Abstract
Introduction
Peripheral artery disease (PAD) affects 8.5 million Americans and thus improving our understanding of PAD is critical to developing strategies to reduce disease burden. The objective of the study was to determine the association of ABO blood type with ankle brachial index (ABI) as well as prevalent and incident PAD in a multi-ethnic cohort.
Methods
The Multi-Ethnic Study of Atherosclerosis includes non-Hispanic White, African, Hispanic, and Chinese Americans aged 45–84. ABO blood type was estimated using ABO genotypes in 6027 participants who had ABI assessed at the baseline exam. Associations with ABO blood type were evaluated categorically and under an additive genetic model by number of major ABO alleles. After excluding those with ABI>1.4, prevalent PAD was defined as ABI≤ 0.9 at baseline and incident PAD as ABI≤0.9 for 5137 participants eligible for analysis.
Results
There were 222 prevalent cases and 239 incident cases of PAD. In African Americans, each additional copy of the A allele was associated with a 0.02 lower baseline ABI (p=0.006). Each copy of the A allele also corresponded to 1.57-fold greater odds of prevalent PAD (95% CI, 1.17–2.35; p=0.004), but was not associated with incident PAD. No associations were found in other racial/ethnic groups for ABI, prevalent PAD, or incident PAD across all races/ethnicities.
Conclusions
Blood type A and the A allele count were significantly associated with baseline ABI and prevalent PAD in African Americans. Further research is needed to confirm and study the mechanisms of this association in African Americans.
Keywords: ABO blood group, ankle brachial index, atherosclerosis, peripheral arterial disease, epidemiology
1. Introduction
Peripheral artery disease (PAD) affects 8.5 million Americans [1]. Differences in the occurrence of PAD are well documented with the highest PAD prevalence in the elderly, African Americans, and women [1–4]. Furthermore, family history of PAD in a first degree relative is a strong risk factor for PAD independent of traditional risk factors [5]. Therefore, improving our understanding determinants of PAD, beyond traditional risk factors, is critical to developing strategies to reduce disease burden.
PAD typically involves the progressive reduction of blood flow to the lower extremities due to the development of atherosclerotic plaques. Atherosclerotic plaques form due to the accumulation of macrophages, cholesterol, and connective tissue in the blood vessels. Thrombus formation further impedes vascular flow [6]. Severity of PAD is associated with increased platelet activation [7]. Previous studies also reported associations between higher plasma levels of von Willebrand Factor (vWF), factor VIII, and the non-O blood groups [8–10]. Differing levels of these hemostatic factors across ABO blood groups may have an effect on the development or progression of PAD.
Several studies examined the associations between ABO blood type and PAD or intermittent claudication and reported that patients with different blood type had varying risks of lower-extremity atherosclerosis [11–17]. However, the associations of ABO blood type with prevalent PAD and incident PAD in non-white populations is not well studied. Likewise, little is known about the associations of ankle brachial index (ABI), the major clinical diagnostic test for PAD, with blood type. To overcome these knowledge gaps, we comprehensively examined the association of ABO blood type and ABO alleles with ABI, prevalent PAD, and incident PAD in the Multi-Ethnic Study of Atherosclerosis (MESA).
2. Methods
2.1. Study participants
The Multi-Ethnic Study of Atherosclerosis (MESA) cohort included 6814 men and women aged 45–84 years and free of clinical cardiovascular disease at enrollment. Participants were enrolled from six field sites in the United States including Los Angeles County, CA (UCLA); Chicago, IL (Northwestern University); Baltimore/Baltimore County, MD (Johns Hopkins University); St. Paul, MN (University of Minnesota); Forsyth County, NC (Wake Forest University); and Northern Manhattan/Bronx, NY (Columbia University). Methods have been previously published [18]. MESA includes 2622 (38%) non-Hispanic white, 1892 (28%) African, 1493 (22%) Hispanic, and 801 (12%) Chinese American subjects with five in-person exams occurring approximately every two years from 2000–2012. The current study includes 6027 participants with ABO blood type phenotype inferred from ABO gene variants that had ABI assessed at Exam 1 and includes data from exams 1, 3, and 5. The study was approved by the Institutional Review Boards at each research center and informed consent was obtained from all participants.
2.2 Ankle brachial index and peripheral artery disease
Methods for the measurement of ABI were described previously [19]. In brief, ABI was measured at exams 1, 3, and 5 and calculated for both the left and right sides as maximum systolic blood pressure in the posterior tibial artery and dorsalis pedis, divided by the average of the left and right brachial pressures. In the event that left and right brachial pressures differed by 10 mmHg or more, the higher of the brachial pressures was used. If a pulse was detected when the cuff was inflated to 300 mmHg, the ABI was classified as “incompressible.” All participants with an ABI > 1.4 (i.e., rigid arteries) at Exam 1 were excluded from further analysis (n = 43). Prevalent PAD was defined by Exam 1 ABI ≤ 0.9, while incident PAD was defined as ABI ≤ 0.9 at Exam 3 or 5 in individuals without PAD at baseline. Participants with incomplete data (i.e., no ABI measured at Exam 3 or 5) or participants with Exam 3 ABI > 1.4 were excluded from analysis.
