Ultra-processed foods

The umbrella review classed the evidence on all-cause mortality as highly suggestive (class II), with a relative risk (the risk in one group relative to another) of 1.21 and a 95% confidence interval (the range the true value probably lies in) of 1.15 to 1.27, while grading its certainty low [1]. Cardiovascular disease mortality showed a stronger association, a relative risk of 1.50, rated convincing (class I) though of very low GRADE certainty [1].

A 2025 meta-analysis (a study that pools the results of earlier studies) of prospective studies (studies that follow people forward in time) found a hazard ratio (a measure of how much more often an event happens in one group) of 1.15 for the highest compared with the lowest intake, with a 95% confidence interval of 1.09 to 1.22 [2]. Risk rose in a straight line with the share of the diet: each 10% increment carried a hazard ratio of 1.10 [2]. A 2026 analysis found a hazard ratio of 1.03 for all-cause mortality per additional 100 g a day [3].

The evidence is observational, tracking existing diets rather than assigning them. The category is broad, covering foods as different as flavoured yogurt and soft drinks, and the authors graded the certainty for all-cause mortality low. People who eat more ultra-processed food also tend to differ in income, smoking and activity. Overlap with fried foods and processed meat is large, and the wider context is described under dietary patterns.