Optimal timing of anticoagulation therapy for venous thromboembolism prevention after intracranial hemorrhage

Abstract

Objective

This study aimed to explore the optimal timing of anticoagulation initiation for the prevention of VTE in patients with ICH.

Methods

This retrospective cohort study enrolled consecutive patients in Nanjing Drum Tower Hospital and The Medical Information Mart for Intensive Care IV database in the United States who were diagnosed with ICH. The restricted cubic spline was fitted to explore the dose-response relationship between anticoagulation initiation time and clinical outcomes. Based on the identified inflection point, patients were stratified into the ≤ 2 days group and the >2 days group according to the timing of anticoagulant initiation. The primary outcome was a composite outcome of VTE and intracranial rebleeding events.

Results

A total of 3,841 patients were divided into ≤ 2 days group (n = 2047) and >2 days group (n = 1794). HRs for ≤ 2 days group and >2 days group of composite outcome, VTE and intracranial rebleeding events were 1.20 (1.02–1.43), 2.01 (1.59–2.55) and 0.68 (0.53–0.87). For the subgroup analysis, significant interactions were observed between gender, race, etiology, location of ICH, and treatment group for the composite outcome (P<0.05). Compared to the all other races, a difference was observed in the relationship between anticoagulation initiation time and outcome events in Asians.

Conclusions

Initiating anticoagulation within 2 days of ICH can balance bleeding and thrombosis risks, and should be administered to patients of all races to optimize outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12959-026-00836-x.

Introduction

Intracranial hemorrhage (ICH) is a serious neurological condition characterized by the accumulation of blood within brain tissue due to the rupture of cerebral blood arteries, leading to compression and eventual injury to cerebral structures. The high mortality rates of ICH pose significant challenges to clinical medicine worldwide [1, 2]. ICH is commonly seen in hypertension, rupture of cerebral aneurysms, adverse reactions to anticoagulants, and trauma, and is a type of acute cerebral stroke [3].

Due to extended bed rest, restricted mobility and the presence of risk factors such as advanced age, obesity, and a hypercoagulable state, patients with ICH are typically at high risk for venous thromboembolism (VTE), mostly occurring within the first two weeks of admission [4, 5]. The occurrence of VTE in patients with ICH, particularly pulmonary embolism (PE), is associated with a notable increase in mortality rates. In two prospective studies, the incidence of deep venous thrombosis (DVT) in patients with ICH during hospitalization reached 20–40%. Simultaneously, DVT can induce pain and swelling in the afflicted extremity, restricting rehabilitation exercises and patient movement, thereby extending hospital duration and recovery time [6, 7]. Consequently, the prevention of VTE in patients with ICH is a critical strategy to enhance treatment results and quality of life.

The application of anticoagulant therapy for the prevention of VTE after ICH has attracted widespread attention, with multiple researches exploring the efficacy and safety of such treatment in patients with ICH [8, 9]. Current guidelines recommended initiating anticoagulation within 24 to 48 h after ICH based on individual risk assessment, yet the data supporting personalized management across different populations were limited [10]. Most existing researches focused on generalized timing windows without in-depth analysis of the etiology (spontaneous or traumatic), locations of ICH and severity. These factors can influence critical pathological characteristics, such as hematoma volume, cerebral edema, and hematoma expansion, all of which may affect the efficacy and safety of anticoagulation therapy. Moreover, there was an increasing recognition of the influence of racial difference on the clinical outcomes of anticoagulant therapy [11]. Racial groupings exhibited variations in drug metabolism, bleeding predispositions, and vascular disease, which may significantly influence the efficacy and safety of anticoagulant therapy. Research indicates that Asian individuals may have an increased risk of drug-related bleeding, a factor that has not been well examined in relation to VTE prophylaxis [12, 13].

Our study aimed to provide a comprehensive analysis of ICH patients by stratifying them based on etiology, location, severity and racial background. This allowed us to assess the impact of these multidimensional factors on the efficacy and safety of anticoagulation therapy for VTE prevention. By identifying the specific needs of various patient subgroups, we intended to contribute to the development of more personalized anticoagulation strategies. Addressing this gap in current guidelines, our research sought to optimize the timing of anticoagulation for VTE prevention following ICH and establish a strong foundation for individualized anticoagulation.

