What Really Matters: Ultra-Processed or Calorie-Dense?
The buzz around ultra-processed foods continues to grow. Much of that buzz swirls around the first-ever randomized and controlled study by Kevin Hall and colleagues. It’s sparking a whole range of reactions. Some offer only praise. Meanwhile, others offer diffuse criticism. Still others are asking sharp questions. But some of the most interesting questions are about ultra-processed foods versus calorie dense foods.
Disbelief and Cheering
Perhaps the only people who weren’t surprised when Hall et al found a big difference in overeating and weight gain with ultra-processed foods were the true believers. The results “do not surprise us” writes Carlos Monteiro in a letter to the editor of Cell Metabolism. Of course, he developed the NOVA system for classifying foods as ultra-processed.
On the disbelief side of the ledger, David Ludwig and colleagues find many problems with the study. “We urge caution,” they write as they cite a host of “controversies” that will need further study. Since Hall has found issues with Ludwig’s studies of the carbohydrate-insulin model, it’s not terribly surprising that Ludwig has some good questions for Hall on his study. Turnabout is fair play.
Cell Metabolism published Hall’s response today. He answered some of the questions, agreed with the need for research to address others (e.g, energy density), and expressed hope for collaboration along the way.
But Is It Really the Ultra-Processing?
The most fundamental question is whether the differences in this study of ultra-processed foods are really attributable to the ultra-processing. As a matter of fact, there was another big difference between the two arms of this study. In the ultra-processed food arm, the solid foods were more calorie-dense. They packed more calories per gram of food – 85 percent more. That difference might be enough to account for all of the differences in how much people ate and the weight they gained.
So for instance, french fries are very calorie dense. Go down to Five Guys and you’ll get fries that are made from nothing but potatoes, salt, and cooking oil. Not ultra-processed, but very caloric. At McDonald’s, you’ll get fries that have those three ingredients, plus stuff like dextrose, sodium acid pyrophosphate, and flavorings made with hydrolyzed wheat and hydrolyzed milk. In other words, an ultra-processed version.
At Five Guys, a serving is 953 calories in a 411 gram serving – 2.3 calories/gram. At McDonald’s, you’ll get 230 calories in a 71 gram serving – 3.2 calories per gram. In either case, that’s a lot of calories per gram. Eating too much of either of them would likely to lead to weight gain.
A Wide Net
As a label and a tool for identifying potentially problematic foods, “ultra-processed” is pretty appealing. But it casts a very wide and sometimes fuzzy net. Clearly, Hall’s study points to some problems with these foods. However, it might be that most of the effect assigned to ultra-processed foods comes from these foods being calorie-dense.
Click here for Hall’s paper, here for Monteiro’s letter, and here for Ludwig’s letter. Hall’s response is online here. For more on ultra-processed foods, click here. Finally, click here for more on the role of cheap, energy-dense foods in obesity.
French Fries in Paris, photograph © Jim Larrison / flickr
Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.
July 1, 2019
July 01, 2019 at 6:54 am, Mary-jo said:
Thank you for the NOVA classification table, Ted! very valid the points made here. As I said in a previous post, I’d love if Kevin would do a similar study comparing the ‘unprocessed’ foods with say, haute cuisine — fresh food, prepared wonderfully, but often calorically dense. Don’t know if the NIH could spring for a Michelin star chef. By the way, that makes me question, what category haute cuisine fits in being fresh, unprocessed foods made with several processed ingredients?
July 01, 2019 at 9:58 am, Ted said:
Very good questions, Mary-Jo. Regarding the calorie density of “fine dining,” I think you’ll find this article very interesting: https://doi.org/10.1038/ijo.2017.18
July 01, 2019 at 8:45 pm, Jennie Brand-Miller said:
Can we class any food with two or more food additives as ultra-processed? And if we remove food additives, are we less prone to obesity? And the calorie density of nuts is one of the highest on record, so should we avoid them. And what about the Mediterranean diet…wouldn’t it be one of the highest in energy density? And yet, one of the best patterns for a host of good outcomes?
July 02, 2019 at 4:29 am, Ted said:
Excellent questions, Jennie. It’s important to find the elusive, definitive answers to them.
July 02, 2019 at 4:14 pm, David Ludwig said:
On behalf of my coauthors, I address a few areas of confusion in Hall’s recent letter, as linked below. On the broader point, let’s distinguish between:
1. Vigorous debate, which is the heart of the scientific process. We need more of it. Totally fair to attack the idea, argument, or (as professionally appropriate) the data.
2. Ad hominem attack. We need less of it, especially on social media. We can always point to bias among those in an opposing camp. But that’s a slippery slope, leading to echo chambers and loss of potential common ground.
3. An end to the “perpetual diet wars”? We’ll need long-term clinical trials before we can or should end debate about potential macronutrients effects on obesity and related chronic disease.
https://www.cell.com/cell-metabolism/fulltext/S1550-4131(19)30318-3#comment-4524341876
July 05, 2019 at 2:27 pm, Doina Kulick, MD said:
“ At Five Guys, a serving is 953 calories in a 411 gram serving – 2.3 calories/gram. At McDonald’s, you’ll get 230 calories in a 71 gram serving – 3.2 calories per gram. In either case, that’s a lot of calories per gram. Eating too much of either of them would likely to lead to weight gain.“
Thus eating the ultra-processed McDonald’s fries provides an extra 0.9 calories per gram , so that’s about 40% more calories per gram than the “home made” type fries. This tells me ultra-processed foods are a lot to blame for, and Dr. Hall maybe right. Maybe the NOVA classification is not perfect, but is a good beginning. Many of the additives used in ultra processed foods were never studied enough to evaluate their effects on metabolism. Take for example a recent study that shows the negative effect on insulin resistance of calcium propionate ( additive community used in ultra-processed baked goods).
July 07, 2019 at 7:25 pm, Paul Ernsberger, Ph.D. said:
Dr. Kulick is right. We need to study individual additives in processed foods rather than make arbitrary classifications –canned vegetables are processed but frozen or not? Salting nuts makes them processed?
Proprioate is synthesized by gut bacteria and probably beneficial. When fed to mice, antioxidant preservatives extend lifespan.
On the negative side of the leger, the natural emulsifier lecithin has been linked to atherosclerosis.