Emulation of the Moderate Alcohol and Cardiovascular Health Trial (MACH15)

The aim of this study is to assess how long-term alcohol consumption influences health risks by emulating the Moderate Alcohol and Cardiovascular Health Trial (MACH15). In the first step, the protocol of the emulation of MACH15, including eligibility criteria, alcohol regimens and assignment, follow-up, endpoints, causal contrasts of interest, and statistical analysis was specified. In the second step, the investigators will emulate an adapted version of MACH15 following the specified protocol using data from the UK Biobank.

Observational data suggests that alcohol consumption lowers the risk of cardiovascular disease (CVD) compared to no consumption. Whether this relationship is truly causal remains uncertain because of the inherent limitations of observational studies, including unmeasured confounding and reverse causation. Mendelian randomization studies using genes as instrumental variables for alcohol are partially protected from these biases and have found no or harmful associations between alcohol consumption and CVD. To date, there has only been one long-term randomized controlled trial to investigate the cardiovascular effects of alcohol consumption: the Moderate Alcohol and Cardiovascular Health Trial (MACH15; NCT Number: NCT03169530). It was, however, terminated shortly after initiation. An alternative to a real randomized trial like MACH15, which must first be completed and is subject to strict eligibility criteria to ensure safety, is to use observational data to emulate a (hypothetical) pragmatic randomized trial. In this study, the investigators will emulate an adapted version of MACH15 using observational data from the UK Biobank, a large prospective cohort study of over 500,000 participants. The cardiometabolic effects of moderate drinking vs quitting, as originally planned in MACH15, as well as the effects of social and heavy/binge drinking on CVD, type 2 diabetes, other alcohol-related health outcomes, and death will be quantified.