ABSTRACT

The predictive values of anthropometric measurements, race, gender, gestational and postnatal ages, and season at birth and at study for the full body dual energy X-ray absorptiometry (DXA)-derived lean mass (LM), fat mass (FM) and fatty mass as a percentage of body weight (%FM) were determined in 214 singleton advisable nascency weight for gestational historic period infants [101 Caucasian (lx boys, 41 girls) and 113 African American (55 boys, 58 girls)]. Gestational ages were 27–42 wk and the infants were studied between nativity and 391 d, weighing between 851 and 13446 g. In addition, predictive value of torso weight, LM and FM for DXA bone measurements was also determined. Browse conquering used Hologic QDR 1000/W densitometer and infant platform and scans without significant movement artifacts were analyzed using software 5.64p. Body weight, length, gender and postnatal age were significant predictors of LM (adapted R 2 >0.94) and FM (adjusted R 2 >0.85). Physiologic variables had fiddling predictive value for %FM except in the newborns (adjusted R ii 0.69). Torso weight was the ascendant predictor of LM and FM, although length had similar predictive value for LM with increasing postnatal age. Female infants had less LM and more than FM throughout infancy (P < 0.01). LM or FM offered no advantage over body weight in the prediction of bone mass measurements. DXA is a useful means with which to make up one's mind torso composition, and our data are of import in the design and assessment of nutritional intervention studies.

Dual energy Ten-ray absorptiometry (DXA)3 is used increasingly equally the method of choice to measure various components of body composition (BC) during infancy (Koo 2000). We reported previously that body mass is the dominant predictor of bone mineral status in newborns (Koo et al. 1996) and older infants (Koo et al. 1998). This is supported by other recent reports of bone mass measurements in infants (Lapillonne et al. 1997, Rigo et al. 1998). Our finding is too consistent with the finding in adults of a positive relation of torso habitus (Aloia et al. 1999, Chumlea and Guo 1999, Ravn et al. 1999) with bone mineral condition, although lean trunk mass (Courteix et al. 1999, Ferretti et al. 1998, Valdimarsson et al. 1999) and fat mass (Courteix et al. 1999, Ferretti et al. 1998) are thought to be stronger determinants of os mass. In contrast, petty is known about the physiologic predictor of soft tissue limerick during infancy, and there are no information to make up one's mind the predictive ability of soft tissue limerick on bone mineral status in infants. The aim of this study was to extend our previous observation on bone mineral status (Koo et al. 1996, and 1998) in newborn infants and throughout infancy to document the differences in soft tissue composition during this period. We aimed to determine the predictive value of anthropometric measurements and other physiologic variables on soft tissue body composition measurements. In addition, the predictive value of soft tissue composition on bone mineral status too was assessed.

SUBJECTS AND METHODS

Subjects.

The total study population included 214 singleton infants with birth weights from 1002 to 3990 g. The subjects' nascency weights were advisable for gestational age (Brenner et al. 1976). Gestational ages of the subjects as determined by maternal menstrual dating and/or ultrasound were from 27 to 42 wk and within 2 wk of gestational age assessment by standard examination (Ballard et al. 1991). Fourscore-5 subjects were preterm with gestational historic period < 38 wk; of these, 53 subjects had low birth weight (≤2500 g). There were 101 Caucasian (60 boys, 41 girls) and 113 African American infants (55 boys, 58 girls). For infants beyond the immediate newborn period, the type of milk and whether the baby was receiving solids were recorded. This report was canonical by the Institutional Review Lath for human subjects at the Academy of Tennessee-Memphis, and written informed consent was obtained from each subject's parent.

Anthropometric measurements.

The nude weight and length of the infant were measured at each study. The weights of the cotton coating that swaddled the baby and the diaper, if used, were also determined. An additional blanket or a large cotton canvas was used in 22 infants to better swaddle the larger infant. A diaper was used in all infants beyond the newborn catamenia. The study weight is the total weight including the nude weight and the weight of blanket/due south and diaper. The weight in grams was adamant with a digital electronic calibration (Air Shields, Vickers, OH). The scales were regularly maintained by the hospital Biomedical Instrumentation personnel and calibrated with known standard weights. Recumbent length was the average of two consecutive measurements within 0.4 cm and was adamant using a standard length board (Ellard Instrumentation, Seattle, WA).

DXA measurements.

All infants were clinically well at the fourth dimension of study, and each infant was studied on one occasion. Scan acquisition of total body (TB) was performed with a single beam whole-body scanner (Hologic QDR k/West densitometer, Hologic, Waltham, MA), with the use of an infant platform. With our densitometer, the typical maximum entry radiation exposure during an babe whole body browse was 3 μSv (1 μSv = 0.1 mrem). The radiation besprinkle at ninety cm from the scanner was <0.3 μSv from 10 min of measurement (Koo et al. 1995a).

All scans for this study were performed with the swaddled discipline placed on peak of the infant platform with a cotton blanket interposed betwixt the subject field and the platform (Koo et al. 1996, and 1998). A rut lamp was used as needed to maintain a satisfactory environmental temperature. All scans were performed without sedation or additional restraint but a pacifier with nonmetallic parts was used as needed. Occasionally, the scanning process was interrupted if motility artifact was noted, and a echo browse was performed when the infant had been pacified.

