Contents - Previous - Next

6.1. Energy expenditure and milk intake in fast-growing preterm infants

We compared the doubly-labeled water method with indirect calorimetry in preterm infants (ROBERTS et al., 1986). The method was also compared with a nutrient balance study for assessment of water intake. Premature infants exhibit a high water turnover and rapid growth in relation to CO₂ production, thus representing a physiologic extreme in which errors associated with the doubly-labelled water method may be large. One particular problem we anticipated in infants with a high water-turnover rate was the possibility of error incurred by not accounting accurately for isotopic fractionation (differential evaporation of H₂O and ²H₂O or H₂O and H₂¹⁸O) or insensible water loss (LIFSON and McCLINTOCK, 1966; COWARD et al., 1986). The methodological details are described elsewhere (ROBERTS et al., 1986; ROBERTS and LUCAS, 1985a). Briefly, 4 well, preterm infants were studied. Two samples of urine were obtained from each infant prior to the start of the isotope study, to determine natural concentrations of ²H and ¹⁸O in body fluid. An accurately weighed dose of sterile isotope solution providing approximately 0.24 g ²H₂O and 0.6 g H₂¹⁸O per kilogram body water was then given nasogastrically. After administration of the isotope, timed urine samples were collected for approximately 5 days. Open-circuit, indirect calorimetry was performed for an average of 77% of the 5-day isotope study. The technique for the nutrient balance study, performed throughout the isotope study, is described elsewhere (ROBERTS and LUCAS, 1985b).

Table 2 shows the comparison of values obtained for rate of CO₂ production, energy expenditure and water intake, determined by the doubly-labelled water method and indirect calorimetry, or a water-balance study. Values for rate of CO₂ production and energy expenditure were not significantly different from those measured using the standard methods (paired t values for isotope-determined rates of CO₂ production and energy expenditure were within 6 and 4%, respectively, of values obtained by calorimetry, and differences between means did not exceed 2%. Published information on RQ (GUDINCHET et al., 1982; REICHMAN et al., 1983; PUTET et al., 1984), rather than the measured value, was used to calculate energy expenditure from CO₂ production rate determined by the isotope method, because the latter information would not normally be available.

Table 2. Rates of CO₂ production energy expenditure, water intake and metabolizable (ME) intake determined by the doubly-labeled water method and indirect calorimetry or a nutrient balance study in four preterm infants

The doubly-labelled water method significantly overestimated water intake by 5.7% (SD 1.4%), but the values were extremely precise. This difference between the results obtained by the two methods for water intake may be due entirely to the exchange of atmospheric water vapor in the lungs (insensible water intake; NAGY, 1980), which contributes to the isotopic estimation but is not a component of the balance-study value. We have used these data to adjust the value for water intake by the isotope technique to provide a value for oral water, and hence milk intake.

Thus, the doubly-labelled water method, previously validated for the measurement of rates of CO₂ production and energy expenditure in adults, has now been validated in 4 preterm infants. In view of the extreme physiological differences among these two types of subjects, it is likely that the method is suitable for use in other physiological groups. The range of error in measuring the CO₂ production rate, although somewhat greater than in adults (ROBERTS et al., 1986), is likely to be substantially lower than that associated with the traditional methods for measuring energy expenditure over long periods (GARROW, 1978; DURNIN, 1984).

Our subsequent studies on full-term infants, which will be discussed later, indicate that the precision of the doubly-labelled water method for measuring energy expenditure is in the range 45%. This value was determined by the method described in detail by COWARD, ROBERTS and COLE (1988) and also takes into account the effect of covariance in the ¹⁸O and ²H postdose enrichments.

6.2. Validation of dose-to-the-baby method for measuring milk intake

In order to assess whether the dose-to-the-baby method was valid for measuring milk intake in full-term babies, we undertook 14 studies, each lasting 7 days, on 8 bottle-fed infants whose milk intake was measured both by a deuterium oxide dilution technique and by weighing the formula consumed together with oral losses. Methods and findings are described in detail elsewhere (LUCAS et al., 1987a and b).

The mean (±SD) values (Table 3) for milk intake, 837 (±98) g/d by weighing and 829 (±108) g/d by isotope, were not significantly different (paired t = 0.75). The correlation coefficient for the two data sets was 0.93 (p < 0.001). The mean difference (bias) between the methods (deuterium dilution minus weighed intake) was -1.1%. The standard deviation of this difference was 5.1%, and thus the limits of agreement of the two methods (we used an approach similar to that of BLAND and ALTMAN, 1986) were -11.3% to +9.1% (mean difference ±2 SD of difference). The highest individual difference in this study was 9.9%. For our sample size of 14 paired observations, the 95% confidence interval for placing the mean difference between methods (deuterium method -1% compared with weighed intake) would be -3.9% to +2.9%. Total body-water estimates derived from a deuterium oxide dilution technique are also shown in Table 3 for reference.

