ABO blood group system and occurrence of ischemic stroke

ABSTRACT Background: Ischemic stroke (IS) is a multifactorial disease that presents high rates of morbimortality in Brazil. Several studies proved that there is a link between the ABO blood group system and the occurrence of thrombotic events. Nonetheless, its association with IS is not well established. Objective: For that reason, the purpose hereof was to investigate the relation between the ABO blood groups and the occurrence of IS in a Brazilian cohort of cerebrovascular diseases. Methods: Five hundred and twenty-nine subjects were included over 12 months, from which 275 presented an IS episode and 254 composed the control group. Blood samples were drawn for direct and reverse serotyping. The control and IS groups were compared regarding the traditional risk factors and the distribution of the ABO blood groups. Results: The IS group presented a higher prevalence of systemic arterial hypertension (SAH), diabetes mellitus, smoking habits, family history, cardiopathy, and sedentary lifestyle in comparison with the control group. The AB blood type prevailed among the patients (5.1 vs. 1.6%; p<0.05) and this group had more SAH cases in comparison with the O type group (92.9 vs. 67.3%; p<0.05). Conclusions: Our results suggest that the occurrence of IS is more frequent among patients of the AB blood type.


INTRODUCTION
Stroke is a common pathological condition caused by changes in blood flow to the brain. It is responsible for high morbidity and mortality rates, resulting in at least 5 million deaths and more than 15 million non-fatal cases per year. It can be ischemic-or hemorrhagic in origin. Hemorrhagic cases are due to rupture of a cerebral blood vessel, whereas ischemic cases are due to obstruction of a vessel by thrombosis or embolism 1,2,3 .
Ischemic stroke (IS) is the most frequent type in Brazil 4 . It is considered a complex disorder caused by a combination of influences of several genes and environmental factors 5 . The risk factors related to the disease can be divided into modifiable and non-modifiable 6 . Non-modifiable factors include age, gender, ethnicity and heredity. On the other hand, modifiable factors include systemic arterial hypertension (SAH), smoking habits, atrial fibrillation (AF), dyslipidemias, sedentary lifestyle, and heart disease 7 .
The ABO system is considered the most important blood group system. It is well established that blood type interferes with the hemostasis because it affects factor VIII and the Von Willebrand factor (VWF) 8 . Subjects of blood group O have lower levels of VWF and factor VIII than subjects of non-O blood groups (A, B, and AB). Approximately 70% of the variation in the plasmatic levels of the VWF and VIII factors is determined by genetics, with 30% of the genetic variation attributable to the subject's blood type. Remarkably, VWF levels are 25% higher in subjects of the non-O blood group 9 .
Besides the relation of the ABO system with blood transfusions compatibility, studies link this system with cardiovascular diseases. The association between non-O blood groups and the risk of developing venous thromboembolism (VTE) is well established 10 . Spiezia et al. reported in their case control study that subjects of the non-O blood groups had a 2.2 times higher risk of developing VTE than O-group subjects. Nevertheless, the link with arterial thrombotic events, such as myocardial infarction and IS was not well established 11 .
Sabino et al. investigated the frequency of ABO blood group polymorphisms and its association with IS and peripheral arterial diseases. The resulting data suggest a role of the non-O blood type in the pathogenesis of thrombotic events and a probable protective effect of the O blood type regarding the occurrence of IS 12 . Williams et al. inferred that the variants of the ABO gene are associated with cardioembolic stroke and large vessel stroke, but not with small vessels stroke 13 . In China, preliminary results from a case control study support the evidence that single nucleotide polymorphisms in the ABO gene may contribute to susceptibility to atherothrombotic IS, without a significant association with lacunar IS 14 . Additionally, Zakai et al. inferred that the AB blood type was linked to an increased ICVA risk, and 60% of such association was mediated by the levels of factor VIII 8 .
In this sense, the purpose hereof was to investigate the association between the ABO blood type and the occurrence of IS in the JOINVASC cohort.