Scans were analyzed using the software (Version 5.64p) developed in conjunction with the manufacturer. In addition to the analysis of the total torso, analyses of different regions were too performed using the same software if the position of the baby allowed adequate delineation of dissever regions. Regional analyses typically involved the caput and each of the four extremities. The residuum region was regarded as the trunk for a total of six regions. Each scan was reviewed past one of two investigators (JW or WK) and was judged technically satisfactory if the external calibration step phantom and the skeletal outline of the subject field lay inside the browse region and without significant movement artifact (Koo et al. 1995b).

Quality control scans were performed daily on a manufacturer-supplied anthropomorphic spine phantom, and the long-term (>3 y) CV for the determination of bone mineral content (BMC), os area (Surface area) and bone mineral density (BMD) using an anthropometric spine phantom are <0.31% for all parameters. The boilerplate annual charge per unit of change for each of these measurements was not significantly unlike from zero. The in vivo replication of TB DXA measurements in fifty infants was significantly correlated [r = 0.99 and P < 0.001 for all parameters, i.e., BMC, Surface area, BMD, lean mass (LM) and fat mass (FM), respectively]. In our laboratory, the standard deviation (SD) of difference (Bland and Altman 1986) between paired DXA measurements for TB BMC was three.8% at a mean of 93 g; TB Area was ii.v% at a hateful of 371 cm2; TB BMD was ii.6% at a hateful of 0.228 m/cm2; TB LM was 2.iii% at a mean of 3184 g; and TB FM was vii% at a mean of 995 thousand, respectively.

Statistical analyses.

The data were treated equally from two separate cohorts to determine the physiologic predictors of BC measurements at birth and postnatally. The "at nascence" data consisted of DXA measurements of infants of all gestational ages studied during the first 8 d after birth, and the "postnatal" accomplice consisted just of infants born at term and studied between birth and throughout infancy.

For the at nativity accomplice, a primary component factor analysis showed that the three measures of weight (nascence weight, study weight, nude weight) were highly interrelated with loadings of 0.994 to 0.998. Thus, any of the weight variables can be used in the regression analysis with equal validity. Nude weight was used in all analyses to minimize the entry of multiple colinear contained weight variables in the analyses. Its utilise has other advantages considering it is the well-nigh consistent weight measurement without concern for the varying amounts of clothing and covering; it also provides consistency with our previous publications (Koo et al. 1996, and 1998).

Stepwise multiple linear regression analyses were performed to make up one's mind the explanatory ability of each of seven independent variables on each of the 6 dependent variables [LM, FM, fat mass as a percentage of trunk weight (%FM), BMC, Expanse and BMD] separately. The independent variables are known to take the potential to affect growth and body limerick; these included race, gender, gestational age, postnatal age at study, nude weight, length and season of birth. The season variable was determined by coding the month of birth at iii monthly intervals beginning at March every bit spring, and was transformed into dummy variables using spring as the reference season. In addition, each of the three variables, LM, FM and nude weight, was entered lone as an contained variable to decide the value of each of these three independent variables in the prediction of BMC, Area and BMD.

For each of the dependent variables, a final model predictive equation was generated, containing but meaning independent variables. This represents a hierarchical modeling process that first determines the most powerful individual predictor of DXA measurements and then determines whether any other set(s) of independent variable(south) either augmented or diminished the model's explanatory capability of the single all-time predictor. Percentiles were also calculated for LM, FM and %FM using Altman'southward method (Altman 1993) and the best-fit curves were plotted on the basis of the actual data.

For the postnatal cohort, written report weight and nude weight were significantly correlated (r ii = 0.99) and just nude weight was entered equally an independent weight variable. Data analysis used the same procedures as described higher up. However, iii additional variables were entered every bit independent variables. These included nascency weight, the flavour at the time of study (derived from the study month and entered into regression model using the same technique described above) and, for infants beyond the immediate newborn menstruum, the blazon of milk intake and apply of solid food on the 24-hour interval of study. Milk intake was transformed into dummy variables before analyses using human being milk equally the reference milk.

All statistical tests were performed with SPSS 9.0 (SPSS, Chicago, IL) for Windows at an adopted significance level of 0.05. Values are means ± SD.

RESULTS

DXA scans were performed between birth and 391 d. The nude weight of the infants at study was between 851 and 13,446 g. The "at nascency" cohort included 150 subjects (85 preterm infants; 82 boys; 78 Caucasian and 72 African American) and DXA scans were performed at 2.1 (±one.six) d afterward nascency. 10 preterm infants were studied betwixt 6 and 7.7 d after they had recovered from transient respiratory illnesses. The "postnatal" cohort included 129 infants born at term with seventy boys and 59 girls, 62 Caucasian and 67 African American. Of infants born at term, 64 were studied beyond the newborn period.

At birth accomplice.

