Epidemiology and Prevention
ABO Blood Group and Risk of Thromboembolic
and Arterial Disease
A Study of 1.5 Million Blood Donors
Senthil K. Vasan, MD, PhD; Klaus Rostgaard, MSc; Ammar Majeed, MD, PhD;
Henrik Ullum, MD, PhD; Kjell-Einar Titlestad, MD, PhD; Ole B.V. Pedersen, MD, PhD;
Christian Erikstrup, MD, PhD; Kaspar Rene Nielsen, MD, PhD; Mads Melbye, MD, DrMedSci;
Olof Nyrén, MD, PhD; Henrik Hjalgrim, MD, DrMedSci; Gustaf Edgren, MD, PhD
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Background—ABO blood groups have been shown to be associated with increased risks of venous thromboembolic and
arterial disease. However, the reported magnitude of this association is inconsistent and is based on evidence from smallscale studies.
Methods and Results—We used the SCANDAT2 (Scandinavian Donations and Transfusions) database of blood donors
linked with other nationwide health data registers to investigate the association between ABO blood groups and the
incidence of first and recurrent venous thromboembolic and arterial events. Blood donors in Denmark and Sweden
between 1987 and 2012 were followed up for diagnosis of thromboembolism and arterial events. Poisson regression
models were used to estimate incidence rate ratios as measures of relative risk. A total of 9170 venous and 24 653
arterial events occurred in 1 112 072 individuals during 13.6 million person-years of follow-up. Compared with blood
group O, non-O blood groups were associated with higher incidence of both venous and arterial thromboembolic events.
The highest rate ratios were observed for pregnancy-related venous thromboembolism (incidence rate ratio, 2.22; 95%
confidence interval, 1.77–2.79), deep vein thrombosis (incidence rate ratio, 1.92; 95% confidence interval, 1.80–2.05),
and pulmonary embolism (incidence rate ratio, 1.80; 95% confidence interval, 1.71–1.88).
Conclusions—In this healthy population of blood donors, non-O blood groups explain >30% of venous thromboembolic
events. Although ABO blood groups may potentially be used with available prediction systems for identifying at-risk
individuals, its clinical utility requires further comparison with other risk markers. (Circulation. 2016;133:1449-1457.
DOI: 10.1161/CIRCULATIONAHA.115.017563.)
Key Words: ABO blood-group system ◼ cardiovascular diseases ◼ cerebrovascular disorders ◼
stroke ◼ thromboembolism ◼ thrombosis
T
he clinical importance of ABO blood groups extends beyond
the direct implications on transfusion and organ-transplantation compatibility. It is increasingly being recognized that
individuals with non-O blood groups may be at elevated risk
of venous thromboembolic events (VTEs), myocardial infarction, cerebrovascular ischemic events, and peripheral vascular
disease compared with individuals with blood group O.1–4 This
risk increase has been attributed to higher concentrations of factor VIII and von Willebrand factor.5,6 Approximately 70% of the
variation in the levels of these factors is genetically determined,
and 30% of the explained variation is attributable to ABO
Clinical Perspective on p 1457
blood groups.7 In addition, Sode and colleagues8 have shown
that ABO blood type has an additive effect on the risk of VTE
when combined with factor V Leiden R506Q and prothrombin
G20210A mutations. A recent genome-wide association study
has identified variants in the ABO locus that are associated with
VTE,9 providing additional clues to the genetic influence on
thromboembolic risk. Available epidemiological data, however,
fail to provide a coherent picture of the association between
ABO blood group and thromboembolic risk. First, there is a
Received May 18, 2015; accepted February 24, 2016.
From Department of Medical Epidemiology and Biostatistics, Karolinska Institutet, Stockholm, Sweden (S.K.V., A.M., O.N., G.E.); Department
of Epidemiology Research, Statens Serum Institute, Copenhagen, Denmark (K.R., M.M., H.H.); Department of Clinical Immunology, Blood Bank,
Rigshospitalet, University Hospital of Copenhagen, Denmark (H.U.); Department of Clinical Immunology, Odense University Hospital, Denmark
(K.-E.T.); Department of Clinical Immunology, Næstved Hospital, Denmark (O.B.V.P.); Department of Clinical Immunology, Aarhus University Hospital,
Denmark (C.E.); Department of Clinical Immunology, Aalborg University Hospital, Denmark (K.R.N.); and Hematology Centre, Karolinska University
Hospital, Stockholm, Sweden (A.M., G.E.).
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.
115.017563/-/DC1.
Correspondence to Senthil K. Vasan, MD, PhD, Department of Medical Epidemiology and Biostatistics, Karolinska Institutet Nobels Väg 12b,
Stockholm, 171 77 Sweden. E-mail [email protected]
© 2016 American Heart Association, Inc.
Circulation is available at http://circ.ahajournals.org
DOI: 10.1161/CIRCULATIONAHA.115.017563
1449
1450 Circulation April 12, 2016
lack of consensus on the magnitude of this association, the
extent to which it may explain population risks, and whether it
affects both arterial and venous thromboembolic disease at similar magnitudes. Second, even though plausible mechanisms
explaining this association have been described, it is unclear if
the risk elevation among non-O blood group carriers is solely a
direct effect or may also be mediated through increased risks of
other diseases. Third, there is a paucity of data concerning associations with recurrent VTE events. To clarify these issues, we
used a large cohort consisting of essentially all healthy blood
donors after 1987 from Sweden and Denmark to investigate
the effect of ABO blood group on venous thromboembolic
and cardiovascular disease.
Methods
Data Sources
The SCANDAT2 (Scandinavian Donations and Transfusions)
database is a computerized combined donation and transfusion
register from Sweden and Denmark. After a recent update, the database includes information on >1.6 million blood donors who have
donated different blood products since 1968 and 1981 in Sweden
and Denmark, respectively.10 Using unique personal national registration numbers assigned to all residents in both countries,11,12 the
database is linked to nationwide population, death, and migration
registers, thus ensuring complete follow-up of all study participants.
The database is also linked to inpatient, outpatient, cause of death,
and cancer registers, allowing long-term follow-up for a range of
health outcomes.
