Contents - Previous - Next

11.4 Practical hints

11.4.1 Administration of dose

The importance of the dosing procedure cannot be over-emphasised. When using a pre-prepared mixture of 2H2O and H218O any errors in calculating the dose given or any spillages or incomplete transfer of the dose to the subject will produce an equivalent percentage error in the estimate of pool spaces and therefore in the final estimate of energy expenditure. If the doses are administered separately an error in the administration of just one of the doses could potentially lead to an even greater error.

If isotope is spilled during administration it is vital not to pretend that this has not occured. The estimate of energy expenditure will be wrong and this will only serve to devalue the whole of your data-set. When it has taken days to recruit the subject, explain the procedures, obtain informed consent and arrange the dosing appointment, and with the prospect of wasting expensive isotope it is very tempting to ignore a less than perfect dosing. This temptation must be avoided, and the run should be terminated immediately in order not to waste any further effort on an invalid measurement. It should be noted that if the subject is redosed before their enrichment has returned to background another pre-dose must be taken to calculate dilution spaces.


Dosing adults presents no real practical difficulties. The main requirements are:

1) Most investigators require that subjects should be fasted for a minimum of 4 hours and preferably overnight.

2) A pre-dose urine sample must be collected to determine the subject's background enrichment. This also ensures that the subjects have voided prior to dosing.

3) An accurate body weight (corrected to nude weight) is required.

4) Care must be taken to avoid any spillage.

5) The dose should be kept sealed for as long as possible to avoid evaporation and fractionation.

6) The exact quantity of dose delivered must be obtained by weighing the dose container before and after administration (accuracy 0.2% or better is desirable).

7) An aliquot of dose should be reserved for dilution and analysis alongside the samples.

8) A small 'chaser' of unlabelled water (50 - 100 ml) is probably advisable to wash the dose into the stomach and avoid immediate evaporation from the mouth or oesophagus.

9) Early protocols insisted that subjects should remain fasted during the 3 - 6 hour equilibration period. The theoretical concern about excessive fluid ingestion during the equilibration period is that it will slightly expand the dilution space. However, the converse can be argued if no fluid is allowed, and recent studies have often permitted moderate food and fluid intake.

10) Collection and analysis of all urine voided during the equilibration period can be performed in order to correct the estimate of dose given by subtraction of dose voided. In practice, however, this correction is usually trivial and can be dispensed with.

Babies and infants

This is much more difficult than dosing adults and many obvious errors (such as estimates of isotope dilution spaces which appear to be greater than the entire body volume) can be traced back to problems with dosing. The actual dosing technique is an art, and success rates increase with experience. Whenever possible it is therefore best to allow a single investigator to develop and practice their technique, and to be responsible for all pediatric dosing. The optimal way to dose very young babies is probably to trickle the dose into the back of the mouth using a fine tube attached to a syringe. Older infants can successfully be dosed by diluting the dose with juice or cordial in a feeding cup. If this is done the entire dose must be consumed and any residue should be washed into the child with a small refill of the cup.

The requirements listed above for adults are all applicable to young children, but the pre-dose fast and equilibration periods can be shorter. Pre-dose fasts as short as 2 hours appear to be satisfactory in children less than 2 years old. Equilibration periods of 2 - 4 hours are usually used. It is often impossible to keep a young infant fasted during the equilibration period, and in practice small comfort feeds have to be tolerated.

In order to avoid having to give a large volume of dose to infants it is best to purchase H218O with at least 10% enrichment and preferably about 15%.

Quantity of dose

The most important constraint when deciding how much dose to give is the cost of 18O. This almost always means that measurements are made using the minimum possible dose. Schoeller 1 has summarised the theoretical basis for determining the amount of dose to give and the reader is referred to his paper for a full derivation of the optimal doses to be employed. In practice the decision is always a 3 way compromise between cost, the need to make a sufficiently long measurement and errors incurred as the final samples approach background enrichment levels. In adult studies many laboratories employ doses of 0.05 g 2H2O/kg and 0.15 g H218O/kg. Irrespective of the absolute amount given it is important to keep the ratio of the two isotopes close to this in most circumstances since, as Schoeller has pointed out (Section 8.4), this minimises the effects of fluctuations in background enrichment by taking full advantage of the natural covariance in such changes.

11.4.2 Choice of physiological fluid to be sampled

Isotope enrichments can be determined in any physiological fluid provided that the same fluid is sampled throughout. Unless there is a specific interest in studying the short-term dynamics of isotope mixing after dosing there is no advantage in using plasma, and there are obvious disadvantages. The choice is therefore between urine and saliva. Urine is usually chosen for adult studies. The main precaution which it is necessary to take is to ensure that the urine has not been collecting in the bladder over a long period since it will then be impossible to ascribe an accurate time to the sample. Saliva is more commonly used in studies of infants since they will not provide a urine sample on demand. Clearly the temporal definition is superior with saliva sampling. However, Schoeller argues that saliva should not be used (especially for the two-point method) since different degrees of fractionation may be present at different sampling times.

11.4.3 Collection and storage of samples

Most adults can be trusted to collect their own samples if provided with a series of numbered universal bottles (or similar), and a record sheet on which to record the time and date of each sample. However, the experimenter must beware of a minority of subjects who confuse the order of the bottles, fill several bottles with aliquots of the same urine sample (presumably because they have forgotten to collect samples on a number of previous days) or even fill the bottles with water!