2.3. Traditional peripheral arterial disease risk factors
Interview and questionnaire data were used for information on age, sex, race/ethnicity, and smoking status. Height was measured while participants were standing without shoes. Body mass index (BMI) was calculated as weight in kilograms per height in meters squared (kg/m). Blood pressure was measured three times using the Dinamap automated blood pressure device and determined by averaging the last two measurements. Triglycerides were measured in plasma by a glycerol blanked enzymatic method. Low-density lipoprotein (LDL) cholesterol was calculated in those with triglycerides <400 mg/dL using the Friedewald formula. High density lipoprotein (HDL) cholesterol was measured using the cholesterol oxidase method after precipitation of non-HDL cholesterol with magnesium/dextran. Diabetes was defined as use of diabetes medication or insulin, or a fasting glucose ≥126 mg/dL. Hypertension was defined as use of anti-hypertensive medications, systolic blood pressure ≥140 mmHg, or diastolic blood pressure ≥90 mmHg.
2.4. Genetic data
ABO blood type was inferred [20] based upon 28 ABO single nucleotide polymorphisms in 6255 MESA participants with available genetic data. Genotype data from the Illumina Exome BeadChip [21], the Illumina Cardio-MetaboChip [22], the Illumina iSelect ITMAT/Broad/CARe (IBC) Chip [23], as well as a custom Sequenom (San Diego, CA) panel, was aggregated under genome build hg19 and haplotype methods were used to estimate participant ABO diplotypes and corresponding allele frequencies, stratified by race/ethnicity. These diplotypes were then matched to genetic alleles from the Blood Group Antigen Gene Mutation Database (BGMUT), an online curated genetic database of ABO alleles [24,25]. Participants were declared to have reliable typing based upon diplotype posterior probability >0.90, resulting in 6202/6255 subjects with successful ABO typing. Each participant was assigned ABO genotypes of major alleles (A, B, O) as well as ABO blood type (A, B, AB, O). Population stratification was assessed using EIGENSTRAT [26] for participants with genome-wide SNP data, and first three ancestry-informative principal components (PCs) were considered for covariate adjustment.
2.5. Statistical analysis
To accommodate the truncated nature of the ABI measurements, truncated regression using the crch R package [27] was applied to evaluate the association between ABO blood group, ABO allele number, and Exam 1 ABI, while associations with prevalent PAD were assessed using logistic regression. Associations with incident PAD were evaluated using Cox proportional hazards regression. Event times for incident PAD were defined as the midpoint between the first exam at which ABI ≤ 0.9 and the preceding exam, and time-to-event analyses excluded all subjects with prevalent PAD (n = 222). Participants with an Exam 3 ABI and no Exam 5 ABI were considered right censored at Exam 3. A total of 92 participants had ABI measurements at Exam 5 but not at Exam 3. Of these, the 4 participants with ABI≤0.9 at Exam 5 but no ABI at Exam 3 were considered to have an event at the midpoint between Exams 1 and 5, while the remaining were assumed to be free of PAD. The assumption of proportional hazards was evaluated for all Cox regression models using a chi-square test on the Schoenfeld residuals. ABO blood type was modeled as a categorical variable with blood type O as the reference (i.e., A, B, AB, O (reference)) and ABO alleles were assessed using an additive genetic model by quantity of each major allele type (i.e., A, B, and O (reference)). For example, a subject who had AB genotype would correspond to covariate values of A = 1 and B = 1. Analyses were stratified by race/ethnicity and adjusted for age, sex, and the first three ancestry-informative PCs. Associations were declared statistically significant with a Bonferroni adjusted p-value threshold of p < 0.05/4 = 0.01, taking into account the number of race/ethnicity strata. Sensitivity analyses were conducted that additionally adjusted for traditional cardiovascular risk factors including smoking, diabetes, and hypertension status, BMI, total cholesterol, HDL cholesterol, and systolic blood pressure. Interactions of race/ethnicity with ABO were evaluated by including the full cohort in the regression analysis adjusting for race/ethnicity and testing the significance of race/ethnicity-ABO interaction terms based upon a likelihood ratio test. Interactions by sex were conducted similarly within all race/ethnicity strata. All analyses were completed in the statistical software R 3.0.1.
3. Results
A total of 6027 MESA participants with ABO blood type inferred from ABO genotypes and had ABI assessed at Exam 1 were included in the study. Baseline characteristics by race/ethnicity are shown in Table 1 including the race/ethnicity specific distribution of blood groups. At Exam 1, 222 participants had prevalent PAD, with African Americans comprising 45% of cases (Table 2). For the 5805 participants without prevalent PAD, 5137 were eligible for the incident PAD analysis and 239 incident cases were identified during follow-up.
