I don't find that study at all compelling. It looks at two genes (but one in particular) that is heavily associated with not drinking which is common in that part of China but rare in Europe. I would be very interested to know how the authors estimate exactly how much these people drink based on their genes, but unfortunately there is little explanation in the study. On the face of it, it doesn't seem possible to say more than that he/she probably doesn't drink at all or drinks very little, but perhaps I am missing something.
The same study does manage to find a link between smoking and lung cancer but with an RR of just 2.26, an order of magnitude lower than all the other evidence suggests.
This is because it is a very blunt tool. It is good at establishing relationships and exposing confounding variables when a gene is very strongly - preferably exclusively - related to a disease, but has so far been pretty useless when it comes to lifestyle factors which involve choice. Asking people whether they smoke or how much they drink is not perfect but it seems to be better than guessing this based on a couple of genes.
I would also be interested in any views on this. The study authors say that their finding that there was no relationship between level of alcohol use and heart disease was based on a fairly small number of cases, so you could say that it gives weak evidence against the link between moderate alcohol use and lower rates of heart disease, which was the focus of Doll's research. But (though I don't have the expertise to be able to judge how robust the study is) it certainly seems to provide good evidence that the J curve is unreliable for some other causes of death.
Really interesting read, and thank you for the link.
Couldn't agree more! The idea that drinking is bad for you is pure fake science. I have a minimum of two drinks a day and my life and health is so good, much better than everyone else's. Same with smoking. This Tim guy reminds me of all the squabbling anti-tobacco ideologues. No, smoking does not cause lung cancer. Yes, the hidden injection of artificial lab carcinogens during "routine bloodwork" while participating in a large government smoking study does. Use your brains people! They would rather murder their own citizens than allow people to enjoy the fine pleasure of smoking. Hitler was the first to say that tobacco is the red man's scourge upon the white man for the gift of hard liquor. Yes, that's right. The so-called scientific anti tobacco campaign is actually a racist crusade that began in nazi germany at nazi "science" institutions. Even to this day, nazi money continues to fund anti-tobacco. I have smoked unfiltered for 30 years and I feel amazing, I've never suffered a single negative side effect. I'm way healthier and more capable than anyone my age due to anti-aging supplements that have taken my blood all the way down to the blood of a 28 year old. For all those reading this... Try taking NMN! It will reverse the wear on your body.
Tobacco is good! Tobacco is not evil! Smoking DOES NOT CAUSE CANCER.
I think most people regard Mendelian Randomisation as the best design for getting at causal relationships when randomised trials are not feasible.
This well conducted MR study seems pretty convincing that the J curve is an artefact https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(18)31772-0/fulltext
From your post I’m guessing this study hasn’t persuaded you. So what evidence would make you change your mind?
I don't find that study at all compelling. It looks at two genes (but one in particular) that is heavily associated with not drinking which is common in that part of China but rare in Europe. I would be very interested to know how the authors estimate exactly how much these people drink based on their genes, but unfortunately there is little explanation in the study. On the face of it, it doesn't seem possible to say more than that he/she probably doesn't drink at all or drinks very little, but perhaps I am missing something.
To elaborate on a point made in the post, MR is very good at *not* finding relationships. It is taken as read in public health that alcohol causes breast cancer but that is not what MR says: https://www.sciencedirect.com/science/article/abs/pii/S2210776220302465
In fact, MR studies suggest that alcohol doesn't cause any form of cancer except lung cancer! https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1003178#sec016
The same study does manage to find a link between smoking and lung cancer but with an RR of just 2.26, an order of magnitude lower than all the other evidence suggests.
This is because it is a very blunt tool. It is good at establishing relationships and exposing confounding variables when a gene is very strongly - preferably exclusively - related to a disease, but has so far been pretty useless when it comes to lifestyle factors which involve choice. Asking people whether they smoke or how much they drink is not perfect but it seems to be better than guessing this based on a couple of genes.
I would also be interested in any views on this. The study authors say that their finding that there was no relationship between level of alcohol use and heart disease was based on a fairly small number of cases, so you could say that it gives weak evidence against the link between moderate alcohol use and lower rates of heart disease, which was the focus of Doll's research. But (though I don't have the expertise to be able to judge how robust the study is) it certainly seems to provide good evidence that the J curve is unreliable for some other causes of death.
Really interesting read, and thank you for the link.
Couldn't agree more! The idea that drinking is bad for you is pure fake science. I have a minimum of two drinks a day and my life and health is so good, much better than everyone else's. Same with smoking. This Tim guy reminds me of all the squabbling anti-tobacco ideologues. No, smoking does not cause lung cancer. Yes, the hidden injection of artificial lab carcinogens during "routine bloodwork" while participating in a large government smoking study does. Use your brains people! They would rather murder their own citizens than allow people to enjoy the fine pleasure of smoking. Hitler was the first to say that tobacco is the red man's scourge upon the white man for the gift of hard liquor. Yes, that's right. The so-called scientific anti tobacco campaign is actually a racist crusade that began in nazi germany at nazi "science" institutions. Even to this day, nazi money continues to fund anti-tobacco. I have smoked unfiltered for 30 years and I feel amazing, I've never suffered a single negative side effect. I'm way healthier and more capable than anyone my age due to anti-aging supplements that have taken my blood all the way down to the blood of a 28 year old. For all those reading this... Try taking NMN! It will reverse the wear on your body.
Tobacco is good! Tobacco is not evil! Smoking DOES NOT CAUSE CANCER.
Cuckoo Cuckoo
Thank you for your insight
Great recap/revisit, thanks