The 'healthy' daily drink was a statistical illusion

Researchers at the Canadian Institute for Substance Use Research reviewed 107 earlier studies covering 4.8 million people to test whether that protective effect was real. Many of those studies had compared drinkers to people who had quit or never drunk, without asking why those people don't drink.

Once the team corrected for that bias, comparing drinkers with lifelong non-drinkers, the apparent protective effect of light to moderate drinking on death from any cause disappeared. Studies with weaker methods tended to show the benefit; studies with stronger designs did not.

Former drinkers and lifelong abstainers include many people who quit or avoided alcohol because of existing health problems, which had been dragging down the comparison group and making moderate drinkers look healthier by contrast. The finding does not mean alcohol is harmless at low doses, only that the protective story was built on a statistical artifact, and some researchers still dispute how the studies should be weighted.

The protection was in the comparison.