The Claim: Skipping breakfast makes you gain weight, while eating it regularly helps you maintain a healthy weight.
The advice is inescapable. The blog for The Dr. Oz Show says that to lose fat, “you can’t skip breakfast,” while WebMD asserts “Lose Weight: Eat Breakfast.” No less than the Mayo Clinic says “regularly eating a healthy breakfast may help you lose excess weight and maintain your weight loss.”
These assertions come from studies as plentiful as diet resolutions on January 1. So much research — from Asia to the United States — has found a connection between eating breakfast and maintaining a healthy weight, or skipping it and being overweight, that the odds of those associations being due to chance are less than 1 in a trillion trillion trillion, according to a 2013 analysis led by David Allison of the University of Alabama at Birmingham. In other words, the breakfast/weight connection looks like a lock.
Gut Check loves the breakfast/weight claim. It has everything you could want to illustrate how a supposedly scientific conclusion is as insubstantial as foam.
With apologies to regular readers, we have to trot out the hoary observational-study criticism. More than 85 percent of breakfast/weight studies compare people who, by their own choice, eat breakfast regularly to those who, also by choice, skip it, then measure the respective rates of obesity or overweight.
But association is not causation. A third factor might cause both breakfast eating and healthy weight, or breakfast skipping and overweight.
Unlike other cases where we repeat that statistical verity, here we have a good idea of what that other factor might be: people who regularly eat breakfast, found a recent study, are more conscientious and more likely to engage in vigorous exercise. Being health-conscious and physically active are both more likely to be true causes of healthy weight, meaning breakfast is a mere proxy for those causal factors.
The weakness of observational studies is only one reason to doubt the breakfast/weight claim. A statistical analysis known as a funnel plot suggests that studies that go against the dogma might not be getting published.
Perhaps most problematic, this field’s researchers seem prone to seeing causality where it hasn’t been demonstrated. Allison’s 2013 analysis found that the authors of 21 of 42 studies that reported associations between skipping breakfast and being overweight, or eating it and having (or reaching) a healthy weight, made unsupported leaps to asserting a causal relationship in the summary of their published paper. That gives a misleading sense of what studies have found, raising doubts about the field’s integrity.
Such deceit extends to cherry-picking data, as has happened with a widely-cited 1992 study. It was a rare randomized controlled experiment, assigning obese women to eat or skip breakfast. Skippers who were assigned to breakfast, and breakfasters assigned to skip, lost more weight than women assigned to keep doing what they had been. That is, change (in either direction) produced the greatest weight loss. But 28 out of 46 later studies that cited the 1992 paper distorted its results, saying it found that breakfast-eaters lost more weight than breakfast-skippers.
All in all, an analysis this year by Christina Economos of Tufts University and her colleagues warned, the breakfast/weight claim is unwarranted due to “methodological limitations and inconsistencies,” problems with study design, and failure to consider whether a third factor explains why breakfast and healthy weight often go together.
These problems might finally be getting addressed. Negative results are being published: one of the largest studies of its kind, although observational, concluded last month that “breakfast consumption was not consistently associated with differences in BMI or overweight/obesity prevalence.” (Gut Check is especially fond of this one, from Canada, because the authors are consultants to Kellogg’s, which has an interest in promoting a cereal/healthy weight claim.)
And a randomized controlled trial from England, also published last month, found no differences in weight change between groups assigned to eat breakfast or to not eat until noon.
If the breakfast/weight claim were true, we’d have rigorous evidence of it by now; we don’t, so it likely isn’t.