Introduction

Spontaneous intracerebral hemorrhage (sICH) is the most lethal type of cerebrovascular disease, with a 30-day mortality rate of 30–50%1,2. The presence of paresis and activated coagulation in patients with sICH may increase the vulnerability to venous thromboembolism (VTE), including deep venous thrombosis (DVT) and pulmonary embolism (PE)3. DVT complication risk estimates vary greatly across diverse intracerebral hemorrhage (ICH) cohorts, which ranges from 2 to 75%, depending on the diagnostic method, time of evaluation, racial disparity, and drug therapy3,4,5,6,7. Clinically significant lower extremity deep venous thrombosis (LEDVT) is a life-threatening complication of post-sICH and is associated with substantial morbidity, mortality, and health-care cost8,9. However, the management strategies have presented a dilemma, particularly in term of how to weigh the neurological risk of hematoma expansion after anticoagulation against the clinical sequelae of an untreated LEDVT. Even worse, an increased risk of LEDVT following ICH can be observed10. Therefore, early identification of patients with sICH who are at increased risk of LEDVT is crucial to prevent DVT and ultimately enhance DVT treatment strategies, as it can lower the occurrence rate of VTE-related morbidity and mortality following sICH.

Previous studies have revealed that LEDVT was associated with inflammatory activation in stroke patients, where inflammatory mediators might trigger endothelial damage and activation of the coagulation cascade, exacerbating thrombus formation, including sICH11,12,13. Among existing inflammatory biomarkers, leukocytes are the most common inflammatory mediators whose predictive value has been widely demonstrated in patients with LEDVT12,14. Leukocytes can be categorized into neutrophils, lymphocytes, monocytes, eosinophils, and basophils, as is commonly understood. The levels of leukocyte subtypes exhibit diverse temporal changes following ICH12,15,16. In addition, each leukocyte subtype carries out various immunological functions and has varying effects on the pathophysiology of cerebrovascular diseases14,17,18.

Basophils account for less than 1.0% of blood leukocytes, which are highly specialized immune system cells involved in inflammatory responses. A study conducted by Karina Gasbrrino, et al. found that in female patients with severe carotid atherosclerosis, a high basophil-to-leukocyte ratio were associated with thrombotic events19. Pizzolo et al. also reported that elevated basophil levels were associated with a higher mortality risk in patients with stable coronary artery disease, as basophils are known to contribute to hemostasis and regulate platelet aggregation, potentially playing a role in thrombosis20. Moreover, increased basophil was associated with enhanced factor II plasma coagulant activity, thereby suggesting underlying prothrombotic mechanisms20. These findings suggest that basophils may play a significant role in thrombotic events. Nonetheless, the association of basophils with LEDVT in patients with sICH remains largely unexplored in the literature. In addition, the majority of literature emphasizes the occurrence of DVT in patients undergoing neurosurgical procedures11,21,22, with little attention given to the DVT status of those sICH patients who do not require surgery (which make up the majority of sICH patients). Hence, this retrospective study aimed to investigate whether the basophil level in peripheral blood (BCPB) upon admission was associated with LEDVT occurrence in non-surgical sICH patients.

Materials and methods

Study population

This retrospective observational study was conducted at The Second Affiliated Hospital of Fujian Medical University on consecutive patients who presented with sICH from January 2021 to August 2023. Data utilized in this study were collected from a prospective database that contained demographic and clinical information. The patients included in the present study had to meet the following criteria: (a) they needed to have a computed tomography (CT) scan for their index ICH admission within 24 h of symptom onset for diagnostic confirmation and ICH hematoma localization; (b) follow-up CT scans, CT angiography (CTA) and/or digital subtraction angiography (DSA) had to be conducted within 24 h of the baseline CT scan; (c) peripheral blood samples had to be collected via venipuncture from patients upon admission; (d) LEDVT had to be confirmed through Doppler ultrasonography. Exclusion criteria included: (a) experiencing surgical treatment, including external ventricular drainage, micro-invasive hematoma removal, craniotomy evacuation of hematoma, and decompressive craniectomy; (b) presence of a malignant tumor; (c) diagnosis of secondary ICH, such as hemorrhagic transformation from brain infarction; (d) absence of blood routine level or blood routine measured > 24 h after ICH onset; (e) history of stroke or mRS score > 2; (f) previous DVT or PE; (g) known thrombophilia; (h) asthma or other allergic disorders. Figure 1 illustrates the study design flowchart.