METHODS
This case control study was originated from the JOINVASC cohort -Population-based Epidemiological Study in Cerebrovascular Diseases in Joinville. During one year, patients that had an IS episode and control subjects (people that never had a stroke episode, also residents of the city of Joinville, Santa Catarina, Brazil), whose age was equal or above 18 years, were included in the study. To classify the IS subtypes, we employed the TOAST (Trial of Org 10172 in Acute Stroke Treatment) diagnosis criteria. Five categories were described: (1) atherosclerosis of large arteries or atherothrombotic; (2) cardioembolic; (3) lacunar; (4) other etiologies; and (5) indeterminate etiology 15 . We collected data on traditional risk factors and compared prevalence levels between the control group and IS group, based on data routinely collected by the JOINVASC cohort 16 . After finding a higher frequency of certain blood types in patients, these individuals were distributed according to traditional risk factors and ischemic stroke subtypes considering the most frequent blood type vs. all other types and vs. O type (some authors indicate a possible protective effect for the incidence of thrombotic events 12 ).
Blood samples from the patients and control subjects were collected by peripheral venous puncture for blood typing. We used the Anti-A, Anti-B, and Anti-AB (Fresenius Kabi, Barueri, Brazil) Soroclone ® kit, according to the manufacturer's instructions. Briefly, a drop of each antiserum and a drop (50 μL) of recently prepared 5% red cell suspension were mixed inside tubes. After homogenization and centrifugation, the presence or absence of agglutination was used to characterize the occurrence of reaction. As a complement, the reverse ABO classification was used to determine serum antibodies using the REVERCEL PLUS ® (Fresenius Kabi) reagent.
In addition to descriptive statistical analysis, means and standard deviations were calculated for quantitative variables and relative frequencies for the qualitative variables. The chi-square test was used to compare proportions and the Student's t-test for verification of significant differences between means. The significance level established was p<0.05.
The study was approved by the Research Ethics Committee of the Universidade da Região de Joinville (protocol 45305615.5.0000.5366; ruling 1.138.314).

RESULTS
During the 12 months of the study, we registered 275 cases of IS (128 women and 147 men) and 254 subjects composed the control group (171 women and 83 men) ( Table 1).  Table 1 also shows the distribution and comparison of the prevalence rates between cases and controls according to the ABO system. The AB blood type was more prevalent among patients (5.1 vs. 1.6%; p<0.05). Table 2 shows the distribution of blood types according to the IS-related risk factors investigated. People with AB blood type had a higher prevalence of SAH compared with people with O type (92.9 vs. 67.3%; p<0.005). There was no significant difference between blood groups for the other risk factors studied.

