3 ms·

I looked at the paper but I don't understand how they got to that conclusion. There is still a J-shaped relationship between alcohol consumption and coronary he

I looked at the paper but I don't understand how they got to that conclusion. There is still a J-shaped relationship between alcohol consumption and coronary heart disease outcomes. Not questioning the conclusion I just don't understand how you get to it from their data (due to my own ignorance).

spekcular 5y ago
Researchers are able to say that the J-shaped relationship is a result of factors correlated with alcohol use, instead of the effects of alcohol itself on the body, using a technique called Mendelian randomization. Here's a summary for the lay reader [0]: > One way to get around these limitations is through genetic studies. Some people carry a genetic variant that disrupts their ability to metabolize alcohol, causing them to develop skin flushing, irritation and other unpleasant symptoms when they drink alcohol. As a result, they tend to abstain or drink very little. If alcohol was good for heart health, these people should in theory have more heart disease compared to others. Instead, as one large analysis published in BMJ in 2014 found, they have “a more favorable cardiovascular profile and a reduced risk of coronary heart disease than those without the genetic variant.” > The study concluded: “This suggests that reduction of alcohol consumption, even for light to moderate drinkers, is beneficial for cardiovascular health.” (One fact not mentioned there is that for this technique to work and identify causality, the distribution of that genetic variant needs to be uncorrelated with all the other relevant variables. To my knowledge, this is true.) [0] https://www.nytimes.com/2020/07/10/well/eat/should-we-be-drinking-less.html https://www.nytimes.com/2020/07/10/well/eat/should-we-be-dri...
- thinkharderdev 5y ago
  Ah, I get it now. Thanks!

3 ms·

I looked at the paper but I don't understand how they got to that conclusion. There is still a J-shaped relationship between alcohol consumption and coronary he

by thinkharderdev 5y ago

spekcular 5y ago
Researchers are able to say that the J-shaped relationship is a result of factors correlated with alcohol use, instead of the effects of alcohol itself on the body, using a technique called Mendelian randomization. Here's a summary for the lay reader [0]: > One way to get around these limitations is through genetic studies. Some people carry a genetic variant that disrupts their ability to metabolize alcohol, causing them to develop skin flushing, irritation and other unpleasant symptoms when they drink alcohol. As a result, they tend to abstain or drink very little. If alcohol was good for heart health, these people should in theory have more heart disease compared to others. Instead, as one large analysis published in BMJ in 2014 found, they have “a more favorable cardiovascular profile and a reduced risk of coronary heart disease than those without the genetic variant.” > The study concluded: “This suggests that reduction of alcohol consumption, even for light to moderate drinkers, is beneficial for cardiovascular health.” (One fact not mentioned there is that for this technique to work and identify causality, the distribution of that genetic variant needs to be uncorrelated with all the other relevant variables. To my knowledge, this is true.) [0] https://www.nytimes.com/2020/07/10/well/eat/should-we-be-drinking-less.html https://www.nytimes.com/2020/07/10/well/eat/should-we-be-dri...
- thinkharderdev 5y ago
  Ah, I get it now. Thanks!