Healthy Nordic diet and cardiovascular disease

During the last five years, there has been an increased interest in the potential health benefits of a healthy Nordic diet. A healthy Nordic diet can be defined as a prudent diet that accord with current dietary guidelines and includes foods that are locally produced or traditionally used in the Nordic countries, for example oily fish such as salmon and herring, wholegrain cereals including rye, barley and oats, berries (e.g. blueberries) and fruits (e.g. apples), vegetables, root vegetables and legumes 1. Controlled trials conducted in various Nordic populations have shown that healthy Nordic diets cause reduction in several key cardiovascular disease (CVD) risk factors 1-6. Overall, the reductions in risk factors were of similar magnitude as those previously reported for, for example, Mediterranean and dietary approach to stop hypertension (DASH) diets 1. In addition, when healthy Nordic diets are provided ad libitum to overweight and obese subjects, they produce weight loss, which probably mediated the reduction in some, but not all, risk factors 1, 3.These results are clinically relevant considering that high blood pressure, cholesterol and glucose could explain around half of the excess risk of coronary heart disease associated with overweight and obesity 7. Also, an isocaloric healthy Nordic diet advised for up to 6 months recently caused downregulation of pro-inflammatory gene expression in adipose tissue 8. In addition to feeding trials using healthy Nordic diets, two large prospective cohort studies conducted in Denmark and Sweden have reported reduced risk of all-cause mortality in people with high adherence to a self-reported Healthy Nordic Food Index (HNFI) 9, 10, an index that may to some extent capture habitual adherence to a healthy Nordic diet. Hitherto, there are, however, no cohort studies that have investigated the link between HNFI scoring and CVD risk, and therefore, a recent large cohort study in Swedish women is of interest 11. Roswall and coworkers are indeed the first to investigate this association using the HNFI 11. Self-reported adherence to the HNFI, using a food frequency questionnaire (FFQ), did, however, not detect a significant association with CVD incidence in that population 11. Considering the above-mentioned beneficial cardiovascular effects caused by healthy Nordic diets in controlled trials, together with the inverse association between HNFI and all-cause mortality, these results may be somewhat surprising. Of note, there was, however, a modest but significantly lower CVD risk at high scoring of HNFI observed amongst former smokers, but as indicated by the authors, this finding may be due to chance and should be interpreted cautiously. The lack of association between HFNI and CVD risk in the Swedish women should be viewed in the light of the limitations of the study including the observational design as such, using a single FFQ assessment to correlate with a composite of various CVD outcomes derived from national disease registries. It is especially important to note that there are several relevant differences between the observational study by Roswall et al., and the controlled trials that need to be considered before discussing study limitations and interpretations. Consequently, the estimated food intake data using the FFQ-based HNFI is not comparable to the healthy Nordic diet used in the trials. Several of these differences have been pointed out by the authors and involve differences in study design (interventional versus observational), but it is especially noteworthy that these studies investigated quite different diets with different food components included 11. In addition, HNFI is only partly reflecting the healthy Nordic diet reported to have benefits on CVD risk factors. For example, in all healthy Nordic diets that showed beneficial effects in trials, rapeseed oil (also known as canola oil) and wholegrain rye were two key components that were missing in the HNFI. This is relevant as improved blood lipids by healthy Nordic diets are to a large extent driven by changes in dietary fat quality 12, and rapeseed oil has pronounced favourable effects on blood lipid profile when compared with, for example, butter 13. Rapeseed-oil-based margarine was also the main food provided in the successful Lyon heart trial which demonstrated a ~70% reduced risk of CVD mortality in patients with myocardial infarction 14. Wholegrain rye is a key factor in healthy Nordic diets and has been ascribed several favourable cardiometabolic health effects 15-17. In a large Danish cohort, wholegrain rye bread was indeed the food item in HNFI most consistently linked to reduced all-cause mortality in men 9. In addition to this, the lack of differentiation between wholegrain crispbread from nonwholegrain crispbread by the FFQ by Roswall et al. may have resulted in inadequate classification of wholegrains and, consequently, could have biased the association between HNFI and CVD. In the intervention trials showing favourable metabolic effects, wholegrains from rye crispbreads and rye bread comprised key foods of the healthy Nordic diets 1-3, 14, 15. Another important difference between the HNFI used in the Swedish study and the healthy Nordic diets used in the trials was that in the latter, they were all designed to follow the Nordic nutrition recommendations. Accordingly, the trials tested a healthy Nordic diet not only aiming at increasing healthy foods (e.g. wholegrain rye, fish, rapeseed oil and legumes and fruits), but also decrease unhealthy foods (e.g. sweets, added sugars, red meats, trans fats, processed foods) 15. Thus, the HNFI only captures a part of a healthy Nordic diet, rather than examining the role of a healthy Nordic diet that is advised by health authorities and dietary recommendations. A previous report from this Swedish cohort suggests additional complexity, as it showed that women scoring higher on HNFI also were more physically active and smoked less, factors that were, however, adjusted for in this study 18, but those women also had higher energy intake (in contrast to trials where the energy intake was lower than controls) and reported higher intakes of food items that are not included in a healthy Nordic diet, that is processed and red meats and sweets 16. It, however, remains unclear whether a higher intake of red meat and sweets amongst women scoring high in HNFI might counteract the potential benefits of the food components in HNFI in this population. In the trials on healthy Nordic diets, red meat as well as added sugars was reduced and consumed in low amounts 17. Furthermore, if compared with the Mediterranean diet score, the HNFI includes only six food items, whereas the Mediterranean score includes nine items as well as negative scoring for food items with potential adverse relationships with health. Thus, a direct comparison between these two dietary patterns using different types of scoring is not meaningful with regard to associations with CVD risk. The strengths of the study by Roswall et al. included the large study population with a large number of CVD cases derived from validated and fairly reliable Swedish registers, as well as state-of-the-art statistical analyses including appropriate adjustment for potentially important confounding lifestyle factors. Residual confounding is, however, a potential problem in most observational studies, and unfortunately, all confounders adjusted for in the study by Roswall et al. were self-reported including smoking, physical activity and BMI. Self-reported BMI is especially relevant as underestimation of BMI in overweight and obese individuals is common and may result in an inadequate adjustment for obesity. Improved weight control does seem to be an important mediator of some of the favourable cardiometabolic effects of a healthy Nordic diet 1. Insufficient or imprecise adjustment for BMI therefore warrants cautious interpretation of the results, considering the strong link between BMI and CVD risk 7. In summary, the lack of association between the HNFI and CVD in a population of Swedish women should be interpreted with caution, partly as it does not accord with several randomized trials that suggest causal effects by healthy Nordic diets on CVD risk factors. Due to important differences in study designs (observational vs. interventional) as well as differences between self-reported HNFI and foods advised in trials, comparison of these results should be avoided. Similarly, the HNFI cannot be directly compared with Mediterranean diet scores with regard to CVD risk, as the latter also contains foods to be avoided in a healthy diet. For these reasons, the HNFI should be used with the above-mentioned considerations in mind. On the other hand, development of new indices that capture the whole healthy Nordic diet could be useful in future studies. Nevertheless, the results reported by Roswall et al. are of interest and should not be neglected, as they do motivate new studies in both sexes and using both an observational and interventional study design to further explore the potential role of a healthy Nordic diet in CVD prevention. The author has no conflict of interest to declare.