Table 1.
MESA Exam 1 Characteristics by race/ethnicity, mean (standard deviation) or N (percentage).
| Characteristics | African American | Chinese American | Hispanic American | Non-Hispanic White American |
|---|---|---|---|---|
| N | 1492 | 741 | 1382 | 2412 |
| Age, years | 61 (10) | 62 (10) | 61 (10) | 62 (10) |
| Sex (female) | 808 (54) | 379 (51) | 721 (52) | 1254 (52) |
| Body mass index, kg/m2 | 30 (6) | 24 (3) | 29 (5) | 27 (5) |
| Hypertension (Yes) | 884 (59) | 280 (38) | 575 (42) | 930 (39) |
| Diabetes mellitus (Yes) | 259 (17) | 99 (13) | 238 (17) | 259 (17) |
| Total cholesterol, mg/dL | 190 (36) | 193 (31) | 199 (38) | 196 (35) |
| HDL cholesterol, mg/dL | 52 (15) | 49 (12) | 48 (13) | 52 (16) |
| LDL cholesterol, mg/dL | 117 (33) | 115 (29) | 120 (33) | 117 (30) |
| Triglycerides, mg/dL | 105 (71) | 142 (82) | 158 (102) | 133 (90) |
| Smoking status | ||||
| Never | 674 (45) | 555 (75) | 758 (55) | 1061 (44) |
| Former | 542 (36) | 145 (20) | 442 (32) | 1070 (44) |
| Current | 272 (18) | 41 (6) | 182 (13) | 280 (12) |
| ABO blood type | ||||
| Type A | 400 (27) | 192 (26) | 420 (30) | 1033 (43) |
| Type B | 265 (18) | 199 (27) | 137 (10) | 245 (10) |
| Type AB | 77 (5) | 65 (9) | 33 (2) | 97 (4) |
| Type O | 750 (50) | 285 (38) | 792 (57) | 1037 (43) |
HDL: high-density lipoprotein, LDL: low-density lipoprotein.
Table 2.
MESA outcomes: Mean (standard deviation) or N (percentage).
| Outcome | Total MESA Cohort | African American | Chinese American | Hispanic American | Non-Hispanic White American |
|---|---|---|---|---|---|
| Exam 1, n | 6027 | 1492 | 741 | 1382 | 2412 |
| ABI | 1.11 (0.11) | 1.07 (0.13) | 1.11 (0.09) | 1.13 (0.11) | 1.12 (0.11) |
| PAD (yes) | 222 (3.7) | 101 (6.8) | 14 (1.9) | 31 (2.2) | 76 (3.2) |
| Exams 3 and 5, na | 5137 | 1216 | 640 | 1165 | 2116 |
| Person years | 41901 | 9576 | 5318 | 9447 | 17560 |
| Incident PAD | 239 | 95 | 16 | 46 | 82 |
| Rate per 1000 person-years | 5.7 | 9.9 | 3.0 | 4.9 | 4.7 |
ABI: ankle brachial index, PAD: peripheral arterial disease.
Includes all subjects with Exam 3 and/or Exam 5 ABI measurements and free of PAD at Exam 1.
Table 3 summarizes the associations between ABO blood type and Exam 1 ABI by race/ethnicity. Compared to those with blood type O, blood type A was associated with lower ABI in African Americans (β = −0.019, p = 0.014). Association analysis by ABO allele count resulted in similar findings in African-Americans, with each additional A allele corresponding to 0.017 lower ABI (p = 0.006) and least square means by allele dosage commensurate with an additive model. No significant associations were identified across any of the other race/ethnicities by either blood type or blood group allele count.
Table 3.
| Pooled Race/Ethnicityd | African American | Chinese American | Hispanic American | Non-Hispanic White American | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| ABO Blood Type | Beta (SE) | p-value | Interaction p-value | Beta (SE) | p-value | Beta (SE) | p-value | Beta (SE) | p-value | Beta (SE) | p-value |
| O (Reference) | – | – | 0.22 | – | – | – | – | – | – | – | – |
| A | −0.002 (0.003) | 0.50 | – | −0.019 (0.008) | 0.014 | −0.007 (0.008) | 0.41 | 0.005 (0.006) | 0.43 | 0.003 (0.005) | 0.51 |
| B | 0.000 (0.004) | 0.99 | – | −0.004 (0.009) | 0.65 | 0.002 (0.008) | 0.76 | 0.007 (0.010) | 0.47 | 0.000 (0.008) | 0.96 |
| AB | −0.005 (0.007) | 0.43 | – | −0.006 (0.015) | 0.72 | 0.010 (0.012) | 0.40 | −0.020 (0.019) | 0.29 | −0.013 (0.012) | 0.28 |
| ABO Allele Count | Beta (SE) | p-value | Interaction p-value | Beta (SE) | p-value | Beta (SE) | p-value | Beta (SE) | p-value | Beta (SE) | p-value |
| O (Reference) | – | – | 0.18 | – | – | – | – | – | – | – | – |
| A | −0.003 (0.003) | 0.21 | – | −0.017 (0.006) | 0.006 | −0.003 (0.006) | 0.58 | 0.004 (0.005) | 0.44 | 0.000 (0.004) | 0.98 |
| B | 0.001 (0.003) | 0.73 | – | −0.001 (0.007) | 0.93 | 0.003 (0.006) | 0.58 | 0.000 (0.008) | 0.97 | −0.005 (0.006) | 0.42 |
ABI: ankle brachial index, SE: standard error.
