Mendelian randomization studies indicate that the J-curve suggesting abstinence is riskier than moderate drinking is false.

“which is that there's you know we we kind of reject the data that says that there's a J curve right so we we have not internally been convinced by the jcurve data which is largely epidemiologic and largely you know suggests that at very low doses I zero alcohol is actually slightly worse than some alcohol and then of course the risk goes up as drinks go up but there's some sweet spot which depending on the study can be actually quite High um for alcohol consumption to produce the lowest all caused mortality um again there are many ways to explain those those those data that I think are a better explanation than alcohol is good for you at some dose I think the mandelian randomizations point to the opposite which is a monotonic change in risk that increases uh consistently from zero and upward”

“which is that there's you know we we kind of reject the data that says that there's a jcurve right so we we have not internally been convinced by the jcurve data which is largely epidemiologic and largely you know suggests that at very low doses I zero alcohol is actually slightly worse than some alcohol and then of course the risk goes up as drinks go up but there's some sweet spot which depending on the study can be actually quite High um for alcohol consumption to produce the lowest all cause mortality”

“both the mandelian randomization makes that not makes it clear that that's not true”

“the J curve is a misnomer”