Role of ABO blood group as a prognostic factor in patients with spontaneous intracerebral hemorrhage


Francesco Dentali, U.O. Medicina I, Ospedale di Circolo, Viale Borri 57, 21100 Varese, Italy.
Tel.: +39 0332 278611; fax: +39 0332 393640.

Dear Sir,

Spontaneous intracerebral hemorrhage (ICH) is a major public health problem with an annual incidence of 10–30 per 1 000 000 population [1,2]. It accounts for about 10–15% of strokes, affecting about 2 million persons worldwide each year [3,4]. Hospital admissions for ICH have increased by 18% in the past 10 years, probably as a result of the increasing age of the population and changes in racial demographics [5].

Spontaneous ICH usually originates from a rupture of a small vessel damaged by chronic hypertension or amyloid angiopathy [1]. Established risk factors for spontaneous ICH include hypertension, excess alcohol consumption, male sex, increasing age and smoking [6].

ICH is associated with a higher mortality rate compared with ischemic stroke, with only about 40% of affected patients surviving the first year [7]. Early case fatality after ICH has not changed over the past two decades. Studies of surgical hematoma evacuation in ICH using a variety of methods have yielded either negative or inconclusive results. Likewise, no medical treatment has been shown conclusively to benefit these patients. Recently, Hemphill et al. [8] developed a simple scale to evaluate the short-term clinical outcome of patients with an acute ICH. However, additional factors may be useful in defining the short and long-term prognosis of these patients.

A large number of studies have described that the ABO blood type has a profound influence on hemostasis, predominantly by determining the plasma levels of von Willebrand factor (VWF) and of factor VIII (FVIII) [9]. Several studies and two recent meta-analyses have analyzed the relationship between ABO blood group and both arterial and venous thrombosis [10,11]. Data available so far suggest also that O blood group is a potentially important genetic risk factor for clinically relevant bleeding [12].

On the other hand, whether the ABO blood group should be useful in predicting the prognosis in patients with spontaneous ICH is yet to be determined.

To explore this issue, we performed a multicenter-retrospective study to investigate whether the ABO blood type plays a role as a prognostic factor in spontaneous ICH. Prognosis has been defined as modified Ranking Score (mRS) at discharge.

We examined all the charts of adult patients who were hospitalized with a diagnosis of ICH between 2006 and 2011 in two hospitals (Varese and Cuneo). Consecutive patients with an objective diagnosis of spontaneous ICH were potentially eligible for inclusion for the purpose of the study. Diagnosis of ICH was made when brain computed tomography or magnetic resonance imaging showing intracranial hemorrhage was available. Conversely, patients with traumatic ICH were excluded, as well as patients with ICH secondary to aneurysms and arteriovenous malformation and patients treated with anticoagulant therapy or a combined therapy with an anticoagulant and an antiplatelet drug. Only patients in whom the ABO blood group was available were considered.

Glasgow coma scale (GCS) score at admission and mRS at the time of discharge were evaluated in all the patients, as well as anamnestic data concerning previous stroke events, history of hypertension and diabetes mellitus. The site of origin, the presence of intraventricular hemorrhage, the volume and the necessity for surgical therapy were also registered. ICH volume was measured according to the ABC/2 method [13]. As per study protocol, for each patient the blood group was then obtained from the patients and hospital documentation as well as the Rhesus status. Death rate and clinical outcome at the time of discharge were evaluated. Patients with a mRS ≥ 4 at discharge were considered as having a poor outcome.

Continuous variables were expressed as mean plus or minus the standard deviation (SD); categorical data are given as counts and percentages. Characteristics of patients with good and poor outcome were compared using Student’s t-test (for continuous variables) and the chi-square or Fisher’s exact test (for dichotomous variables). Only variables found significant in the univariate analysis were used as covariates in the multivariate analysis. Results of multivariate analysis were presented as odds ratios (ORs) and corresponding 95% confidence intervals (CIs).

The present study was approved by local institutional review boards and patient information was codified to ensure anonymity.

A total of 398 patients were potentially eligible for the study. Blood group was not available in 169 patients (43.4%). Thus, 229 patients (mean age 69.7 ± 14.1 years, 91 women) with a spontaneous ICH were included in the study. Baseline characteristics of included patients and their ABO blood group distribution are summarized in Table 1. About 43% of included patients were of O blood group. Mean GCS at presentation was 10.4 (±4.7). Twenty-nine ICHs (12.3%) were infratentorial and 97 (42.4%) had a volume ≥ 30 mL according to the ABC/2 method. At the end of hospitalization 70 patients (30.6%) had died and 148 patients (64.6%) were judged to have a poor prognosis (mRS ≥ 4). GCS at the time of hospitalization was similar in patients with different blood groups (10.5 in patients with O blood group vs. 10.2 in patients with non-O blood group, P = 0.63). Furthermore, ABO blood group distribution was similar in patients with a good or poor outcome at discharge (O blood group, 44.4% of patients with a good outcome vs. 42.6% of patients with a poor outcome, P = 0.89) and in patients who died or were alive at discharge (O blood group, 48.6% in patients who had died at discharge time vs. 40.9% in patients alive at discharge time, P = 0.28). At univariate analysis, age (P = 0.008), female sex (P = 0.006), GCS (P < 0.001) lobar location (P = 0.011), basal nuclei location (P = 0.002), intraventricular hemorrhage (P = 0.0019), hemorrhage volume > 30 mL (P < 0.001) and surgical hematoma evacuation (P = 0.04) were significantly different in patients with a good outcome and patients with a poor outcome at discharge, whereas anamnesis of hypertension or diabetes mellitus, infratentorial or thalamus location and antiplatelet therapy were not (P > 0.05). When all the significant variables at univariate analysis were included in a multivariate model, GCS ≤ 8 (OR 12.8, 95% CI 4.9, 33.4), hemorrhage volume > 30 mL (OR 3.4, 95% CI 1.5, 7.8), female sex (OR 2.7, 95% CI 1.2, 5.8), age ≥ 80 years (OR 2.5, 95% CI 1.0, 5.8), intraventricular hemorrhage (OR 2.4, 95% CI 1.0, 5.6) and basal nuclei location (OR 0.2, 95% CI 0.1, 0.5) remained significant. (Appendix 1).

