dementia; baseline clinical and laboratory parameters, including lower GCS score and platelet count; as well as postoperative complications like postoperative bleeding. The existing literature on AAT in patients with CSDH remains inconsistent, particularly concerning perioperative risks. Earlier studies suggested that antithrombotic therapy may increase perioperative morbidity, mainly due to a higher incidence of bleeding complications.^13,14 However, our findings contrast these results and align with most recent publications in multimorbid patient populations,^7,10,15–19 which similarly found no association between AAT and in-hospital mortality.

Furthermore, our analyses demonstrated that discontinuation of AAT was not an independent predictor of poor functional outcome at discharge. Instead, unfavorable functional recovery was multivariately associated with higher age and ASA status, comorbidities such as epilepsy and dementia, baseline functional and hematologic indicators (higher preoperative mRS score and lower hemoglobin level), and postoperative complications including seizure and bleeding. Few studies have systematically assessed functional outcomes following AAT discontinuation. Notable investigations by Poon et al. ⁴ and Younsi et al., ¹² consistent with our results demonstrated that although comorbid conditions and perioperative complications may contribute to poorer outcomes, AAT itself is not an independent predictor.

In contrast to in-hospital mortality and functional outcomes, our findings on recurrence revealed that, compared with patients without prior AAT use, preoperative discontinuation of aspirin in patients with prior aspirin therapy was independently associated with lower recurrence rates. Earlier studies indicated a possible association between perioperative use of antithrombotic therapy and an increased risk of CSDH recurrence.^14,17,20,21 This relationship was further highlighted in a large meta-analysis by Wang et al., ¹⁰ which showed that perioperative AAT use significantly raises recurrence rates without finding any substantial difference between antiplatelet and anticoagulant therapy in the subgroup analyses. Considering these previous findings together with our results, focusing specifically on patients with prior aspirin exposure who underwent temporary perioperative discontinuation, our data suggest a potential protective effect of prior aspirin therapy. A similar trend was observed, though not explored in detail, in the study by Ridwan et al. ²² This observation accords with hypotheses derived from cerebrovascular pathology, where aspirin’s combined anti-inflammatory and antithrombotic properties have been linked to protection against growth, bleeding, and rebleeding in conditions such as intracranial cavernomas^23–25 and aneurysms.^26–30 Moreover, the mechanistic insights reinforce this clinical observation. CSDH recurrence and resolution are strongly influenced by angiogenesis and inflammation. Increased cyclooxygenase-2 (COX-2) activity promotes vascular endothelial growth factor production and angiogenesis; both associated with elevated prostaglandin E2 (PGE2) levels and higher recurrence rates.^31–34 The observed protective effect of aspirin may stem from its inhibition of COX-2 enzyme and subsequent suppression of PGE2 synthesis, thereby attenuating inflammatory and angiogenic cascades from the onset of CSDH through diagnosis. Theoretically, this mechanism could reduce microvascular injury within the CSDH membrane during hematoma formation and expansion and help explain the lower recurrence rates observed in CSDH patients under aspirin therapy. The Surgical Evacuation of Chronic Subdural Hematoma and Aspirin randomized trial ³⁵ further demonstrated that prior aspirin therapy—whether continued or discontinued perioperatively—did not affect recurrence rates, conveying that the proposed protective effect of aspirin on CSDH membrane occurs well before diagnosis and surgery, particularly during the development and expansion of CSDH. In particular, the extent of microvascular injury due to inflammation and rupture of fragile microvessels within the CSDH membrane appears to be an intrinsic factor that is largely irreversible once the damage has occurred, except by interrupting its vascular supply through middle meningeal artery embolization, which has proven effective in preventing recurrence.^36–39 Given the possibility that recurrence risk is, to some extent, predetermined by the degree of microvascular injury within the CSDH membrane at the time of diagnosis—a concept that may represent a paradigm shift in both the understanding of recurrence and the perspective from which its treatment strategies are developed—future research should aim to identify reliable surrogate markers that can inform neurosurgeons about the degree of the CSDH membrane microvascular pathology and guide individualized recurrence risk assessment.

Limitations of the study

This study has certain limitations. First, the analysis was conducted on a single-center cohort, which may limit the generalizability of the findings to broader populations. Nevertheless, our observation of lower recurrence rates in patients with prior aspirin therapy is consistent with an early report—although this prior study did not address the key mechanistic questions explored here. Second, the predictive model assessing the effect of aspirin demonstrated limited explanatory power, as reflected by the low Nagelkerke R² value and suboptimal AUC in the ROC analysis. This limited power may reflect heterogeneity in clotting propensity within the cohort of patients with prior aspirin therapy. These findings underscore the need for further investigations, ideally incorporating subgroup analyses based on relevant biological markers that capture individual variability in clotting propensity.

Conclusion

This large-scale retrospective study supports the safety of AAT discontinuation in patients with CSDH. Our findings indicate no significant independent association between prior AAT use and in-hospital mortality, poor functional outcomes, or recurrence. Intriguingly, prior aspirin therapy was linked to a reduced recurrence risk, suggesting a potential protective effect. Although AAT is often perceived as a factor increasing surgical risk and recurrence, our results challenge this notion, emphasizing that treatment decisions should weigh individual risk and recurrence profiles rather than default to empiric management.

Transparency, Rigor, and Reproducibility Statement

All supporting data are available within this article. Anonymized data will be shared upon request from any qualified investigator and after approval of the Institutional Review Board of University Hospital Essen.

Authors’ Contributions

S.S. acquired, analyzed and interpreted the data, drafted the initial article, and reviewed and revised the final version of the article. S.A.-R., M.A., and E.T. acquired the data, reviewed and revised the final version of the article. J.R., F.D., O.G., R.J., and P.D. analyzed and interpreted the data, reviewed and revised the final version of the article. U.S. reviewed and revised the final version of the article for important intellectual content. M.C. conceptualized and designed the study, analyzed and interpreted the data, drafted the initial article, revised the article for important intellectual content and reviewed and revised the final version of the article. All authors approved the final article as submitted and agree to be accountable for all aspects of the work.

Abbreviations Used

AAT

antiplatelet and anticoagulation therapy

AF

atrial fibrillation

aOR

adjusted odds ratio

ASA

American Society of Anesthesiologists

aspirin

Acetylsalicylic acid

AUC

area under the curve

CCI

Charlson comorbidity index

CHD

chronic heart disease

CI

confidence intervals

CKD

chronic kidney disease

COX-2

cyclooxygenase-2

CSDH

chronic subdural hematoma

CVA/TIA

cerebral vascular accident or transient ischemic accident

DOAC

direct oral anticoagulants

GCS

Glasgow coma scale

mRS

modified Rankin scale

PGE2

prostaglandin E2

ROC

receiver operating characteristic

SD

standard deviation

Cite this article as: Salmanian S, Rodemerk J, Al-Rubaiey S, et al. Perioperative discontinuation of antiplatelet and anticoagulation therapy in chronic subdural hematoma: Impact on in-hospital mortality, functional outcome, and recurrence, Neurotrauma Reports, 2026, 7, 2689288X261444630, doi: 10.1177/2689288X261444630.