Table 1. Characteristics of the Study Population
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Non-O Blood Groups
Blood Group O
Combined
Subjects, n (% of total)
665 952 (59.9)
446 120 (40.1)
1 112 072 (100.0)
Female, n (%)
341 307 (51.3)
230 533 (51.7)
571 840 (51.4)
Sweden
430 186 (64.6)
273 315 (61.3)
703 501 (63.3)
Denmark
235 766 (35.4)
172 805 (38.7)
408 571 (36.7)
≤39 y
477 990 (71.8)
319 947 (71.7)
797 937 (71.8)
40–59 y
Country, n (%)
Age at study entry, n (%)
180 068 (27.0)
120 666 (27.0)
300 734 (27.0)
≥60 y
7 894 (1.2)
5 507 (1.2)
13 401 (1.2)
Mean (SD) age, y
32.5 (11.2)
32.5 (11.2)
32.5 (11.2)
Overall
12.6 (6.5–17.2)
12.4 (6.3–17.0)
12.6 (6.4–17.1)
Sweden
13.7 (6.8–19.3)
13.6 (6.7–19.2)
13.7 (6.7–19.3)
Denmark
11.1 (6.1–14.9)
11.0 (6.0–14.9)
11.1 (6.1–14.9)
628 503 (94.4)
421 947 (94.6)
1 050 450 (94.5)
37 449 (5.6)
24 173 (5.4)
61 622 (5.5)
Median duration of follow-up (IQR), y
Country of birth, n (%)
Sweden/Denmark
Immigrants
ABO blood group, n (%)
A
486 297 (73.0)
486 297 (43.7)
AB
55 733 (8.4)
55 733 (5.0)
B
123 922 (18.6)
123 922 (11.1)
O
446 120 (100.0)
446 120 (40.1)
549 023 (82.4)
364 488 (81.7)
913 511 (82.1)
Atrial fibrillation
8 833 (1.3)
5 761 (1.3)
14 594 (1.3)
Cancer
26 260 (3.9)
17 533 (3.9)
43 793 (3.9)
Diabetes mellitus
8 889 (1.3)
5 434 (1.2)
14 323 (1.3)
Fractures
4 811 (0.7)
3 148 (0.7)
7 959 (0.7)
196 287 (29.5)
131 624 (29.5)
327 911 (29.5)
Rhesus antigen positivity, n (%)
Comorbidity during follow-up, n (%)
Surgery
Data represented as mean (SD) for normally distributed variables and as median and interquartile range (IQR) for skewed variables. Comorbidity
includes only established major risk factors for thromboembolism. Diabetes mellitus includes both types 1 and 2. Fractures include major fractures of
the hip or pelvis, spinal cord injury, multiple fractures, and fractures of long bones. Cancer includes cancers of the lung, gastrointestinal tract, pancreas,
female genitals, or prostate or leukemia. Surgery includes coronary bypass grafting, surgery for gynecological malignancies, major urological surgeries,
neurosurgery, hip and knee replacement surgeries, surgery for major fractures of long bones, and multiple trauma and spinal cord surgeries.
Vasan et al ABO Blood Group and Risk of Vascular Events 1451
Table 2. Risk of Thromboembolism and Vascular Disease in Relation to Blood Group, Presented Overall and Stratified by Age
Non-O Blood Groups
Classification
Events
PYs
Blood Group O
IRR (95% CI)
Events
PYs
IRR (95%CI)
1.00
(Referent)
PAR% (95% CI)
All VTE
All ages combined
6710
8 122 072
1.80 (1.71 to 1.88)
2460
5 380 683
≤30 y
732
2 041 723
2.04 (1.76 to 2.36)
238
1 353 341
32 (30 to 35)
31–60 y
4326
5 475 888
1.92 (1.81 to 2.03)
1471
3 621 708
36 (33 to 38)
≥61 y
1652
604 461
1.46 (1.34 to 1.59)
751
405 634
22 (17 to 26)
3302
8 147 684
1.75 (1.64 to 1.86)
1252
5 389 285
38 (31 to 45)
Pulmonary embolism
All ages combined
1.00
(Referent)
31 (28 to 34)
≤30 y
302
2 043 556
1.95 (1.56 to 2.43)
105
1 353 970
31–60 y
2025
5 494 218
1.93 (1.77 to 2.10)
690
3 627 786
36 (32 to 40)
≥61 y
975
609 910
1.42 (1.27 to 1.58)
457
407 529
20 (14 to 26)
36 (25 to 46)
Deep vein thrombosis
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
All ages combined
3514
8 141 852
1.92 (1.80 to 2.05)
1181
5 388 128
≤30 y
314
2 043 339
2.18 (1.73 to 2.75)
94
1 353 925
31–60 y
2422
5 489 299
2.03 (1.87 to 2.21)
764
3 626 847
38 (34 to 42)
≥61 y
778
609 214
1.56 (1.37 to 1.78)
323
407 357
25 (18 to 32)
1.00
(Referent)
36 (32 to 39)
42 (31 to 51)
Other venous thromboembolic events
All ages combined
364
8 164 054
1.53 (1.27 to 1.85)
157
5 394 829
≤30 y
53
2 044 432
1.99 (1.17 to 3.40)
18
1 354 291
31–60 y
222
5 504 635
1.46 (1.15 to 1.85)
100
3 631 239
22 (8 to 34)
≥61 y
89
614 988
1.50 (1.03 to 2.18)
39
409 299
23 (2 to 41)
325
3 778 388
2.22 (1.76 to 2.78)
96
2 507 085
1.00
(Referent)
24 (14 to 34)
37 (9 to 59)
Pregnancy/abortion-related VTE
All ages combined
≤30 y
144
1 130 400
2.24 (1.59 to 3.16)
42
750 426
31–60 y
181
2 647 987
2.20 (1.62 to 2.98)
54
1 756 659
6476
8 126 796
1.10 (1.05 to 1.14)
3909
5 372 408
1.00
(Referent)
42 (31 to 52)
43 (26 to 56)
42 (27 to 54)
Myocardial infarction
All ages combined
≤30 y
1.00
(Referent)
6 (3 to 8)
31
2 044 611
0.88 (0.52 to 1.50)
24
1 354 313
31–60 y
3950
5 486 610
1.18 (1.12 to 1.24)
2235
3 621 162
10 (7 to 13)
N/A
≥61 y
2495
595 575
1.00 (0.94 to 1.06)
1650
396 933
N/A
4401
8 142 440
1.07 (1.02 to 1.12)
2736
5 380 992
92
2 044 432
1.55 (1.07 to 2.24)
40
1 354 239
31–60 y
2301
5 494 906
1.07 (1.00 to 1.14)
1433
3 625 208
4 (0 to 8)
≥61 y
2008
603 102
1.06 (0.98 to 1.13)
1263
401 544
3 (−1 to 7)
3784
8 146 862
1.06 (1.01 to 1.12)
2389
2389
71
2 044 238
0.85 (0.60 to 1.21)
57
57
31–60 y
1766
5 497 183
1.10 (1.02 to 1.18)
1085
1085
9 (2 to 15)
≥61 y
1947
605 441
1.04 (0.97 to 1.12)
1247
1247
4 (−3 to 11)
672
8 161 732
1.55 (1.35 to 1.78)
286
286
Cerebrovascular stroke
All ages combined
≤30 y
1.00
(Referent)
4 (1 to 7)
25 (4 to 43)
Peripheral vascular disease
All ages combined
≤30 y
1.00
(Referent)
6 (1 to 11)
N/A
Other arterial thrombosis
All ages combined
1.00
(Referent)
35 (26 to 44)
≤30 y
40
2 044 501
1.79 (0.99 to 3.25)
15
15
31–60 y
415
5 503 149
1.68 (1.40 to 2.02)
163
163
40 (29 to 50)
≥61 y
217
614 081
1.32 (1.05 to 1.66)
108
108
24 (5 to 40)
44 (−1 to 69)
Event is defined as occurrence of first venous thromboembolic or cardiovascular outcome. Adjusted model was adjusted for age, sex, calendar period, country, and
comorbidity. CI indicates confidence interval; IRR, incidence rate ratio; PAR%, population-attributable risk percent; PY, patient-year; and VTE, venous thromboembolic event.