Nude weight consistently proved to be the single best predictor of LM, FM and %FM with an adjusted R two of 0.978, 0.837 and 0.632, respectively. Gender and length were the but boosted predictors that could be forced into a predictive equation for these dependent variables on the basis of statistical significance, although the additional contributions to the prediction of LM, FM and %FM were 0.five, three.2 and 6.2%, respectively. Female infants had significantly lower LM but college FM and %FM (P < 0.01). The final regression equations (P < 0.001 for all models) for the prediction of TB LM, FM and %FM including all significant predictors are equally follows:

DXA LM (g) = −714 + 0.626 nude weight (grand) + 29.94 length (cm) − 39.7 gender

Adjusted R 2 = 0.983, SEE 76 m

DXA FM (chiliad) = 644 + 0.347 nude weight (thousand) − 25.ix length (cm) + 33.three gender

Adapted R ii = 0.869, SEE 70 g

DXA FM (%) = 22.0 + half-dozen.525E-03 nude weight (g) − 0.581 length (cm) + i.3 gender

Adjusted R ii = 0.694, Run across 2.1%

where gender = 0 for male infants and i for female person infants and Meet is the standard fault of estimate.

Percentiles for DXA measurements of TB LM, FM and %FM in newborn infants based on nude weights are shown in Figure 1A–C. Information technology should be noted that percentiles are descriptive, not predictive, and draw attention to the increasing variability of the dependent variables equally nude weights increased. The standard error of estimate for a predictive equation is a function of the dependent variable and represents the strength of the correlation between independent and dependent variables. The capability of the predictive equation across the body weight range of our newborn accomplice was determined by computing predictive equations based on ii nude weight ranges divided approximately at the midpoint of the full weight range. The resultant r and SEE values of the prediction equations generated from the total cohort and from the ii subpopulations are shown in Table ane. With increasing torso weight from 1.v to 3.5 kg, the average proportion of TB LM decreased from 90.8 to 81.8% and the average proportion of TB FM increased from 7.5 to xvi.2%.

Table 1

Correlation and standard mistake of estimate (SEE; expressed as Z-score) for predicting fat and lean tissue mass from nude body weights of infants i

Body weights Newborn
All weights ≤two.four kg >2.4 kg
r SEE r Meet r SEE
Lean mass, g 0.99 0.15 0.99 0.xv 0.94 0.35
Fatty mass, grand 0.91 0.41 0.83 0.56 0.79 0.61
Fatty mass, g/100 thou 0.lxxx 0.61 0.51 0.86 0.58 0.82
Torso weights Newborn
All weights ≤ii.4 kg >2.4 kg
r Run across r SEE r SEE
Lean mass, g 0.99 0.xv 0.99 0.fifteen 0.94 0.35
Fatty mass, 1000 0.91 0.41 0.83 0.56 0.79 0.61
Fat mass, g/100 g 0.80 0.61 0.51 0.86 0.58 0.82
Body weights Postnatal
All weights ≤half dozen kg >6 kg
r Meet r Encounter r Come across
Lean mass, g 0.98 0.20 0.94 0.35 0.91 0.42
Fat mass, g 0.95 0.32 0.92 0.40 0.seventy 0.72
Fat mass, g/100 g 0.75 0.66 0.81 0.59 NS —
Body weights Postnatal
All weights ≤6 kg >6 kg
r SEE r SEE r SEE
Lean mass, g 0.98 0.twenty 0.94 0.35 0.91 0.42
Fatty mass, g 0.95 0.32 0.92 0.40 0.seventy 0.72
Fat mass, g/100 g 0.75 0.66 0.81 0.59 NS —

1

No significant deviation (Student t exam, two-tailed P = 0.49–0.97) in mean residuum values for predicting lean mass (−0.21 one thousand vs. 0.25 chiliad for newborn accomplice; −eleven.7 1000 vs. xl.four one thousand for postnatal accomplice) or fat mass (0.61 1000 vs. −0.42 thou for newborn cohort; 1.ix g vs. −17.two g for postnatal cohort) using equations derived from subjects in the lower or upper half of the body weight range, respectively.

Table 1

Correlation and standard error of estimate (Meet; expressed every bit Z-score) for predicting fatty and lean tissue mass from nude body weights of infants 1

Body weights Newborn
All weights ≤two.4 kg >2.4 kg
r SEE r SEE r SEE
Lean mass, g 0.99 0.15 0.99 0.15 0.94 0.35
Fat mass, g 0.91 0.41 0.83 0.56 0.79 0.61
Fat mass, grand/100 grand 0.80 0.61 0.51 0.86 0.58 0.82
Trunk weights Newborn
All weights ≤2.iv kg >2.4 kg
r Meet r SEE r See
Lean mass, g 0.99 0.fifteen 0.99 0.15 0.94 0.35
Fat mass, one thousand 0.91 0.41 0.83 0.56 0.79 0.61
Fat mass, g/100 g 0.80 0.61 0.51 0.86 0.58 0.82
Torso weights Postnatal
All weights ≤half dozen kg >6 kg
r SEE r Come across r SEE
Lean mass, g 0.98 0.twenty 0.94 0.35 0.91 0.42
Fatty mass, g 0.95 0.32 0.92 0.xl 0.70 0.72
Fat mass, g/100 one thousand 0.75 0.66 0.81 0.59 NS —
Body weights Postnatal
All weights ≤six kg >vi kg
r Encounter r Meet r See
Lean mass, k 0.98 0.20 0.94 0.35 0.91 0.42
Fat mass, one thousand 0.95 0.32 0.92 0.40 0.70 0.72
Fat mass, g/100 thousand 0.75 0.66 0.81 0.59 NS —