DISCUSSION
Our results suggest that the occurrence of IS is more frequent among patients of the AB blood type. On the other hand, the occurrence of SAH seems to be more significant among these patients.
Several authors have compared the distribution of IS occurrence among ABO system blood groups. Blood donors in Denmark and Sweden were followed for diagnosis of thromboembolism and arterial events between 1987 and 2012. Compared to blood group O, non-O blood groups were associated with higher incidence of both venous and arterial thromboembolic events. Among arterial events, incidence rate ratios (IRRs) were generally low, but still significantly increased, with IRRs of 1.10 (95%CI 1.05-1.14) for myocardial infarction and 1.07 (95%CI 1.02-1.12) for stroke in non-O blood groups individuals compared to blood group O 17 . Zakai et al., in their REGARDS study to access the racial and regional differences on stroke and with 30,239 participants of the USA between the years of 2003 and 2007, reported a hazard ratio (HR) of 1.18 for the AB group for the occurrence of stroke in comparison with the O group. The authors also evaluated factor VIII in plasma by an immunoenzymatic test and they reported a higher level in the AB group subjects and estimated that factor VIII mediated 60% of the association between the AB group and stroke risk 8 . To our knowledge, this was the only study to date that reported an association between the AB blood type and IS. Such link was expected based on the studies that reported a higher occurrence of thrombotic events in subjects of the non-O blood groups.
Sabino et al., evaluating polymorphisms of the ABO system blood groups and the association with IS risk and peripheral artery disease, found data suggesting a possible protective factor of the O group in the pathogenesis of thrombotic events 12 . In another study assessing the association between venous thromboembolism and blood types, Sode et al. found the ABO system responsible for 20% of the risk attributable to venous thromboembolism 10 . In this same study, which comprised 2,279 events of venous thromboembolism, the HR of 1.4 (95%CI 1.3-1.5) was found for venous thromboembolism in subjects of the non-O groups in comparison with the O-group subjects 10 . The EUROCLOT study was designed in three stages to identify common variations that affect the fibrin structure and function in the normal population and significant single nucleotide polymorphisms linked to the occurrence of stroke. The authors reported that non-O blood groups were associated to a higher risk of peripheral vascular disease, heart attacks, and IS 13 . Consequently, although the link between the ABO system with thrombotic events is well established in the literature, further studies are still required with more diverse populations to confirm the association between the AB blood group and the occurrence of IS.
As expected, in our study, a higher prevalence of SAH was found among the patients with IS than in control subjects. The association between IS and hypertension is well established, and SAH is considered the most important modifiable risk factor for IS. The meta-analysis "Prospective Studies Collaborations" of 61 studies and a million participants found a strong link between blood pressure and mortality rate by IS by age. This study presented a log-linear increase of the mortality rate by IS according to age for increases of 20 and 10 mmHg in the systolic and diastolic pressures, respectively 18 . The same association was found by Rapsomanik et al. in their cohort that electronically gathered the data of 1.25 million patients without cardiovascular diseases at first and with a minimal age of 30 years, resulting in HR of 1.35 (95%CI 1.28-1.42) for the occurrence of IS with an increase of 20 mmHg in the systolic pressure 19 .
In addition to hypertension, we also found a higher prevalence of DM, cardiopathy, sedentary lifestyle, smoking habits, and positive family history among patients in comparison with the control subjects. We know that some heart diseases can increase the risk of IS, from which AF is considered the most important and treatable pre-IS cardiopathy 20 . The Framingham study associated non-valvular AF with a three-to five-time higher risk of IS. Additionally, the IS risk decreased with age for all other cardiovascular conditions, except for AF. In subjects aged 80 to 89 years, AF was the only cardiovascular condition that independently affected IS risk 20 . The higher prevalence of DM among patients was also expected, since diabetic subjects are more susceptible to atherosclerosis and have a higher frequency of atherogenic risk factors. The relationship between DM and IS was also reported in the Framingham study, in which glucose-intolerant subjects had twice the risk of cerebral infarction than non-diabetic subjects 20 . As for the modifiable risk factors such as smoking habits and sedentary lifestyle, studies have reported their association with IS risk. Regarding smoking habits, both the Framingham study and the Nurses Health study -a cohort that collected 12 years of data from 117,006 female nurses aged 30 to 55 years in 11 of the most populous states in the USA -reported a reduction in IS risk with smoking cessation 20,21 . Therefore, all differences found between patients and control subjects in modifiable and non-modifiable risk factors related to IS were already expected and support the association of our results with well-established findings in the literature 3,22,23 .
Our results suggest a higher prevalence of SAH in the AB group subjects in comparison with the O-group subjects. We found no other study supporting the link between hypertension and the ABO system's blood groups. Additionally, there was no evidence of an association between blood groups and any of the other risk factors investigated. However, Zakai et al. reported a relationship with diabetes, with higher prevalence in the AB group than in the O group 8 . However, this association was not found in our study.
Our main finding -AB blood group patients had a higher frequency of IS -could be affected by the small number of subjects with this blood type, which is the first limitation of the study, followed by the fact that it was conducted in a single center. However, we know that the prevalence of the AB blood group in the Brazilian population is actually low. Studies conducted in different regions of the country with blood donors identified prevalences that ranged from 3.13 to 4% 24,25 . In addition, because of the limited number of individuals in the patient and control groups, as well as the lack of homogeneity of baseline characteristics between groups, it is not possible to perform inferenceseg, by multiple logistic regression, about the AB blood group being a risk factor for the ischemic event. To overcome these limitations and confirm the results presented here, we recommend increasing the number of evaluated subjects and replicating the study in other centers. In addition, the clinical utility of ABO system for IS prediction warrants further study, since the determination of blood group phenotypes is easy and robust and there is no influence of acute phase response on blood group assessment.