1452 Circulation April 12, 2016
Table 3. Risk of Thromboembolism and Vascular Disease in Relation to A1/A2 Subtype
A1
A1B
Classification
Events
PYs
IRR (95% CI)
Events
PYs
IRR (95% CI)
All venous thromboembolism
1894
1 799 802
2.01 (1.88–2.14)
175
146 903
2.24 (1.91–2.61)
Pulmonary embolism
794
1 808 452
1.82 (1.65–2.01)
75
147 681
2.07 (1.64–2.63)
Deep vein thrombosis
1184
1 804 553
2.21 (2.03–2.41)
108
147 427
2.43 (1.99–2.96)
Other VTEs
90
1 812 398
1.86 (1.39–2.48)
2
148 225
0.50 (0.12–2.04)
Pregnancy/abortion-related VTE
87
748 100
2.69 (1.97–3.67)
11
63 127
4.25 (2.25–8.02)
Myocardial infarction
1529
1 803 251
1.07 (1.00–1.14)
150
147 287
1.26 (1.07–1.49)
Cerebrovascular stroke
986
1 807 213
1.04 (0.96–1.12)
91
147 661
1.16 (0.94–1.44)
Peripheral vascular disease
728
1 809 546
1.03 (0.94–1.13)
59
147 941
0.99 (0.76–1.29)
Other arterial thrombosis
169
1 811 802
1.76 (1.42–2.17)
14
148 103
1.72 (1.00–2.96)
(Continued)
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Event is defined as occurrence of first venous thromboembolic or cardiovascular event. Adjusted model was adjusted for age, sex, calendar period, country, and comorbidity.
CI indicates confidence interval; IRR, incidence rate ratio; PAR%, population-attributable risk percent; PY, patient-year; and VTE, venous thromboembolism.
Study Design
The primary objective was to analyze the association between ABO
blood group and risk of VTEs (both incident and recurrent) and
arterial thromboembolic events after the first donation. Although
the SCANDAT database dates back to 1968 and 1981, for the purpose of the present study, we used data on all blood donors with
known blood group and at least 1 allogeneic donation of whole
blood, plasma, or platelets only after January 1, 1987 (for Sweden,
when the Swedish inpatient register attained full national coverage), or January 1, 1994 (for Denmark, when the International
Classification of Disease [ICD]. Revision 10 was implemented),
until 2012. We used ICD-9 and ICD-10 codes (Table I in the
online-only Data Supplement) to identify all inpatient and outpatient healthcare encounters with a diagnosis of venous thromboembolic or cardiovascular disease, as well as all related deaths.
The ICD-9 and ICD-10 codes for primary discharge diagnoses and
comorbidities are classified by the treating physician. Information
on diagnosis and were obtained by linking the SCANDAT2 database to the respective patient registers.
Donors with a history of venous thromboembolic or cardiovascular disease (in the national patient registers) before the first registered blood donation were excluded from the study. To allow the
assessment of the possible influence of other established risk factors
for venous thromboembolic and cardiovascular disease,13 we also
extracted data on healthcare encounters with a diagnosis of diabetes
mellitus (either type 1 or type 2), atrial fibrillation, fractures or orthopedic procedures (surgery for fractures of the hip or leg, replacement
arthroplasty, major trauma, hip and knee replacement, and fractures
involving spinal cord injury), major surgery (abdominal surgery, cardiovascular surgery such as coronary artery bypass graft surgery, surgery for cancer, and gynecological surgery), and cancer (Table I in the
online-only Data Supplement).
Statistical Analysis
Differences in baseline characteristics between subjects with non-O
blood groups and those with blood group O were assessed with t tests
for continuous variables and with χ2 tests for categorical variables.
In analyses of first occurrences (ie, incident cases) of VTEs or
cardiovascular events, we followed up donors from the date of first
electronically recorded blood donation during the study period until
the date of first VTE or cardiovascular event, emigration, death, or
end of follow-up (December 31, 2012), whichever came first. In the
analyses of pregnancy/abortion-related venous thromboembolism,
we terminated follow-up at 60 years of age. The main analyses for
each outcome were run separately following up subjects only for a
particular outcome of interest and not censoring for the occurrence
of other outcomes. Therefore, subjects could have 1 event of each
type. A person who was diagnosed with both pulmonary embolism
and a deep venous thrombosis therefore contributed events in both of
these outcome categories, as well as 1 event in the combined category
for all venous thromboembolism. Realizing that the different types
of outcomes share common risk factors, we also conducted sensitivity analyses in which follow-up was allowed only for the first event,
regardless of the type.14
We defined recurrent venous thromboembolism as deep vein
thrombosis or pulmonary embolism, after any first recorded venous
thrombotic event. During the follow-up period, these individuals were censored at date of emigration, death, or end of followup (December 31, 2012), whichever came first. Subjects with >1
venous event at first presentation (eg, both deep venous thrombosis and pulmonary embolism) were also included in the recurrence
analyses. We only followed up subjects for the first recurrent event
to avoid effects of lifelong antithrombotic prophylaxis, which is
commonly used for repeat venous thrombosis. Because the patient
registries do not specify diagnoses as incident or recurrent and
because patients may have several healthcare contacts for the same
thromboembolic event, we delayed the start of follow-up for recurrent thromboembolism until 180 days after the first healthcare contact for the first event.
We also considered the association between ABO blood group
and the risk of provoked thromboembolic event, which was defined
as occurring in an individual with at least 1 identifiable transient risk
factor (trauma, fracture, major surgeries) no more than 6 months
before any venous thrombotic event or a permanent risk factor (diabetes mellitus, cancer, atrial fibrillation) diagnosed any time before a
thromboembolic event.15
Baseline characteristics of continuous variables are presented
as mean and SD and categorical variables as frequencies and percentages. The relative risk of VTEs or cardiovascular events in
individuals with non-O blood groups compared with individuals
with blood group O was expressed as incidence rate ratios (IRR),
estimated with log-linear Poisson regression models. For each
of the different outcomes, we fitted 2 separate models. Model 1
had adjustment for attained age (in 1-year intervals), sex, country
(Sweden/Denmark), and calendar period of observation (in 1-year
intervals). In model 2, we additionally adjusted for comorbidity.
In both models, attained age and calendar period were treated as
time-dependent factors, allowing individuals to move between
categories over time, and were modeled as restricted cubic splines
with 5 knots. We defined comorbidity as a history of a diagnosis of 1 of the following previously reported concomitant risk
Vasan et al ABO Blood Group and Risk of Vascular Events 1453
Table 3. Continued
A2
A2B
O
Events
PYs
IRR (95% CI)
Events
PYs
IRR (95% CI)
Events
PYs
IRR
(95% CI)
309
494 902
1.19 (1.05–1.34)
67
68 241
1.86 (1.46–2.38)
1794
3 559 340
1.00 (Referent)
122
496 426
1.01 (0.84–1.22)
33
68 507
1.99 (1.40–2.82)
809
3 566 735
1.00 (Referent)
186
495 711
1.26 (1.08–1.48)
38
68 438
1.87 (1.35–2.59)
1007
3 564 476
1.00 (Referent)
21
496 961
1.59 (0.99–2.55)
4
68 714
2.16 (0.79–5.88)
95
3 570 651
1.00 (Referent)
13
207 944
1.45 (0.80–2.62)
0
29 705
0.00 (n/a)
74
1 626 072
1.00 (Referent)
432
494 336
1.09 (0.98–1.21)
60
68 360
1.11 (0.86–1.44)
2440
3 556 810
1.00 (Referent)
263
495 659
0.98 (0.86–1.11)
32
68 531
0.87 (0.61–1.23)
1687
3 561 380
1.00 (Referent)
180
496 163
0.91 (0.78–1.07)
37
68 576
1.37 (0.99–1.90)
1233
3 565 348
1.00 (Referent)
34
496 787
1.28 (0.89–1.85)
6
68 692
1.62 (0.72–3.65)
174
3 569 887
1.00 (Referent)
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
factors: diabetes mellitus,16 major surgery, cancer, or fractures.
