How Healthy Are Government Dietary Guidelines?
Part 2: Are the Current Government Dietary Guidelines Fit For Purpose?
Jane Philpott MA (Oxon), MSc, PhD
Introduction
From the exploration of the origin and development of government dietary guidelines (in Part 1Summer edition of Nutrition Practitioner), it is clear that the aims of dietary standards and guidelines have changed over time.
Initially, they were proposed as a guide for preventing scurvy; then for preventing diseases associated with starvation; then to feed the army and the nation; then to maintain health and working capacity; then to integrate health and agriculture; and finally to maintain ‘perfect’ health into advanced old age.
So are the current government dietary guidelines achieving their stated objective of encouraging healthy longevity in the population?
Crisis In Public Health
According to a 2008 report from the government’s Scientific Advisory Committee on Nutrition (SACN), the prevalence of obesity (Body Mass Index (BMI) above 30 kg/m2) or overweight (BMI above 25 kg/m2) increased from 45% of men and 36% of women who were obese or overweight in 1986/87 to 66% of men and 53% of women in 2000/0156. This trend is confirmed by data from the Health Survey for England57. Data from the International Obesity Taskforce show that England has one of the worst levels of obesity in Europe (Figure 2).
Obesity raises the risks of many chronic diseases including cardiovascular disease, certain cancers, osteoporosis, diabetes and liver disease. For young adults, the risk of an earlier death for someone with a BMI of 30 kg/m2 is about 50% higher than that for someone with a BMI in the range 20 to 25 kg/m2. The UK National Audit Office estimated that 7% of all deaths in England were attributable to obesity in 2002 (compared to about 10% due to smoking, and less than 1% from road accidents)58. In the UK, the economic costs of obesity are conservatively estimated at £3.3–3.7 billion per year and of obesity plus overweight at £6.6–7.4 billion.59
Figure 2: Obesity levels in Europe
According to the World Health Organisation, these health problems arise predominantly from elevated consumption of energy-dense, nutrient-poor foods that are low in fibre and high in fat, sugar and salt and reduced levels of physical activity at school, work, and home; and smoking.60
Long-term studies by the Harvard School of Public Health have concluded that an optimally nutritious diet combined with regular exercise and not smoking can prevent 80% of heart disease, 90% of type 2 diabetes, 70% of stroke and some cancers, as well as substantiallyreduce the incidence of a host of other chronic diseases and health ailments61. Similarly, the World Cancer Research Fund reports that changing what people eat could help to reduce the risk of about a third of all cancers.62
Figure 3: Comparison of prevalence of obesity (body mass index greater than 30 kg/m2) in adults in 1986/7 and 2000/1.
This crisis in public health suggests that something is wrong. Is the population failing to comply with the government’s dietary guidelines, or are the dietary guidelines themselves at fault, or both?
What Is The UK Population Eating?
A series of cross-sectional surveys aiming to provide detailed quantitative information on the diet, nutritional status and related characteristics of the British population are conducted at approximately three-yearly intervals. The programme is split into four surveys of different population age groups. Four surveys were conducted in the National Diet and Nutrition Survey programme between 1992 and 2001.63
Nutrient intakes are compared with COMA Dietary Reference Values and other COMA and SACN recommendations. Dietary Reference Values (DRV) for total fat, saturated and trans fatty acids and nonmilk extrinsic sugars are the recommended maximum contribution these nutrients should make to the population average diet, expressed as percentage of energy intake. For total carbohydrate, cis monounsaturated fatty acids and non-starch polysaccharide (NSP), the DRVs are population averages, i.e. the average contribution, as a percentage of energy, that total carbohydrate and cis monounsaturated fatty acids should make and the average intake of NSP in grams per day. For energy, the DRV is the Estimated Average Requirement (EAR), that is, the intake that meets the energy requirement of 50% of the population group. Finally, DRVs for protein, vitamins and minerals are expressed as Lower Reference Nutrient Intakes (LRNI) (not protein) and Reference Nutrient Intakes (RNI), the intakes at which the requirements of 2.5% and 97.5% of the population group are met.
Results of the surveys are summarized in Table 1 and compared with the dietary guidelines.