Methods

This retrospective cohort study included consecutive patients with ICH in Nanjing Drum Tower Hospital and The Medical Information Mart for Intensive Care IV (MIMIC-IV, version 2.2) database in the United States. To obtain the data from MIMIC-IV, we successfully completed the National Institutes of Health (NIH) training course on safeguarding human research participants and passed the Collaborative Institutional Training Initiative (CITI) examinations. A waiver of informed consent was approved as the database lacked protected health information and all patient data was anonymized. The study utilizing data from our medical center was approved by the Institutional Ethics Committees of Nanjing Drum Tower Hospital (Approval Number: 2024-837-02), and the study protocol met the ethical principles of Declaration of Helsinki.

Study design and patients

Inclusion criteria were as follows: (1) New ICH confirmed by assessment of symptoms and imaging (CT or MRI); (2) Age ≥ 18 years; (3) Anticoagulant therapy administered during hospitalization (unfractionated heparin [UFH] or low-molecular-weight heparin [LMWH]); (4) Availability of complete clinical case data and relevant laboratory results.

Exclusion criteria included: (1) VTE had already occurred at the time of initiating anticoagulant therapy; (2) Presence of active clinical conditions requiring systemic anticoagulation (e.g., atrial fibrillation, valvular heart disease); (3) Death or discharge within 72 h of admission; (4) History of DVT or PE within the past 3 months; (5) Severe renal insufficiency (eGFR < 15 mL/min/1.73 m²).

Data extraction and covariates

Data extraction was performed using Navicat Premium (Version 17.0.4) with structured query language (SQL). The study examined various covariates categorized as follows: (1) Demographic characteristics: Age, gender and race. (2) Complicating diseases: Hypertension, diabetes, coronary heart disease (CHD), ischemic stroke (IS), heart failure (HF) and cancer. (3) Etiology: Spontaneous (Hypertension, antithrombotic therapy, vascular malformations, brain tumors, coagulopathies, etc.) and traumatic (Trauma, fractures of the skull and coup-contrecoup injury, etc.) ICH. (4) location of ICH: Intraparenchymal (IPH), intraventricular hemorrhage (IVH), epidural hemorrhage (EDH), subdural hemorrhage (SDH), and subarachnoid hemorrhage (SAH), etc. (5) Severity: Glasgow Coma Scale (GCS) score was a neurological assessment tool used to evaluate the level of consciousness, particularly in patients with head trauma or other etiologies of impaired consciousness. The severity was classified into severe (score: 3–8), moderate (score: 9–12) and mild (score: 13–15). The scores were extracted within the first 24 h of admission. (6) Laboratory indicators: Platelet, hemoglobin (HB) and creatinine. The baseline was extracted using the first record after admission. The estimated glomerular filtration rate (eGFR) was calculated using the MDRD formula, and renal function was classified as normal renal function (eGFR ≥ 90 mL/min/1.73 m²), mild renal insufficiency (eGFR: 60–89 mL/min/1.73 m²), moderate renal insufficiency (eGFR: 30–59 mL/min/1.73 m²), and severe renal insufficiency (eGFR: 15–30 mL/min/1.73 m²). (7) Anticoagulation initiation time: The interval from hospital admission to the initiation of anticoagulation. (8) Anticoagulant type and dosage: The anticoagulants used in this study were low molecular weight heparin (LMWH) or unfractionated heparin (UFH). The LMWH dosage was 0.4 mL (nadroparin: 0.4mL/4000AXaIU; enoxaparin: 0.4mL/4100AXaIU) once daily; the UFH dosage was 5000 IU subcutaneously every 12 h.

Study outcomes

The patients were followed up for 3 months. The primary outcome was defined as a composite outcome of VTE and intracranial rebleeding events. The secondary outcomes were the occurrence of VTE and intracranial rebleeding events. VTE events were defined as a composite of symptomatic or asymptomatic DVT of the lower extremities and PE, diagnosed by color Doppler ultrasound, venography, or pulmonary angiography. Intracranial rebleeding included hematoma propagation confirmed by imaging and development of a new bleeding unrelated to original site.