ABI at Exam 1 excluding ABI > 1.4.
ABI = ankle brachial index; SE = standard error
Adjusted for age, sex, and leading ancestry informative PCs.
Pooled analyses adjusted for race/ethnicity using PCs; Interaction tests based on self-reported race/ethnicity.
Associations of ABO and prevalent PAD are presented in Table 4. We observed an increased odds of prevalent PAD in blood Type A for African (OR = 1.78; p = 0.016) and Chinese (OR = 5.31, p = 0.05) Americans. When modeling A allele dosage, increased odds of prevalent PAD for each additional A allele present was observed in African American (OR = 1.57; p = 0.004) and Chinese American (OR = 1.98; p = 0.12). In pooled analyses, each additional A allele increased the odds of prevalent PAD (OR = 1.24, p = 0.065) but the association appears to be driven by African Americans (race/ethnicity interaction p = 0.064). Furthermore, no significant interactions were observed within African Americans by sex (p = 0.17). In contrast, only in Hispanic Americans did we observe an increased risk of incident PAD in those with AB blood type compared to O blood type (HR = 4.97; p = 0.01) and ABO blood type and allele count were not significantly associated with incident PAD in the race/ethnicity pooled analyses (Table 5).
Table 4.
Associations between ABO Blood Group and ABO allele counts and prevalent Peripheral Artery Disease by race/ethnicity.a,b,c
| Pooled Race/Ethnicityd | African American | Chinese American | Hispanic American | Non-Hispanic White American | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| ABO Blood Type | OR (95% CI) | p-value | Interaction P-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value |
| O (Reference) | – | – | 0.079 | – | – | – | – | – | – | – | – |
| A | 1.17 (0.86,1.58) | 0.32 | – | 1.78 (1.11,2.84) | 0.016 | 5.31 (1.16,38.7) | 0.051 | 0.56 (0.20,1.35) | 0.23 | 0.80 (0.48,1.32) | 0.39 |
| B | 0.67 (0.40,1.06) | 0.97 | – | 0.62 (0.28,1.23) | 0.20 | 2.90 (0.51,22.8) | 0.25 | 0.22 (0.12,1.11) | 0.15 | 0.83 (0.33,1.79) | 0.65 |
| AB | 1.36 (0.69,2.47) | 0.33 | – | 1.48 (0.49,3.70) | 0.44 | 1.54 (0.07,18.5) | 0.74 | 0.85 (0.42,5.26) | 0.89 | 1.59 (0.53,3.87) | 0.35 |
| ABO Allele Count | OR (95% CI) | p-value | Interaction p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value | OR (95% CI) | p-value |
| O (Reference) | – | – | 0.064 | – | – | – | – | – | – | – | – |
| A | 1.24 (0.98,1.57) | 0.065 | – | 1.57 (1.17,2.35) | 0.004 | 1.98 (0.83,4.70) | 0.12 | 0.63 (0.25,1.35) | 0.27 | 0.99 (0.66,1.43) | 0.97 |
| B | 0.81 (0.56,1.14) | 0.24 | – | 0.71 (0.40,1.17) | 0.20 | 1.06 (0.30,3.28) | 0.92 | 0.39 (0.61,1.31) | 0.20 | 1.15 (0.61,1.98) | 0.65 |
CI: confidence interval, OR: odds ratio.
Prevalent PAD defined as ABI ≤ 0.9, excluding ABI > 1.4.
CI = confidence interval; OR = odds ratio.
Adjusted for age, sex, and leading ancestry informative PCs.
Pooled analyses adjusted for race/ethnicity using PCs; Interaction tests based on self-reported race/ethnicity.
Table 5.