1

No pregnant difference (Educatee t test, two-tailed P = 0.49–0.97) in mean balance values for predicting lean mass (−0.21 g vs. 0.25 g for newborn cohort; −11.7 1000 vs. 40.4 g for postnatal cohort) or fat mass (0.61 thou vs. −0.42 one thousand for newborn cohort; 1.nine g vs. −17.two g for postnatal cohort) using equations derived from subjects in the lower or upper one-half of the body weight range, respectively.

FIGURE i

Individual values and percentile curves for total body: (A) lean mass; (B) fat mass; and (C) fat mass as a percentage of body weight, in 150 human newborn infants according to nude weight at study. Lines represent 10th, 25th, 50th, 75th and 90th percentiles on best-fit curves for the actual data.

Individual values and percentile curves for total body: (A) lean mass; (B) fat mass; and (C) fatty mass as a percentage of body weight, in 150 human newborn infants according to nude weight at study. Lines represent 10th, 25th, 50th, 75th and 90th percentiles on best-fit curves for the actual data.

FIGURE one

Individual values and percentile curves for total body: (A) lean mass; (B) fat mass; and (C) fat mass as a percentage of body weight, in 150 human newborn infants according to nude weight at study. Lines represent 10th, 25th, 50th, 75th and 90th percentiles on best-fit curves for the actual data.

Private values and percentile curves for full body: (A) lean mass; (B) fat mass; and (C) fat mass as a percent of torso weight, in 150 human newborn infants according to nude weight at study. Lines represent 10th, 25th, 50th, 75th and 90th percentiles on all-time-fit curves for the actual data.

Regional distribution of soft tissue mass (upper and lower extremities, and body) was as well well predicted by nude weight, with adjusted R ii from 0.87 to 0.93 for LM and from 0.78 to 0.83 for FM. With increasing body weight from one.five to 3.5 kg (an increase of 133%), there was an average increment in LM at the upper and lower extremities, and the torso of 121, 122 and 212%, respectively, although the trunk:extremities ratio for LM remained stable at ∼1.85. The average increase in FM at these regions was 540, 528 and 345%, respectively, whereas the torso:extremities ratio for FM decreased from 1.03 to 0.79.

Postnatal cohort.

For term infants during infancy, length was the nearly dominant predictor of LM, with an adjusted R ii of 0.915. However, nude weight became the dominant predictor for LM with an adjusted R 2 of 0.958 if length was excluded from the regression. Nude weight was the dominant predictor of FM with an adjusted R 2 of 0.738. There was no single predictor of %FM that resulted in an adapted Rtwo of >0.20. Gender and postnatal (study) age were the boosted predictors that could be forced into a predictive equation for these dependent variables on the footing of statistical significance, although the additional contribution to the prediction of LM, FM and %FM was <3, <12 and half dozen.2%, respectively. Female infants had significantly lower LM but higher FM and %FM (P < 0.001). Incorporating whatsoever other independent variable including blazon of milk intake (10 infants were fed human milk, 9 infants were fed homogenized whole cow'due south milk and the others were fed infant formulas) and the utilize of solids in the diet concurrent with DXA assessment failed to improve prediction. The final regression equations (P < 0.001 for all models) for the prediction of TB LM, FM and %FM including all meaning predictors are as follows:

DXA LM (1000) = −1319 + 0.278 nude weight (g) + 64.59 length (cm) − 307 gender + two.473 historic period (d)

Adjusted R 2 = 0.944, SEE 338 g

DXA FM (yard) = 908.4 + 0.706 nude weight (g) − 53.0 length (cm) + 358.five gender − 3.057 historic period (d)

Adapted R two = 0.856, Run across 345 g

DXA FM (%) = 9.57 + 0.0037 nude weight (g) + 4.56 gender − 0.0538 age (d)

Adjusted R 2 = 0.403, SEE 4.vii%

where gender = 0 for male infants and 1 for female infants.

Percentiles for DXA measurements of LM, FM and %FM in term infants during infancy based on nude weights are shown in Figure 2A–C. Adequacy of the predictive equation across the trunk weight range of our postnatal cohort was determined every bit for the newborn accomplice (Table 1). After birth, the proportion of TB LM continued to decrease, whereas the TB FM increased. The TB LM and TB FM averaged 66.3 and 31.4%, respectively, at the body weight of 10.5 kg.