Statistical analysis

Continuous variables were reported as the mean ± standard deviation or median and interquartile range (IQR), whereas categorical variables were presented as quantity and frequency. Variables with over 10% missing values were omitted. Variables exhibiting missing values were imputed using the multiple imputation of chained equations with baseline characteristics.

The restricted cubic spline (RCS) with four knots was fitted to explore the dose-response relationship between anticoagulation initiation time and clinical outcomes. Based on the identified inflection point, patients were stratified into different groups according to anticoagulation initiation time.

To balance baseline differences and evaluate the differences in clinical outcomes between two groups, multivariable Cox proportional hazards regression models were employed. Three Cox models were fitted with incremental adjustments for confounders. Model 1 did not adjust for any variables. Model 2 additionally accounted for age, gender, and complicating diseases. Model 3 further adjusted for etiology and location of ICH, GCS score, anticoagulation duration, and laboratory indexes based on Model 2. To explore the association between different anticoagulation initiation time groups with clinical outcomes, the cumulative incidences of 3-month outcome events were estimated using Kaplan-Meier curves and detected by the log-rank test. Statistical analysis was performed with R version 4.2.2 statistical software (The R Foundation for Statistical Computing). Hazard ratios (HRs) with 95% confidence intervals (CIs) were calculated to compare the risks of clinical outcomes between the different groups.

Risk of clinical outcomes in specified subgroups (age, gender, race, GCS score, etiology, location of ICH and eGFR) was assessed. For subgroup analysis, multivariate Cox proportional hazards regression was employed. The significance of interaction between treatment group and subgroup was defined as P for interaction < 0.05.

To assess the potential impact of immortal time bias, we conducted a landmark analysis using a 2-day post-admission time point. Specifically, we included only patients who remained alive and free of the primary outcome event at 2 days after hospitalization. Follow-up was restarted from this landmark time point, and patients were classified according to their anticoagulation status at that time. Subsequently, the same multivariate Cox proportional hazards model was applied to evaluate the association between anticoagulation timing and subsequent clinical outcomes. This analysis was performed to assess the robustness of the primary findings.

Results

Patient characteristics

A total of 3841 patients diagnosed with ICH were included in this study. The nonlinear relationship between anticoagulation initiation time and clinical outcomes was illustrated in Fig. 1, with the inflection points observed in the curve at approximately 2 days. Based on the inflection point of composite outcome, patients were divided into two groups: the ≤ 2 days anticoagulation initiation group (≤ 2 days group) and the > 2 days anticoagulation initiation group (> 2 days group).

Fig. 1.

Relationship between anticoagulation initiation time (days) and composite outcome (A), VTE (B) and intracranial rebleeding events (C)

Baseline characteristics of patients between the two groups were presented in the Table 1. A total of 2047 patients initiated anticoagulation therapy within 2 days, while 1794 patients started therapy after 2 days. Compared to the ≤ 2 days group, the > 2 days group had a higher proportion of male patients and individuals of Asian populations, as well as lower GCS scores, platelet counts, anticoagulation duration, and a lower prevalence of HF and cancer. Moreover, there were statistically significant differences between the two groups in the locations of ICH and eGFR.

Table 1.

Baseline characteristics of patients with anticoagulation initiation times of the ≤ 2 days vs. the > 2 days

