More Bad Science: Cereal for Breakfast

Your friends in the grain business have been busy promoting the latest study that supposedly shows that eating "whole grain breakfast cereal" prevents heart failure.

Whole Grains Do a Heart Good

As reported in U.S. News and World Report: "Compared to those who ate no whole-grain cereal, men who consumed 2 to 6 servings per week saw their risk of heart failure fall by 21 percent, while those who ate 7 or more servings per week reaped a 29 percent reduction in risk, the researchers reported in the Oct. 22 issue of the Archives of Internal Medicine."

What's wrong with this study?

Well, for starters, we know nothing else about the weight, diets, ethnic heritage, and lifestyle of these people who ate whole grain cereal for breakfast, but it is very likely that they ate those breakfasts because they have been touted as "healthy" and that eating the cereal rather than causing the outcome was a marker for a lifestyle high in exercise, weight control, ethnic background, and other behaviors likely to impact health.

Using the same logic, you might be able to prove that people who owned a specific brand of running shoe had a lower "risk" of heart failure. Did owning the running shoe cause the health outcome? No. If you buy that running shoe and wear it while watching TV you won't improve your health.

The other big problem here, as usual is that you are comparing people eating one high carb breakfast with others eating an even higher carb breakfast. So yes, I'm willing to believe that eating granola for breakfast might be healthier than eating Sugar Frosted Flakes, but I'd really like to see what happens when you compare people eating the grain breakfasts with those eating NO carbs at breakfast.

Most importantly, the study did not separate out people with Diabetes from those with normal glucose tolerance. We know for a fact that anyone with Type 2 diabetes who eats a grain cereal for breakfast is likely to see their highest blood sugar of the day after breakfast, because of the natural increase in insulin resistance we all experience in the morning. And we also know that cumulative exposure to high blood sugar leads to bad health outcomes. In addition, people with diabetes are more likely to experience heart attacks--not being measured here--rather than heart failure, which usually develops in people who survive heart attacks.

This raises an important point: The condition being studied was "heart failure"--a condition in which people suffer weakened heart muscles--rather than heart attack, and that the measurement being used was "risk of" which is very different from "incidence of".

The Risk statistic is used to amplify very slight differences between the incidence of a condition in large groups of people, and is almost always used to make a very small effect look more significant. If you reduce the number of people in a group of 100,000 who get a heart attack from 10 to 8 you've made a huge difference in risk but a very small difference in incidence. But if you are trying to promote a drug or product, citing risk rather than incidence makes your product look like it makes a huge difference when it doesn't.

These studies seem to be more of the same kind of "let's prove what we believe to be true by ignoring rigorous analysis of the data" research that Gary Taubes highlights so brilliantly in his book, Good Calories, Bad Calories.

By the way, the nutritionists in the article recommend that you eat your high carb whole grain cereal with high carb low fat skim milk and fruit. More proof, if you needed it, that religious beliefs can not be challenged by logically supported argument. Sadly, no matter how densely documented the truth might be (and Taubes documents it to the point where it exhausts even ME, which makes me wonder exactly who his target reader must be!) the belief in the value of the high carb/low fat diet will not go away no matter how much more rigorous research shows it to be a flop.

Meanwhile, if you have diabetes and are tempted to breakfast on granola with skim milk and bananas, by all means do so. Just be sure to measure your blood sugar at 1 hour and 2 hours after eating. If you can get normal blood sugar values at those times (under 140 at 1 hour, under 120 at 2 at a bare minimum--under 100 at 2 hours is better) keep eating like that. If you can't, consider switching to a low carb breakfast of ground flax, protein powder pancakes, eggs, or meat.