Fig. 1
figure 1

Diagram flowchart of inclusion and exclusion criteria of the study.

Full size image

Baseline data collection

Demographic information, medical history, laboratory findings, complications, and outcomes were extracted from the original perspective database to analyze the enrolled patients. The patient’s whole blood samples were collected via venous puncture within 2 h of admission and then stored at room temperature in a vacuum container tube with EDTA. Subsequently, the counts of each leukocyte subtype were examined with an automated blood analyzer. The measurements were all conducted by laboratory personnel blinded to patients’ clinical condition. Two radiologists with 5 or 6 years of experience diagnosing ICH manually measured the hematoma volume and calculated it using the ABC/2 method, as described previously23. Deep ICH is characterized by its selective impact on the thalamus, basal ganglia, internal capsule, and deep periventricular white matter. In contrast, lobar ICH is typically associated with the cortex and cortical-subcortical junction23,24. ICH hematoma expansion was characterized as an increase in hematoma size of 33% or 6 ml on the follow-up CT scan conducted at 24 h25,26,27.

The primary outcome measure was the presence of LEDVT within 14 days after the acute ICH episode. LEDVT was identified as proximal thrombosis of the iliac or superficial femoral vein, or thrombosis of at least the upper third part of the deep calf veins28, as confirmed by lower-extremity ultrasonography demonstrating non-compressibility of a distal or proximal vein29. Thrombosis found in the popliteal vein or above is considered proximal DVT, whereas thrombosis below the popliteal vein is considered distal DVT30,31,32.

Proximal and distal DVT are regarded as mixed LEDVT. Prophylaxis and treatment of LEDVT are performed in conjunction with previous reports in the literature and our institution’s treatment protocol11,22,33,34. Our institution’s treatment protocol is as follows. The Caprini risk assessment score is utilized to improve VTE prevention and determine the necessary thromboprophylaxis strategies accordingly35. If a patient with ICH presents with leg paralysis symptoms, it is important to start intermittent pneumatic compression (IPC) upon admission. Patients diagnosed with ICH who remain bedridden for more than three days are classified as high-risk for lower extremity thrombosis28. Along with increasing IPC therapy, we closely monitor coagulation functions and conduct lower extremity Doppler ultrasound in such cases. When patients exhibited signs or symptoms of LEDVT (such as swelling or pain), the diagnosis of LEDVT was confirmed through color Doppler ultrasound (CDU) waveforms and images from the thigh to the ankle32. CDU is regularly carried out weekly for patients in the neuro ICU and bi-weekly for patients in neurosurgery. If a patient shows abnormal blood coagulation function or elevated d-dimer levels, it is important to promptly perform a CDU. After securing the diagnosis of LEDVT in patients with sICH using CDU and evaluating the risk of rebleeding, treatment with low-molecular-weight heparin (LMWH, 100 U/kg, Q12H) was initiated11. The evaluation of rebleeding/hematoma expansion included clinical factors, such as patient stability and coagulation profiles, as well as imaging studies. Large baseline hematoma size, and CT image markers, such as black hole sign, island sign, and blend sign, are indicators of increased risk for rebleeding/hematoma expansion25,26,27, thus necessitating careful evaluation before administering LMWH. LMWH was typically initiated on the 3rd to 4th day following ICH and maintained for 14 days in sICH patients who were diagnosed with LEDVT at our center, which is in line with previous studies34. Using a retrievable filter as a temporary measure for patients with acute ICH and proximal LEDVT who are not yet suitable for anticoagulation until anticoagulation can be initiated is a reasonable approach36.

Statistical analysis

The statistical analyses utilized R software (version 4.2.1, R Foundation for Statistical Computing, Vienna, Austria) and MedCalc version 20.0.4 (MedCalc Software, Ostend, Belgium). Data were expressed as means (± standard deviations, SD) or medians (interquartile ranges, IQR), while categorical variables were expressed as counts (percentages). Continuous data with normality and non-normality were compared using Student’s t-tests and Mann–Whitney tests. Differences between categorical variables are assessed by employing the χ2 or Fisher’s exact tests. To address potential selection bias, a weighted univariate analysis was undertaken to identify the risk factors for LEDVT occurrence. Next, given that the primary outcome variable (LEDVT) is binary-coded, we adjusted for multiple covariates in our model to account for potential confounders. We used weighted multivariable logistic regression models to analyze the association basophil counts and the occurrence of LEDVT. The receiver operating characteristic (ROC) curve was generated to calculate the sensitivity, specificity, and area under the ROC curve (AUC). Patients were divided into two groups based on the ROC curve according to their optimal cut-off value.