Associations between ABO Blood Group and ABO allele counts and incident peripheral artery disease by race/ethnicity.a,b,c
| Pooled Race/Ethnicityd | African American | Chinese American | Hispanic American | Non-Hispanic White American | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| ABO Blood Type | HR (95% CI) | p-value | Interaction p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value |
| O (Reference) | – | – | 0.30 | – | – | – | – | – | – | – | – |
| A | 1.05 (0.78,1.41) | 0.74 | – | 1.39 (0.86,2.26) | 0.18 | 0.60 (0.15,2.38) | 0.46 | 1.45 (0.74,2.85) | 0.28 | 0.73 (0.46,1.17) | 0.19 |
| B | 1.20 (0.84,1.73) | 0.32 | – | 1.39 (0.81,2.37) | 0.23 | 0.68 (0.20,3.66) | 0.55 | 1.63 (0.59,3.85) | 0.27 | 0.90 (0.44,1.87) | 0.79 |
| AB | 1.08 (0.57,2.08) | 0.81 | – | 1.24 (0.49,3.13) | 0.65 | 0.45 (0.20,2.34) | 0.45 | 4.97 (1.47,16.8) | 0.010 | 0.25 (0.03,1.86) | 0.18 |
|
| |||||||||||
| ABO Allele Count | HR (95% CI) | p-value | Interaction p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value | HR (95% CI) | p-value |
| O (Reference) | – | – | 0.44 | – | – | – | – | – | – | – | |
| A | 1.02 (0.81,1.28) | 0.86 | – | 1.16 (0.80,1.68) | 0.43 | 0.77 (0.29,2.08) | 0.61 | 1.38 (0.80,2.40) | 0.25 | 0.81 (0.56,1.17) | 0.27 |
| B | 1.16 (0.87,1.53) | 0.31 | – | 1.20 (0.81,1.78) | 0.36 | 0.67 (0.24,1.87) | 0.45 | 1.93 (1.01,3.67) | 0.045 | 0.84 (0.45,1.56) | 0.59 |
CI, confidence interval; HR, hazard ratio.
Incident PAD is defined as ABI ≤ 0.9 at Exam 3 or Exam 5, excluding ABI > 1.4 and prevalent PAD cases.
CI = confidence interval; HR = hazard ratio.
Adjusted for age, sex, and leading ancestry informative PCs.
Pooled analyses adjusted for race/ethnicity using PCs; Interaction tests based on self-reported race/ethicity.
To evaluate potential differences of association between the two major A alleles (A1 and A2) and ABI, we reparameterized the A allele count into two separate variables corresponding to the respective counts of each allele. Parameter estimates for association with ABI in African Americans were comparable in magnitude and direction (βA1 = −0.016;βA2 = −0.015). No new significant associations were identified in the other race/ethnicities using these A subtypes. Additionally, adjustment for traditional risk factors did not meaningfully change the ABO association results with ABI or PAD (Supplementary Tables S1–S3).
4. Discussion
The primary goal of the current study was to investigate whether ABO blood type is associated with ABI as well as prevalent and incident PAD in a multi-ethnic cohort. Our findings suggest that blood group A and A allele count are associated with lower ABI and a 1.57 greater odds of PAD prevalence among African Americans independent of traditional PAD risk factors. In contrast, ABO blood type and ABO allele count were not significantly associated with ABI and prevalent PAD among Chinese, Hispanic, or non-Hispanic white Americans. Similarly, ABO blood type and ABO allele count were not associated with incident PAD in race/ethnicity pooled analyses.
While the relationship between ABO blood types and PAD risk has been investigated, the role of race/ethnicity remains elusive, as most previous studies have primarily focused on non-Hispanic white populations. The association between non-O blood type and PAD/intermittent claudication was first noted in the 1970’s by Hall and later Kingsbury [11,15]. Subsequent to these case-control studies, the Framingham Heart Study prospectively investigated the association of ABO blood types and the development of intermittent claudication over a 10-year follow-up period and observed an increased risk in those with non-O blood types in both men and women [12]. These previous studies were limited by evaluating only subjects of European ancestry with symptomatic intermittent claudication, whereas the current study included four racial/ethnic groups exclusively with ABI measurements. In contrast, we did not observe ABO blood type associations with ABI, prevalent PAD, or incident PAD in non-Hispanic whites. Furthermore, by leveraging MESA’s multi-ethnic cohort, we demonstrate that significant differences in ABI and PAD prevalence rate for blood group A compared to blood type O are exclusive to the African American population representing an independent risk factor in this group.
While there is no conclusive explanation on how ABO blood type influences PAD risk or ABI, there is evidence that the effect may be mediated by ABO antigenic determinants by blood type expressed on carbohydrate structures of vWF that are thought to influence the clearance rate of vWF [28]. Thus, the survival of vWF is longer in those with non-O blood types and circulating levels of vWF and Factor VIII release bioactive molecules under stress which have been shown to participate in the development of atherosclerosis and PAD related to endothelial dysfunction and inflammatory response [29,30]. In addition, African Americans have demonstrated higher levels of vWF and Factor VIII despite higher proportions of O blood type when compared to non-Hispanic whites and a higher prevalence of PAD compared to other race/ethnicities [3,4,31,32]. Therefore, we explored potential mediation by Factor VIII in African Americans. We found that Factor VIII is significantly associated with both A and B allele dosage, respectively corresponding to estimates of 14% (p < 0.001) and 34% (p < 0.001) increases in Factor VIII per additional allele carried. However, adjustment for Factor VIII did not attenuate the association of blood group A and lower ABI and prevalent PAD in African Americans. Likewise, prior work has shown that traditional and novel risk factors such as interleukin-6, fibrinogen, D-dimer, and homocysteine did not account for the race/ethnic specific associations with PAD [33]. Other possible explanations may include involvement of other blood antigen phenotypes that are highly represented in subjects of African descent. For example, the Duffy Fy(a-b-) null phenotype results in the lack of Duffy antigen and is present in ~70% of African Americans, but very rare in subjects of European descent.