Binary variables for each of these factors were included as timedependent variables in the model. Because atrial fibrillation might
trigger thrombosis formation, increasing the risk of stroke, myocardial infarction, and arterial emboli, as well as being associated with risk of venous thromboembolism independently17 or in
association with concomitant stroke,18 we additionally included
atrial fibrillation in the comorbidity-adjusted model. Wald tests
were used to compute 95% confidence intervals (CI). Interactions
between blood group and sex, country, calendar period, age, and
the respective comorbidities were individually tested. Because the
observed IRRs did not differ materially between models 1 and 2,
we present only estimates from model 2. We also tested for differences in risk among a subset of Swedish donors with known A1/
A2 subtypes compared with blood group O. Finally, we calculated
the population-attributable risk percent (PAR%), that is, the percentage of events in the entire donor cohort that could be attributed
to non-O blood groups. CIs for the PAR estimates were constructed
with the substitution method.19
All data processing and statistical analyses were done with SAS
statistical analysis software (version 9.4, SAS Institute, Inc, Cary,
NC). Values of P<0.05 were considered statistically significant.
Ethics
The creation of the SCANDAT2 database and the conduct of this
study were approved by the regional ethics committees in Stockholm,
Sweden, and the Danish Data Protection Agency in accordance with
national legislations.
Results
A total of 1 112 072 blood donors were included in the analysis.
Mean age at entry was 33 years. Median duration of follow-up
was 12.6 years. Female donors constituted 51.4% (n= 571 840).
Apart from differences in the proportion of Swedish/Danish
donors (P<0.001) and immigrated donors (P<0.001), as well
as of the distribution of Rhesus positivity (P<0.001), baseline
demographics did not show any statistically significant difference
between individuals with non-O and O blood groups (Table 1).
The analysis of incident VTEs and cardiovascular events
included 665 952 individuals with non-O blood groups (A, B,
or AB) and 446 120 with blood group O. Throughout followup, we observed 9170 VTEs and 24 653 cardiovascular events.
The IRRs were highest for the venous events, with all venous
thrombotic events combined having an IRR of 1.80 (95% CI,
1.71–1.88) in individuals with non-O blood group compared
with blood group O. The risk patterns were largely similar for
pulmonary embolism and deep vein thrombosis, but a somewhat higher IRR was observed for pregnancy-related thromboembolic events (IRR, 2.22; 95% CI, 1.76–2.78). Among
arterial events, IRRs were generally lower yet still statistically
significant 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 individuals in non-O blood groups compared with those in blood
group O (Table 2). Age demonstrated a significant interaction
with ABO blood group (P<0.001 for all outcomes), with IRRs
generally decreasing with older age (Table 2). For all venous
outcomes except other venous thrombosis, blood group also
exhibited significant interactions with sex, with higher IRRs
in men, although there was no evidence of interaction between
blood group and sex for the arterial outcomes (Table II in the
online-only Data Supplement). There was no evidence of statistical interaction between blood group and country (Table III
in the online-only Data Supplement) or between blood group
and calendar time (data not shown).
The PAR%s conferred by the non-O blood groups were considerably higher for VTEs than for the cardiovascular events,
except for cerebrovascular stroke in young subjects and other
arterial thrombosis. For venous vascular events, the PAR%
ranged from 24% (95% CI, 14–34) for other venous thrombosis to 42% (95% CI, 31–52) for pregnancy/abortion-associated
VTE. In addition, the PAR% was generally higher in younger
subjects (Table 2). Virtually identical results were reached in
the sensitivity analyses in which we censored individuals upon
having their first vascular event of any type (data not shown).
Further analyses of the 122 003 and 35 738 Swedish donors
with known A1 and A2 subtype revealed increased risks of
both venous and arterial events for individuals with the A1
subtype, whereas among individuals with the A2 subtype, significant risk was observed only for deep venous thrombosis
(Table 3).
1454 Circulation April 12, 2016
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
A consistent pattern was observed when the non-O blood
group was subclassified as blood groups A, AB, and B, with
the highest rate ratios for VTEs consistently seen in individuals with blood group AB and relatively lower risks observed
with blood groups A and B compared with blood group O. The
risk was slightly lower but statistically significant for arterial events compared with the venous group (Table IV in the
online-only Data Supplement). Sensitivity analysis restricted
to individuals with atrial fibrillation did not show any change
in IRR estimates for acute myocardial infarction, stroke, and
arterial embolic disease.
Table 4 shows IRR estimates for recurrent and provoked
pulmonary embolism and deep venous thrombosis. In both
sets of analyses, the incidence was again higher in donors
with non-O blood groups than in donors with blood group
O, but relative risk estimates were generally lower than in
the analyses of incident events. In individuals with non-O
blood groups, the risks for recurrent pulmonary embolism
(IRR, 1.47; 95% CI, 1.30–1.66) and deep venous thrombosis provoked by a comorbid illness (IRR, 1.67; 95% CI,
1.51–1.85) were higher compared with those with blood
group O.
Finally, in analyses restricted to events associated with
inpatient care (as opposed to inpatient and outpatient visits),
the estimates remained largely similar, although the number of
events decreased (data not shown).
Discussion
Using a large database comprising almost all Swedish and
Danish blood donors in the past 2 decades, we were able
to show that individuals with non-O blood groups are at
increased risks of both venous thromboembolic and cardiovascular disease compared with individuals with blood
group O. Strikingly, although IRRs were comparatively
modest, the incidence differences translated to generally
high PARs, exceeding 30% for VTEs in this healthy population of donors. Although absolute incidence rates were
higher for recurrent and provoked venous events than for
first-time events, the proportion attributable to the non-O
blood group was smaller. This presumably may reflect a
combination of the use of thromboprophylaxis after the first
thromboembolic event and a greater prevalence of other
important risk factors, with the latter explanation supported
by higher absolute rates in both of these groups.
Our study has several important strengths. It is by far
the largest study ever to have investigated the association
between ABO blood group and risk of both venous thromboembolic and cardiovascular disease in a large, well-defined
cohort with a wide age range and data on the most important
comorbidities. The robust study design and complete followup data from nationwide population and health registers are
additional strengths of the study. Because the blood group
data were recorded prospectively in registers that are used
Table 4. Risk of Recurrent and Provoked Thromboembolism (Pulmonary Embolism and Deep Venous Thrombosis) in Relation to
Blood Group, Presented Overall and Stratified by Sex
Non-O Blood Group
Blood Group O
Events
PYs
IRR (95%CI)
Events
PYs
IRR (95%CI)
Overall
1082
168 426
1.47 (1.30–1.66)
356
83 147
1.00 (Referent)
Male
669
108 039
1.58 (1.35–1.85)
207
53 329
Female
413
60 387
1.31 (1.08–1.58)
149
29 818
Overall
2278
162 174
1.44 (1.32–1.56)
752
81 411
Male
1385
104 478
1.57 (1.41–1.75)
416
52 422
Female
893
57 696
1.26 (1.11–1.43)
336
28 989
Overall
1346
587 740
1.56 (1.42–1.73)
564
391 884
Male
758
304 126
1.59 (1.39–1.81)
316
203 607
Female
588
283 614
1.53 (1.32–1.78)
248
188 277
Overall
1325
587 292
1.67 (1.51–1.85)
505
392 123
Male
779
303 788
1.76 (1.54–2.01)
284
203 767
Female
546
283 504
1.55 (1.33–1.82)
221
188 356
Recurrent thrombosis*
Pulmonary embolism
Deep venous thrombosis
1.00 (Referent)
Provoked thrombosis†
Pulmonary embolism
1.00 (Referent)
Deep venous thrombosis
1.00 (Referent)
*Recurrent thrombotic event defined as occurrence of DVT or PE 180 days after discharge for any incident venous or arterial thrombotic event.