Figure 2

Individual values and percentile curves during infancy for total body: (A) lean mass; (B) fat mass; and (C) fat mass as a percentage of body weight in 128 infants born at term.* All body composition measurements were expressed according to nude weight at study. Lines represent 10th, 25th, 50th, 75th and 90th percentiles on best-fit curves for the actual data. *One infant with body weight of 13,446 g was not shown in the figure.

Private values and percentile curves during infancy for total trunk: (A) lean mass; (B) fat mass; and (C) fat mass every bit a percentage of body weight in 128 infants born at term.* All body composition measurements were expressed according to nude weight at study. Lines stand for 10th, 25th, 50th, 75th and 90th percentiles on best-fit curves for the actual data. *One infant with body weight of thirteen,446 g was not shown in the figure.

FIGURE 2

Individual values and percentile curves during infancy for total body: (A) lean mass; (B) fat mass; and (C) fat mass as a percentage of body weight in 128 infants born at term.* All body composition measurements were expressed according to nude weight at study. Lines represent 10th, 25th, 50th, 75th and 90th percentiles on best-fit curves for the actual data. *One infant with body weight of 13,446 g was not shown in the figure.

Individual values and percentile curves during infancy for total body: (A) lean mass; (B) fat mass; and (C) fat mass every bit a percent of body weight in 128 infants born at term.* All body limerick measurements were expressed according to nude weight at study. Lines represent tenth, 25th, 50th, 75th and 90th percentiles on best-fit curves for the actual data. *One infant with body weight of 13,446 chiliad was non shown in the figure.

Regional distribution of soft tissue mass (upper and lower extremities, and trunk) was also well predicted by written report nude weight, with adjusted R ii from 0.92 to 0.96 for LM and from 0.82 to 0.96 for FM. With increasing body weight from 3.five to x.5 kg (an increase of 200%), in that location was an average increase in LM at the upper and lower extremities and trunk of 113, 194 and 155%, respectively, although the trunk:extremities ratio for LM remained stable at ∼i.90. The boilerplate increase in FM at these regions was 485, 573 and 365%, respectively, whereas the trunk:extremities ratio for FM decreased from 0.82 to 0.57.

Body weight vs. soft tissue mass prediction of DXA bone measurements.

When each of the variables (nude weight, LM and FM) was entered independently into the regression model, nude weight consistently provided the best predictive value at nascency and throughout infancy for DXA os measurements compared with LM and FM (Tabular array 2). Details of the DXA os measurements are reported elsewhere (Koo et al. 1996, and 1998).

Table ii

Predictive value of written report nude weight, lean mass (LM) and fat mass (FM) on dual energy X-ray absorptiometric bone measurements in human infants

At birth
Study nude weight LM FM
1 Adjusted R ii 2 Adjusted R 2 one Adjusted R ii 2 Adjusted R 2 1 Adjusted R 2 two Adjusted R two
Bone mineral content 0.950 0.950 0.913 0.905 0.955 0.884
Bone area 0.946 0.940 0.930 0.923 0.944 0.050
Os mineral density 0.840 0.832 0.812 0.026 0.854 0.061
At birth
Study nude weight LM FM
1 Adjusted R 2 ii Adjusted R 2 i Adapted R 2 ii Adjusted R 2 i Adapted R two 2 Adjusted R 2
Bone mineral content 0.950 0.950 0.913 0.905 0.955 0.884
Bone area 0.946 0.940 0.930 0.923 0.944 0.050
Bone mineral density 0.840 0.832 0.812 0.026 0.854 0.061
Postnatal
Study nude weight LM FM
1 Adjusted R two two Adjusted R 2 1 Adjusted R 2 2 Adjusted R two one Adjusted R 2 2 Adjusted R 2
Bone mineral content 0.975 0.965 0.952 0.931 0.969 0.015
Bone area 0.984 0.981 0.969 0.004 0.982 0.019
Bone mineral density 0.867 0.834 0.851 0.000 0.866 0.012
Postnatal
Study nude weight LM FM
1 Adjusted R ii 2 Adjusted R 2 1 Adjusted R two 2 Adjusted R 2 1 Adjusted R 2 2 Adjusted R two
Bone mineral content 0.975 0.965 0.952 0.931 0.969 0.015
Os area 0.984 0.981 0.969 0.004 0.982 0.019
Os mineral density 0.867 0.834 0.851 0.000 0.866 0.012

1

Variance accounted for by the complete model including one specific contained variable (written report nude weight, LM or FM) entered alone with other independent variables. (For newborn cohort: race, gender, gestational age, season of birth, study historic period and study length. For postnatal cohort: additional independent variables included nascence weight, flavour at study, and blazon of milk consumed and intake of solids on the day of study.)

two

Variance deemed for by 1 specific variable: study nude weight, LM or FM.

