Commentary: Mendelian randomization, testosterone, and cardiovascular disease

摘要: In this issue of the International Journal Epidemiology, Dr. Zhao and colleagues used a Mendelian randomization (MR) analysis to examine causal effects testosterone on two cardiovascular risk factors, prolonged QT interval elevated heart rate. Previously they have investigated relationship between other factors such as blood pressure, low-density lipoprotein (LDL) cholesterol, high-density (HDL) cholesterol fasting glucose, using similar analysis. Understanding disease its is critical. Testosterone use has dramatically increased since 1993 (by roughly 5-fold), especially in USA UK, remains leading cause morbidity mortality globally. Previous observational studies shown associated with but inferring causality based these results problematic are susceptible residual confounding reverse causality; even careful study design statistical adjustment, incorrect inference can persist. Randomized controlled trials (RCTs) well-established gold standard for inference; however, due concerns regarding safety androgen supplementation, there no well-powered RCTs examining potential testosterone. Alternatively MR, which uses genetic variants instrumental variables assess causality, serves post hoc analytical method dealing This work by et al. provides good example implementing MR Since often modest exposure interest, typically requires very large sample sizes gain adequate power. multiple (i.e. weighted score) try address issue. Using method, association was observed majority except LDL, HDL, rate-corrected interval. The null associations found here potentially reflect continued deficiency power their studies. Whereas heritability serum approximately 65% adult men, only limited number testosterone-related been identified. Thus, it expected that an increasing will be identified future studies; additional then further improve One thing keep mind ethnicity population influence significantly associate Future should take advantage architecture ethnically diverse populations identify novel testosterone-influencing variants. separate-sample variable (SSIV) estimate MR. An SSIV incorporates measurement from one outcome separate independent study. Here investigators developed prediction rule dataset applied participants larger obtain genetically predicted factors. allows flexibility