†Provoked thrombotic event defined as presence of either transient (6 mo) or permanent risk factor before occurrence of new (incident) thromboembolic event. IRR
(95%CI): Incidence rate ratio (95% confidence interval). IRR adjusted for age, sex, calendar period, income, country of birth and comorbidity. PY indicates person years.
Vasan et al ABO Blood Group and Risk of Vascular Events 1455
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
in transfusion management, there should be little or no misclassification of blood group. Importantly, although blood
donors may be aware of their blood group, it is unlikely that
such knowledge will have any effect on behavior; therefore,
significant confounding by lifestyle factors is unlikely.
At the same time, because the analyses are based on data
on a healthy blood donor population, the results may not be
entirely generalizable to other settings where comorbidity is
more prevalent. This is exemplified by the lower relative risk
estimates in the analysis of provoked venous thromboembolic disease and among older subjects. Similarly, because
our study is limited to individuals of mostly European
descent, it may be difficult to extrapolate our findings to
other populations. This is especially relevant for the PAR
calculations, which are affected by the prevalence of other
risk factors. We also acknowledge that the SCANDAT2 database is based on administrative databases that were not created for research purposes. However, because these data are
routinely used in clinical transfusion practice, the validity
of the blood group data should be very high; therefore, we
strongly believe that the blood group data are sufficiently
accurate for this study purpose. The use of hospital registers
for outcome ascertainment is likely to have resulted in some
degree of underreporting and possibly some misclassification for the outcomes that do necessarily warrant hospital
care or may remain undiagnosed, for example, deep venous
thrombosis. However, the degree of such underreporting or
misclassification is unlikely to be related to the donor blood
group, which is a genetic factor, so it should affect only the
power of investigation. Therefore, we do not believe that
outcome misclassification would have an important effect on
our observations, and it has previously been shown that the
validity of VTEs and arterial events in the inpatient register
has high positive predictive value.20
The association between ABO blood group and the risk of
vascular disease has long been recognized, but previous studies have usually been limited to specific hospital settings with
marked variation in study quality. To the best of our knowledge, this is the first study to provide precise estimates of
venous thrombosis and arterial disease risk with ABO blood
groups using a large cohort of blood donors, and our results
are consistent with previous reports.3,21–23 However, a recent
meta-analysis reported higher relative risk estimates for both
venous and arterial events compared with our study, although
the included studies demonstrate a high degree of statistical
heterogeneity.4 Scattered hospital-based reports using small
samples have previously shown inconsistent results attributable to ABO blood group on the risk of thromboembolic
events after surgery,24–26 trauma,27,28 cancer,29,30 and diabetes mellitus,31 and these studies show a marked variation in
study quality. The overall risk pattern observed in our study
appeared to be similar for all subgroups of venous thromboembolic disease, with similar risk elevations for individuals with blood groups A and B who may be carriers of an O
allele, and a uniformly higher risk for individuals with AB
group who do not carry any O alleles. This pattern is compatible with the risk being correlated with the probability of
carrying an O allele; however, because the SCANDAT 2 database does not include information on ABO genotyping, we
can only speculate about the dose-response effect between
the amount of O(H) antigen and risk of thromboembolism
as reported previously.32 Our observation of associations in
both crude and comorbidity-adjusted analyses suggests that
ABO blood groups act independently of other risk factors for
thromboembolic and vascular events. The only significant
interaction observed was that between age and blood group,
and we believe that this is driven mostly by the gradually
increasing baseline risk occurring with older age and is not
necessarily reflective of a biological effect. The greater risk
increase of both arterial and venous events seen in individuals with A1 subtype compared with those with A2 subtype
indicates a greater hypercoagulable state for the A1 allele,
presumably through higher levels of von Willebrand factor
and factor VIII compared with other genotypes.33,34
We would like to caution interpretations on 2 specific associations. The effect conferred by blood group B on other VTEs seems
most likely to be related to the relatively small sample sizes and
the heterogeneous mixture of several minor venous thrombosis
categories, which might reflect some residual confounding from
specific comorbidities related to ABO blood group, for example,
cancer. The relatively smaller magnitude of our relative risk estimates for incident and recurrent events compared with previous
studies29,35 might be attributed to the fact that we were not able
to exclude individuals on anticoagulation therapy. In addition,
our criteria for identifying subjects with a diagnosis of recurrent
thromboembolic event could have led to missing early recurrent
thrombosis cases and misclassifying care for the already prevalent thrombosis as new events. However, again, we believe that
such misclassification is unlikely to relate to the donors’ blood
group, making it less likely to affect the risk estimates. That said,
if prevalent thrombosis was often misclassified as a recurrent
event, this may have contributed to the generally lower relative
risk estimates for recurrent thrombosis. Nevertheless, the risk
estimates in our study suggest that in the absence of ultrasound or
D-dimer measurement, non-O blood groups may still potentially
benefit as a simple proxy biomarker in the assessment and planning of anticoagulation therapy for residual venous thrombosis.
Possible
mechanisms
underlying
the
ABOthromboembolism association include excess concentrations
of or interaction with factor VIII and von Willebrand factor in
individuals with non-O blood36,37; a unified action of immunedominant blood group antigens present in the lining endothelium of blood vessels, along with the production of von
Willebrand factor by the endothelial cells38,39; and genetic variation in the ABO locus, affecting serum levels of inflammatory markers such as soluble intracellular adhesion molecule
1,40 tumor necrosis factor,41 and soluble E- and P-selectin.42,43
Conclusions
Our study provides robust evidence of a consistent association between the non-O blood groups and both VTEs and
cardiovascular events, with uniformly greater risk for venous
events. Given that non-O blood groups confer an overall
increased risk of incident, recurrent, and provoked thromboembolism, ABO blood group may have a role in thrombosis
risk assessment and could potentially be added to available clinical prediction systems, given the relative ease and
robustness of assessing blood group phenotypes and lack of
1456 Circulation April 12, 2016
influence of acute-phase response on blood grouping. This
is specifically likely to be the case for unprovoked venous
events. However, its clinical utility warrants further comparison with other risk markers.
Acknowledgments
We are thankful to all the blood banks in Sweden and Denmark for
both collecting and contributing data to this study.
Sources of Funding
The assembly of the SCANDAT database was made possible through
grants from the Swedish research council, the Swedish Heart-Lung
Foundation, the Swedish Society for Medical Research, and the
Danish Research Council.
Disclosures
None.
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
References
1. Dentali F, Sironi AP, Ageno W, Crestani S, Franchini M. ABO blood group
and vascular disease: an update. Semin Thromb Hemost. 2014;40:49–59.
doi: 10.1055/s-0033-1363460.
2. Franchini M, Liumbruno GM. ABO blood group: old dogma, new perspectives. Clin Chem Lab Med. 2013;51:1545–1553. doi: 10.1515/
cclm-2013-0168.
3. Dentali F, Sironi AP, Ageno W, Turato S, Bonfanti C, Frattini F, Crestani
S, Franchini M. Non-O blood type is the commonest genetic risk factor
for VTE: results from a meta-analysis of the literature. Semin Thromb
Hemost. 2012;38:535–548. doi: 10.1055/s-0032-1315758.
4. Wu O, Bayoumi N, Vickers MA, Clark P. ABO(H) blood groups and vascular disease: a systematic review and meta-analysis. J Thromb Haemost.
2008;6:62–69. doi: 10.1111/j.1538-7836.2007.02818.x.
