A limitation of genetic epidemiological analysis when associations are genuinely J-shaped

A paper published in The Lancet in 2019 received wide media coverage when it reported that even light drinking patterns increased blood pressure and the chances of having a stroke. The study, based on 500,000 people, used mendelian analysis to counter previous claims that one or two drinks a day could in fact be protective.
In an editorial to be published in the International Journal of Epidemiology Sir Nicholas Wald, of University College London, and Chris Frost, of the London School of Hygiene and Tropical Medicine, claim that the 2019 analysis was flawed.
Wald and Frost created a hypothetical population in which there was a genuine J-shaped relationship, and then applied the genetic epidemiological analysis used in the 2019 paper. It failed to uncover the J-shaped relationship.
Wald and Frost comment “We conclude that the observation in many studies that light drinking reduces the risk of stroke but heavier drinking increases it, is not necessarily disproved by the genetic analysis of Millwood and colleagues.
“It is important for public health policy that the true relationship between alcohol consumption and vascular disease is recognized.”
Source: Chris Frost, Nicholas Wald, A limitation of genetic epidemiological analysis when associations are genuinely J-shaped illustrated using a prospective study of alcohol consumption and vascular disease, International Journal of Epidemiology, 2021. doi.org/10.1093/ije/dyab162