We demonstrated that ABO blood type was not associated with incident PAD in any of the African, Chinese, and non-Hispanic white Americans. Similar to our findings, Folsom et al., observed variations in PCSK9 that were associated with prevalent but not incident PAD among African Americans in the Atherosclerosis Risk in Communities Study [34]. This observation may be, in part, explained by the definition of PAD that was strictly based on ABI measurements. Therefore, we conducted sensitivity analyses that included the nine additional definite and probable cases of PAD adjudicated to date to the ABI defined cases. The association results were unchanged. We also examined whether differences in the risk factor profile of prevalent verses incident subjects contributed to these findings. The supplementary Table S4 provides the comparison between groups for all major PAD risk factors that demonstrates similar risk profiles. We examined bias due to missing information and observed that after adjusting for age there was a higher chance of participant dropout at exam 3 if baseline ABI measurement between 0.9 and 0.95 (p = 0.0002). This finding suggests bias may be contributing to the differences in the association of ABO blood type with prevalent and incident PAD. Finally, despite similar number of prevalent and incident events (n = 101 vs n = 95), limited power could be an issue. Only in Hispanic Americans did we observe an increased risk of incident PAD in those with AB blood type compared to O blood type (HR = 4.97; 95% CI, 1.47–16.80, p = 0.010). This result should be interpreted cautiously given that of the 46 Hispanic Americans had AB blood type, 3 participants had incident PAD.
Strengths of our study include the investigation of a large multi-ethnic cohort of both men and women from four race/ethnicity groups. ABI measurement was available at 3 times over 10 years using standard methods and the MESA cohort was followed for PAD events adjudicated using standard protocols [18]. Some limitations need to be acknowledged; first, for each subsequent exam, some participants were lost to follow-up. Prior studies of blood type and PAD/intermittent claudication focused in part or whole on symptomatic patients whereas the current study used exclusively ABI measurements for the main analyses. Furthermore, PAD is rare in Chinese and Hispanic Americans and thus power may limit our ability to detect associations in stratified analyses. Although vWF was measured in a subset of 1000 MESA participants at baseline, only 173 African Americans had available ABI, ABO, and vWF data, and thus mediation analysis could not be fully explored. In general the ABO glycosyltransferase only produces A antigens, B antigens, or no antigens (blood type O), though rare ambiguous variants exist. Finally, bias arising from the fact that MESA subjects were free of CVD at baseline may have impacted our ability to detect associations.
5. Conclusions
This study showed that the ABO blood groups are associated with ABI and prevalent PAD in African Americans. Further research is needed to explore the biological mechanisms that mediate these associations and investigate reasons for differences in the association of prevalent and incident PAD.
Supplementary Material
Highlights.
Risk of lower-extremity atherosclerosis differs by blood type.
Type A was associated lower ankle brachial index in African Americans.
Each copy of the A allele corresponded to 1.57-fold greater odds of PAD.
Acknowledgments
This work was supported by contracts N01-HC-95159, N01-HC-95160, N01-HC-95161, N01-HC-95162, N01-HC-95163, N01-HC-95164, N01-HC-95165, N01-HC-95166, N01-HC-95167, N01-HC-95168, N01-HC-95169, and R01HL98077 from the National Heart, Lung, and Blood Institute and by grants UL1-TR-000040 and UL1-TR-001079 from NCRR. The authors thank the other investigators, the staff, and the participants of the MESA study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org.
Role of the Funding Source
The funding source had no involvement in preparation of this manuscript.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Author contributions
Study conception and design: M. Pike, S.J. Bielinski, N. B. Larson, P.A. Decker, C.L. Wassel.
Data Collection: N.Q. Hanson, M.Y. Tsai, S.J. Bielinski, N. B. Larson.
Data analysis: N.B. Larson, P.A. Decker.
First draft: M. Pike, S.J. Bielinski, N. B. Larson, P.A. Decker, K.P. Cohoon Critical revision of the manuscript for intellectual content: M. Pike, N. B. Larson, C.L. Wassel, K.P. Cohoon, M.Y. Tsai, J.S. Pankow, N.Q. Hanson, P.A. Decker, C. Berardi, K.S. Alexander, M. Cushman, N.A. Zakai, S.J. Bielinski.