Characteristic	≤ 2 days (n = 2047)	>2 days (n = 1794)	SMD
Ages, yesrs	65.28 ± 17.65	64.02 ± 16.98	0.072
Male, n(%)	1119 (54.7)	1130 (63.0)	0.170
Race, n(%)			0.716
Caucasian/African American	1321 (64.5)	717 (40.0)
Asians	292 (14.3)	807 (45.0)
Others	434 (21.2)	270 (15.1)
GCS, scores	13.82 ± 2.37	12.85 ± 3.38	0.331
Etiology, n(%)			0.077
Spontaneous	1357 (66.3)	1123 (62.6)
traumatic	690 (33.7)	671 (37.4)
Location, n(%)			0.241
IPH	398 (19.4)	387 (21.6)
SDH	449 (21.9)	280 (15.6)
SAH	455 (22.2)	319 (17.8)
Multifocality	655 (32.0)	739 (41.2)
Others	90 (4.4)	69 (3.8)
Hypertension, n(%)	1084 (53.0)	985 (54.9)	0.039
Diabetes, n(%)	229 (11.2)	221(12.3)	0.035
CHD, n(%)	166 (8.1)	170 (9.5)	0.048
IS, n(%)	393 (19.2)	309 (17.2)	0.051
HF, n(%)	232 (11.3)	117 (6.5)	0.169
Cancer, n(%)	321 (15.7)	197 (11.0)	0.139
eGFR, n(%)			0.170
≥ 90 mL/min/1.73m2	940 (45.9)	963 (53.7)
60–89 mL/min/1.73m2	717 (35.0)	552 (30.8)
30–59 mL/min/1.73m2	316 (15.4)	242 (13.5)
15–30 mL/min/1.73m2	74 (3.6)	37 (2.1)
Anticoagulation duration, days	12.01 ± 11.91	10.61 ± 12.97	0.112
Platelet,10^9/L。	211.47 ± 85.64	201.99 ± 91.16	0.107
HB, g/L	119.82 ± 27.34	119.92 ± 28.97	0.003

Clinical outcomes

In the comparison between the ≤ 2 days and > 2 days groups, the incidences of composite outcome, VTE and intracranial rebleeding events were 12.75% vs. 23.80%, 5.13% vs. 18.90% and 8.45% vs.7.30%, respectively. The cumulative risk curves for clinical outcomes between the two groups were shown in Fig. 2. Most events occurred within one month of follow-up period. Patients in the > 2 days group exhibited significantly higher incidence of composite outcome compared to those in ≤ 2 days group (log-rank P < 0.05). The HRs for Cox proportional hazards models 1, 2, and 3 were presented in Fig. 3. After adjusting for age, gender, and complicating diseases (Model 2), the HRs and 95% CIs for composite outcome, VTE and intracranial rebleeding events in the > 2 days group were 1.27 (1.08–1.50), 2.11 (1.67–2.66) and 0.75 (0.59–0.95), respectively. After further adjusting for etiology and location of ICH, GCS score, anticoagulation duration, and laboratory indexes (Model 3), the HRs and 95% CIs for composite outcome, VTE and intracranial rebleeding events in the > 2 days group were 1.20 (1.02–1.43), 2.01 (1.59–2.55) and 0.68 (0.53–0.87), respectively. In the three Cox proportional hazards models, as more variables were included, the HRs and 95% CIs for the composite outcome and VTE events between the two groups remained stable. Only the risk for intracranial rebleeding events slightly reduced after adjusting for variables (Model 1: HR: 0.85, 95% CI: 0.67–1.06; Model 2: HR: 0.75, 95% CI: 0.59–0.95; Model 3: HR: 0.68, 95% CI: 0.53–0.87).

Fig. 2.

The cumulative risk curves for composite outcome (A), VTE (B) and intracranial rebleeding events (C) in the ≤ 2 days group and the > 2 days group

Fig. 3.

Adjusted HRs and 95%CIs of clinical outcomes for different models in the > 2 days group compared to the ≤ 2 days group

Subgroup analysis

To further explore whether the relationship between different initiation groups and clinical outcomes across different conditions, subgroup analysis was conducted for gender, age, race, GCS score, etiology, location of ICH and eGFR (Fig. 4). The correlation between different initiation groups and composite outcome was not statistically significant regardless of the age, GCS score or eGFR subgroups. For the gender, race and etiology stratification, the > 2 days group was associated with a higher risk of composite outcome in female patients (HR = 1.58, 95%CI: 1.20–2.07), the Asians (HR = 1.54, 95%CI: 1.17–2.03) and traumatic etiology (HR = 1.54, 95%CI: 1.13–2.11) compared with the ≤ 2 days group, with significant interactions between the initiation groups and subgroups (P < 0.05). The Asians also showed similar result in terms of VTE outcome. For the location of ICH stratification, there was significant interaction between the location of ICH and subgroups (P < 0.001) for composite outcome that the > 2 days group increased the risk of composite outcomes in patients with IPH (HR = 1.55, 95%CI: 1.02–2.34) and SAH (HR = 1.74, 95%CI: 1.14–2.66).

