Public health applications of cardiovascular genomics

• Main Author: Holmes, M. V.
• Other Authors: Casas, J. P. ; Hingorani, A. D. ; Kivimäki, M.
• Published: University College London (University of London) 2014
• Subjects: 614.4
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.626727

Background Genetic epidemiology is at the interface of translational and basic research and the pace of progress has been unprecedented, with findings representing some of the most robust available in the Scientific literature. However, how we can translate this high- fidelity genomic information into improvements in health of the population? Two dis¬tinct translational opportunities include personalized medicine (pharmacogenetics) and using Mendelian randomization to investigate disease aetiology to inform public health policy and develop new therapies. Objectives In this PhD thesis, I investigated the evidence base underlying the well-publicized use of the pharmacogenetic biomarker CYP2C19 genotype to predict the response to clopi¬dogrel, a widely prescribed antiplatelet drug. Second, I used Mendelian randomization to investigate the role of an endogenous biomarker, secretory phospholipase A2-IIA (sPLA2-IIA), thought to be a pro-atherogenic enzyme and a potential drug target for the prevention of cardiovascular disease (CVD). Third, I used Mendelian randomization to investigate the relationship between alcohol, an exogenous exposure, and cardiovas¬cular traits and disease events. Results CYP2C19 and cardiovascular disease I identified 32 studies of 42,016 patients reporting 3545 CVD events. Only 6 studies were set within randomized trials (“effect-modification” design) and the remaining 26 reported individuals exposed to clopidogrel (“treatment-only” design). In treatment- only studies, possession of one or more *2-*8 CYP2C19 alleles was associated with lower cytochrome P450 C19 (CYP2C19) enzyme activity and a higher risk of CVD events (RR 1.18; 95%CI:1.09, 1.28), however, there was strong evidence of small-study bias (Harbord test P=0.001) and, when restricted to large studies (≥200 events), the association of CYP2C19 *2-*8 carrier status with CVD was null (RR 0.97; 95%CI: 0.86, 1.09). In the effect-modification studies, CYP2C19 genotype did not modify the effect of clopidogrel on CVD end-points. These findings cast doubt on whether information on CYP2C19 genotype would be helpful to guide selection of the dose of clopidogrel or use of an alternative antiplatelet agent. The role of secretory phospholipase A2-IIA (sPLA2-IIA) in CVD I used Mendelian randomization to make causal inference on the role of sPLA2-IIA in CVD. I identified a single nucleotide polymorphism (SNP) in PLA2G2A (rs11573 156) that was specific for and had a very strong impact on circulating levels of the sPLA2-IIA isoform. Using data from 36 studies and over 100,000 participants, instrumental variable analysis found no association between sPLA2-IIA with incident, prevalent or recurrent CVD events. These findings suggest sPLA2-IIA is not a valid therapeutic target for CVD prevention, which was in keeping with a phase III randomized clinical trial that was halted for futility in 2012 (during this thesis). Alcohol and CVD I used a SNP in ADH1B to investigate the relationship between alcohol and coronary heart disease (CHD) in >260,000 participants. The genetic variant (ADH1B rs1229984 A-allele) showed very strong association with reduced alcohol consumption when evalu¬ated as volume of alcohol consumed, binge drinking and abstaining from alcohol. The A-allele of rs1229984 showed associations with SBP, CRP, IL-6, BMI and waist circum¬ference that were all directionally concordant with a reduced risk of CHD. Indeed, when the clinical outcome CHD was investigated, individuals carrying the A-allele (who con¬sumed less alcohol than non-carriers) had a reduced risk of CHD at all levels of alcohol consumption. No evidence of a cardioprotective association of alcohol with CHD was identified. Conclusions My investigation into use of CYP2C19 genotype as a pharmacogenetic biomarker for clopidogrel response did not identify evidence to support its clinical use and limitations were identified that could apply to other pharmacogenetic tests. Use of Mendelian randomization revealed no evidence to support a causal role of sPLA2- IIA in CVD, which paralleled findings from a phase III randomized clinical trial, and provides support for the use of Mendelian randomization studies more widely to in¬form drug development. Finally, using the ADH1B gene to interrogate the relationship of alcohol yielded findings that argue against a cardioprotective effect of alcohol con¬sumption. These findings should encourage rethinking of public health advice about the cardiovascular benefits of moderate levels of alcohol consumption.