Is 'ultraprocessed' a matter of physical processing? Chemical? Or is it actually about superstimuli, which may or may not actually experience a lot of processing? Cheese is a great example: all cheese is chemically processed milk, pasteurized cheese is also heat treated, aged cheese has more chemical changes, and finally "processed" cheese is also emulsified. Where's the line?
The article gives us "industrial food formulations made up mostly or entirely of ingredients... that are not found in a similar form and combination in nature." So a wedge of aged Parmesan is clearly ultraprocessed - even the natural lactose is gone! Somehow, I don't think that's what they're blaming obesity on.
It also references "frosted snack cakes and ready-to-eat meals from the supermarket freezer", which is one of the least helpful examples I've ever seen. Frosted snake cakes probably date back a few hundred years, but this obviously means to include Ho-Hos and exclude what you'd get at a tea shop. Ready-to-eat stir-fry can be normal stir-fry tossed in the freezer, perhaps with a stabilizer to sell it in stores. Or you could have a TV dinner with chicken nuggets, fitting exactly the same description.
The study described is interesting, and does control for some suspects like energy density. But "ultraprocessed" is not a chemical, or even a defined term. It's hard to see this as anything but preliminary work showing that there's a problem somewhere in a large set of foods, which now needs to be narrowed down.
> Unfortunately, there's no clear definition of ultraprocessed. Like is factory-made whole wheat break ultraprocessed? Canned pinto beans? Artisan cheese? Factory cheese?
The article defines "processed" and "unprocessed":
Processed foods add a few substances such as sugar, fat, and salt to natural food products, with the goal of improving preservation or sharpening taste. The category includes canned vegetables and fish, cured and salted meats, cheeses, and fermented drinks such as wine and beer.
Unprocessed foods are the edible parts of plants (such as seeds or roots or leaves) and animals (such as meat and eggs). The main processing of this food type is freezing, drying or pasteurizing to extend storage life. Salts, sugars, oils and fats are not added.
And the nature of ultraprocessed food:
Ultraprocessed foods often contain a combination of nutritive and nonnutritive sweeteners that, Small says, produces surprising metabolic effects that result in a particularly potent reinforcement effect. That is, eating them causes us to want more of these foods.
This... doesn't really define ultraprocessed. The article tries several times, but the definitions conflict with each other and with the examples given.
The first definition of ultraprocessed is "industrial food formulations made up mostly or entirely of ingredients... that are not found in a similar form and combination in nature". That includes a lot of merely 'processed' foods. A highly aged cheese will have basically no naturally-occurring ingredients left, and any bread with leavening and refined flour fails "form and combination found in nature". Similarly, the fat and salt flavors of strong cheese, or the sweetened flavors of a pastry with sugar, go beyond "sharpening taste" to creating entirely new nonnatural flavors.
We can all apply the Potter Stewart test to say that Twinkies are ultraprocessed and sharp cheddar isn't, so I don't meant to be pedantic here. But I think this loose definition points to genuinely important unresolved questions.
Where on the scale from sausage to Twinkie does the problem start?
Is it really true that an Entenmann's coffee cake causes vastly different eating habits than equal access to a home-made version? Can we isolate the difference?
Above all, which differences actually drive this? 'Ultraprocessed' is not a food additive but a loose class of recipes, and it'd be nice to reduce that to a distinction we could put on a label.
The problem probably doesn't isolate to any one additive or alteration, but we should be able to find something more concrete than a vague naturalistic appeal. The study in question has some promising work in that direction, like controlling for energy density. Outright added calories are a long-standing suspect (since we're likely to use the same amount of e.g. spaghetti sauce despite store-bought versions having far more sugar), but Hall's work suggests that the problem persists even without that. I'd very much like to see more of these studies to replicate the effect and extract a more substantive definition of 'ultraprocessed'.
It's a lot easier to define unprocessed than processed. Unprocessed is just how it is from plants or animals with little or no alteration. Processed is altered, like through mechanical action or chemical alteration, and also different things mixed together that are not in nature. The problem here is there are so many directions to go doing that, and so many degrees.