Fig. 4.

Adjusted HRs of composite outcome (A), VTE (B) and intracranial rebleeding events (C) according to various subgroups in the ≤ 2 days group and the > 2 days group

Racial differences

In the subgroup analysis, a significant interaction was observed between different initiation groups and racial subgroups, indicating that the impact of race on clinical outcomes was not uniform. Based on these findings, we further examined the nonlinear relationship between anticoagulant initiation and clinical outcomes across different racial groups (Fig. 5). The RCS curve revealed that the inflection points for the Caucasian/African American and other racial groups observed in the curve at approximately 2 days, while the Asians showed a distinct inflection point around 7 days. For the Asians, initiating anticoagulation within 7 days increased the risk of composite outcome and VTE event as the initiation time was prolonged. Beyond 7 days, the risk of composite outcome and VTE event remained stable, while the risk of intracranial rebleeding event decreased with the extension of anticoagulation initiation time.

Fig. 5.

Relationship between anticoagulation initiation time(days) and composite outcome, VTE, and intracranial rebleeding events adjusted for confounding factors in patients of the Caucasian/African American, Asians, and other races (A. Caucasian/African American and composite outcome; B. Caucasian/African American and VTE; C. Caucasian/African American and intracranial rebleeding; D. Asians and composite outcome; E. Asians and VTE; F. Asians and intracranial rebleeding; G.Others and composite outcome; H. Others and VTE; I. Others and intracranial rebleeding)

Sensitivity analysis

In this landmark analysis, delayed initiation of anticoagulation (> 2 days) remained significantly associated with a higher risk of the composite outcome (HR 1.39, 95% CI 1.16–1.67; P < 0.001) and VTE (HR 2.50, 95% CI 1.94–3.24; P < 0.001), while being associated with a lower risk of intracranial rebleeding (HR 0.74, 95% CI 0.57–0.96; P = 0.024) (Table S1). These findings were consistent with the primary analysis, supporting the robustness of the main results.

Discussion

At present, the optimal timing for initiating this therapy remained a matter of considerable clinical debate. While early initiation can effectively mitigate the risk of thrombosis formation, it also posed a heightened risk of intracranial rebleeding. In this multicenter study of consecutive admissions for all-cause intracerebral hemorrhage, we observed that the incidence of VTE among ICH patients was strikingly high at 11.56% and a significant nonlinear relationship was identified between the timing of anticoagulation initiation and the occurrence of composite outcome. Moreover, compared to early administration, delay heparins administration beyond 48 h after admission was associated with an increased risk of composite outcomes, including VTE and intracranial rebleeding events.

Previous studies have yielded inconsistent findings. For instance, initiating anticoagulation within 24 h in patients with SAH has been linked to an increased risk of rebleeding. In contrast, Faust et al. reported that initiating anticoagulation within 48 h in patients with spontaneous ICH did not raise the risk of rebleeding [14, 15]. Furthermore, Qian et al. found that anticoagulation administered 24 h post-admission did not contribute to hematoma expansion, while Wasay et al. observed that anticoagulation initiated within 1 to 6 days could aid in preventing DVT, albeit with limited efficacy [16, 17]. Variations in study designs, characteristics of the patient populations have contributed to the lack of consensus or high-level recommendations regarding the optimal timing of anticoagulant therapy. The Canadian Best Practices for Stroke Care guidelines recommend initiating anticoagulant therapy 48 h after the onset of spontaneous ICH to prevent DVT. Meanwhile, the American Heart Association/American Stroke Association (AHA/ASA) guidelines advocate for initiating therapy within 24 to 48 h after onset to reduce the risk of VTE in patients with spontaneous ICH. Similarly, Chinese guidelines for diagnosis and treatment of acute intracerebral hemorrhage 2019 recommend starting anticoagulation within 1 to 4 days to prevent DVT in patients with spontaneous ICH [10, 18, 19].

