Light Drinking and Breast Cancer Link: Statistics Are Complicated

Alcohol

One of the less-talked about risk factors for breast cancer is alcohol consumption. Past research has noted a distinct dose-response relationship between drinking and breast cancer—this means that the risk goes up among heavier drinkers. However, while this makes the risk of heavier drinking easy to spot and study, the effects of light drinking is not as clear. A recent systematic review has tried to pool together past research on light drinking and breast cancer and comes away with the idea that light drinking is a large cause of the disease. Unfortunately, these conclusions are not as reliable as a first glance might suggest.

Alcohol and Breast Cancer Systematic Review: Summary

  • The systematic review looked at 15 meta-analysis studies on the risk between drinking and breast cancer, 13 of which showed a dose-response relationship.
  • An additional population-level analysis was performed which used a method called Population-Attributable Fraction methodology to estimate the impact of alcohol on breast cancer cases and deaths.
  • The analysis found that 144,000 breast cancer cases and 38,000 breast cancer deaths in 2012 (globally) were attributable to alcohol and that 27,072 (18.8%) of these cases and 6,650 (17.5%) of the deaths were from light drinking specifically.
  • The conclusion is that, due to the findings and global alcohol consumption, this shows a strong relationship between light drinking and breast cancer and that light drinking is a large portion of alcohol-attributable breast cancer cases and deaths.
  • In other words, the conclusion is that even light drinking can create a distinct risk increase and is behind a large portion of alcohol-attributable breast cancers.

Breaking Things Down

There are two key parts of the review: the systematic review itself and the Population-Attributable Fraction analysis, both of which need to be addressed in order to explain why these results are not as clear as the researchers seem to think. Unfortunately, the full study requires a subscription but there is enough in the press release and abstract that certain observations can be made.

The Systematic Review

A systematic review is when a body of research is looked at and broader conclusions are drawn from the combined findings and trends that are observed. A meta analysis is a statistical approach that combines the data from the multiple studies in order to perform larger evaluations. A simpler way to think of this is that the systematic review conducted here was a study of studies about smaller studies.

Advertisement

The thing to pay attention to when the authors describe their systematic review is that they only describe the studies as showing a dose-response relationship. Any particular effects or levels of risk that may be found from light drinking is not brought up. In other words, while the review may have found that heavier alcohol consumption leads to higher breast cancer risk, this does not automatically mean that a significant risk increase also exists at light drinking levels.

In other words, there is not anything in the way the systematic review is described that suggests anything different than what groups like the American Cancer Society currently says—that light alcohol consumption creates a small but not significant risk increase of breast cancer.

The Population Analysis

A Population-Attributable Fraction (PAF) is a specific type of calculation that is being used here to produce the number of breast cancer cases and deaths that are attributable to alcohol and, subsequently, which of those alcohol cases and deaths are attributable to light drinking. This is, simply put, not how a PAF works.

A Population-Attributable Fraction (PAF) is a specific type of calculation that is being used here to produce the number of breast cancer cases and deaths that are attributable to alcohol and, subsequently, which of those alcohol cases and deaths are attributable to light drinking. This is, simply put, not how a PAF works.

Advertisement

PAF is, speaking generally, used to determine the differences in risks between a population at large and the amount of that population that hasn’t been exposed to a specific risk factor. It has certain uses but one thing that it is not suited for is trying to determine an attributable cause of a disease, especially one as multifaceted as cancer. This problem gets even worse when the risk factor being examined has a large exposure, like being over a certain age or drinking alcohol. This issue has been noted by the CDC which concluded that using PAF methods in “attempts to partition [assign] causality [attribute] when multiple forces act together to produce the outcomes are meaningless.”

Bottom Line

  • The description of the systematic review noted a dose-response relationship between alcohol consumption and breast cancer but not whether there was meaningful risk at light drinking levels.
  • The population analysis used the wrong type of calculation and its results are highly unreliable at best and meaningless at worst.
  • While the full study may clarify the issue with the systematic review, right now there isn’t anything that useful to be found.

Sources:
“What are the risk factors for breast cancer?,” American Cancer Society web site, last editing May 4, 2016; http://www.cancer.org/cancer/breastcancer/detailedguide/breast-cancer-risk-factors, last accessed June 3, 2016.
Shield, K., et. al., “Alcohol Use and Breast Cancer: A Critical Review,” Alcoholism: Clinical and Experimental Research, 2016; 10.1111/acer.13071
Levine, B., “What Does the Population Attributable Fraction Mean?,” Preventing Chronic Disease, 2007; http://www.cdc.gov/pcd/issues/2007/jan/06_0091.htm.