Introduction

Hematological diseases often present with cerebrovascular complications, including ischemic stroke, intracranial hemorrhage (ICH), microbleeds, posterior reversible encephalopathy syndrome, and dural sinus and cerebral vein thrombosis [1]. A previous study found that the prevalence of stroke in patients with hematological diseases ranged from 0 to 7% [2]. Various hematological disorders may play a direct or indirect role in neurological complications and ICH [3, 4]. Previous studies have already identified hypertension, vessel wall abnormality, invasion or compression of vessels by a tumor in or adjacent to the brain, low platelet count or platelet dysfunction, coagulation factor deficiency, disseminated intravascular coagulation, sepsis, and hyperleukocytosis as independent risk factors for ICH in cancer patients [5,6,7]. However, the clinical characteristics and prognostic factors in ICH patients with a hematological disease remain controversial.

The pathogenesis of ICH differs in the general and hematological disease populations. The pathogenic mechanisms include abnormal platelet counts or function, coagulation disorders, hyperleukemia, sepsis, and abnormal vessel walls [6,7,8]. ICH may occur at various timepoints after the diagnosis of hematological disease, especially for patients who firstly experienced hematological disease [9, 10]. Moreover, most hemorrhages in patients with a hematological disease occur in the brain parenchyma, with variable clinical characteristics [11, 12]. The prognosis of ICH patients with hematological diseases has improved because of the development of modern neuroimaging and advances in hematology and neurology [12, 13].

However, epidemiological data on ICH patients with hematological diseases in China are lacking. Therefore, we performed a retrospective single-center study to describe the clinical characteristics and prognostic factors in ICH patients with a hematological disease in China.

Methods

Patient population

A total of 238 imaging-confirmed ICH patients with hematological disease attending the Institute of Hematology and Blood Diseases Hospital, Tianjin, China, from January 2015 to April 2020, were analyzed retrospectively. The study was approved by the Institutional Review Board of the Institute of Hematology and Blood Diseases Hospital.

The types of hematological diseases included acute lymphoblastic leukemia, acute myeloid leukemia (AML), acute mixed phenotype leukemia, myelodysplastic syndrome (MDS), atypical chronic granulocytic leukemia (aCML), chronic myelocytic leukemia (CML), chronic myelomonocytic leukemia (CMML), aplastic anemia (AA), immune thrombocytopenia (ITP), hemophilia, acquired hemophilia, vitamin-dependent coagulation factor deficiency, hereditary coagulation factor VII deficiency, non-Hodgkin’s lymphoma (NHL), multiple myeloma (MM), pure red blood cell aplastic anemia (PRCA), Evans syndrome, Fanconi anemia, connective tissue disease combined with abnormal blood cell counts, and hypersplenism. These were grouped into three: Firstly, the MDS/MPN group included MDS, aCML, CML, and CMML. Secondly, the congenital/acquired coagulation factor deficiency (CCFD/ACFD) group included hemophilia, acquired hemophilia, vitamin-dependent coagulation factor deficiency, hereditary coagulation factor VII hemophilia, acquired hemophilia, vitamin-dependent coagulation factor deficiency, and hereditary coagulation factor deficiency. Thirdly, the other group included NHL, MM, PRCA, Evans syndrome, Fanconi anemia, connective tissue disease with abnormal blood counts, and hypersplenism.

Cranial computerized tomography was performed in all patients with a hematological disease, and their cranial hemorrhage site; hemorrhage characteristics, signs, and symptoms; and the presence of any co-infections were recorded. However, the extent of hemorrhage was not known, because most of the hemorrhages were diffuse and multifocal.

ICH included parenchymal hemorrhage (PCH), subarachnoid hemorrhage (SCH), subdural hemorrhage (SH), epidural hemorrhage (EH), and multi-site hemorrhage (MSH). PCH was defined as a single-site hemorrhage that occurred in the brain parenchyma, while MSH was defined as hemorrhages in multiple parts of the brain parenchyma combined with hemorrhages in other parts of the brain parenchyma.

The signs and symptoms included headache, disturbance of consciousness, mental disorder, meningeal irritation sign, hemiplegia, epilepsy, aphasia, and dysopia.

