Rationale: Several definitions of the Nordic Healthy Diet havebeen proposed, but there is no consensus on the role of dairy products. We aim to study the impact of the Nordic diet on all- cause mortality in a population of 70 year-olds, exploring different versions of a Nordic Healthy Diet Score (NHDS) with focus on the specific role of key dairy products.

Methods: Cox proportional hazard models, adjusted for potential confounders, were used to assess the prediction of all-cause mortality using two NHDS definitions (calculated either including low-fat or without dairy products). In separate Cox models, cheese intake and the combined intakes of milk, soured milk and unsweetened yoghurt were tested as well as intake of fat from these two sources.

Results: Neither of the standard versions of NHDS was associated with mortality. However, cheese intake was pro- tective of total mortality, whereas milk products (including soured milk and unsweetened yoghurt) predicted increased risk of mortality. Cheese fat intake was also inversely associated with mortality, whereas fat from the other dairy products positively predicted the outcome only when the model was adjusted by energy intake.

Conclusion: The NHDS, as previously defined, was not associated with all-cause mortality in this population of elderly[GT1] Swedes. Intake of milk, soured milk and/or unsweetened yoghurt was positively associated with all-cause mortality whereas cheese intake was inversely associated with this outcome.