Taken together, the results of these four National Diet and Nutrition Surveys show that, over the last fifteen years, there has been:
· A decrease in energy intakes
· A fall in total fat and saturated fat intakes
· A reduction in the consumption of red meat, processed meat and meat-based dishes
· An increase in fruit and vegetable consumption
· An increase in salt consumption
· An increase in sugar consumption
· An increase in overweight and obesity
The Scientific Advisory Committee on Nutrition concluded:
“Although there have been positive changes in the diets of British adults over the last fifteen years…..there is still room for improvement.”
The areas for improvement are being targeted by current government policy64:
· Low fruit & vegetable consumption
o Despite an encouraging increase in fruit and vegetable consumption, it still remains below the recommendations in all age groups and is associated with low biochemical micronutrient status.
· High total/saturated fat intakes
o The proportion of dietary energy derived from fat and saturated fat has decreased since the 1986/87 adults survey.
o Mean intakes of total fat are close to recommendations in all population groups, though intakes of saturated fat exceed recommendations in all groups.
· High salt intakes
o Mean adult salt intake was higher in the 2000/01 than in 1986/87. A survey in 2005/06 showed a small fall in mean intake but intakes remain well in excess of the 6g/day recommended maximum.
· Obesity
o It would appear that mean energy intakes fell below the EARs in all population groups although the number of obese individuals is increasing. This apparent paradox reflects under-reporting of intake and possibly overestimation of energy requirements for the level of physical activity.
o Under-reporting would also affect intakes for other nutrients. Further research to characterise the diets of the obese and overweight groups would assist work to reduce of the prevalence of obesity.
· Specific “at risk” groups
o Young adults, older adults living in institutions and people in lower socioeconomic groups were identified as having a high prevalence of low nutrient intake and low biochemical status.
· Poor dietary patterns
o The group with the lowest mean intakes and biochemical status of all nutrients (except iron) consumed more soft drinks, savoury snacks and alcoholic beverages.
o A higher consumption of sugar, preserves and confectionery was associated with low nutrient intake and biochemical status.
· High sugar-rich food consumption
o The proportion of energy intake derived from non-milk extrinsic sugars (NMES) exceeds the recommendation in most age groups, particularly amongst children and young adults, as well as older adults.
o Groups with the lowest mean intakes and biochemical status of almost all nutrients had the highest consumption of soft drinks.
· Low intake of non-starch polysaccharides
o Intake of non-starch polysaccharides was low. No groups met the recommendation for adults.
· Low fish (especially oily fish) consumption.
o Mean consumption of oily fish was below the recommendation in all age groups even though it has increased in certain groups over the past 15 years.
o The group with the highest consumption of fish and fish dishes had the highest mean intakes/biochemical status of most nutrients.
Specific population groups identified as most at risk of poor dietary variety and low nutrient intake and biochemical status were:
• Children aged 18 years and under
• Young adults aged 19-24 years
• Smokers
• People in lower socio-economic groups
• Adults aged 65 years and over living in institutions
Table 1: Summary of findings of National Diet and Nutrition Surveys compared with dietary guidelines65
It can therefore be seen that, although there has been progress towards gaining greater compliance with government dietary guidelines, there are still substantial numbers of people who are not eating in accordance with the scientific advice available. This is likely to be contributing to the current public health crisis, but is it the whole story? Are the dietary guidelines themselves still valid?
Evidence Base for Dietary Guidelines
By their own admission, there was insufficient data for the Committee on the Medical Aspects of Food Policy to establish dietary reference values with great confidence66. The first reason for this is that it is technically difficult to conduct meaningful research on nutrition, the diets of populations and their effects on health. Secondly, funding for nutrition research has historically been much lower than that for drug and other medical research and this remains the case.
In medical research, there is a hierarchy of evidence:
Grade / EvidenceA / Systematic review of all randomised control trials (RCTs) - meta-analyses
B / Properly designed RCTs or well-designed pseudo-RCTs
C / Cohort studies
D / Case control studies or interrupted time-series with a control group
E / Comparative studies with a historical control
F / Case series
G / Other relevant information, such as reports of expert committees
Randomised control trials (RCTs) are considered to be the ‘gold standard’ by which other studies are judged. In these carefully controlled studies, half of a group of volunteers is randomly assigned to the experimental diet or treatment, and the other half is assigned to the standard diet or treatment (the control) or possibly to no treatment at all. After a preset time, the number of people in the control group who have developed the predetermined ‘endpoint’ – death, heart attack, cancer, and so on – is compared with the number in the experimental group.