Table 3. Comparison of milk formula intake by direct weighing and by deuterium dilution

These data suggest that for population studies the dose-to-the-baby deuterium method is sufficiently precise for the measurement of milk intake. Indeed the results would be satisfactory for many types of study on individuals.

6.3. Determination of milk energy content using the doubly-labelled water method

Employing the principles described above for estimating milk energy content using stable isotopes (LUCAS et al., 1987b), a study of 29 exclusively formula-fed infants, aged 10-12 weeks, was undertaken over a 7-day period, so that the isotope-determined value for milk energy content could be compared with that derived from the known composition of the formula using standard Atwater factors.

Total body water at the start of each study period was calculated by isotope dilution as described previously (ROBERTS et al., 1986). In order to calculate energy expenditure, RQ was taken to be 0.87. On the basis of water balances (unpublished), we took a figure of 13% of water output as being subject to fractionation (i.e., evaporative loss). To calculate energy stored, the following assumptions were made. Firstly, the decline in total body water (% body weight) from the start of another study at 4-6 weeks to the start of this study at 10-12 weeks, was assumed to be linear in each subject (the data of FOMON et al., 1982, suggest a near linear decline at this age). Thus, total body water at the end of each one-week study period could be calculated. Secondly, published values were used (FOMON et al., 1982) for the water and protein content of lean body mass: 80.0% and 15.7% respectively at 11 weeks (published data on both sexes combined). These numerical assumptions are less critical than might be supposed, since the data are used to detect a change in body composition during the study and not an absolute value.

Increases in body fat, lean body mass and protein over the one-day study periods were calculated as follows: lean body mass at the start or end of the study = total body water/proportion of water in lean body mass, and body fat = body weight - lean body mass; fat accretion = difference in body fat between start and end of the period; protein accretion = lean body mass gain x (published) proportion of protein in lean body mass. Using conversion factors (9.25 for fat and 5.6 for protein) energy stored as fat and protein during each study was derived (carbohydrate storage was assumed negligible).

Milk intake was calculated from the isotope-determined value for water intake, taking account of the water content of the milk and making adjustments for the water derived from complete oxidation of its fat, carbohydrate and protein and for the water in fat and protein stored in new tissue (as derived above).

Data obtained on energy expended and stored (as fat, protein and total) are shown in Table 4, and hence metabolisable energy intake (MEI) calculated (energy expended + energy stored). Mean milk energy density derived from MEI/milk volume intake was 66 (SE ±3) kcal/100 mL; 95% confidence interval of 60-72 kcal/100 mL. The manufacturers' stated figure for the energy density of the formula, 68 kcal/100 mL, lay within one standard error of the isotope-determined value and was therefore quite symmetrically placed within the confidence interval.

Table 4. The metabolisable energy content of a formula determined by the doubly-labelled water method, compared with the value for energy content derived from macronutrient composition. Based on 29 formula-fed infants at 3 months of age, studied for a 7-day period

This validation study is confounded by the lack of certainty over the precision of the Atwater conversion factors used to calculate metabolisable energy from milk macronutrient composition. Nevertheless, our data suggest that the sequence of calculations and assumptions used for the isotope procedure were broadly appropriate and would give confidence to data obtained in breast-fed infants.

7. Metabolisable energy and energy content of breast milk determined by the doubly-labelled water method

7.1. A study and its results
7.2. Critique of findings

7.1. A study and its results

Having obtained satisfactory validation data, we studied metabolisable energy and the energy content of breast milk in 20 healthy breast-fed babies at 4-6 weeks and 10-12 weeks postnatally using the same procedures. The findings are shown in Table 5, together with the corresponding values for the energy density of the formula in 29 formula-fed babies studied at these ages.

Table 5. Metabolisable energy consumed and energy content of breast milk (SE) determined by doubly-labelled water method at 4-6 and 10-12 weeks of age. For comparison, energy content of a formula also derived by the isotope method

The energy density of the formula based on conventional conversion factors for metabolisable energy was 68 kcal/100 mL.

At 10-12 weeks mean metabolisable energy was 87 (SE ±3) kcal/kg (95% confidence interval: 81 to 93); mean breast-milk energy was 58 (SE ±2) kcal/100 mL (95% confidence interval: 5462 kcal/100 mL). The value of 58 kcal/100 mL was significantly lower (p < 0.05) than that in the formula-fed group at this age (66 ±3 kcal/100 mL).

At 4-6 weeks, metabolisable energy was similar at 89 (SE ±4) kcal/kg (95% confidence interval: 81 to 97). Surprisingly, however, values for the energy content of milk were significantly lower than at 10-12 weeks: 60 ±3 and 53 ±2 kcal/100 mL for formula- and breast-fed infants, respectively. Nevertheless, the value in the formula-fed group remained significantly lower than that in breast-fed infants (p < 0.05).

Contents - Previous - Next