Weighted multivariable logistic regression models determined correlations between BCBP and LEDVT. Model I was unadjusted. Model II adjusted for variables with a P-value < 0.05 in weighted univariate analysis. Statistical significance was determined in all analyses by a two-sided P-value less than 0.05.

Results

Out of the 315 patients with sICH who met the inclusion criteria, 47 (14.92%) patients experienced LEDVT. This included 39 (82.98%) patients with distal LE DVT, five (10.64%) patients with proximal LEDVT, and three (6.38%) patients with mixed LEDVT. The mean age of these patients was 60.7 ± 11.1 years, and the majority (63.8%) were male (Table 1). The ROC curves were generated to compare further the predictive power of different leukocyte subtypes, including neurocytes, lymphocytes, monocytes, eosinophils, and basophils (Table 2 and Fig. S1). The count of other leukocyte cell subtypes did not reach statistical significance for LEDVT except basophils in ROC analysis (Fig. S1A–E). Basophils showed the best power to predict LEDVT (Table 2 and Fig. S1E). The ROC curve determined that the optimal cutoff value of basophils for predicting LEDVT at 100 /µL (AUC 0.707, 95% confidence interval [CI] 0.653–0.757, p < 0.001; the sensitivity was 61.70%, and the specificity was 75.37%; Fig. 2) in sICH patients. We categorized BCPB ≥ 100/µL as basophil-rich and lower BCPB values (< 100/µL) as basophil-poor. One hundred and ninety-five patients with sICH had a basophil-rich BCPB status. The clinical characteristics of sICH with basophil-rich and basophil-poor status are depicted in Table 3. The baseline Glasgow Coma Scale (GCS) score was significantly associated with basophil-rich status (Table 3). Patients with basophil-rich status were found to have a significantly increased risk of developing LEDVT among sICH patients.

Table 1 Baseline clinical and radiographical characteristics of sICH in 315 non-surgical patients.
Full size table
Table 2 Area under the curve of different types of leukocytes associated with LEDVT.
Full size table
Fig. 2
figure 2

ROC curve of basophils for LEDVT.

Full size image

The optimal cut-off value of basophils calculated by the ROC curve to predict LEDVT was 100/µL (AUC 0.707, 95% CI 0.653–0.757, p < 0.001; the sensitivity was 61.70%, and the specificity was 75.37%) in non-surgical sICH patients.

Table 3 Clinical characteristics of 315 sICH patients with basophils-poor and basophils-rich status.
Full size table

The occurrence of LEDVT in our dataset was relatively low at 14.92%, resulting in an imbalance between the LEDVT and non-LEDVT groups. Therefore, weighted logistic regression was utilized to address this imbalance and ensure that both groups’ influences were properly captured. Risk factors for LEDVT occurrence were investigated using weight logistic regression analysis (Tables 4 and 5). In the weighted univariate analysis, age, smoking status, alcohol, baseline ICH volume, baseline GCS score, and basophil-rich status were significantly associated with LEDVT occurrence. Since the primary outcome variable (LEDVT) was binary-coded, we included multiple covariates in our model to control for potential confounders. Next, weighted multivariable logistic regression models were performed to explore the association between basophil levels and LEDVT.

Table 4 Weighted univariable logistic regression analysis of predictors of LEDVT in the patients with sICH.
Full size table

Table 5 presents the results of the weighted multivariable logistic regression models examining the relationship between basophil levels and LEDVT incidence. LEDVT was significantly associated with basophil-rich status in models I and II. sICH patients with basophil-rich status had a higher risk of LEDVT compared to those with basophil-poor status, with odds ratios (OR) and confidence intervals of 4.2 (1.8–9.6) for model I, 2.7 (1.1–6.8) for model II, respectively (Table 5). Consequently, the regression models indicated that basophil-rich peripheral blood was associated with an increased risk of LEDVT.