Data collection and definition

The baseline characteristics recorded included age, sex, hypertension, diabetes mellitus (DM), intracranial hemangioma, and cranial trauma. Patients with acute promyelocytic leukemia (APL) received plasma or fibrinogen infusions if fibrinogen ≤ 1.5 g/L, while platelets were transfused if platelet ≤ 30 × 109/L in these patients. For non-APL patients, red blood cells were infused when hemoglobin ≤ 70 g/L, plasma or fibrinogen was infused if fibrinogen ≤ 1.0 g/L, and platelets were transfused if platelets ≤ 10 × 109/L. Thrombocytopenia was defined as platelets ≤ 30 × 109/L in patients with APL, platelets ≤ 20 × 109/L in patients with acute leukemia other than APL, and platelets ≤ 10 × 109/L in patients with diseases other than these two categories. Coagulation abnormalities were defined as international normalized ratio (INR) > 1.4, and/or partial thromboplastin time (APTT) > 50 s, and/or fibrinogen concentration ≤ 1.5 g/L. Leukocytosis was defined as leukocytes > 100 × 109/L. Serum urea and creatinine levels were also recorded, and patients were accordingly divided into normal and abnormal groups. All data were collected within 3 days for hematological disease patients diagnosed with ICH.

ICH management and prognosis

Medical management of ICH in patients with a hematological disease included conservative and surgical treatments. Conservative treatment strategies included red blood cell, platelet, plasma, and fibrinogen support therapy; the use of hemostasis and dehydration to reduce cranial pressure; and control of the original disease as well as comorbidities. Surgical treatment was determined by the neurosurgeon after consultation and evaluation of the patient’s condition. Mortality events and the times of mortality were recorded after follow-up to 30 days from the inception of ICH.

Statistical analysis

Continuous variables are presented as means (standard deviations) or medians (quartiles), according to the data distribution, and categorical variables are presented as numbers (percentages). The differences among groups were calculated using analysis of variance and chi-squared tests. The Kaplan–Meier method was used for survival analysis. Univariate analysis was performed using the Log-rank approach. A Cox proportional hazards model was used to identify the prognostic factors for mortality in ICH patients with a hematological disease. All reported P values were two-sided, and statistical significance was set at P < 0.05. All statistical analyses were conducted using SPSS software (Version: 22.0; Chicago, IL, USA).

Results

Patients’ characteristics

Of the 238 included patients, 123 patients had AL, 20 had MDS/MPN, 35 had AA, 29 had ITP, 19 had CCFD/ACFD, and 12 had other types of hematological diseases. ICH occurred during induction therapy in 218 patients, while 20 patients presented ICH during consolidation therapy. With respect to the bleeding sites, 121 patients presented with MSH, 58 with PCH, 23 with SCH, 33 with SH, and three with EH. We noted that the bleeding sites were different in patients with different types of hematological diseases (P < 0.001; Table 1).

Table 1 The characteristics of recruited patients
Full size table

Clinical characteristics according to the type of hematological disease

The characteristics of recruited patients according to the types of hematological diseases are presented in Table 2. Age (P = 0.036), sex (P = 0.002), thrombocytopenia (P < 0.001), INR (P = 0.011), APTT (P < 0.001), FIB (P < 0.001), white blood cell count (P < 0.001), intracranial hemangioma (P = 0.006), cranial trauma (P = 0.004), disease status (P < 0.001), coagulation abnormalities (P < 0.001), co-infection (P < 0.001), and serum urea (P = 0.026) differed significantly among the patients with different types of hematological diseases. However, the distribution of hypertension, diabetes, headache, disturbance of consciousness, mental disorder, meningeal irritation sign, hemiplegia, epilepsy, aphasia, dysopia, and serum creatinine levels were not significantly different among patients with the different types of hematological diseases.

Table 2 The clinical characteristics of patients according to the type of hematological disease
Full size table

Clinical characteristics according to bleeding sites

The characteristics of the recruited patients according to the bleeding sites are presented in Table 3. APTT (P = 0.009), FIB (P = 0.022), white blood cell count (P < 0.001), cranial trauma (P < 0.001), disease status (P = 0.007), headache (P = 0.006), disturbance of consciousness (P = 0.002), meningeal irritation sign (P < 0.001), hemiplegia (P < 0.001), and co-infection (P = 0.002) were significantly associated with bleeding sites. However, the distribution of age, sex, thrombocytopenia, INR, hypertension, diabetes, intracranial hemangioma, mental disorder, epilepsy, aphasia, dysopia, coagulation abnormalities, serum urea, and serum creatinine did not show a significant association with any bleeding site.

