Chronic disease risk climbs with every daily serving of UPFs

Clinical takeaway: Counsel gradual reduction of ultra-processed intake toward 10% to 20% of daily energy, favoring swaps over restriction alone. 

Primary care rarely treats one disease at a time. Diabetes often arrives with depression, cardiovascular disease with obesity, and each added condition multiplies visits, medications, and cost. That pattern has pushed attention toward exposures the conditions might share. Diet sits high on the list: ultra-processed foods supply roughly half the calories US adults take in each day, a share that has barely moved in two decades. 

Whether diet helps explain disease clustering has been hard to pin down. Individual studies have linked ultra-processed intake to single outcomes. Each answered its own narrow question, and none could say whether any consumption level is safe or where risk begins to build. A new pooled analysis of prospective cohort evidence takes on both questions at once, measuring how risk scales with intake across the span of conditions a generalist manages, and its answer is not reassuring. 

Compared with the lowest ultra-processed intake, the highest was associated with risks: 31% higher for digestive diseases, 27% for depression or anxiety, 24% for cardiovascular events, 24% for metabolic syndrome or diabetes, 23% for overweight or obesity, 18% for death from any cause, and 12% for cancer. In the cardiovascular cohorts, the highest consumers took in a median of roughly 59 servings a week against almost 10 in the lowest. Hypertension was the lone exception: the highest-versus-lowest comparison fell short of significance. 

Each additional two daily servings was linked to 14% higher risk of cardiovascular events, 4% higher for cancer, 2% higher for metabolic syndrome or diabetes, and 3% higher for death from any cause. Measured as a share of the diet, each 10% increase in the ultra-processed proportion was associated with a 16% higher risk of metabolic syndrome or diabetes. The researchers went looking for a safe consumption level and did not find one. 

The analysis pooled 51 prospective cohort studies following 8,819,894 adults for two to 32 years across the Americas, Europe, Asia, and Oceania. Intake came from food frequency questionnaires, 24-hour recalls, or dietary histories, standardized to 50-gram servings. Certainty ran low for most outcomes, moderate for overweight or obesity, and very low for hypertension. 

NOVA classifies foods by degree of processing rather than nutrition, grouping wholegrain bread and unsweetened yogurt with soda and processed meat. Pooling foods so unlikely to act alike may dilute the harms of the most damaging subtypes, and subtype-specific data would tell clinicians which substitutions matter most. Policy is already moving ahead without that sorting: Chile's front-of-package warning labels were associated with sustained drops in household purchases of sugar and sodium, and Singapore now grades beverages by sugar content. In UK trials, warning labels steered shoppers away from unhealthier choices, population-level movement that could reshape diets before a patient reaches the exam room. 

"Given that UPF related chronic diseases often cluster as multimorbidity, reducing consumption of UPFs and replacing them with minimally processed alternatives may be relevant for chronic disease prevention and primary care management," the researchers conclude. 

Source: Liu X, et al. (2026 Sep 22) Fam Med Community Health. Ultra-processed food consumption and risk of multiple chronic diseases among 8,819,894 adults from 51 prospective cohorts: a dose-response meta-analysis