There are many advantages of this type of study. If it is large enough, the randomisation process makes sure that the people in the experimental group are very similar to those in the control group in terms of age, health, exercise and other relevant factors. This means that the only difference between the two groups is the diet or treatment.
Unfortunately, randomised trials are often impossible to do when it comes to nutrition research. Ensuring that people prepare and eat special meals for a long time is difficult. It is equally difficult to persuade people to take a vitamin pill or placebo for maybe a decade or more. Given the large numbers of people needed, the cost of running a randomised control trial can be astronomical. The Women’s Health Initiative, which is primarily testing the impact of reducing dietary fat to 20 % of calories and increasing fruits and vegetables on the development of breast cancer, will cost more than US$1 billion and probably still will not yield clear answers67. Randomised control trials investigating the effect of individual nutrients on health typically use supplements that may be administered at much higher doses than would be likely from consuming the nutrient in food form, or which may not be absorbed by the body as well as food-form nutrients. Food contains a wide range of nutrients that usually work in concert, so the effect of one will usually depend on the presence, concentration and effect of others. If one nutrient is present at higher than normal physiological concentrations, the regulation and action of other nutrients may be distorted. In addition, the effect of individual nutrients will depend on the background nutrient status of an individual. Consuming vitamin B pills whilst living on a diet of junk food, for example, will almost certainly have a different effect from consuming vitamin B pills whilst on a healthier diet. It is very difficult to control for this as every individual eats different types and amounts of food, particularly if the study is over a long period.
Cohort studies involve following large groups of people – described by epidemiologists as ‘free-living humans’ – for long periods. These cohort studies start with a group of people who often have something in common, such as an occupation or a place of residence. They are asked about their diets, smoking and drinking habits, education, occupation, medical conditions and other relevant questions. The group is then followed, usually for a decade or more, either directly with occasional check-ups and mailed questionnaires or by monitoring death certificates. When the study has been in progress for sufficient time, researchers examine the data collected to test their hypotheses. For example, do the people in the cohort who eat the most fibre have the lowest incidence of colon cancer? Such long-term cohort studies have yielded some of the best information so far on the link between diet and health. Well-known cohort studies include the EPIC Study68, the Nurses’ Health Study69 and the Health Professionals Follow-up Study.70
Cohort studies have three main weaknesses: the estimate of food intake, confounding, and practical duration of follow-up. For such studies food-frequency questionnaires are commonly collected by postal survey with no interviewer's help; they do not correlate well with more direct measurements of actual food intakes or with dietary records. A single snapshot of diet does not reflect the gradual changes that take place over years in most people's eating habits. Cohort studies with periodic estimates of food intake, backed by biochemical tests of key nutrients, are higher quality than once-only questionnaires. Dietary components associated (or not) with a disease (e.g., dietary fibre) may accompany a feature of lifestyle that was not tested but is more directly related. This confounding can never be eliminated. Lastly, cohort studies are short-term compared with the average human life span. Most cohorts are of middle-aged people in whom there is a fair risk of the disease developing, so such studies cannot reveal effects of the diet in childhood and adolescence.71
Case-control studies gather data from a group of people who have developed particular diseases (the cases) and a similar group of people who are free of the disease (the controls) and compares the two groups for differences in diet, exercise or other life-style factors. Case control studies do not work well for research on diet, when only small differences are likely to be seen from person to person. They also have the drawbacks of uncertainty about how long the likely incubation period is between the possible causative diet and onset of the chronic disease and of inaccuracy in summoning up past dietary habits. Case-control studies are also more prone to error and bias than cohort studies. They can be performed more quickly and cheaply than either randomized control trials or cohort studies; thus they were used to supply the evidence for many of the early recommendations about diet and health. As more information emerges from long-term cohort studies, it is evident that conclusions from case-control studies were not always correct.