Table 1.   Baseline characteristics
  1. GCS, Glasgow Coma Scale; SD, standard deviation; n, number.

Mean age (SD)69.1 (14.1)
Previous stroke, n (%)20 (8.7)
Diabetes mellitus, n (%)37 (16.2)
Hypertension, n (%)137 (59.8)
 Tentorial, n (%)29 (12.6)
 Talamic, n (%)22 (9.6)
 Lobar, n (%)115 (50.2)
 Basal ganglia, n (%)40 (17.4)
Surgery, n (%)43 (18.7)
Antiplatelet therapy, n (%)66 (28.8)
Intraventricular hemorrhage, n (%)63 (27.5)
Blood type
 A group, n (%)96 (41.9)
 B group, n (%)25 (10.9)
 AB group, n (%)9 (3.9)
 0 group, n (%)99 (43.2)
Rhesus type +, n (%)195 (85.1)
Mean GCS at admission, n (SD)10 (4.7)
GCS ≤ 8 at admission, n (%)87 (38.0)
Death, n (%)71 (30.6)

ABO blood group is a major determinant of hemostasis affecting plasma levels of VWF and, consequently, of FVIII [9]. Thus, ABO blood group could influence the individual risk of cardiovascular and bleeding events [10–12]. In particular, as plasma VWF levels are 25–35% lower in subjects with type-O blood group than in non-O individuals [9], the former subjects could have an increased bleeding tendency.

However, in our study we failed to find any significant association between the blood group and clinical presentation or short-term prognosis in patients with spontaneous ICH.

After a multivariate analysis, other potential risk factors such as GCS ≤ 8, hemorrhage volume, female sex, age and intraventricular hemorrhage were significantly associated with a poor outcome. On the other hand, patients who had an ICH located at basal nuclei seemed to have a better prognosis. Interestingly, previous use of antiplatelet therapy was not significantly associated with a poor outcome. Although aspirin and other antiplatelet therapies could worsen the outcome from ICH by promoting bleeding, a similar prognosis in terms of poor functional outcome was found in a recent meta-analysis of the literature, after adjustment for potential confounders, in patients on antiplatelet therapy and in patients not on antiplatelet therapy at the time of ICH [14].

Taken together our results suggest that ABO blood group should not be included among the parameters used to evaluate the prognosis of patients with an acute ICH. Other studies should evaluate if this parameter may be useful in defining the prognosis in particular subgroups of ICH patients, including patients with ICH secondary to other risk factors such as trauma and use of oral anticoagulants.

Our study has some potential limitations. First, the retrospective design of the study limits inferences that can be drawn and increases the risk of bias. However, to minimize this risk, we paid meticulous attention to collecting patients’ characteristics and potential prognostic factors and we have no missing values in our data collection. Second, because of the relatively small number of included patients we can not definitively exclude that our results are due to a type two error (false negative result). However, clinical presentation and prognosis in patients with different blood groups were very similar, making this possibility extremely unlikely. Third, blood group was not available for all the potentially eligible patients and the presence of a selection bias could not be excluded. However, this does not seem to be the case, because patients for whom blood group was available had similar baseline characteristics and prognosis to patients for whom blood group was not available (data not shown). Last, patients were not followed-up after discharge and we only have data on their clinical outcome during hospitalization. Thus, we could not exclude a role of ABO blood group as a long-term prognostic factor.

In conclusion, according to the results of our study, ABO blood group was not significantly different in patients with spontaneous ICH with different clinical outcomes and should not be used as a prognostic risk factor in these patients.


F. Dentali, F. Pomero and G. Bono conceived and designed the article; F. Dentali analyzed and interpreted the data; F. Dentali, F. Pomero., F. Annoni, A.M. Maresca, L. Fenoglio, M. Franchini and G. Bono drafted and provided critical revision of the article for important intellectual content; F. Annoni and A.V. Giraudo collected and assembled the data. All the authors provided final approval of the article.

Disclosure of Conflict of Interests

The authors state they have no conflict of interest.


Appendix 1. Results of univariate and multivariate analyses

Risk factorResult of univariate analysis (P value)Result of multivariate analysis
  1. *For age ≥ 80 years.

Age0.008OR 2.5 (95% CI 1.03, 5.8)*
Female gender0.006OR 2.7 (95% CI 1.2, 5.8)
Previous stroke0.50
Diabetes mellitus0.537
 Tentorial, n (%)0.753
 Talamic, n (%)0.736
 Lobar, n (%)0.011
 Basal ganglia, n (%)0.002OR 0.2 (95% CI 0.1, 0.5)
Surgery, n (%)0.043OR 2.1 (95% CI 0.7, 6.8)
Antiplatelet therapy0.573
Intraventricular hemorrhage0.0019OR 2.4 (95% CI 1.0, 5.6)
Hemorrhage volume > 30 mL< 0.001OR 3.4 (95% CI 1.5, 7.8)
Blood type
 0 group, n (%)0.893
Rhesus type +0.553
Glasgow Coma Scale at admission ≤ 8< 0.001OR 12.8 (95% CI 4.9, 33.4)