Table 3 The clinical characteristics of patients according to the bleeding sites
Full size table

Prognostic factors

The prognostic factors for 30-day mortality in ICH patients with a hematological disease are presented in Table 4. Univariate analysis showed that SCH (hazards ratio [HR]: 0.101; 95% confidence interval [CI]: 0.013–0.809; P = 0.031) and SH (HR: 0.187; 95% CI: 0.041–0.858; P = 0.031) were associated with an improvement in the 30-day mortality, as compared with PCH. Moreover, patients with M3 hematological disease had better 30-day mortality than those with non-M3 hematological disease (HR: 0.425; 95% CI: 0.209–0.866; P = 0.018). Furthermore, patients with AA (HR: 0.294; 95% CI: 0.135–0.639; P = 0.002) and ITP (HR: 0.135; 95% CI: 0.038–0.477; P = 0.002) had a better prognosis than those with AL. Finally, disturbance of consciousness (HR: 2.231; 95% CI: 1.294–3.846; P = 0.004) and coagulation abnormalities (HR: 4.823; 95% CI: 1.014–22.935; P = 0.048) were associated with a poor prognosis in terms of 30-day mortality.

Table 4 The prognostic factors of 30-day mortality
Full size table

Multivariate analysis showed that SH was associated with a better 30-day mortality (HR: 0.230; 95% CI: 0.053–0.996; P = 0.049), while MSH was associated with a poor 30-day mortality, as compared with PCH (HR: 1.891; 95% CI: 1.147–3.117; P = 0.012). Moreover, a low white blood cell count was associated with a lower risk of 30-day mortality (HR: 0.563; 95% CI: 0.348–0.910; P = 0.019). Furthermore, AA (HR: 0.408; 95% CI: 0.203–0.821; P = 0.012) and ITP (HR: 0.197; 95% CI: 0.061–0.640; P = 0.007) were associated with an improved 30-day mortality, as compared with AL. Finally, increased age (HR: 1.012; 95% CI: 1.001–1.022; P = 0.034) and presentation with a disturbance of consciousness (HR: 1.989; 95% CI: 1.269–3.117; P = 0.003) were associated with an increased risk of 30-day mortality.

Discussion

ICH is more common in patients with a hematological disease than in the general population, and is a serious complication and cause of death in these patients [14,15,16,17,18]. In this study, 238 patients with a hematological disease who presented with ICH were retrospectively recruited and their characteristics observed. Most patients with hematological disease who developed ICH had MSH. The type of hematological disease in patients with ICH differed by age, sex, thrombocytopenia, INR, APTT, FIB, white blood cell count, intracranial hemangioma, cranial trauma, disease status, coagulation abnormalities, co-infection, and serum urea. Furthermore, bleeding sites differed by APTT, FIB, white blood cell count, cranial trauma, disease status, headache, disturbance of consciousness, meningeal irritation sign, hemiplegia, and co-infections in patients with a hematological disease. Finally, the 30-day mortality for ICH patients with hematological disease may be influenced by age, white blood cell count, AA, ITP, SH, MSH, and presentation with a disturbance of consciousness.

This study described the clinical characteristics of ICH in Chinese patients with a hematological disease, which have not been reported previously, and identified prognostic factors for the 30-day mortality in these patients. A previous study conducted by Chen et al. recruited 2574 adult patients with hematological malignancies and found that the incidence of ICH was 2.8% and that the risk of ICH in patients with AML was higher than that in patients with other hematological malignancies [19]. They found that patients who presented with central nervous system–involved lymphomas had a higher risk of ICH than patients with central nervous system–involved AL [19]. Furthermore, the type of hematological malignancies was not associated with mortality [19]. However, that study did not focus on ICH patients with a hematological disease, and the prognostic factors for the 30-day mortality in ICH patients with a hematological disease were not identified. Therefore, we performed this retrospective study to describe the clinical characteristics of ICH patients with a hematological disease and to identify the prognostic factors for 30-day mortality in these patients.