In ICH patients, clinical decisions regarding the timing of anticoagulant initiation were profoundly influenced by factors such as the etiology, location of ICH and severity, and the management of comorbid conditions. The etiology of spontaneous and traumatic hemorrhage may significantly influence the optimal timing for anticoagulant initiation after ICH. Spontaneous hemorrhage, often linked to hypertension, cerebral small vessel disease, or anticoagulant use, typically involved deep brain regions like the basal ganglia and featured a gradual hematoma formation process with a high recurrence risk. In contrast, traumatic hemorrhage, resulting from physical injury and commonly involving epidural or subdural hematomas associated with skull fractures, followed a mechanical course where anticoagulation timing depended on hematoma resolution and stabilization [3]. The timing of anticoagulant initiation was also influenced by location of ICH. IPH, often occurring in deep brain structures like the basal ganglia or thalamus, involved large and slowly expanding hematomas that exerted prolonged pressure on brain tissue, with higher risks of rebleeding and neurological decline [20]. SAH, typically due to aneurysmal rupture, rapidly spread into the cerebrospinal fluid space, often leading to complications such as increased intracranial pressure, vasospasm, and cerebral edema [21]. EDH and SDH, generally caused by trauma, presented lower expansion risks and better outcomes when promptly managed [22]. Hemorrhage severity was crucial in determining the timing of anticoagulant initiation. Key factors included hematoma volume, expansion rate, brain tissue compression, and neurological impairment [23]. Larger, rapidly expanding hematomas generally required delaying anticoagulant therapy to reduce bleeding risks. For patients with multiple hemorrhages or acute intracranial hypertension, surgical stabilization was often needed before considering anticoagulation.

This study further explored the impact of these factors on clinical outcomes through subgroup analysis. The results of the subgroup analysis revealed that initiating anticoagulation more than 2 days after admission increased the risk of composite outcome in female patients, patients with traumatic hemorrhage, and patients with IPH or SAH. These findings were consistent with previous studies. Kawase et al., identified a correlation between female and an increased incidence of DVT following ICH. Additionally, studies by Wu et al. and Jakob et al. demonstrated that initiating anticoagulation therapy within 48 h significantly reduced the incidence of VTE in patients with traumatic hemorrhage and SAH without increasing the risk of rebleeding [7, 24, 25]. These findings emphasized the critical importance of prioritizing VTE prevention in high-risk populations and initiating anticoagulation therapy as early as possible-preferably within 2 days of admission, to strike an optimal balance between preventing thrombus formation and minimizing the risk of rebleeding.

This study indicated significant racial differences in the relationship between the timing of anticoagulation initiation after ICH and the risk of composite outcome. Subgroup analysis revealed that the anticoagulant initiation (> 2 days) increased the risk of composite outcome in the Asian population, a finding not observed in other racial groups. Moreover, the actual incidence of composite outcomes was significantly higher in the Asians (Table S2). Although the RCS curve for the Asians suggested an inflection point at 7 days, we further examined the relationship between anticoagulation initiation timing and the risk of composite outcomes using 2 days as the inflection point (Figure S1). The results showed that the risk of composite outcome rapidly increased as time extended before 2 days, with hazard ratio of 1.79 (1.00-3.18) per day, which differed from the findings with a 7-day inflection point. This suggested that the risk of composite outcome in the Asians primarily occurred mainly in the first 2 days, the incidence distribution of composite outcome further supported this (Figure S2). Thus, the significant increase in composite outcome risk after 2 days in the Asian subgroup emphasized the critical importance of early initiation of anticoagulation therapy. In contrast, the more moderated changes for the risk of composite outcome in other racial groups may reflect a lower composite outcome risk or a more stable balance between thrombotic and hemorrhagic risks over time. These racial differences had important clinical implications that initiating anticoagulation therapy within 2 days after ICH appeared to be more beneficial for the Asians. Conversely, in other racial groups, a more flexible approach to anticoagulation timing was considered feasible, with adjustments based on individual risk assessments. Baseline characteristics of patients between different races were presented in the Table S3.

