This article arrives at interesting suggestions about the potential benefits and risk of the long term use of metformin.
“Conclusion: The MR study discovered from a genetic standpoint that metformin may lower the risk of hypertrophic cardiomyopathy and valvular heart disease, yet it could elevate the risk of myocardial infarction.”
Why even post a trash study like this? An obscure Chinese study.
“Background: The cardiovascular effects of metformin continue to be a subject of debate within the medical community.”
“Conclusion: The MR study discovered from a genetic standpoint that metformin may lower the risk of hypertrophic cardiomyopathy and valvular heart disease, yet it could elevate the risk of myocardial infarction.”
Calling this a “trash study” is scientifically unjustified. Of course, one needs to understand the usefulness and risks of a MR study. As for the implication that the study’s use of the terms “could” and “may” might somehow make it weaker or less useful, quite the opposite is true. Studies that assert certainty in generalizations based on findings, even RCTs, are suspect in my assessment because we do not get that kind of certainty in the human medical and behavioral sciences. Anyone can observe this by noting how rapidly perspectives and so-called facts on the ground change, sometimes reversing themselves 180 degrees. The disruptive analytic methodology I posted using red meat as an exemplar demonstrates that dynamic in definitive if somewhat complicated terms.
Overall, it is my personal opinion that the shared interests of this group are advanced by sharing a wide range of potentially relevant (especially refereed) research, even though we may choose to ignore many of these studies because we find them irrelevant to us or too early in the research cycle. Others will feel differently and that is a good thing, again, IMO. It would, however, be reasonable for those familiar to debate the scientific merit of MR studies in more objective ways.
It’s also less easy to cheat or do fraudulent MR studies as anyone can replicate the analysis as the data is public and if others publish MR papers showing something different their will be incentives to call out the first group of authors and scrutinize their work
For animal or cell studies someone would have to generate their own data to replicate the study, but cannot just verify the actual data.
The study reaches no conclusions, other than might or could.
You make a good point on the degree of transparency in MR studies, greater than most of the other types of studies we see and not subject to the risks of analytic bias (not necessarily intentional but in some cases just that) almost universal to other studies we discuss here. When I review a MR study, i generally look for hints as to straighter causal lines than we get with other studies. Many limitations and risks for sure but more robust than most on initial paths of causality.