Measuring Food Intake 269Very few validation studies have attempted to eval-
uate how well the different methods rank individuals.
It has been observed that the correlation between
urinary nitrogen and dietary nitrogen measured by
diet records was better (0.65 and 0.79) than between
urinary nitrogen and dietary nitrogen measured by
food frequency questionnaires (0.15 and 0.24).
Doubly labeled water method
The DLW technique allows the measurement of
energy expenditure in free-living respondents over
several days with minimal inconvenience to the
respondent and with a high level of accuracy and
precision. Under controlled conditions the DLW
method gives a small overestimate of 2–3% compared
with whole body calorimetry, and under fi eld condi-
tions bias is not expected to exceed 5%.
The DLW method requires that the respondent
drinks a small measured dose of water enriched with
naturally occurring stable isotopes of deuterium and
(^18) O. The two isotopes disperse throughout the body,
and are metabolized and then gradually lost from the
body. Since the deuterium labels the body water pool,
and the^18 O labels both the water and the bicarbonate
pools, the difference between the disappearance rates
of deuterium and^18 O can be used to calculate carbon
dioxide production. The level of both isotopes is
determined, using mass spectrometry, in a small
sample of urine collected each day for between 5 and
28 days. Energy expenditure is calculated from carbon
dioxide production using calorimetric equations.
Further details of the DLW technique and the main
factors infl uencing its accuracy and precision are pro-
vided in Chapter 3.
Using the DLW technique several investigators
have compared self-reported dietary energy intake
with energy expenditure based on the equation:
Energy expenditure (EE) = Energy intake (EI)
± Change in the body energy store
Differences between measured energy intake and
expenditure varied from –44% to +28% depending
on the population subgroup studied. This fi nding
confi rms the need to include one or more indepen-
dent measures of validity in all dietary studies
to ascertain the level applicable to the particular
group under study, since it is not readily predicted
on the basis of gender, age, or body mass index
(BMI).
The main advantage of the DLW method is that it
makes minimal demands on the respondents and
does not in any way interfere with their normal daily
activities and therefore their habitual level of energy
expenditure. Its main disadvantage is that the cost of
the DLW required for each estimate is exceedingly
high and the method also requires access to sophisti-
cated laboratory equipment for mass spectrometric
analysis. It is, therefore, not available for use on a
routine basis for the validation of dietary intake
data.
Ratio of energy intake to basal metabolic rate
Because of the limitations of the DLW method,
another approach that is used compares the energy
intake (EI) reported from published studies with the
presumed requirements for energy expenditure, both
intake and expenditure being expressed as multiples
of the basal metabolic rate (BMR). The relevant
equation is:
EI:BMR = EE:BMR (PAL)
where PAL is the physical activity level. To determine
whether reported energy intake is a “plausible”
measure of actual diet during the measurement period
(i.e., represents either the habitual diet or is a low/
high energy intake obtained simply by chance) an
equation was developed by Goldberg and colleagues
(1991) to calculate the 95% confi dence limits of
agreement between EI:BMR and PAL. This equation
allowed for variation in EI, BMR, and PAL and also
for the length of the dietary assessment period and
study sample size.
For a group, if mean reported EI:BMR is below the
lower 95% confi dence limit (cut-off ) for the given
study period and sample size, then there is defi nitely
bias to the underestimation of energy intake.
However, the identifi cation of individual under-
reporters is much more diffi cult, since reported EI can
deviate quite markedly from energy expenditure (EE)
before it falls outside the limitations of the methods.
Figures 10.7 and 10.8 illustrate the limitations of both
techniques. Figure 10.7 shows the energy intake and
DLW EE of 264 women. The solid lines indicate EI:
EE of 0.76 and 1.24. These are the 95% confi dence
limits of agreement between EI and EE, allowing for
day-to-day variation in food intake and within-subject
variation on repeat DLW measurements. Only women
with an EI:EE ratio above 1.24 can be confi dently