Introduction

Although Nutrient Reference Values (NRVs) are determined on the basis of needs for sustenance and avoidance of deficiency disease, it is obviously most beneficial if nutrient intakes are also compatible with intakes that may reduce chronic disease risk. Whilst there is an extensive and growing data base related to diet and chronic disease risk in humans, the population methodologies generally employed have a number of limitations in relation to identifying a specific level of intake that is optimal for reducing the risk of chronic disease.

The major tools used for associating food intake patterns and specific nutrients with chronic disease risk are:

  • Ecologic studies in which national per capita food intakes are correlated with national health statistics relating to the incidence, prevalence and mortality of diseases.

  • Case-control studies in which food intake patterns in individuals who have contracted a disease under study are gathered retrospectively and compared with the food intakes of appropriately chosen individuals who do not have the disease.

  • Cohort studies in which food intake patterns in many study subjects are recorded while they are all free of the disease(s) of interest – and after an appropriate efflux of time (usually many years) the dietary patterns of those who develop disease(s) are compared with those who are still disease free.

  • Intervention studies (randomised controlled trials, RCTs), in which subjects at (high) risk of the disease under study are randomly allocated to either a modified dietary/supplemental regime or a control regime – and the two study groups are compared with respect to their subsequent disease incidence or progression.

Evidence from individual studies may be supported by additional meta-analyses of a number of related studies that increase the power to detect associations and/or by systematic review of individual studies.

Ecologic studies provide the weakest evidence, since a diversity of other explanations may account for any observed association. The striking association between per capita consumption of fat and breast cancer mortality in women is a salutary example of a hypothesis generated by an ecologic study that could not be sustained in subsequent analytic studies. A case-control study is the most popular analytic tool for investigating chronic disease aetiology as it is easier to undertake (in terms of finance and time) in than cohort studies or RCTs, but it is extremely vulnerable to biases arising from either inappropriate selection of control subjects, or from selective recall bias by case subjects of the foods they ate prior to the diagnosis of their condition.

While cohort studies are far less vulnerable to the problems of the case-control study, the huge numbers of study subjects required to ensure the future accrual of a sufficient number of ‘cases’ and the time required for the disease to develop has meant that very few cohort studies have ever been conducted. Finally, in a few instances, there has been sufficient confidence in the disease-preventive capacity of a specific food component or micronutrient to initiate an intervention trial. Modelled on the RCT, an intervention study provides the most reliable information for confirming a direct causal relationship between a dietary component and a chronic disease outcome. However, most interventions that have been undertaken to date have, for pragmatic reasons, involved use of supplements rather than dietary change. In many instances, supplement mixes (eg of antioxidant micronutrients) have been used and there is some evidence that mixes of micronutrients may be more effective than single nutrient approaches.

It is salutary to note, however, that at least in the case of the nutrient ß-carotene, very promising data from both case-control and cohort studies indicating a protective effect for certain cancers was not confirmed in subsequent interventions studies and, indeed, indicated some potential harm. Earlier work had shown that dietary ß-carotene could prevent DNA-damaging steps in the genesis of cancer, but intervention studies involving its use in people at high risk of cancer of the lung, colon and cervix either found no effect, or were discontinued due to the apparent impetus to progression of cancer in the study subjects. Recent work has shown similar concerns in relation to other nutrients given in supplemental form.

With these provisos in mind, there is some evidence that a range of nutrients could have benefits in chronic disease aetiology at levels above the RDI or AI. This is discussed in detail in the publications of the Food and Nutrition Board: Institute of Medicine as part of the reviews of the US:Canadian DRIs, notably those published in 1998, 2000 and 2002. It is not the purpose of this NRV review to revisit this extensive database of studies, but to acknowledge its existence and its complementarity to the NRV recommendations, and to summarise key findings from some of these studies and the intervention trials.

The nutrients for which higher than RDI and AI intakes have been linked to benefits for chronic disease risk include the antioxidant vitamins such as vitamin C, vitamin E and vitamin A (primarily its precursor, ß-carotene) as well as selenium and nutrients such as folate, omega 3 fats and dietary fibre. These nutrients have been assessed in relation to heart disease and cancer as well as degenerative eye diseases such as cataract formation or macular degeneration, and conditions like Alzheimer's or cognitive decline.

The balance and type of macronutrients in the diet have also been studied extensively. The role of the various types of carbohydrates (starches, sugars, high vs low-glycaemic carbohydrates, resistant starch or RS, dietary fibres), fats (saturated, polyunsaturated, monounsaturated) and protein (animal, plant-based) have been variously assessed in relation to risk of conditions such as coronary heart disease (CHD), certain cancers, diabetes or insulin sensitivity and risk of obesity.