But in the end it doesn't matter, since it is becoming so overwhelmingly clear that it is best to stick with unprocessed.
> Unprocessed is just how it is from plants or animals with little or no alteration. Processed is altered, like through mechanical action or chemical alteration, and also different things mixed together that are not in nature.
That's a clear definition, but it's extremely sweeping! It's also very clearly not the definition being used by this study, which had an "unprocessed" menu that including wheat flour pasta, cooked food, frozen food, food with herbs and spices from all over the world, food that had undergone chemical changes, etc.
> But in the end it doesn't matter, since it is becoming so overwhelmingly clear that it is best to stick with unprocessed.
By your definition, this study does not support that conclusion.
>By your definition, this study does not support that conclusion
Not this study alone, but a great many studies to which this study adds one more bit of evidence. Which is how it generally goes in research.
When I said "in the end it doesn't matter" I meant having a precise definition of processed didn't matter for the question what our diet should be. I should have made that clearer.
> to which this study adds one more bit of evidence.
This study does not support your conclusion, and so adds no evidence to support that conclusion. It's not even studying the sort of diet you're talking about!
> I meant having a precise definition of processed didn't matter for the question what our diet should be
No, it really does matter. If you're trying to argue that some specific food or processing technique is bad, then a study that had all participants consume it can't be used to show that it's actually harmful. And by your definition, that's what this study did - it had everyone consume highly processed foods, and some of them had good outcomes. That undermines your conclusion; it does not support it.
Perhaps the above comment is accurate in the sense that a raw, paleo diet is healthiest for humans, or even the only truly healthy diet. (Though I've yet to see any strong studies claim that it beats a Mediterranean diet, much less a great many studies.)
But even if that's true, it's an entirely different question from "what's up with modern ultraprocessed foods?" This study was contrasting foods like "normal" pasta with canned ravioli and finding a difference. Since "everyone only eat raw food" is an unlikely and unpopular outcome, it's absolutely worth finding what's actually problematic within the enormously broad sweep of "processed".
It's much easier to define unprocesesd, definitely, but what concerns me is that processed food in general doesn't seem to be the issue. There are cuisines which have depended very heavily on processed foods for centuries, and yet the obesity question is new within the last 100 years. In particular, bread, cheese, cured meats, and pickled anything are all clearly processed. And yet historically Norwegians haven't ended up with obesity than temperate cultures eating things like unprocessed rice and fresh fish, or minimally-processed corn tortillas.
Perhaps this will just come down to modern diets having a higher percentage of processed foods and more calorie availability. Even in Norway in the winter, potatoes and some fresh game would be part of the diet. And on a ship or snowbound wheat farm with only processed food available, food might be scarce enough that people weren't freely eating to satiation.
But even wealthy urbanites with plentiful access to processed calories don't seem to have become obese with the consistency of modern people, and this study and theory aren't endorsing a raw or paleo diet. Bread, yogurt, and even pasta and butter are technically processed foods. Canned ravioli and store-bought cake are being contrasted with their home-made counterparts. And I suspect that's correct: even if a raw/paleo diet turns out to be best, 20th century ultraprocessed foods have caused new problems not seen with earlier processing. That means there's another definition worth asking about.
If the "mixed signals" theory is right, there won't be a single culprit. Sugar alcohols are pleasant and create anticipation, but don't follow through with energy. Low-fat yogurts don't trip fat-based satiation fully, and yet substituted sugar is higher-calorie per bite than fat. Perhaps the rule we'll find that degree of processing doesn't matter, but any food 'impersonating' something else is a risk for overeating.
Regardless, I'd like to know what aspect of "industrial" food is causing different outcomes than merely non-natural foods.
>Unprocessed foods are the edible parts of plants (such as seeds or roots or leaves) and animals (such as meat and eggs). The main processing of this food type is freezing, drying or pasteurizing to extend storage life.
But the example picture includes white flour, which is wheat processed to make it more palatable. The removal of the wheatgerm does increase shelf-life, but this is only a side effect. If the processing was really intended to maximize shelf-life then it would be processed into polished grains like rice or barley is, not into flour.
