ABO Blood Type and the Risk of Cardovascular Events: An Analysis of Framingham Data and Meta-Analysis

Objective An association between ABO blood group and cardiovascular risk has been postulated and explored in the literature but results have been inconsistent. A possible mechanism for variability in ischemic risk by blood group was elucidated in the late 1980s and early 1990s when evidence emerged that levels of von Willebrand factor varied depending upon blood type. The lowest levels were observed in people with two O genes, the highest level in people who were genotypically AA, BB, or AB, and intermediate levels with AO and BO. Additionally a study which quantified plasma levels of factor VIII-VWF showed they are 25% higher in non-O individuals than in those with type O. Based upon this preliminary supportive data we sought to to examine this relationship using Framingham data with the full cohort and more than 30 years of follow up.

Methods and Results A large prospective cohort study consisting of 4018 subjects without prior cardiovascular events was conducted with 38 years of follow up. All data collected in the Framingham Heart Study cohort were reviewed for use as potential confounders. We used the log-rank test for equality of survivor functions to assess the unadjusted association between ABO blood group and MI, stroke, and the combined endpoint of non-fatal MI or death associated with CHD (coronary heart disease). We compared blood group O to non-O (combining the A,B, and AB groups) and to (A or B) and AB individually with the goal of comparing those patients that should be least likely to clot in a VWF-dependent manner to those with intermediate and high risk. We designated A and B as intermediate risk because genotyping was unavailable and therefore phenotypic blood groups A and B included a mixture of AO, AA, BO, and BB which should place them between O and AB individuals in terms of VWF levels and risk of cardiovascular events. The reported hazard ratios describe the risk of events in the non-O blood groups compared to O. We used a Cox regression model to adjust for potential confounders. In analyses adjusted for cardiovascular risk factors and comparing participants with blood group O to those with A, B, or AB, there was no statistically significant relationship found between blood group and any cardiovascular outcome. Comparing the most disparate groups, non-O vs O found the correlation to be MI (HR =1.10 with confidence interval 0.96 to 1.25), stroke(1.03, confidence interval 0.89 to 1.19) , or the combined end point of non-fatal MI and death (HR of 1.04 with confidence interval of 0.93 -1.16).

A meta-analysis of 5 prior studies plus this cohort demonstrated that non-O blood types are associated with an increased risk of cardiovascular events [OR 1.12 (1.06 to 1.18).

Conclusion Our analysis yielded a non-significant effect but when included in a meta-analysis demonstrated 12% increased risk in cardiovascular events for people with non-0 compared with O blood types. ABO blood group could be considered when evaluating risk factors for primary prevention for cardiovascular disease.