Final approval or submission: M. Pike, N. B. Larson, C.L. Wassel, K.P. Cohoon, M.Y. Tsai, J.S. Pankow, N.Q. Hanson, P.A. Decker, C. Berardi, K.S. Alexander, M. Cushman, N.A. Zakai, S.J. Bielinski.
Conflicts of Interest
None
References
- 1.Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke statistics–2014 update: a report from the American Heart Association. Circulation. 2014;129(3):e28–e292. doi: 10.1161/01.cir.0000441139.02102.80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Fowkes FG, Rudan D, Rudan I, et al. Comparison of global estimates of prevalence and risk factors for peripheral artery disease in 2000 and 2010: a systematic review and analysis. Lancet. 2013;382(9901):1329–1340. doi: 10.1016/S0140-6736(13)61249-0. [DOI] [PubMed] [Google Scholar]
- 3.Criqui MH, Vargas V, Denenberg JO, et al. Ethnicity and peripheral arterial disease: the San Diego Population Study. Circulation. 2005;112(17):2703–2707. doi: 10.1161/CIRCULATIONAHA.105.546507. [DOI] [PubMed] [Google Scholar]
- 4.Allison MA, Ho E, Denenberg JO, et al. Ethnic-specific prevalence of peripheral arterial disease in the United States. Am J Prev Med. 2007;32(4):328–333. doi: 10.1016/j.amepre.2006.12.010. [DOI] [PubMed] [Google Scholar]
- 5.Wassel CL, Loomba R, Ix JH, Allison MA, Denenberg JO, Criqui MH. Family history of peripheral artery disease is associated with prevalence and severity of peripheral artery disease: the San Diego population study. J Am Coll Cardiol. 2011;58(13):1386–1392. doi: 10.1016/j.jacc.2011.06.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Singh RB, Mengi SA, Xu YJ, Arneja AS, Dhalla NS. Pathogenesis of atherosclerosis: A multifactorial process. Exp Clin Cardiol. 2002;7(1):40–53. [PMC free article] [PubMed] [Google Scholar]
- 7.Rajagopalan S, McKay I, Ford I, Bachoo P, Greaves M, Brittenden J. Platelet activation increases with the severity of peripheral arterial disease: implications for clinical management. J Vasc Surg. 2007;46(3):485–490. doi: 10.1016/j.jvs.2007.05.039. [DOI] [PubMed] [Google Scholar]
- 8.Tirado I, Mateo J, Soria JM, et al. The ABO blood group genotype and factor VIII levels as independent risk factors for venous thromboembolism. Thromb Haemost. 2005;93(3):468–474. doi: 10.1160/TH04-04-0251. [DOI] [PubMed] [Google Scholar]
- 9.Souto JC, Almasy L, Muniz-Diaz E, et al. 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(8):2024–2028. doi: 10.1161/01.atv.20.8.2024. [DOI] [PubMed] [Google Scholar]
- 10.Song J, Chen F, Campos M, et al. Quantitative influence of ABO blood groups on Factor VIII and its ratio to von Willebrand factor, novel observations from an ARIC Study of 11,673 Subjects. PLoS One. 2015;10(8):e0132626. doi: 10.1371/journal.pone.0132626. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hall R, Bunch GA, Humphrey CS. The frequencies of ABO blood groups and of secretors of ABH group substances in peripheral arteriosclerosis. Atherosclerosis. 1971;14(2):241–246. doi: 10.1016/0021-9150(71)90053-0. [DOI] [PubMed] [Google Scholar]
- 12.Garrison RJ, Havlik RJ, Harris RB, Feinleib M, Kannel WB, Padgett SJ. ABO blood group and cardiovacular disease: the Framingham study. Atherosclerosis. 1976;25(2–3):311–318. doi: 10.1016/0021-9150(76)90036-8. [DOI] [PubMed] [Google Scholar]
- 13.Horby J, Gyrtrup HJ, Grande P, Vestergaard A. Relation of serum lipoproteins and lipids to the ABO blood groups in patients with intermittent claudication. J Cardiovasc Surg (Torino) 1989;30(4):533–537. [PubMed] [Google Scholar]
- 14.Cronenwett JL, Davis JT, Jr, Garrett HE. ABO blood group and serum lipids in female atherosclerosis. J Cardiovasc Surg (Torino) 1983;24(6):658–661. [PubMed] [Google Scholar]
- 15.Kingsbury KJ. Relation of ABO blood-groups to atherosclerosis. Lancet. 1971;1(7692):199–203. doi: 10.1016/s0140-6736(71)90945-7. [DOI] [PubMed] [Google Scholar]
- 16.Norrgard O, Beckman G, Cedergren B. HLA antigens, blood groups and serum protein groups in patients with intermittent claudication. Hum Hered. 1989;39(4):192–195. doi: 10.1159/000153859. [DOI] [PubMed] [Google Scholar]
- 17.Sabino Ade P, Ribeiro DD, Domingheti CP, et al. ABO blood group polymorphisms and risk for ischemic stroke and peripheral arterial disease. Mol Biol Rep. 2014;41(3):1771–1777. doi: 10.1007/s11033-014-3026-8. [DOI] [PubMed] [Google Scholar]
