By Eric Crampton 28/06/2016 2


University of Otago has come out with another study on cancer risks from moderate drinking.

I just can’t get too excited about these. Moderate alcohol consumption reduces your risk of some stuff and increases your risk of other stuff. On net, drinking about a standard drink per day reduces your risk of dying of anything by about 15% relative to baseline.

Drinking more than about four standard drinks per day increases your risk of dying relative to baseline. And it’s not like they get that number from adding up all the bad things and netting out the good things: they just look at alcohol consumption and who died, then run the risk curves correcting for as much as they can.

So it doesn’t really matter whether cancer turns out to be much worse or less bad than expected. However bad it is is already in the mortality stats, along with however good it is on other stuff.

Eileen Goodwin at the ODT got in touch with me asking for comment on the story. Here’s what I told her:

“Alcohol consumption has known cancer risk, even for moderate consumption. But moderate consumption of alcohol also has offsetting health benefits in other areas. And so more interesting work looks at the effects of alcohol consumption on overall mortality risk. And that work finds that drinking about a drink a day reduces your risk of dying, from any cause, by about 15%.
Drinking more than about 3 or 4 standard drinks per day increases your mortality risk. So, if you have a family history of cancer, knowing about the cancer-specific risks should encourage you to avoid drinking to excess. But it would be a shame if health-conscious people switched from a drink a day to abstinence, fearing the cancer risk, and wound up at higher overall risk from heart disease because of it.”

Here are the results from DiCastelnuovo et al’s metastudy looking across many different studies tracing out the mortality risks of different levels of drinking. The baseline risk, 1.0, is established with reference to non-drinkers. When no former drinkers are in the reference group (dark line, panel A below), mortality risk is minimised at just under a standard drink per day. The “reduces risk of dying by about 15%” claim in my quote above is based on that dark line.

Screen shot 2016-06-28 at 3.29.37 PM
Relative risk of total mortality (99% confidence interval) and alcohol intake stratified according to type of reference category (A), sample size at baseline (B), year of publication (C), and follow-up duration (D).
I guess there wasn’t room for the “reduces your risk of dying by 15%” if the space was needed for the “right-wing think-tank” part. Here’s what was used:

Seeking comment, the Otago Daily Times contacted brewing giant Lion, where a spokeswoman suggested the newspaper contact right-wing think-tank The New Zealand Initiative.Initiative head of research Dr Eric Crampton said people who cut their alcohol consumption risked losing potential benefits. Dr Crampton said he accepted moderate alcohol consumption was a cancer risk.

“But moderate consumption of alcohol also has offsetting health benefits in other areas.

“And so more interesting work looks at the effects of alcohol consumption on overall mortality risk.

“It would be a shame if health-conscious people switched from a drink a day to abstinence, fearing the cancer risk, and wound up at higher overall risk from heart disease, because of it,” Dr Crampton said.

Featured image: CC flickr


2 Responses to “Moderate drinking linked to cancer risk: NZ study reviewed”

  • In the phrase”mortality risk is minimised at just under a standard drink per day”, should the “at” be replaced with “when your average alcohol consumption is” ? I assume the “Drinks per Day” in the graphs is an average, although it could be something else entirely, eg mode, median, maximum.

  • Always a balancing act in having it terse enough to not be cut substantially in editing, and not have it be wrong. Folks wishing to minimise death risk should aim at about a standard drink a day, or perhaps a bit more than that with a day off every couple days, for an average of about a standard drink a day. Drinking 14 every other Friday would not be helpful.