TABLE two

Predictive value of written report nude weight, lean mass (LM) and fat mass (FM) on dual energy X-ray absorptiometric bone measurements in human infants

At birth
Study nude weight LM FM
ane Adapted R 2 2 Adapted R 2 one Adjusted R 2 ii Adjusted R ii 1 Adjusted R ii 2 Adjusted R 2
Bone mineral content 0.950 0.950 0.913 0.905 0.955 0.884
Bone area 0.946 0.940 0.930 0.923 0.944 0.050
Bone mineral density 0.840 0.832 0.812 0.026 0.854 0.061
At nativity
Study nude weight LM FM
1 Adjusted R 2 2 Adjusted R 2 1 Adapted R 2 2 Adapted R ii 1 Adapted R 2 2 Adjusted R ii
Bone mineral content 0.950 0.950 0.913 0.905 0.955 0.884
Os expanse 0.946 0.940 0.930 0.923 0.944 0.050
Bone mineral density 0.840 0.832 0.812 0.026 0.854 0.061
Postnatal
Written report nude weight LM FM
1 Adjusted R 2 2 Adjusted R two 1 Adjusted R two ii Adjusted R two 1 Adjusted R 2 2 Adjusted R 2
Bone mineral content 0.975 0.965 0.952 0.931 0.969 0.015
Bone expanse 0.984 0.981 0.969 0.004 0.982 0.019
Bone mineral density 0.867 0.834 0.851 0.000 0.866 0.012
Postnatal
Written report nude weight LM FM
1 Adjusted R ii 2 Adjusted R ii ane Adjusted R 2 2 Adjusted R two 1 Adapted R 2 2 Adjusted R 2
Bone mineral content 0.975 0.965 0.952 0.931 0.969 0.015
Bone area 0.984 0.981 0.969 0.004 0.982 0.019
Bone mineral density 0.867 0.834 0.851 0.000 0.866 0.012

1

Variance deemed for by the complete model including one specific independent variable (study nude weight, LM or FM) entered lone with other independent variables. (For newborn cohort: race, gender, gestational age, season of birth, written report age and written report length. For postnatal cohort: boosted independent variables included nascence weight, season at study, and type of milk consumed and intake of solids on the day of study.)

2

Variance deemed for past 1 specific variable: written report nude weight, LM or FM.

Word

Changes in BC tin can have numerous functional implications in health and in affliction. For case, the amount of lean tissue mass affects positively the ventilatory function, whereas increased proportion of torso fatty has the opposite consequence (Lazarus et al. 1997); low os mass increases the fracture run a risk independently of age (Kanis et al. 1994, Koo and Steichen 1998); low body mass index associated with low trunk fat increases the risk for bone loss (Ravn et al. 1999) and possibly the development of osteoporosis and its complications. In contrast, the very high trunk mass index associated with high trunk fatty predisposes the private to the numerous complications of obesity (Dietz 1998). Thus, an increased agreement of the relation between physiologic determinants of BC in infants may lead to a greater agreement of the role of genetic and ecology influence on changes in BC and may exist critical to the management of healthy and sick infants, specially in the design and assessment of the role of nutritional intervention (Shetty 1999).

In infants, body weight can predict diverse aspects of BC during the newborn period (Koo et al. 1996, and 1998, Lapillonne et al. 1997, Rigo et al. 1998), but no data exist apropos the predictive effect of various physiologic parameters on BC across this catamenia. This study demonstrated that trunk weight contributed heavily to the model'southward explanatory power for soft tissue (LM and FM) composition during infancy. Length becomes the dominant predictor of LM with increasing postnatal historic period, although the predictive value of torso weight on soft tissue composition remains pregnant throughout infancy considering length and weight are colinear.

Information technology is well documented that in children (Taylor et al. 1997) and adults (Frisancho 1993), females have more FM and less LM than males. Females are shorter and weigh less than males at birth and throughout infancy (Brenner et al. 1976, Hamill et al. 1979), but footling is known about the earliest onset of gender-related difference in FM and LM. Just i report in newborn infants specifically showed greater FM in females compared with males (Rigo et al. 1998). Our data demonstrated that gender has an contained predictive consequence on the amount of LM and FM at nascence. In addition, the gender difference in FM and LM increased throughout infancy. The increase in FM in females was accompanied by a similar decrease in LM. In contrast, there is no gender-related departure in bone mass measurements throughout infancy (Koo et al. 1996, and 1998). The consistency and persistence of the gender-related difference in soft tissue limerick is as well reflective of the standardized technique in scan acquisition, including the consistency in the type and corporeality of roofing used for each infant, thus minimizing any interference with DXA soft tissue measurements from variable types and amounts of clothing and covering.

On the basis of differences in adapted R 2 values in the statistical models, our study demonstrated that the independent physiologic variables, i.east., weight, length and gender, appear to exist stronger predictors for the amount of LM than for FM. The ability of physiologic variables to predict FM and in detail %FM decreases with increasing postnatal age. This is presumably related to the increased role of environmental influences such as dietary intake (and physical activity in older children) on fat mass accumulation compared with lean mass (Barlow and Dietz 1998, Grandjean 1999, Shetty 1999). In this written report, the type of milk intake and the use of solids on the 24-hour interval of DXA assessment did not contribute to the determination of trunk limerick in infants. However, this report was not designed to determine the influence of dietary intake considering no details on the duration or quantity of specific intake were bachelor.

