8th European Beer and Health Symposium

Luc Djoussé received his medical degree (Dr. Med) from the University of Saarland, Germany; a Master of Public Health from Boston University, and a Doctor of Science degree (Epidemiology) from Boston University. He trained in internal medicine (Germany) and preventive medicine (USA). Completed a preventive cardiology fellowship in Framingham Heart Study, MA. He currently serves as Director of Research in the Division of Aging, Department of Medicine at Brigham and Women’s Hospital and Associate Professor of Medicine at Harvard Medical School. He served as adjunct professor of epidemiology at Massachusetts College of Pharmacy and Health Science University, Boston and is also adjunct faculty at Tufts University school of Medicine, Boston. He is a renowned cardiovascular epidemiologist with research focus on the role of diet, genetics, and their interactions on cardiovascular health. Over the past 20 years, he has served as principal investigator on numerous grants from the National Institute of Health and industry and has mentored several preventive cardiology fellows. His recent projects focus on the role of omega-3 fatty acids, free fatty acids, and moderate alcohol consumption on the risk of heart failure and its predictors. He is Editor-in-Chief of Current Nutrition Reports, a member of the Scientific Executive Committee and Advisory Board of the International Academy of Cardiology and has held several leadership positions within the Council on Epidemiology of the American Heart Association. He has published 240 scientific papers and book chapters.

Different Concepts in Alcohol Research: Are the Observed Protective Health Effects of Moderate Beer Consumption Still Valid?

A large body of evidence from observational studies supports beneficial health effects of moderate alcohol consumption. However, not all studies have supported such benefits. Some of the inconsistencies can be attributed to the choice of inappropriate reference group, inadequate control of major confounders, or methodologic issues. One of the challenges in observational studies is the inability to randomly assign subjects to a particular exposure category, thereby leading to unbalanced distribution of measured confounders. More importantly, investigators are not able to quantify the impact of unmeasured confounding on the effect estimate. To address these shortcomings in observational study design, the use of Mendelian randomization (MR) built on instrumental variable (IV) has become popular in recent years. A good IV must not relate to confounders of the exposure-outcome being studied and must be associated with the outcome only via the exposure of interest. Given the random assortment of alleles during meiosis, genetic markers appear suitable and are often used as IV. In the absence of large randomized trials of alcohol intake on complex traits such as cardiovascular disease, researchers have used single nucleotide polymorphisms in the genes encoding enzymes responsible for alcohol metabolism as IV in an attempt to minimize confounding when assessing alcohol effects on health. However, results of those studies have been inconsistent. Such inconsistency might be partially due to violation of IV assumptions and/or limitations of MR including confounding by population admixture, pleiotropy of genes, linkage disequilibrium, and canalization. Furthermore, the use of a weak instrumental variables or not utilizing complete genetic information in the presence of genome-wide association data might play a role. This presentation first discusses common sources of bias in alcohol research, followed by IV and MR concepts, strengths and limitations, and future considerations of MR in the context of observational epidemiologic research. While MR is a great tool to minimize confounding when appropriately applied, it can lead to erroneous inference when key assumptions are ignored or when weak instruments are used. With the availability of genome-wide association data, consideration should be given to multivariable MR that accounts for multiple genetic loci as instrumental variables rather than a simple approach focused on a single nucleotide polymorphism. Studies results based on MR should be interpreted in the context of above limitations.

Pin It on Pinterest

Share This