Can processed foods explain our obesity epidemic?
An ambitious and meticulously controlled trial from the NIH and Dr. Kevin Hall may shed light on the question of why ultra-processed foods are so bad for us.
On the one hand, some may see this study as a no-brainer. Ultra-processed foods are bad, they have contributed to our obesity and diabetes epidemics, and we should focus on eating real foods. That isn’t earth shattering.
Yet, on the other hand, others are interpreting this study to “prove” that the type of calories (carbohydrates, fats, or proteins) doesn’t matter as long as we are eating real foods. That would be remarkable if it was what the study showed, but it did not.
Thus the interpretation lies somewhere in between.
First, let’s look at the details of this small but impressive study. Researchers enrolled 20 volunteers (average age 31 without metabolic disease at baseline) willing to live in a controlled NIH lab for four-weeks where every meal and snack was provided for them. They were randomized to either an ultra-processed food group or an unprocessed (really it should be “less processed”) food group. After two weeks, they switched groups. Importantly, the food groups were matched for the amount of fat (37%), protein (14%), carbohydrates (48%), sugar and salt. They also wanted the fiber to be the same but since the ultra-processed food had much less fiber, they had to add it to the drinks to make it equal. Last, and this is crucial, the subjects were allowed to eat as much as they wanted across all meals.
Many of the ultra-processed foods were fairly standard, westernized fare such as Honey Nut Cheerios or flavored yogurt. Most came from cans and boxes and had more than five ingredients, whereas the less processed foods were much closer to basic ingredients. (There is an excellent visual depiction of the meals in the study and in The New York Times article.)
Investigators found that those eating the ultra-processed meals ate on average 500 more calories per day, and after just two weeks, participants gained an average of 2 pounds when assigned to the ultra-processed cohort. Those eating the less processed meals, on the other hand, lost an average of 2 pounds over two weeks. Thus, study authors concluded that the ultra-processed foods somehow stimulated the subjects to eat more. Although they have theories as to why that was the case, the study did not prove a mechanism. They did conclude, however, that it did not appear to be due to macronutrient make up since the percentage of carbs, fats and proteins were held roughly equal.
Instead, they postulated the cause may be hormonal. When eating the less processed foods, subjects had higher levels of appetite suppressing hormones (PPY) and lower levels of hunger signaling hormones (ghrelin), and the reverse was the case when subjects were eating ultra-processed foods.
Or it may have to do with the speed of eating. The ultra-processed food group ate much faster and therefore may have continued to eat before they experienced the gut-to-brain connection signaling fullness.
Others suggest it may have been the fiber in the less-processed food making subjects feel full.
All of these are potentially viable hypothesis, and they all need to be put into perspective.
I fear this study may fall into the same trap as prior whole grain and fiber studies. Studies that looked at a high-carb diets with high fiber versus low fiber or whole grains versus refined grains clearly show more fiber and whole grains are better. But that does not mean they are better for everyone and are by definition “healthy.” Instead, they are simply better than the comparison foods. We are missing the other important comparisons such as a diet with whole-grain compared to a low-carb diet with vegetables instead of whole grains or a low-carb diet with plenty of fiber compared to a low-carb diet with little fiber.
I would argue the same holds true for this study. It is clear that when comparing these two high-carbohydrate diets (composed of 48% carbohydrates), the less-processed foods were better even though they had the same percentages of macronutrients as the processed meals. In no way does this prove that macronutrients do not matter. Once again, we are missing the comparison group of a low-carbohydrate real food diet.
I don’t mean to detract from the impressive work done by Dr. Hall and colleagues; however, we do have to keep the interpretation of the results in-line with what was compared in the study. Processed foods undoubtedly have contributed to our obesity epidemic, but that does not mean there aren’t other factors, such as macronutrient composition, that we need to consider as well.
When eating a high-carbohydrate diet, avoiding ultra-processed foods and eating real foods with natural fiber is a good thing. It keeps you fuller and helps you eat less. Interestingly, those are the same findings on an LCHF diet compared to higher carbohydrate diets. Maybe that is the next study Dr. Hall will take on? (Hint, hint….)
Thanks for reading,
Bret Scher, MD, FACC