It is important to note that the large range of LM, FM and %FM at any given body weight shown in the figures represents biologic variability expressed equally percentile channels rather than predictive value of body weight on these DXA measurements. Withal, stability of the correlations (r-values) for the prediction of LM whether from body weights of total or subpopulations supports the adequacy of our model based on the full population of subjects in each cohort. Lack of meaning differences in the residuals derived from the prediction equations based on subpopulations also supports the contention that the predictive value of nude weight on LM is independent of the range of body weights within each cohort. We presented See in standard Z-score class to reverberate the office of correlation in determining the Meet considering the standard deviation of the dependent variable is unity in the Z-score measure. The observed stability of correlation across body weight ranges means that a divergence in Come across between two ranges of torso weights was due to the increased variability in dependent measure, non to a change in correlation. Our data support a similar determination for FM prediction, although the predictive ability of trunk weight for FM decreased somewhat in heavier postnatal infants. FM%, a calculated value, is poorly predicted past trunk weight specially in the postnatal cohort whether the prediction equation was based on the full or subpopulations.

In contrast to the well-defined racial differences in BC of children (Aloia et al. 1999, Chumlea and Guo 1999, Gilsanz et al. 1991) and adults (Aloia et al. 1999, Chumlea and Guo 1999, Ortiz et al. 1992), our study showed that at that place is no racial effect on soft tissue composition in this historic period grouping once body weight and length are taken into account. This is consistent with our previous reports on TB DXA bone measurements (Koo et al. 1996, and 1998), and other reports on skeletal weight, density and percentage of ash (Trotter and Hixon 1974), and distal radial BMC (Namgung et al. 1994) during infancy. Racial deviation in BC found in older ages presumably is also related to the increasing importance of ecology influence and peradventure the genetic and environmental interaction. Similarly, season did non touch soft tissue composition during infancy.

In the range of torso weights studied, changes in TB LM can be represented by linear modeling but the changes in FM were represented by both linear and nonlinear models depending on the trunk weight range. The pattern of aggregating of TB LM and FM in our birth and postnatal cohorts reflects the rapid growth during the concluding trimester and after nascence, peculiarly the accumulation of TB FM during the late in utero and postnatal periods. With increasing trunk weight, there was a greater range of LM and FM, especially of FM, supporting the greater biologic variability and increasingly of import role of environmental influences such as differences in dietary intake in larger and older infants.

Our data are comparable to other reports using the same DXA technique for newborn (Lapillonne et al. 1997, Rigo et al. 1998) and older (Mehta et al. 1998) infants. Nonetheless, strict comparison amongst studies is hard because of the unlike populations studied. Some reports included infants with birth weights of >iv kg, thus raising the possibility of large-for-gestational-historic period infants in the written report accomplice (Lapillonne et al. 1997, Rigo et al. 1998); boosted small differences may be related to the use of different models of DXA densitometer (Abrahamsen et al. 1995) and different versions of software (Picaud et al. 1999), even those provided past the same manufacturer. Nevertheless, despite the limitations associated with all in vivo techniques of BC measurement (Koo 2000), there appears to be general agreement in the overall absolute and relative changes in the soft tissue composition amid the various reports of BC based on the aforementioned DXA technique.

None of the in vivo DXA data are directly comparable with the chemical assay of cadavers (Widdowson 1975, Ziegler 1976) because the techniques used in deriving the LM and FM are not comparable with the in vivo reports (Koo 2000). Furthermore, BC extrapolated from chemical assay may not exist truly representative of normal infants considering most of the subjects reported were below the fiftieth percentile on the growth bend, the causes of decease, especially those that may have afflicted growth and BC, were not available, and there is a lack of cadaver data beyond the newborn period.

Our data show that body weight is also a major predictor of regional DXA soft tissue composition, although its predictive ability is somewhat weaker than that for TB soft tissue. Like to our findings for DXA bone measurements (Koo et al. 1996, and 1998), at that place was extensive variation in the amount of LM and FM among dissimilar regions (upper and lower extremities, trunk); the changes in these regions during in utero and postnatal growth equally indicated past our "at birth" and "postnatal" cohorts, were not directly proportional to the changes in torso weight or the soft tissue composition of the whole body. The relative difference in regional BC was as well reflected in a proportionately greater increase in FM at the extremities compared with the torso as body weight increases, whereas the proportion of LM between trunk and extremities remained steady throughout infancy.

Caution is required in the interpretation of DXA BMD, an areal density based on BMC divided by skeletal area (Nelson and Koo 1999, Prentice et al. 1994). The reasons for this circumspection include the dissimilar rate of increase in BMC and skeletal surface area during infancy and childhood, and the technical difficulty in obtaining an authentic TB surface area in a swaddled infant/child. To let better interpretation of DXA bone mass measurements, attempts take been made to normalize the DXA bone measurements on the ground of the reports in adults that LM is a adept predictor of bone mass as BMC (Ferretti et al. 1998, Valdimarsson et al. 1999) or as BMD (Courteix et al. 1999, Valdimarsson et al. 1999) and that FM is a skilful predictor of BMC (Ferretti et al. 1998) and BMD (Courteix et al. 1999). All the same, nosotros showed that in healthy infants, LM is an independent predictor of TB BMC throughout infancy and TB Surface area in newborns, merely is consistently weaker than the use of body weight to predict these measurements. Furthermore, LM has minimal predictive value on TB BMD, and FM has minimal predictive value for any DXA os measurement except BMC in the newborn period. These findings suggest that weight-bearing and impact-loading exercise critical to the increase of LM and bone mass in older subjects (Courteix et al. 1999, Pettersson et al. 1999) are not well adult in younger subjects. Thus the utilize of LM or FM provides no advantage over body weight in the prediction of skeletal os mineral status during infancy.