Table 5 Association of basophils with the occurrence of LEDVT in the weighted multivariable logistics regression models.
Full size table

Discussion

Our results demonstrated that admission BCPB was associated with an increased risk of LEDVT in the non-surgical sICH patients. Despite adjusting for potential variables, patients exhibiting a basophil-rich status remained 2.7 times more at risk for LEDVT than those with a basophil-poor status. By far as we know, this is the first retrospective study focusing on the association between admission BCPB and LEDVT in patients with sICH. More attention and further medical care are needed for spontaneous ICH patients with a higher BCPB.

DVT is an independent predictor for unfavorable outcomes in patients with sICH, contributing to worse functional recovery throughout hospitalization and at the 3-month and 1-year follow-ups3,5,7. The incidence of LEDVT in the present study was 14.92%, which indeed appears higher than the DVT prevalence reported in some studies but is consistent with others3,4,5,6,7. It is important to note that the majority of DVT cases occur in the lower extremities (known as LEDVT), yet there is limited literature focusing exclusively on this subgroup. The reported risk of DVT complications varies significantly across different ICH cohorts, with estimates ranging widely from 2–75%3,4,5,6,7. Furthermore, several prospective studies have reported the incidence of DVT detected by ultrasonography during hospitalization for sICH to range from 21–40.4%6,7,37,38. A systemic review also revealed that the occurrence of DVT among immobilized post-stroke patients ranged from 10 to 75%4. These variations in reported prevalence may result from differences in factors such as the timing and use of early chemoprophylaxis, the application of IPC devices, racial or ethnic disparities, diagnostic methods, and the timing of evaluation during hospitalization.

Several studies have indicated that heightened inflammation markers are independent risk factors for acute LEDVT in patients with sICH4,11,22,31. Basophils, accounting for less than of 1% circulating leukocytes, are mainly identified as pro-inflammatory effector cells that play a role in inflammation related to allergies39. Basophils are essential contributors to IgE-mediated reactions, as they enter inflamed sites, release pro-inflammatory agents like histamine and leukotriene C4, and modulate cytokines such as interleukin (IL)-4 and IL-1340. To date, there have been few investigations into the response of basophils post-sICH, and the association between basophils and DVT still needs to be well-established. Since basophils account for less than 1% of leukocytes and exhibit similarities to mast cells’ morphology and function, researchers have largely overlooked them.

However, recent literature has illustrated that elevated basophil levels strongly affect venous thrombosis in patients with vascular diseases20,41,42. A recent literature revealed that elevated basophil counts correlated with coronary no-reflow, a phenomenon that serves as an independent predictor of major adverse cardiac events in patients with acute myocardial infarction42. Additionally, findings from the UK Biobank study cohort revealed that those who died from either cardiovascular or non-cardiovascular causes had slightly higher basophil counts in univariate analysis41. High basophil level was an independent predictor of total and cardiovascular mortality in in coronary artery disease, and corelated with enhanced factor II plasma coagulant activity20. These findings indicate a correlation between basophils and thrombotic events. Although the differences in disease backgrounds, these studies seem to support our research findings. Their studies did not explore the mechanism by which elevated basophils lead to venous thrombosis, but we suspect it may be attributed to inflammation. Our research revealed a noteworthy correlation between baseline GCS score and basophil-rich status. This association suggests that basophil-rich status may serve as a marker of disease severity, further supporting previous findings that inflammation is linked to disease severity following sICH26. Additionally, ICH volume and baseline GCS score were found to be significant in predicting LEDVT in a univariate analysis, as patients with more severe ICH and lower GCS scores may experience greater immobility, which increases the risk of LEDVT. An acute ICH can potentially trigger the systemic inflammatory response43. Following an acute ICH episode, a systemic inflammatory response is triggered, causing an increase in peripheral blood inflammatory indicators44.

Conversely, the ICH hematoma component initiates an inflammatory response in the surrounding brain tissue. This, coupled with the ICH hematoma itself, worsens the damage to brain tissue and further amplifies the inflammatory response in the peripheral blood26. Coagulation activation tends to rise in ICH patients due to inflammatory responses. The presence of inflammation on the blood vessel wall can initiate the formation of a thrombus in the vein45. Venous wall inflammation is the probable initial cause of DVT formation46. Gang Wang et al. have found that leukocytes upon admission can independently predict the risk of LEDVT in elderly patients with primary ICH11, indicating that inflammation plays a significant role in the development of LEDVT12,47.