5. Souto JC, Almasy L, Muñiz-Diaz E, Soria JM, Borrell M, Bayén L, Mateo
J, Madoz P, Stone W, Blangero J, Fontcuberta J. Functional effects of the
ABO locus polymorphism on plasma levels of von Willebrand factor, factor VIII, and activated partial thromboplastin time. Arterioscler Thromb
Vasc Biol. 2000;20:2024–2028.
6. Miller CH, Haff E, Platt SJ, Rawlins P, Drews CD, Dilley AB, Evatt B.
Measurement of von Willebrand factor activity: relative effects of ABO
blood type and race. J Thromb Haemost. 2003;1:2191–2197.
7. Orstavik KH, Magnus P, Reisner H, Berg K, Graham JB, Nance W. Factor
VIII and factor IX in a twin population: evidence for a major effect of
ABO locus on factor VIII level. Am J Hum Genet. 1985;37:89–101.
8. Sode BF, Allin KH, Dahl M, Gyntelberg F, Nordestgaard BG. Risk of
venous thromboembolism and myocardial infarction associated with
factor V Leiden and prothrombin mutations and blood type. CMAJ.
2013;185:E229–E237. doi: 10.1503/cmaj.121636.
9.Heit JA, Armasu SM, Asmann YW, Cunningham JM, Matsumoto
ME, Petterson TM, De Andrade M. A genome-wide association
study of venous thromboembolism identifies risk variants in chromosomes 1q24.2 and 9q. J Thromb Haemost. 2012;10:1521–1531. doi:
10.1111/j.1538-7836.2012.04810.x.
10. Edgren G, Rostgaard K, Vasan SK, Wikman A, Norda R, Pedersen OB,
Erikstrup C, Nielsen KR, Titlestad K, Ullum H, Melbye M, Nyrén O,
Hjalgrim H. The new Scandinavian Donations and Transfusions database
(SCANDAT2): a blood safety resource with added versatility. Transfusion.
2015;55:1600–1606. doi: 10.1111/trf.12986.
11.Ludvigsson JF, Otterblad-Olausson P, Pettersson BU, Ekbom A. The
Swedish personal identity number: possibilities and pitfalls in healthcare
and medical research. Eur J Epidemiol. 2009;24:659–667. doi: 10.1007/
s10654-009-9350-y.
12.Schmidt M, Pedersen L, Sørensen HT. The Danish Civil Registration
System as a tool in epidemiology. Eur J Epidemiol. 2014;29:541–549.
doi: 10.1007/s10654-014-9930-3.
13.Anderson FA Jr, Spencer FA. Risk factors for venous thromboem
bolism. Circulation. 2003;107(suppl 1):I9–I16. doi: 10.1161/01.
CIR.0000078469.07362.E6.
14. Ageno W, Becattini C, Brighton T, Selby R, Kamphuisen PW. Cardiovascular
risk factors and venous thromboembolism: a meta-analysis. Circulation.
2008;117:93–102. doi: 10.1161/CIRCULATIONAHA.107.709204.
15.Iorio A, Kearon C, Filippucci E, Marcucci M, Macura A, Pengo V,
Siragusa S, Palareti G. Risk of recurrence after a first episode of symptomatic venous thromboembolism provoked by a transient risk factor: a
systematic review. Arch Intern Med. 2010;170:1710–1716. doi: 10.1001/
archinternmed.2010.367.
16. Petrauskiene V, Falk M, Waernbaum I, Norberg M, Eriksson JW. The
risk of venous thromboembolism is markedly elevated in patients
with diabetes. Diabetologia. 2005;48:1017–1021. doi: 10.1007/
s00125-005-1715-5.
17. Saliba W, Rennert G. CHA2DS2-VASc score is directly associated with
the risk of pulmonary embolism in patients with atrial fibrillation. Am J
Med. 2014;127:45–52. doi: 10.1016/j.amjmed.2013.10.004.
18. Noel P, Gregoire F, Capon A, Lehert P. Atrial fibrillation as a risk factor for
deep venous thrombosis and pulmonary emboli in stroke patients. Stroke.
1991;22:760–762.
19. Daly LE. Confidence limits made easy: interval estimation using a substitution method. Am J Epidemiol. 1998;147:783–790.
20. Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall
C, Heurgren M, Olausson PO. External review and validation of the
Swedish national inpatient register. BMC Public Health. 2011;11:450.
doi: 10.1186/1471-2458-11-450.
21. Gallerani M, Reverberi R, Manfredini R. ABO blood groups and venous
thromboembolism in a cohort of 65,402 hospitalized subjects. Eur J Intern
Med. 2014;25:e27–e28. doi: 10.1016/j.ejim.2013.10.002.
22. Bronte-Stewart B, Botha MC, Krut LH. ABO blood groups in relation to
ischaemic heart disease. Br Med J. 1962;1:1646–1650.
23. Jick H, Slone D, Westerholm B, Inman WH, Vessey MP, Shapiro S, Lewis
GP, Worcester J. Venous thromboembolic disease and ABO blood type: a
cooperative study. Lancet. 1969;1:539–542.
24.White RH, Murin S, Wun T, Danielsen B. Recurrent venous
thromboembolism after surgery-provoked versus unprovoked
thromboembolism. J Thromb Haemost. 2010;8:987–997. doi:
10.1111/j.1538-7836.2010.03798.x.
25. Kodama J, Seki N, Masahiro S, Kusumoto T, Nakamura K, Hongo A,
Hiramatsu Y. D-dimer level as a risk factor for postoperative venous
thromboembolism in Japanese women with gynecologic cancer. Ann
Oncol. 2010;21:1651–1656. doi: 10.1093/annonc/mdq012.
26. Bakker EJ, Valentijn TM, Hoeks SE, van de Luijtgaarden KM, Leebeek
FW, Verhagen HJ, Stolker RJ. ABO blood type does not influence the risk
of cardiovascular complications and mortality after vascular surgery. Eur J
Vasc Endovasc Surg. 2013;45:256–260. doi: 10.1016/j.ejvs.2012.11.022.
27. Carter YM, Caps MT, Meissner MH. Deep venous thrombosis and ABO
blood group are unrelated in trauma patients. J Trauma. 2002;52:112–116.
28. Geerts WH, Code KI, Jay RM, Chen E, Szalai JP. A prospective study
of venous thromboembolism after major trauma. N Engl J Med.
1994;331:1601–1606. doi: 10.1056/NEJM199412153312401.
29.Ohira T, Cushman M, Tsai MY, Zhang Y, Heckbert SR, Zakai NA,
Rosamond WD, Folsom AR. ABO blood group, other risk factors and
incidence of venous thromboembolism: the Longitudinal Investigation of
Thromboembolism Etiology (LITE). J Thromb Haemost. 2007;5:1455–
1461. doi: 10.1111/j.1538-7836.2007.02579.x.
30. Streiff MB, Segal J, Grossman SA, Kickler TS, Weir EG. ABO blood
group is a potent risk factor for venous thromboembolism in patients
with malignant gliomas. Cancer. 2004;100:1717–1723. doi: 10.1002/
cncr.20150.
31. Jager A, van Hinsbergh VW, Kostense PJ, Emeis JJ, Yudkin JS, Nijpels
G, Dekker JM, Heine RJ, Bouter LM, Stehouwer CD. von Willebrand
factor, C-reactive protein, and 5-year mortality in diabetic and nondiabetic subjects: the Hoorn Study. Arterioscler Thromb Vasc Biol.
1999;19:3071–3078.
32.Morelli VM, De Visser MC, Vos HL, Bertina RM, Rosendaal FR.
