Following the revision of the French dietary guidelines in 2017, the Programme National Nutrition Santé – guidelines score (PNNS-GS), built upon previous recommendations released in 2001, needed to be updated. This cross-sectional study thus aimed to develop and validate the PNNS-GS2, a predefined food-based dietary index based on the 2017-revised French nutritional guidelines. A total of 80 965 participants recruited among French adults (≥18 years old) in the NutriNet-Santé web-based prospective cohort were included. Collected data included repeated 24 h-dietary records over a 2-year period, sociodemographic and, for 16 938 subjects, clinical and biological data. Weighting and cut-offs of the PNNS-GS2 components were collegially arbitrated by nutrition experts who participated in the 2017 revision of the guidelines. Sociodemographic, nutritional and clinical and biological factors were investigated according to quintiles (Q) of PNNS-GS2 (theoretical ranging −17 to +13·5). Mean PNNS-GS2 was 2·1 (sd 3·1) in women and −0·3 (sd 3·6) in men. Higher PNNS-GS2 (higher adherence to 2017 dietary guidelines) was positively associated with (mean difference between Q5 and Q1 in women/men) age (+8·4/+4·7 years), education (+3·9/+7·4 % of university level), physical activity (+13·3/+3·5 % of ≥60 min/d) and non-smoking (+9·7/+13·7 %), and was negatively associated with mean blood pressure (−3·0/−2·8 mmHg), plasma LDL-cholesterol (−0·07/−0·06 g/l) and TAG (−0·10/−0·16 g/l) concentrations. Higher PNNS-GS2 was also associated with higher intake of favourable nutrients, e.g. n-3 PUFA (+0·2/+0·2 % of energy intake), fibres (+8·7/+10·7 g) and vitamin C (+36·6/+43·8 mg). Associations between PNNS-GS2 and sociodemographic and nutritional factors arguing for its validation are coherent. Further studies are needed to evaluate its association with mortality and morbidity.

We aimed to examine the association between the Alternative Healthy Eating Index updated in 2010 (AHEI-2010), the Dietary Inflammatory Index (DIITM) and risk of mortality in the Whitehall II study. We also conducted a meta-analysis on the DII-based results from previous studies to summarise the overall evidence. Data on dietary behaviour assessed by self-administered repeated FFQ and on mortality status were available for 7627 participants from the Whitehall II cohort. Cox proportional hazards regression models were performed to assess the association between cumulative average of AHEI-2010 and DII scores and mortality risk. During 22 years of follow-up, 1001 participants died (450 from cancer, 264 from CVD). Both AHEI-2010 (mean=48·7 (sd 10·0)) and DII (mean=0·37 (sd 1·41)) were associated with all-cause mortality. The fully adjusted hazard ratio (HR) per sd, were 0·82; 95 % CI 0·76, 0·88 for AHEI-2010 and 1·18; 95 % CI 1·08, 1·29 for DII. Significant associations were also observed with cardiovascular and cancer mortality risk. For DII, a meta-analysis (using fixed effects) from this and four previous studies showed a positive association of DII score with all-cause (HR=1·04; 95 % CI 1·03, 1·05, 28 891deaths), cardiovascular (HR=1·05; 95 % CI 1·03, 1·07, 10 424 deaths) and cancer mortality (HR=1·05; 95 % CI 1·03, 1·07, n 8269).The present study confirms the validity to assess overall diet through AHEI-2010 and DII in the Whitehall II cohort and highlights the importance of considering diet indices related to inflammation when evaluating all-cause, cardiovascular and cancer mortality risk.

Tools, called ‘diet/dietary quality indices’, evaluate the level of adherence to a specified pattern or a set of recommendations in populations. Yet, there are no review studies providing unanimous comprehensive results of dietary indices on obesity. We reviewed observational studies, focusing on the association of diet quality indices with general obesity or abdominal obesity in adults. We systematically conducted a search in all English language publications available on MEDLINE, ISI Web of Science and Embase between January 1990 and January 2016. Among the wide variety of indices and weight-derived variables, studies with dietary-guideline-based indices and mean changes for weight gain or OR for general obesity and abdominal obesity were selected. From a total of 479 articles, thirty-four studies were selected for the current review, ten of which had prospective designs and twenty-six had cross-sectional designs. Associations of weight status with the original Healthy Eating Index (HEI) and other versions of the HEI including alternative HEI, HEI-2005 and HEI-05 were examined in thirteen studies, with ten studies revealing significant associations. The HEI was a better general obesity predictor in men than in women. Diet scores lacked efficacy in assessing overall diet quality and demonstrated no significant findings in developing countries, in comparison with US populations. In addition, indices based on dietary diversity scores were directly associated with weight gain. Despite the insufficient evidence to draw definitive conclusions about the relation between dietary indices and obesity, HEI was found to be inversely associated with obesity and diversity-based indices were positively associated with obesity.