P1-4 ALDH2 is a valid instrument for mendelian randomisation studies in alcohol epidemiology—the Guangzhou Biobank Cohort Study

Introduction Western observational studies show moderate alcohol use positively associated with many health outcomes. Moderate alcohol users differ systematically from others, making these observations susceptible to residual confounding. A randomised controlled trial of moderate alcohol use is unlikely to be possible. A Mendelian randomisation design using a genetic polymorphism affecting alcohol use ( ALDH2 ) as an instrumental variable offers an alternative approach for establishing causality. However, such an approach is only valid if ALDH2 polymorphisms are unrelated to the relevant health outcomes. In a virtually non-drinking population where ALDH2 polymorphisms are common, older Southern Chinese women, we examined the association of ALDH2 with biological cardiovascular risk factors (blood pressure, HDL- and LDL-cholesterol and glucose) in the Framingham score. Methods We used DNA from 833 women never or occasional alcohol users in the Guangzhou Biobank Cohort Study. We extracted DNA using TIANamp Blood DNA Kit and ALDH2 (rs671) was genotyped externally using Sequenom MassARRAY system at CapitalBio. We used multivariable linear regression to assess the association of ALDH2 polymorphisms with cardiovascular risk factors. Results The genotype frequencies of ALDH2 (*2*2,*1*2,*1*1) were 9.5%, 38.1%, and 52.5% as expected. Adjusted for age or additionally for education, job type, income, physical activity and smoking status, there was no association of ALDH2 with these cardiovascular risk factors. Conclusion The lack of association of ALDH2 polymorphisms with cardiovascular risk factors suggests that ALDH2 polymorphisms do not have independent effect on cardiovascular risk factors, hence ALDH2 is a potential instrument for Mendelian randomisation studies of alcohol use.