ABO blood group genotypes and the risk of venous thrombosis:
effect of factor V Leiden. J Thromb Haemost. 2005;3:183–185. doi:
10.1111/j.1538-7836.2004.01071.x.
33. O’Donnell J, Boulton FE, Manning RA, Laffan MA. Amount of H antigen
expressed on circulating von Willebrand factor is modified by ABO blood
group genotype and is a major determinant of plasma von Willebrand factor antigen levels. Arterioscler Thromb Vasc Biol. 2002;22:335–341.
34. Tirado I, Mateo J, Soria JM, Oliver A, Martínez-Sánchez E, Vallvé C,
Borrell M, Urrutia T, Fontcuberta J. The ABO blood group genotype and
factor VIII levels as independent risk factors for venous thromboembolism. Thromb Haemost. 2005;93:468–474. doi: 10.1160/TH04-04-0251.
35. Gándara E, Kovacs MJ, Kahn SR, Wells PS, Anderson DA, Chagnon I, Le
Gal G, Solymoss S, Crowther M, Carrier M, Langlois N, Kovacs J, Little
Ma J, Carson N, Ramsay T, Rodger MA. Non-OO blood type influences
Vasan et al ABO Blood Group and Risk of Vascular Events 1457
the risk of recurrent venous thromboembolism: a cohort study. Thromb
Haemost. 2013;110:1172–1179. doi: 10.1160/TH13-06-0488.
36. Koster T, Blann AD, Briët E, Vandenbroucke JP, Rosendaal FR. Role of
clotting factor VIII in effect of von Willebrand factor on occurrence of
deep-vein thrombosis. Lancet. 1995;345:152–155.
37.Jenkins PV, O’Donnell JS. ABO blood group determines plasma von
Willebrand factor levels: a biologic function after all? Transfusion.
2006;46:1836–1844. doi: 10.1111/j.1537-2995.2006.00975.x.
38. Ruggeri ZM. Von Willebrand factor, platelets and endothelial cell interactions. J Thromb Haemost. 2003;1:1335–1342.
39. Goto S, Ikeda Y, Saldívar E, Ruggeri ZM. Distinct mechanisms of platelet
aggregation as a consequence of different shearing flow conditions. J Clin
Invest. 1998;101:479–486. doi: 10.1172/JCI973.
40. Paré G, Chasman DI, Kellogg M, Zee RY, Rifai N, Badola S, Miletich
JP, Ridker PM. Novel association of ABO histo-blood group antigen
with soluble ICAM-1: results of a genome-wide association study of
6,578 women. PLoS Genet. 2008;4:e1000118. doi: 10.1371/journal.
pgen.1000118.
41.Melzer D, Perry JR, Hernandez D, Corsi AM, Stevens K, Rafferty I,
Lauretani F, Murray A, Gibbs JR, Paolisso G, Rafiq S, Simon-Sanchez J,
Lango H, Scholz S, Weedon MN, Arepalli S, Rice N, Washecka N, Hurst
A, Britton A, Henley W, van de Leemput J, Li R, Newman AB, Tranah G,
Harris T, Panicker V, Dayan C, Bennett A, McCarthy MI, Ruokonen A,
Jarvelin MR, Guralnik J, Bandinelli S, Frayling TM, Singleton A, Ferrucci
L. A genome-wide association study identifies protein quantitative trait
loci (pQTLs). PLoS Genet. 2008;4:e1000072. doi: 10.1371/journal.
pgen.1000072.
42. Barbalic M, Dupuis J, Dehghan A, Bis JC, Hoogeveen RC, Schnabel RB,
Nambi V, Bretler M, Smith NL, Peters A, Lu C, Tracy RP, Aleksic N,
Heeriga J, Keaney JF Jr, Rice K, Lip GY, Vasan RS, Glazer NL, Larson
MG, Uitterlinden AG, Yamamoto J, Durda P, Haritunians T, Psaty BM,
Boerwinkle E, Hofman A, Koenig W, Jenny NS, Witteman JC, Ballantyne
C, Benjamin EJ. Large-scale genomic studies reveal central role of ABO
in sP-selectin and sICAM-1 levels. Hum Mol Genet. 2010;19:1863–1872.
doi: 10.1093/hmg/ddq061.
43. Paterson AD, Lopes-Virella MF, Waggott D, Boright AP, Hosseini SM,
Carter RE, Shen E, Mirea L, Bharaj B, Sun L, Bull SB; Diabetes Control
and Complications Trial/Epidemiology of Diabetes Interventions and
Complications Research Group. Genome-wide association identifies the
ABO blood group as a major locus associated with serum levels of soluble E-selectin. Arterioscler Thromb Vasc Biol. 2009;29:1958–1967. doi:
10.1161/ATVBAHA.109.192971.
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
CLINICAL PERSPECTIVE
ABO blood groups have previously been found to be associated with arterial and venous thrombosis, with consistently
increased risks observed among individuals with blood group A, AB, or B compared with those with blood group O.
However, the reported risk magnitudes from previous studies have been inconsistent and often come from small studies.
In the present study, we used data on 1.1 million healthy blood donors from the binational SCANDAT2 (Scandinavian
Donations and Transfusions) database, which contains nationwide data on blood donations and transfusions from Sweden
and Denmark, to investigate the relationship between ABO blood group and arterial/venous thrombotic events. Our results
confirm that individuals with non-O blood groups are at significantly increased risks of both venous and arterial thromboembolic events, with the highest relative risks observed for venous events. As a result of the high prevalence of the non-O
blood groups, as much as 30% to 40% of venous events in this healthy population can be attributed to factors related to these
blood groups. Associations persisted also for recurrent and provoked venous thrombosis, but with generally lower relative
risk estimates. The proposed underlying mechanisms driving these associations include higher concentrations of factor VIII
and von Willebrand factor in individuals with non-O blood groups. Our results indicate that ABO blood group could perhaps
be used along with the established risk factors both during the assessment of thromboembolic disease and during decision
making for secondary prevention with prophylactic antithrombotic therapy.
ABO Blood Group and Risk of Thromboembolic and Arterial Disease: A Study of 1.5
Million Blood Donors
Senthil K. Vasan, Klaus Rostgaard, Ammar Majeed, Henrik Ullum, Kjell-Einar Titlestad, Ole
B.V. Pedersen, Christian Erikstrup, Kaspar Rene Nielsen, Mads Melbye, Olof Nyrén, Henrik
Hjalgrim and Gustaf Edgren
Downloaded from http://circ.ahajournals.org/ by guest on June 17, 2017
Circulation. 2016;133:1449-1457; originally published online March 3, 2016;
doi: 10.1161/CIRCULATIONAHA.115.017563
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2016 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/133/15/1449
Data Supplement (unedited) at:
http://circ.ahajournals.org/content/suppl/2016/03/02/CIRCULATIONAHA.115.017563.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located,
click Request Permissions in the middle column of the Web page under Services. Further information about
this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/
SUPPLEMENTAL MATERIAL
Supplementary Table 1. Swedish and Danish versions of the ICD-9 and ICD-10 codes used for disease
classification.
