1st BATCH OF FILES AVAILABLE NOW!

To help support their claim that there's "no safe level" of drinking, advocates increasingly focus on cancer, pointing to research suggesting increased risk at even low levels of consumption. But a pair of recent studies complicate the story. Not that you'd know it from most media coverage. "Deadly cancer risk spikes with certain level of alcohol consumption," some headlines blared last month, while others alerted readers in a more level-headed tone: "lifetime alcohol use linked to higher risk of colorectal cancer." While those stories emphasized findings about increased risk among heavy drinkers, they mostly ignored that the same study also found substantially lower risk among lighter drinkers. A separate, widely ignored study found "inconsistent" results, including "moderate-to-weak evidence" that alcohol causes cancers long associated with drinking.Risk Nearly Doubles for Lifetime Heavy Users, Drops 20-40% for Moderate Drinkers Most media outlets did not question the emphasis placed on increased risk for colorectal cancer (CRC) identified among heavy drinkers by a press release announcing a study published in the journal Cancer in late January. Folks who reported consuming an average of 14+ drinks per week across their lifetime saw a 25% increased risk of CRC, including a 95% increase in rectal cancer risk, compared to folks reporting less than 1 drink per week, researchers found. Those stats got all the attention, often alongside reminders that CRC cases have increased in recent years.Researchers also highlighted findings that, in their view, point to the benefits of quitting. All former drinkers, regardless of how much they used to drink, appeared to have a significantly lower risk of developing precancerous tissue that can lead to CRC, according to the study. "These numbers are not guarantees, but signals to do something before it's too late," one doctor told ABC News about the results.As usual, stories skipped over other key findings, despite a handy summary that even made the study's abstract: "Current drinkers averaging from seven to less than 14 drinks compared with less than one drink per week had a lower risk of CRC." How much lower? A 21% drop overall, so almost as large as the increased risk among heavy drinkers. And for one type of colon cancer, risk was 36% lower among moderate drinkers. The authors do not plot these relative risks onto a graph, but if they did, it would look a bit like a "J." Ref 1The study also tracked impacts of screening, which reduced risks even further among folks reporting "moderate average lifetime intake" (7-14 drinks/week), "suggesting that screening has the potential to modify the association between alcohol drinking and CRC risk." In other words, screening could mitigate even the increased risks seen among heavier drinkers.Comparisons and Lifetime Measurement Correct for Familiar "Problems" Notably, the authors of this study worked to correct for the kind of issues critics often highlight in similar studies about alcohol's impact on health. For one, they separated out all former drinkers from non-drinkers. And instead of comparing current drinkers to non-drinkers, they compared them to very light drinkers, folks reporting less than 1 drink per week.Furthermore, participants in the study estimated their lifetime consumption patterns, reporting average consumption during four different life periods (ages 18-24, 25-39, 40-54 and 55+). Obviously, self-reporting offers the usual challenges and problems for researchers. But at least these authors attempted to mitigate concerns about studies only asking about consumption at one specific point in time.Only Consistently Heavy Drinkers Saw Higher Risk Given those lifetime measurements, authors focused on the fact that "always heavy" drinkers showed a 91% increase in CRC risk compared to "always light" drinkers. The latter group is made up of women who always consumed up to one drink per day and men up to two. But the latter group includes women consistently consuming over two drinks per day and men over three, across all life periods. Researchers found no statistically significant increase in risk among folks whose drinking habits changed. There was no pattern among "occasionally moderate" or even "occasionally heavy" drinkers, meaning folks who drank above 1 or 2 drinks per day at some point in their lives didn't always show a higher risk of CRC than lighter drinkers.Rising CRC cases in recent years have understandably garnered increasing attention, with many researchers trying to understand the causes. A year ago, upon the release of the Surgeon General's Advisory on Alcohol and Cancer Risk, some folks suggested that upticks in consumption during the lead-in to and peak years of the Covid-19 pandemic could be part of the explanation. This study would appear to suggest that such a short-term change in drinking patterns is unlikely to have been a primary driver."Inconsistent" Results in Study Based on Genetic Predictors Another branch of research attempts to correct for another persistent "problem" with studying alcohol: the challenges arising from self-reported consumption (and the tendency for folks to under-report how much they drink). Over the last decade or so, studies started to use genetics instead, employing an analysis called "Mendelian randomization" (named for Gregor Mendel, the 19th century scientist who pioneered studying how traits are passed down through genes). The idea is simple: identify genetic markers statistically associated with a certain trait or behavior and then examine outcomes of folks with those genes rather than depending on self reports. It's become a go-to tool for researchers fed up with traditional observational studies of drinkers, for instance, seeking an arguably less biased (or fungible) metric.But a recent examination of the link between cancer incidence among folks with a set of "95 genetic variants" associated with alcohol consumption turned up "inconsistent" results. Overall, researchers found "no association" among "genetically-predicted" drinkers and any cancer risk. In fact, they measured a slightly lower risk (but just 4%). They found a clear increase in risk for head/neck cancer (51%), and more modest yet statistically insignificant increases for colorectal and esophageal cancers. But there was not a statistically significant increase in risk for liver cancer or breast cancer. And they found significantly decreased risk for a slew of other cancers. Indeed, "the strongest associations we observed were in the protective direction," authors wrote. Ref 2Skepticism is Warranted, Authors Acknowledge If you're already wondering how well our genes can predict our drinking, you're not alone. Even the authors of this study recognize that "the genetic variants [used in the analysis] may not be valid instrumental variables." In general, the authors "find that readers of our papers on Mendelian randomization are less skeptical about our findings than we are ourselves." And frankly, "genetic variants may explain only a small proportion of variance in alcohol consumption."While the authors pose that as merely possible in their conclusion, the paper itself shows that it is absolutely true. That group of 95 genetic markers "explained around 0.2% of the variability in alcohol consumption," they wrote. You read that right. Less than a quarter of one percentage point of the differences in how much or whether people drink can be explained by these genes. And yet, somehow it met a statistical threshold for this kind of research, allowing these scientists to plow forward with an analysis that turned up some not so useful results. Is now a good time to point out that the analysis is based on genetic and health data for almost 1.5 million individuals in the US and Europe?To find such paltry statistical power with such a large number of people doesn't just raise fundamental questions about the value of using genetic models to prove that alcohol causes cancer. It undermines the use of Mendelian randomization (MR) to prove just about anything related to alcohol, as UK-based writer and economist Christopher Snowdon put so eloquently: "It's probably time to stop using MR in alcohol research. You could say that would be throwing the baby out with the bathwater, but it's not much of a baby and the bathwater stinks."We'll highlight more of Snowdon's commentary and unpack how MR research has been used to support the "no safe level" rhetoric used by public health advocates and the WHO next issue.

Cancer Story Gets Complicated: Light Drinking Linked to Lower Colorectal Risk

Publishing Info