Epidemiological studies typically show a flat mortality curve for low alcohol levels, with mortality increasing as consumption rises, while Mendelian randomization suggests any increase in alcohol consumption leads to an increase in mortality.

“so the epidemiology usually shows kind of a flat curve for low levels of alcohol and then a ramp up of mortality as alcohol creeps up again different studies and different cohorts are going to find different places um I generally tell patients that I think conservatively one drink a day is at least according to the Epi a minimal increase in Risk um whereas you know that jamama study found you could get up to two drinks a day maybe even three it was only at three when you started to see the uptick um but the mandelian randomization which again is a technique where we look at genes that control a trait so you might look at genes that control cholesterol or genes that control in this case alcohol consumption because again um we know that there are certain genes that make it very difficult for people to drink alcohol so if you if you believe that um possessing those genes can speak to the phenotype of drinking and I think this is a decent example of where mandelian randomizations work there are some where it doesn't um the mrss show that at any increase in the consumption of alcohol there is indeed an increase in mortality so they um they show an increasing increasing level so we say that that means that the first and second derivative are positive so any you know standard deviation and increase in the consumption of of alcohol leads to a greater increase in the risk of everything from hypertension to dementia to cardiovascular disease to cancer to all cause mortality”