All venous thromboembolism
Pulmonary embolism
Deep vein thrombosis
Other venous thrombosis
Pregnancy/abortion-related
venous thromboembolism
Myocardial infarction
Cerebrovascular stroke
Peripheral vascular disease
Other arterial thrombosis
Diabetes mellitus
Atrial fibrillation
Fractures
Cancer
ICD-9
415, 451B, 4511, 453A, 453C,
453D, 4530, 4532, 4533, 452,
671C,671D,671E, 6712, 6713, 6714
ICD-10
415
451B, 4511
453A, 453C, 453D, 4530, 4532,
4533, 452
I26
I801, I802
671C,671D,671E, 6712, 6713, 6714
O087, O223, O871, O882
410
434A, 434B, 4340, 4341, 435
440, 441, 442, 443W, 443X,4438,
4439
433C, 593W, 444X, 444B,
444C,444W, 557A, 4332, 5938,
4449, 4441, 4442, 4448, 5570
I21
I63
250
427
808, 820, 806, 827, 823A, 823B,
823C, 823D, 952, 820, 821
140 to 208
E109, E119, E149
I48
I26, I801, I802, I81, I820, I822, I823,
I636, I676, O087, O223, O871, O882
I81, I820, I822, I823, I636, I676
I70, I71, I72, I739,
I74, N280
952,S31, T12, S72, S127
C00 to C96
Supplementary Table 2. Risk of thromboembolism and vascular disease in relation to blood group,
presented stratified by sex.
Non-O blood group
Classification
Events
Person years
IRR (95%CI)
Blood group O
Events
Person years
IRR (95%CI)
All venous thromboembolism
Males
3749
4 109 397
1.90 (1.78-2.02)
1297
2 712 137
Females
2961
4 012 675
1.67 (1.56-1.79)
1163
2 668 546
Males
1914
4 121 574
1.88 (1.72-2.05)
672
2 715 965
Females
1388
4 026 110
1.59 (1.44-1.75)
580
2 673 320
Males
2109
4 118 048
2.03 (1.87-2.22)
668
2 715 078
Females
1405
4 023 804
1.76 (1.59-1.95)
513
2 673 050
1.00 (ref)
Pulmonary embolism
1.00 (ref)
Deep vein thrombosis
1.00 (ref)
Other venous thromboembolic events
Males
180
4 131 143
1.51 (1.16-1.98)
78
2 718 800
Females
184
4 032 911
1.54 (1.18-2.01)
79
2 676 029
Males
5222
4 099 970
1.10 (1.06-1.15)
3136
2 700 461
Females
1254
4 026 826
1.07 (0.98-1.17)
773
2 671 947
Males
2921
4 116 321
1.04 (0.98-1.10)
1871
2 709 327
Females
1480
4 026 120
1.14 (1.05-1.24)
865
2 671 665
1.00 (ref)
Myocardial infarction
1.00 (ref)
Cerebrovascular stroke
1.00 (ref)
Peripheral vascular disease
Males
2641
4 119 122
1.08 (1.01-1.15)
1641
2 711 312
Females
1143
4 027 740
1.03 (0.94-1.13)
748
2 672 174
1.00 (ref)
Other arterial thrombosis
Males
410
4 129 472
1.45 (1.22-1.72)
187
2 717 878
Females
262
4 032 260
1.74 (1.38-2.20)
99
2 675 648
1.00 (ref)
Supplementary Table 3. Risk of thromboembolism and vascular disease in relation to blood group,
presented stratified by country.
Non-O blood group
Classification
Events
Person years
IRR (95%CI)
Blood group O
Events
Person years
IRR (95%CI)
All venous thromboembolism
Denmark
1574
2 500 194
1.71 (1.56-1.87)
666
1 821 343
Sweden
5136
5 621 878
1.82 (1.72-1.92)
1794
3 559 340
Denmark
1031
2 503 155
1.68 (1.50-1.88)
443
1 822 551
Sweden
2271
5 644 529
1.78 (1.64-1.93)
809
3 566 735
Denmark
457
2 505 908
1.90 (1.60-2.27)
174
1 823 652
Sweden
3057
5 635 944
1.92 (1.79-2.06)
1007
3 564 476
1.00 (ref)
Pulmonary embolism
1.00 (ref)
Deep vein thrombosis
1.00 (ref)
Other venous thromboembolic events
Denmark
115
2 507 828
1.35 (0.99-1.83)
62
1 824 177
Sweden
249
5 656 226
1.65 (1.30-2.09)
95
3 570 651
1.00 (ref)
Pregnancy/abortion-related venous thromboembolism
Denmark
71
1 208 073
2.36 (1.46-3.81)
22
881 014
Sweden
254
2 570 314
2.17 (1.68-2.81)
74
1 626 072
Denmark
2229
2 495 053
1.09 (1.02-1.16)
1469
1 815 598
Sweden
4247
5 631 744
1.10 (1.05-1.16)
2440
3 556 810
1.00 (ref)
Myocardial infarction
1.00 (ref)
Cerebrovascular stroke
Denmark
1583
2 500 823
1.08 (1.00-1.17)
1049
1 819 612
Sweden
2818
5 641 617
1.06 (1.00-1.13)
1687
3 561 380
1.00 (ref)
Peripheral vascular disease
Denmark
1776
2 498 572
1.10 (1.02-1.19)
1156
1 818 138
Sweden
2008
5 648 290
1.03 (0.96-1.10)
1233
3 565 348
1.00 (ref)
Other arterial thrombosis
Denmark
219
2 507 070
1.41 (1.12-1.77)
112
1 823 639
Sweden
453
5 654 661
1.64 (1.38-1.95)
174
3 569 887
1.00 (ref)
Supplementary Table 4. Risk of thromboembolism and vascular disease in relation to ABO blood group.
Blood group A
Classification
All venous
thromboembolism
Events
Personyears
4853
5 935 111
Pulmonary embolism
2370
5 954 048
Deep vein thrombosis
2531
5 949 312
Other venous
thromboembolic events
269
5 965 522
Pregnancy/abortionrelated VTE
254
2 744 542
Myocardial infarction
4787
5 937 886
Cerebrovascular stroke
3143
5 950 309
Peripheral vascular
disease
2727
5 953 268
478
5 963 978
Other arterial thrombosis
IRR (95%CI)
1.77
(1.69-1.86)
1.72
(1.61-1.84)
1.89
(1.76-2.02)
1.55
(1.28-1.89)
2.37
(1.87-2.99)
1.11
(1.07-1.16)
1.05
(1.00-1.11)
1.06
(1.00-1.12)
1.52
(1.31-1.76)
Blood group AB
Events
Personyears
658
686 547
317
689 147
361
688 544
32
690 905
27
325 192
549
687 629
402
688 817
329
689 302
64
690 691
IRR
(95%CI)
2.05
(1.88-2.24)
1.94
(1.72-2.20)
2.31
(2.06-2.60)
1.57
(1.07-2.29)
2.19
(1.43-3.36)
1.10
(1.00-1.20)
1.14
(1.02-1.26)
1.06
(0.94-1.19)
1.71
(1.31-2.25)
Blood group B
Events
Personyears
1199
1 500 414
615
1 504 489
622
1 503 996
63
1 507 627
44
708 654
1140
1 501 282
856
1 503 315
728
1 504 292
130
1 507 063
IRR
(95%CI)
1.73
(1.62-1.86)
1.75
(1.59-1.93)
1.86
(1.69-2.05)
1.43
(1.07-1.92)
1.63
(1.14-2.32)
1.03
(0.97-1.11)
1.11
(1.02-1.19)
1.09
(1.00-1.18)
1.62
(1.31-1.99)
Blood group O
Events
Personyears
2460
5 380 683
1252
5 389 285
1181
5 388 128
157
5 394 829
96
2 507 085
3909
5 372 408
2736
5 380 992
2389
5 383 485
286
5 393 526
IRR
(95%CI)
1.00
(ref)
1.00
(ref)
1.00
(ref)
1.00
(ref)
1.00
(ref)
1.00
(ref)
1.00
(ref)
1.00
(ref)
1.00
(ref)