In our study, most patients had MSH; this was not consistent with findings from prior studies [20,21,22], which found that the most common site of ICH in hematological disorders was a single cortico-parietal area, with much lower incidences of subcortical and cerebellar hemorrhages. Moreover, the most common site of cerebral hemorrhage in patients with hypertension was a hemorrhage in the basal ganglia region. We noted that the sites of ICH varied among the different hematological diseases: (1) The most common bleeding sites for AL patients were MSH and PSH; (2) MSH and SCH were the most common sites in MDS/MPN and AA patients; (3) the most common bleeding site in ITP patients was SH; and (4) MSH and SH were the most common sites in CCFD/ACFD patients. Finally, the symptoms of headache, nausea, vomiting, lateral motor and sensory disturbances, and disturbance of consciousness were relatively common, while mental disorders, meningeal irritation sign, epilepsy, aphasia, and dysopia were relatively uncommon.

Our study found there were significant differences in the age, sex, thrombocytopenia, INR, APTT, FIB, white blood cell counts, intracranial hemangioma, cranial trauma, disease status, coagulation abnormalities, co-infections, and serum urea levels, among patients with the various types of hematological diseases. However, these results were obtained from univariate analysis, and the patients included had various types of hematological diseases, which could have caused some instability in the results. Furthermore, the bleeding sites in patients with a hematological disease may have been affected by APTT, FIB, white blood cell count, cranial trauma, disease status, headache, disturbance of consciousness, meningeal irritation sign, hemiplegia, and co-infection. However, the vascular risk factors did not affect the bleeding sites. These factors included hypertension, diabetes, intracranial hemangioma, and cranial trauma, which were significantly different from the general population and reflect the unique pathophysiological characteristics of ICH in patients with a hematological disease [23,24,25]. This may have been due to the elevated levels of the white blood cells, resulting from infiltration of the vascular wall, as these patients are prone to vascular endothelial injuries, and release of procoagulant substances and inflammatory markers, which could increase the risk of ICH [26, 27]. Moreover, the APTT and FIB values reflected coagulation abnormalities, which was associated with an increased risk of bleeding [28, 29]. The signs and symptoms of patients may reflect the severity of disease, with the distribution being related significantly to the bleeding sites.

This study found that the 30-day mortality in ICH patients with a hematological disease was 41.2%, which was significantly higher than that in the general population with ICH [15, 30]. The prognostic factors for the 30-day mortality in ICH patients with a hematological disease included age, white blood cell count, AA, ITP, SH, MSH, and a disturbance of consciousness. This may have been because ICH was regarded as secondary to vascular rupturing caused by the leukemia and thrombosis and necrosis of the ischemic tissue [31]. Moreover, AA and ITP are considered to be benign hematological diseases, and prognosis for ICH patients with AA and ITP is relatively better than that of AL patients with ICH, which may be attributed to AL patients being associated with and being more prone to co-infections, coagulation abnormalities, and elevated white blood cell levels. Furthermore, the bleeding sites and the presence of a disturbance of consciousness have been significantly related to the severity of the disease, which may have affected the prognosis of ICH patients with a hematological disease [12]. Finally, infection has been reported to be the main cause of death in patients with hematological malignancies [32]. In our study, most death events (82, 34.5%) occurred within 4 days in patients diagnosed with ICH, while progression of disease or infection contributed less to death events within 4 days.

Certain shortcomings of this study should be acknowledged. First, the current study was a retrospective study, and thus, selective and recall biases were inevitable. Second, the number of patients in the various hematological disease groups was not balanced. Third, the chronic worsening thrombocytopenia and transfusion refractoriness in patients with MDS or CMML resulted in a poor treatment effect and poor prognosis. Fourth, data on the treatment strategies used for hematological diseases, which may have affected the prognosis of ICH patients with a hematological disease, were not available. Finally, stratified analyses were not performed due to the small number of patients with a specific bleeding site.

Conclusions

The findings of this study systematically describe the clinical characteristics of Chinese ICH patients with a hematological disease. Most of our patients had MSH, and age, sex, thrombocytopenia, INR, APTT, FIB, white blood cells, intracranial hemangioma, cranial trauma, disease status, coagulation abnormalities, co-infection, and serum urea differed among patients with various types of hematological diseases. Moreover, APTT, FIB, white blood cell, cranial trauma, disease status, headache, disturbance of consciousness, meningeal irritation sign, hemiplegia, and co-infection were significantly associated with the bleeding sites. Finally, age, white blood cells, AA, ITP, SH, MSH, and disturbance of consciousness were significantly associated with 30-day mortality in ICH patients with a hematological disease. Further prospective studies should be performed to verify the findings of this study and to assess the prognostic factors in patients with a specific hematological disease.