Collection of saliva samples from infants can conveniently be achieved by swabbing the mouth with a small sponge or some absorbent cotton wool on the end of an orange stick. When saturated this is immediately transferred to a 1 ml syringe in order to express the saliva into a small sampling tube using the plunger.

All sampling and storage procedures should observe the following rules to avoid isotopic fractionation: a) the sample should only be exposed to the atmosphere for the minimum possible time; b) containers should be absolutely air-tight; c) there should be a minimum of air-space above the sample to minimise the possibility of isotopic exchange with any trapped atmospheric moisture; d) although some mass spectrometer procedures only require microlitre samples it is prudent to collect several ml of urine and at least 0.5 ml of saliva in order to minimise the chances of significant fractionation during pre-analytical manipulations such as transfer to auto-sampler vials.

Samples can be stored indefinitely and should preferably be frozen although this is not an absolute requirement under difficult field conditions. There is no objection to samples being frozen, defrosted (for example in transit) and then refrozen.

11.4.4 Sampling regimes

The pros and cons of various sampling regimes have been discussed fully in the body of the report. The basic rule governing the duration of a measurement is that it should last between 2 and 3 biological half-lives of each isotope. This tends to yield measurement periods of; 6-9 days in babies and children; 12-16 days in normal sedentary adults in temperate climes; 8-12 days in exceptionally active adults or people living in tropical climates where high rates of water turnover reduce the half-lives; and 16-20 days in elderly subjects.

In the two-point method the critical issue relating to sample timing is the most appropriate choice of the equilibration period in order to ensure that the sample obtained is on the plateau.

In the multi-point method it has been shown in Section 5.5 that the most efficient sampling regime uses a cluster of samples at each end of the measurement period. However, this demonstration is based on random errors and does not account for possible step changes in flow rates. If no samples are collected in the middle of the measurement period the process of inspecting residuals for deviations from the Lifson model becomes much less sensitive, and it is more difficult to decide which data should be rejected.

11.4.5 Collection of subsidiary data

Environmental conditions

Although it is impossible to make accurate predictions of the fractionated proportion 'x' (see Chapter 6) the working group recommended that information should be provided on average temperature and humidity encountered during each study. This will provide the reader with at least some means of assessing whether they endorse the fractionation assumptions selected by the investigators.

Fractionation corrections

In most circumstances it is probably futile to attempt to collect direct information on the appropriate value for 'x'. The variables required are: total water turnover, respiratory water losses and non-sweating insensible water losses. The first of these is known (rH2O) and the others cannot be determined in free-living people. Measurement of urinary output is not helpful in this respect. Figure 6.3 contains data on insensible water losses measured under cool conditions (ie with the children assumed not to be sweating) in a whole-body respiration chamber. This data is useful in indicating that 'x' is likely to be very low, but the insensible losses were measured in sleeping children and may not be entirely representative of 24-hour values in real life.

Changes in isotopic background

If substantial changes in isotopic background are anticipated it may be advisable to enrol an unclosed control group in order to quantify the average change by analysis of serial samples. Examples of such a situation may be when subjects change their primary water source (eg Arctic explorers), or when the enrichment of the main water source may change (eg during seasonal recharge of drinking water wells in the tropics). As discussed in Chapter 8 any changes in isotopic background have to be quite severe before they become a major concern to DLW particularly if Schoeller's advice concerning isotope ratios in the loading dose is adhered to.

Food quotients

As a very minimum it is necessary to have some qualitative knowledge of the fat: carbohydrate ratio in the diet in order to decide what RQ to assume when ascribing an energy equivalence to the estimate of carbon dioxide production. The reader is referred to Chapter 9 and the publication of Black et al for methods of calculating the energy equivalent of CO2.

Assessing changes in body composition

It would be useful to have an accurate assessment of changes in body composition during a measurement period for the following reasons: a) to assess possible isotope sequestration (see Chapter 7); b) to correct the food quotient by allowing for endogenous fat oxidation or fat deposition; and c) to assess agreement between simultaneous estimates of energy intake and expenditure. In theory this could be achieved by making a second estimate of total body water at the end of the measurement (using 2H2O for economy). However, in practice this is of limited value since, even if total body water can be assessed to within +0.5% at the beginning and end of the run this feeds through to an uncertainty in the region of 1 MJ/day with regard to fat oxidation or deposition. Changes in body weight therefore tend to be as useful as any refined attempts to assess fat store changes.

Additional physiological measures

Measurement of basal metabolic rate in subjects studied by DLW allows calculation of the Physical Activity Level (PAL) as TEE/BMR. An alternative way of expressing the same information is to calculate the energy cost of activity-plus-thermogenesis as TEE - BMR. Both of these are extremely valuable indices. It therefore seems false economy ever to make a DLW measurement which is not matched by an estimate of BMR particularly in view of the large investment inherent in DLW measurements.

11.5 References

1. Schoeller DA (1983) Energy expenditure from doubly-labelled water: some fundamental considerations in humans. Am J Clin Nutr; 38: 999-1005.

2. Black AK, Prentice AM & Coward WA (1986) Use of food quotients to predict respiratory quotients for the doubly-labelled water method for measuring energy expenditure. Human Nutr: Clin Nutr; 40C: 381-391.

Contents - Previous - Next