Are diets with a greater environmental impact less healthy? This is a key question for nutrition policy, but previous research does not provide a clear answer. To address this, our objective here was to test whether American diets with the highest carbon footprints predicted greater population-level mortality from diet-related chronic disease than those with the lowest.

Baseline dietary recall data were combined with a database of greenhouse gases emitted in the production of foods to estimate a carbon footprint for each diet. Diets were ranked on their carbon footprints and those in the highest and lowest quintiles were studied here. Preventable Risk Integrated Model (PRIME), an epidemiological modelling software, was used to assess CVD and cancer mortality for a simulated dietary change from the highest to the lowest impact diets. The diet–mortality relationships used by PRIME came from published meta-analyses of randomised controlled trials and prospective cohort studies.

USA.

Baseline diets came from adults (n 12 865) in the nationally representative 2005–2010 National Health and Nutrition Examination Survey.

A simulated change at the population level from the highest to the lowest carbon footprint diets resulted in 23 739 (95 % CI 20 349, 27 065) fewer annual deaths from CVD and cancer. This represents a 1·83 % (95 % CI 1·57 %, 2·08 %) decrease in total deaths. About 95 % of deaths averted were from CVD.

Diets with the highest carbon footprints were associated with a greater risk of mortality than the lowest, suggesting that dietary guidance could incorporate sustainability information to reinforce health messaging.