- 18.Bild DE, Bluemke DA, Burke GL, et al. Multi-Ethnic Study of Atherosclerosis: objectives and design. Am J Epidemiol. 2002;156(9):871–881. doi: 10.1093/aje/kwf113. [DOI] [PubMed] [Google Scholar]
- 19.Wassel CL, Berardi C, Pankow JS, et al. Soluble P-selectin predicts lower extremity peripheral artery disease incidence and change in the ankle brachial index: the Multi-Ethnic Study of Atherosclerosis (MESA) Atherosclerosis. 2015;239(2):405–411. doi: 10.1016/j.atherosclerosis.2015.01.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Larson NB, Bell EJ, Decker PA, et al. ABO blood group associations with markers of endothelial dysfunction in the Multi-Ethnic Study of Atherosclerosis. Atherosclerosis. 2016;251:422–429. doi: 10.1016/j.atherosclerosis.2016.05.049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Huyghe JR, Jackson AU, Fogarty MP, et al. Exome array analysis identifies new loci and low-frequency variants influencing insulin processing and secretion. Nat Genet. 2013;45(2):197–201. doi: 10.1038/ng.2507. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Voight BF, Kang HM, Ding J, et al. The metabochip, a custom genotyping array for genetic studies of metabolic, cardiovascular, and anthropometric traits. PLoS Genet. 2012;8(8):e1002793. doi: 10.1371/journal.pgen.1002793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Keating BJ, Tischfield S, Murray SS, et al. Concept, design and implementation of a cardiovascular gene-centric 50 k SNP array for large-scale genomic association studies. PLoS One. 2008;3(10):e3583. doi: 10.1371/journal.pone.0003583. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Patnaik SK, Helmberg W, Blumenfeld OO. BGMUT database of allelic variants of genes encoding human blood group antigens. Transfus Med Hemother. 2014;41(5):346–351. doi: 10.1159/000366108. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Patnaik SK, Helmberg W, Blumenfeld OO. BGMUT: NCBI dbRBC database of allelic variations of genes encoding antigens of blood group systems. Nucleic Acids Res. 2012;40:D1023–1029. doi: 10.1093/nar/gkr958. Database issue. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Price AL, Patterson NJ, Plenge RM, Weinblatt ME, Shadick NA, Reich D. Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet. 2006;38(8):904–909. doi: 10.1038/ng1847. [DOI] [PubMed] [Google Scholar]
- 27.Messner JW, Zeileis A, Broecker J, Mayr GJ. Probabilistic wind power forecasts with an inverse power curve transformation and censored regression. Wind Energy. 2014;17(11):1753–1766. [Google Scholar]
- 28.Gallinaro L, Cattini MG, Sztukowska M, et al. A shorter von Willebrand factor survival in O blood group subjects explains how ABO determinants influence plasma von Willebrand factor. Blood. 2008;111(7):3540–3545. doi: 10.1182/blood-2007-11-122945. [DOI] [PubMed] [Google Scholar]
- 29.Heinen Y, Stegemann E, Sansone R, et al. Local association between endothelial dysfunction and intimal hyperplasia: relevance in peripheral artery disease. J Am Heart Assoc. 2015;4(2) doi: 10.1161/JAHA.114.001472. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Reriani MK, Lerman LO, Lerman A. Endothelial function as a functional expression of cardiovascular risk factors. Biomark Med. 2010;4(3):351–360. doi: 10.2217/bmm.10.61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.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(8):1717–1723. doi: 10.1002/cncr.20150. [DOI] [PubMed] [Google Scholar]
- 32.Folsom AR, Cushman M, Tsai MY, et al. A prospective study of venous thromboembolism in relation to factor V Leiden and related factors. Blood. 2002;99(8):2720–2725. doi: 10.1182/blood.v99.8.2720. [DOI] [PubMed] [Google Scholar]
- 33.Allison MA, Criqui MH, McClelland RL, et al. The effect of novel cardiovascular risk factors on the ethnic-specific odds for peripheral arterial disease in the Multi-Ethnic Study of Atherosclerosis (MESA) J Am Coll Cardiol. 2006;48(6):1190–1197. doi: 10.1016/j.jacc.2006.05.049. [DOI] [PubMed] [Google Scholar]
- 34.Folsom AR, Peacock JM, Boerwinkle E. Variation in PCSK9, low LDL cholesterol, and risk of peripheral arterial disease. Atherosclerosis. 2009;202(1):211–215. doi: 10.1016/j.atherosclerosis.2008.03.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.