Our study demonstrated a positive association between elevated basophil levels and an increased risk of LEDVT in non-surgical sICH patients, which is the first report to connect basophil with prothrombotic diathesis and LEDVT risk. While the underlying mechanisms require further investigation, several potential pathways might contribute to this phenomenon. There are molecular grounds that could potentially support the biological plausibility of this relationship. First, sICH triggers a significant inflammatory cascade, and basophils, known as inflammatory cells, are implicated in this process48. Basophils release various pro-inflammatory mediators, including IL-4, IL-6, IL-13, and histamine, which can activate endothelial cells, promoting a pro-coagulant state and adhesion of platelets and white blood cells to the vessel wall40, ultimately increasing the risk of LEDVT formation39. Basophils can stimulate the release of Neutrophil Extracellular Traps (NETs) from neutrophils. NETs are web-like structures composed of DNA and antimicrobial proteins, and their presence in the vasculature can contribute to thrombosis formation49,50,51.

Additionally, basophil granules contain polyphosphate52, a procoagulant player in hemostasis that can speed up blood clotting by activating the contact pathway and facilitating the activation of factor V53. Basophils have the ability to produce platelet-activating factor, which controls platelet aggregation and is involved in various pathophysiological processes such as thrombosis and tissue ischemia54,55. Increased basophil can further impede the natural clearance of any thrombi formed, potentially contributing to LEDVT development in patients with sICH. Third, basophils have been shown to express tissue factor, a key initiator of the coagulation cascade56. Upregulation of tissue factors in response to sICH might contribute to the increased risk of LEDVT observed in patients with basophil-rich status. Patients with sICH often experience reduced mobility due to paralysis, weakness, or sedation, a well-established risk factor for LEDVT. Notably, basophils can secrete a lower amount of heparin, which may not be enough to have a direct impact on blood clotting similar to a drug. Their heparin release is usually localized and limited, potentially not directly affecting overall coagulation. Finally, immobility leads to venous stasis, or slow blood flow, particularly in the lower extremities57,58,59. This stasis promotes platelet aggregation, activation of the coagulation cascade, and impaired fibrinolysis. Platelets are more likely to clump together and form clots in stagnant blood55. Stasis activates clotting factors, further promoting clot formation. Reduced blood flow can impair the delivery of plasminogen, the inactive form of plasmin, the enzyme responsible for clot breakdown, to the site of a clot, hindering its removal57,58,59.

Despite shedding light on the potential association between elevated basophil levels and LEDVT risk in sICH patients, it is crucial to acknowledge its limitations. First, this study was a single-center retrospective investigation with a limited sample size. Our study establishes a correlation, but it cannot definitively prove causation. We could not conclude that elevated basophil levels directly LEDVT in sICH patients. A larger sample size would increase the generalizability of the findings and reduce the possibility of chance associations. Additionally, the scope of this study was limited to investigating the association between basophil levels upon admission and the occurrence of LEDVT without considering the dynamic fluctuations in basophil levels about LEDVT. Third, detailed data on the duration of anticoagulation therapy and specific mobility scores (e.g., mRS) during hospitalization were lacking, which could significantly influence the risk of clot formation. Obesity is a well-established risk factor for DVT. However, our study did not include detailed information on body habitus or BMI classification. Additionally, the NIHSS extremities score and GCS motor score were not available for inclusion in our analysis. Due to the retrospective nature of our study and limitations in the available clinical data, we were unable to incorporate these variables into our analysis. In future studies, we will aim to include body habitus, BMI classification, the National Institute of Health stroke scale (NIHSS) extremities score, and GCS motor score to better assess their potential impact on DVT formation. Forth, it is worth noting that certain sICH patients with LEDVT exhibited admission basophil levels within the normal range, which adds complexity to the direct interpretation of the results. Finally, while we discussed potential mechanisms, the study did not directly investigate them. Further research using in vitro and in vivo models is needed to confirm these pathways.

Conclusions

Our study demonstrated that elevated basophils in the peripheral blood upon admission were independent risk factors for the occurrence of LEDVT in non-surgical sICH patients, highlighting their potential to identify high-risk patients with LEDVT. Furthermore, it possesses substantial clinical importance, underscoring the necessity of early LEDVT prevention measures. Nevertheless, the fundamental mechanisms should be extensively assessed in basic research studies.