Dietary assessment and physical activity measurements toolkit

Dietary assessment - Biomarkers

Nutritional biomarkers are constituents in the blood or urine that can be used to estimate nutrient intake or compare nutrient intake to that estimated by dietary assessment. Biomarkers can provide objective validation of a variety of dietary assessments as they reflect, but are independent of food intake.  There are three categories of validation studies using biomarkers:

  • Recovery which are considered the gold standard.

Examples include: Doubly labelled water and urinary nitrogen and urinary potassium

  • Concentration which correspond to but do not reflect total intake.

Examples include: Plasma vitamin C or plasma carotenoids

  • Replacement which serve as a proxy for intake when it is not possible to capture intake due to limited nutrient databases to assess intake.

Examples are: Sodium, phytoestrogens, polyphenols or aflatoxin

(Fraser, 2003 and Livingstone & Black, 2003).  

The underlying assumption of a biomarker applied to validate a measure of nutrient intake is that it responds to intake in a dose-dependent way.  

Biomarkers in nutritional epidemiology
Arab and Akbar (2002) have provided a useful summary of the types of biomarker in nutritional epidemiology which is presented below.

  • Biomarkers of dietary exposure

- Indicate prior consumption
- Ranking: measures that identify high and low consumers such as adipose tissue levels of fat soluble vitamins
- Quantification: markers that reflect total intake over a defined time e.g. 24-hour urinary calcium, sodium, vitamin B metabolites or doubly labelled water

  • Biomarkers of compliance
  • Biomarkers of susceptibility
  • Biomarkers of pre-clinical disease
  • Biomarkers for validation

It is important to appreciate that the time period of exposure reflected by the biomarker differs from that of dietary intake; the biomarker reflects both intake and the processes involved in utilisation of the nutrient in question (Arab & Akbar, 2002).

Use of biomarkers to validate dietary intake data
It is problematic to ‘validate’ one dietary assessment method against another due to the risk of correlated error between the methods (see Basic Concepts). This means that a high correlation coefficient is due to a high level of agreement in the errors.  Strictly speaking such studies should be referred to as calibration studies.  The ideal way to validate a dietary assessment tool is to compare it with an objective measure of intake, which is independent of intake i.e. quantitative recovery biomarkers such as doubly labelled water or 24-hour urine collections; but they are expensive methods (Bingham, 2002). Circulating concentrations of nutrients in blood such as plasma ascorbic acid can be used to evaluate the performance of a dietary assessment method but they be influenced by behaviours such as smoking or use of supplements therefore they do not reflect absolute dietary intake.  They can therefore only be used to interpret the lower limit of the true validity of the dietary assessment (McKeown et al, 2001).

Validation of dietary intake data with biomarkers has demonstrated substantial differences in the extent of measurement error from those derived by comparison with other methods of dietary assessment (Bingham, 2002).   

The gold standard to validate energy intakes is to compare them to measures of energy expenditure made by doubly labelled water (see doubly labelled water section).  Alternatively, energy expenditure can be estimated by the use of indirect calorimetry or standard equations to measure resting metabolic rate, and then physical activity level estimated using standard values or by a measure of physical activity (see indirect calorimetry and under-reporting sections).

The Observing Protein and Energy Nutriton (OPEN) Study was conducted in the US from 1999-2000 and sought to assess the dietary measurement error of two self-reported instruments; the FFQ and the 24-hour recall (Subar et al 2003).  Doubly labelled water and urinary nitrogen were used as unbiased biomarkers of energy and protein intakes in nearly 500 men and women.  Results showed greater under-reporting of both energy and protein in both men and women by the FFQ compared to the 24-hour recall.  

Other researchers have used urinary measures of intakes of nitrogen, potassium and sodium to compare the performance of a semi-quantitative FFQ and a 7-day diet diary as part of the EPIC Norfolk study (Day et al, 2001). In this study 123 individuals completed a FFQ and 7-day diet diary twice, with a 12 month interval between.  Single day urine collections were made on six occasions over a 6-9 month period, covering the time that the dietary assessments were completed.  Results of this study showed that that the 7-day diet diary performed consistently better than the FFQ; the correlated error between the two reports of intake was also demonstrated.

Examples of biomarkers to measure dietary exposure of validated dietary intake

Urinary nitrogen
The use of biomarkers was first described by Isaakson in 1980 when he proposed urinary nitrogen as an independent measure of protein intake; it remains one of the most common biomarkers used.  The measure is based on the following equation: reported protein intake (g) = (24h urinary N + 2) X 6.25 (g) (Ruithauser and Black, 2002).  Like a single 24-hour recall, a single 24-hour urine collection does not reflect usual intake but nitrogen intake has been shown to be less variable than protein intake, such that only 8 days of collection are required compared to 16 days of dietary intake data to assess habitual protein intake.  

The utility of urinary nitrogen was demonstrated in a metabolic study lasting 28 days using 24-hour urinary nitrogen to estimate protein intake for eight individuals living in a metabolic suite.  Duplicate diets were used.  Urinary nitrogen underestimated intake at higher levels of intake, and over estimated at lower levels but a constant factor for faecal and skin loss can be used to counteract this. Output for urine was found to be 81% of intake (Bingham & Cummings, 1985); this level of 80% has been confirmed by a meta-analysis in a larger population (Kipnis et al, 2001).