The observational nature of our study necessitates careful consideration of potential biases. First, the decision to initiate anticoagulation therapy and the timing of its initiation were determined by clinician discretion. This approach may have introduced sample selection bias, as only patients who received anticoagulation therapy were included, a bias that may underestimate early mortality or severe complications in the overall ICH population. Clinically, patients not receiving anticoagulation are often deemed unsuitable for anticoagulation due to hematoma expansion, worsening neurological function, or early mortality. Therefore, the conclusions of this study may be more applicable to ICH patients with clinically stable who meet the clinical criteria for anticoagulation. Second, the exclusion of patients with early death or discharge could introduce immortal time bias. To assess the impact of this bias, we conducted a landmark sensitivity analysis at 2days post-admission. The consistency in the direction and magnitude of the hazard ratios for delayed initiation observed in the landmark analysis supports the robustness of our primary findings with respect to this potential bias.

Despite the robustness checks above, several limitations must be acknowledged. As a retrospective study, it is susceptible to information bias and confounding factors, which could compromise the accuracy and reliability of the findings. Furthermore, we lacked systematic monitoring data on anticoagulation intensity. The use of different anticoagulant agents may also introduce residual confounding. UFH and LMWH differ in pharmacologic characteristics and monitoring requirements, which could influence bleeding and thrombotic risk and were not fully accounted for in this retrospective analysis. To address these limitations, future research should involve multicenter, large-scale clinical trials or prospective studies to further validate and expand upon these findings. Such studies would allow for a more comprehensive evaluation of the impact of anticoagulation initiation timing on the prognosis of ICH patients.

Conclusion

In general, Initiating anticoagulation therapy within 2 days of ICH can effectively balance the risks of thrombus formation and rebleeding. However, an individualized treatment approach is essential, taking into account the cause of ICH, location of ICH, and the patient’s racial differences. Particularly for the Asian population, despite the theoretically lower risk of thrombus formation and the higher risk of bleeding, our study indicated that delaying anticoagulation treatment appeared to increase the incidence of VTE in the Asian population. Furthermore, the risk of composite outcome rapidly increased as time extended before 2 days. Therefore, we suggested that anticoagulation therapy should still be initiated as early as possible within 2 days of ICH in Asian patients to optimize clinical outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Q.N. contributed to data curation, formal analysis, investigation, methodology, software, writing–original draft, writing–review and editing, conceptualization, and visualization. Q.H.W. contributed to data curation, formal analysis, investigation, and writing–original draft. J.L.Q contributed to data curation, methodology, and writing–review and editing. J.C. contributed to writing–review and editing, investigation, supervision, and funding acquisition. T.Q. contributed to resources, supervision, writing–review and editing, validation, funding acquisition, and project administration. B.Y.W contributed to data curation, methodology, software, supervision, writing–review and editing, validation, and project administration.

Funding

The author(s) declare that financial support was received for the research, authorship, and/or publication of this article. This study was supported by funding from the Clinical Trials from the Affiliated Drum Tower Hospital, Medical School of Nanjing University (2021-LCYJ-PY-19).

Data availability

The raw data supporting the conclusion of this article will be made available by the corresponding authors, without undue reservation.

Declarations

Ethical approval

A waiver of informed consent was approved as the database lacked protected health information and all patient data was anonymized. The study utilizing data from our medical center was approved by the Institutional Ethics Committees of Nanjing Drum Tower Hospital (Approval Number: 2024-837-02), and the study protocol met the ethical principles of Declaration of Helsinki.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Consent to publish

Not applicable.

Statements and declarations

Confirm that manuscript complies with all instructions to authors. Confirm that authorship requirements have been met and the final manuscript was approved by all authors. Confirm that this manuscript has not been published elsewhere and is not under consideration by another journal. Confirm the use of reporting checklist.

Competing interests

The authors declare no competing interests.