Diet and Health: The Need for New and Reliable Approaches

Andrew Mente; Mahshid Dehghan; Salim Yusuf


Eur Heart J. 2020;41(28):2641-2644. 

In This Article

Abstract and Introduction


Nutrition has long been related to health. Until the mid-1950s, the predominant concern in most countries was undernutrition but, since the 1960s, the focus has shifted towards the potential harmful effects of excess consumption of certain foods, especially animal fats.[1] Given the large increases in cardiovascular disease (CVD) in the mid- and late 1950s in Western countries, the focus has been on CVD with specific emphasis on coronary heart disease (CHD). However, apart from the apparent protective effects of fruit and vegetables for stroke or CHD, the associations of other aspects of diet with CVD have been inconsistent. Why is this?

Since the 1960s, conventional thinking on nutrition and CHD was fuelled by epidemiological studies such as the Seven Countries Study.[2] Although influential, this study had significant methodological problems including selection biases, limited data on diet (recorded in only ~5% of the sample), using non-standardized and non-validated methods, and inconsistent methods of follow-up. Subsequently better designed large prospective cohort studies were initiated, mostly in the USA and Europe. Given that in these countries, diseases of 'undernutrition' (e.g. vitamin deficiencies or protein calorie malnutrition) were uncommon by the 1960s, the focus was largely on whether overconsumption of specific foods such as animal fats (e.g. red meat which contains saturated fats, or eggs, a major source of dietary cholesterol) was harmful. However, these studies ignored the fact that some of these foods also contained potentially beneficial nutrients such as high-quality proteins, minerals, folate, vitamin D, vitamin Bs, and monounsaturated fats. Nevertheless, because of 'theoretical' concerns that these foods (as well as dairy products) may adversely affect blood lipids, guidelines have recommended limiting meat, eggs, or dairy (especially whole fat) consumption. However, these recommendations did not consider that there may be a critical need for many natural foods and that a minimal amount of intake may be needed.

Data that did not fit in with the conventional, linear thinking (increased animal fats causes CHD irrespective of levels of intake) were largely ignored or criticized. Randomized trials, animal studies, and mechanistic studies have shown that dietary cholesterol has little effect on the entire blood lipid pattern (e.g. the total cholesterol/HDL cholesterol ratio or the apolipoprotein B/apolipoprotein A1 raio) and the impact on other markers (e.g. glucose or insulin) was ignored.[3,4] Most cohort studies show that egg consumption has a neutral effect on mortality and CVD.[5–7] Further, several recent reviews showed that the impact of even relatively large differences in saturated fat intake in the population (e.g. the top vs. the bottom quintile of intakes) on LDL cholesterol is very small[8,9] and that moderate (e.g. three servings per week) red meat consumption is not harmful.[10–13] This work has gradually changed conventional positions. For instance, the 2015 US guidelines no longer include a limit for dietary cholesterol, and recommend eggs as part of a healthy diet.[14,15]

In a recent analysis published in this issue of the European Heart Journal, Tong et al.[16] examine the associations of 21 different foods with stroke in the prospective EPIC (European Prospective Investigation into Cancer and Nutrition) study. The authors found that higher intake of fruits and vegetables and dairy products is associated with lower risk of total stroke and ischaemic stroke, a finding consistent with several cohort studies. Higher dairy intake (especially whole-fat or fermented dairy foods) has been shown to reduce blood pressure in randomized trials and so a lower risk of strokes from dairy consumption is plausible.

In contrast, in EPIC, higher egg consumption was associated with a higher risk of haemorrhagic stroke, but this is not supported by several large studies[17–19] and so the findings might be due to residual confounding or chance. Red meat showed a neutral effect on stroke after adjusting for other diet components, a finding consistent with recent reviews.[10–13] Taken together, moderate consumption of animal-based foods including red meat, dairy, and eggs is probably not harmful.

The EPIC study is among the best that can be performed with current methods of nutrition epidemiology for chronic diseases. The strengths of EPIC include its large size (400 000 people), prospective design with a long follow-up, the use of validated FFQs (food frequency questionnaires), and adjustment for known confounders. But are these methods sufficiently robust to reliably detect the sorts of plausible moderate benefits or harms (e.g. a 10–20% lower or higher risk) of most foods on health?