Five landmark randomized trials of ultra-processed foods did not cleanly isolate processing effects, a new Science analysis found, because the diets often differed in calorie density, fiber, saturated fat, sodium and texture.
Those gaps matter because the ultra-processed meals were frequently softer and more energy-dense—traits already known to speed eating and raise intake—so harms blamed on processing may reflect other nutritional factors.
The review also argues many trials made an "extreme contrast" by pitting NOVA group 4 foods against mostly unprocessed group 1 diets, rather than testing whether ultra-processing is uniquely harmful versus other processed foods.
Kevin Hall, whose trial helped shape the field, said the critique offers plausible alternative explanations but should lead to better experiments, not abandonment of the question.
The findings could ripple beyond academia as U.S. officials and food-labeling programs increasingly use the ultra-processed concept to guide policy, reformulation and consumer choices.