How much does ultra-processed food increase your risk of dying?

People who eat the most ultra-processed food die earlier than people who eat the least. How large that difference is depends strongly on which study you consult. An overarching review of meta-analyses estimates the additional mortality risk at 21%. A large American cohort study with more than 30 years of follow-up found only 4%. And for each additional serving of 50 grams per day of ultra-processed food, the risk rises by roughly 2%. That wide margin is not a calculation error, but an honest reflection of how much uncertainty still exists.

The sharpest link to early death lies with processed meat products: sausage, ham and ready-made meat products show the strongest and most consistent associations in that same cohort study. For cancer mortality and cardiovascular mortality, that large study found no elevated risk at all; the strongest link was with 'other causes' such as respiratory diseases and neurodegenerative conditions, with a 9% higher risk. Other meta-analyses do report higher mortality from heart disease, but the quality of that evidence has been rated 'very low' by the researchers themselves.

A major question mark hangs over all these figures: do the associations disappear once you also account for the overall dietary pattern? In the Harvard cohort study, that was indeed the case among people with an otherwise healthy diet. It therefore cannot be ruled out that 'ultra-processed' is simply a proxy for an unhealthy dietary pattern in general, and that the processing itself is not the direct culprit. Causality has not been proven anywhere.

What the studies do firmly support is the link with type 2 diabetes. For every 10% increase in ultra-processed food in your daily diet, the risk of diabetes rises by 12%. That association is consistent across multiple analyses and has the strongest quality of evidence of all outcomes examined. There are also indications of higher rates of anxiety symptoms and obesity, but the quality of evidence for those is low and reverse causality -- people with complaints eating more ultra-processed food -- cannot be ruled out.