Abstract
Our aim was to estimate and rank 12 food groups according to disability-adjusted life years (DALYs) from coronary heart disease (CHD), stroke, type 2 diabetes (T2D), and colorectal cancer (CRC) in 16 European countries. De novo published non-linear dose-response meta-analyses of prospective studies (based on 297 primary reports), and food consumption data from the European Food Safety Authority Comprehensive European Food Consumption Database in Exposure Assessment, and DALY estimates from the Institute for Health Metrics and Evaluation were used. By implementing disease-specific counterfactual scenarios of theoretical minimum risk exposure level (TMRELs), the proportion of DALYs attributed to 12 food groups was estimated. In addition, a novel modelling approach was developed to obtain a single (optimized) TMREL across diseases. Four scenarios were analysed (A: disease-specific TMRELs/all food-disease associations; B: disease-specific TMRELs/only significant food-disease associations; C: single TMREL/all food-disease associations; D: single TMREL/only significant food-disease associations). Suboptimal food intake was associated with the following proportions of DALYs; Scenario A (highest-estimate) and D (lowest-estimate): CHD (A: 67%, D: 52%), stroke (A: 49%, D: 30%), T2D (A: 57%, D: 51%), and CRC (A: 54%, D: 40%). Whole grains (10%) had the highest impact on DALYs, followed by nuts (7.1%), processed meat (6.4%), fruit (4.4%) and fish and legumes (4.2%) when combining all scenarios. The contribution to total DALYs of all food groups combined in the different scenarios ranged from 41-52% in Austria to 51-69% in the Czech-Republic. These findings could have important implications for planning future food-based dietary guidelines as a public health nutrition strategy.