Urinary potassium
In healthy people, urine is the major route of excretion of potassium.  Faecal excretion of potassium constitutes between 5 to 13 mmol per day in Western populations, or 11-15% of dietary intake (Bingham, 2002).  There can be good correlation between intake and excretion of potassium, even when dietary data are analysed using food tables, provided an adequate number of samples are obtained.  It has been shown that at least eight complete 24-hour urine collections have correlations of at least 0.7 between calculated intake and excretion (Bingham, 2002).  A biomarker for potassium is useful given that it is present in a reasonably wide variety of foods.

Urinary sodium
A 24-hour urine collection can also provide an assessment of sodium intake.  Given the widespread presence of sodium in manufactured products at varying levels and individual use of salt in cooking and at the table, it is not possible to assess sodium intake via dietary assessment methods with accuracy. Urinary sodium is the recommended way to measure sodium intake.

Urinary thiamine
A novel study has investigated 24-hour urinary thiamine as a potential biomarker for thiamine intake for use in validation studies to assess the validity intake data collected by self-report methods (Tasevska et al, 2008).  Individuals were housed for 30 days in a metabolic suite under strict conditions, prior to this 4x7 estimated diet records were kept to ascertain usual diet and determine the food served while in the metabolic suite.  Intake was weighed at this time.  Results showed a highly significant correlation between thiamine intake and secretion.

Use of PABA to verify completeness of urine collections
It is essential that urine collections are complete when undertaking biomarker studies.  Researchers in Cambridge UK, developed the PABA check test in the 1980s to verify the completeness of a 24-hour urine collection.  P-aminobenzoic acid (PABA) is a precursor of PAHA (p-amino hippuric acid) which is secreted by the renal tubules and used to assess plasma flow.  Typically oral doses of 3x80mg of PABA are taken by respondents at each main meal; in humans, the maximum between-person range in PABA excretion is only 15% and average urinary recovery is 93% of the administered dose in single 24-hour collections and 100% in sequential ones (Bingham, 1987).  The start and end time of urine collection should be recorded, along with the time the PABA tablets were taken.  Any lost specimens and a list of any medications should also be recorded.

Plasma measures
Plasma vitamin C
Plasma vitamin C measured by flourometric assay has been shown to be a marker of foods rich in vitamin C (Ness et al, 1999).  Plasma vitamin C has been used in the EPIC Norfolk study to investigate the association between plasma vitamin C (as a biomarker of fruit and vegetable intake) and mortality from cardiovascular disease, ischaemic heart disease and cancer, in nearly 20,000 men and women (Khaw, 2001).  Results showed an inverse relationship with urinary concentrations and mortality from both cardiovascular and ischaemic heart disease (Khaw et al, 2001).

Plasma concentrations of carotenoids have been shown to reflect dietary intake of fruits and vegetables (Van Het Hof, 2001) despite individual variability in absorption, availability and metabolism (Block et al, 2001 and Brevik et al, 2004). It is recommended that a range of carotenoids measured as a single biomarker is unlikely to be meaningful because of the diverse phytochemical composition of plant foods (Campbell et al, 1994).  Validation studies of dietary assessment tools with carotenoids have yielded a wide range of correlation and values rarely exceed r=0.4.

A dose-response relationship between carotenoid intake and appearance in plasma has been shown (Rock et al, 1992), making carotenoids a reasonable biomarker of intake. The most commonly measured carotenoids in studies to date include the provitamin A compounds: α-carotene, β-carotene, cryptoxanthin and also lycopene and lutein. Validation studies using carotenoids as biomarkers have largely been carried out in adults (Anderson et al, 2005 and McNaughton et al, 2005), with relatively few studies undertaken in children.  Previously published studies in children have found body weight to be a cofounder of plasma carotenoid concentration levels with obese children reported to have lower levels of α and β-carotene compared to normal weight children (Decisi et al, 1997).

Total fat intake
There is no good biomarker to measure total fat intake there are biomarkers sensitive to change in fat intake and biomarkers that reflect exposure to, or consumption of essential and non-essential exogenously produced fatty acids.  Body fat stores tend to reflect long-term fat consumption, except in individuals who are fasting or are not weight stable.

Fatty acid intake
In individuals in energy balance, adipose tissue is an ideal measure of long-term intakes of exogenous fatty acids.  It is not possible to relate to a quantitative intake at a definite time point so it is a ranking or concentration marker (Arab & Akbar, 2001).  Adipose tissue aspiration has been likened to phlebotomy in terms of safety (Handelman et al, 1988). An understanding of fatty acid metabolism, exogenous factors and the contributions of various body pools is important when interpreting these biomarkers (Arab, 2003).  For more detail about the technical advances in the quantitative measure of individual fatty acids and related issues see Arab & Akbar, (2001) and Arab (2003).

A recent review has focused on the identification and validation of individual polyphenols, or their metabolites, that may represent useful biomarkers of the intake of polyphenols in humans (Spencer et al, 2008).

To date some biomarkers for the validation of dietary assessment methods have been developed but the area is limited by the lack of biomarkers to reflect wider aspects of diet.  This is an important research priority. Doubly labelled water, urinary nitrogen and potassium are recognised as routine methods in validation studies. Circulating concentrations provide a less robust, but nonetheless helpful validation method.

Web design by Studio 24