We conclude that in healthy infants, body weight is the ascendant predictor of LM and FM, although length has the aforementioned or stronger predictive value for LM with increasing postnatal age. Physiologic variables have footling predictive value for %FM across the newborn period. Gender difference in LM and FM tin be demonstrated at birth and increases throughout infancy. The use of LM or FM offers no advantage over torso weight in the prediction or normalization of bone mass measurements during infancy. Our data on the precision of DXA measurement and the physiologic factors that influence BC are important to the design and assessment of nutritional intervention studies in infants under various physiologic and pathologic situations.

Nosotros appreciate the help provided by the nursing staff of the Memphis Clinical Research Center.

LITERATURE CITED

Abrahamsen

,

B.

,

Gram

,

J.

,

Hansen

,

T. B.

 &

Beck-Nielsen

,

H.

 (

1995

)

Cross calibration of QDR 2000 and QDR 1000 dual energy X-ray densitometers for os mineral and soft tissue measurements

.

Bone

16

:

385

–

390

.

Aloia

,

J. F.

,

Vaswani

,

A.

,

Mikhail

,

M.

 &

Flaster

,

E. R.

 (

1999

)

Body limerick by dual-energy 10-ray absorptiometry in black compared with white women

.

Osteoporos. Int.

10

:

114

–

119

.

Altman

,

D. G.

(

1993

)

Construction of age-related reference centiles using absolute residuals

.

Stat. Med.

12

:

917

–

924

.

Ballard

,

J. Fifty.

,

Khoury

,

J. C.

,

Wedig

,

Yard.

,

Wang

,

50.

,

Eilers-Walsman

,

B. L.

 &

Lipp

,

R.

 (

1991

)

New Ballard score, expanded to include extremely premature infants

.

J. Pediatr.

119

:

417

–

423

.

Barlow

,

S. Due east.

&

Dietz

,

W. H.

(

1998

)

Obesity evaluation and treatment: expert committee recommendations

.

Pediatrics

102

:

e29

.

Bland

,

J. Thou.

&

Altman

,

D. K.

(

1986

)

Statistical methods for assessing agreement betwixt two methods of clinical measurement

.

Lancet

1

:

307

–

310

.

Brenner

,

W. E.

,

Edelman

,

D. A.

&

Hendricks

,

C. H.

 (

1976

)

A standard of fetal growth for the U.s.a. of America

.

Am. J. Obstet. Gynecol.

126

:

555

–

564

.

Chumlea

,

Due west. C.

&

Guo

,

S. S.

(

1999

)

Body mass and bone mineral quality

.

Curr. Opin. Rheumatol.

xi

:

307

–

311

.

Courteix

,

D.

,

Lespessailles

,

E.

,

Jaffre

,

C.

,

Obert

,

P.

 &

Benhamou

,

C. Fifty.

 (

1999

)

Bone mineral conquering and somatic development in highly trained girl gymnasts

.

Acta Paediatr

88

:

803

–

811

.

Dietz

,

W. H.

(

1998

)

Health consequences of obesity in youth: babyhood predictors of adult disease

.

Pediatrics

101

:

518

–

525

.

Ferretti

,

J. L.

,

Capozza

,

R. F.

,

Cointry

,

G. R.

,

Garcia

,

S. L.

,

Plotkin

,

H.

,

Alvarez Filgueira

,

Grand. L.

 &

Zanchetta

,

J. R.

 (

1998

)

Gender related differences in the human relationship between densitometric values of whole-body bone mineral content and lean body mass in humans between 2 and 87 years of age

.

Os

22

:

683

–

690

.

Frisancho

,

A. R.

(

1993

)

Anthropometric standards

.

Frisancho

,

A. R.

eds.

Anthropometric Standards for the Assessment of Growth and Nutritional Status

1st ed. :

37

–

118

The University of Michigan Press Ann Arbor

,

MI

.

Gilsanz

,

5.

,

Roe

,

T. F.

,

Mora

,

S.

,

Costin

,

G.

 &

Goodman

,

W. G.

 (

1991

)

Changes in vertebral os density in blackness girls and white girls during babyhood and puberty

.

N. Engl. J. Med.

325

:

1597

–

1600

.

Grandjean

,

A.

(

1999

)

Nutritional requirements to increase lean mass

.

Clin. Sports Med.

18

:

623

–

632

.

Hamill

,

P.V.V.

,

Drizd

,

T. A.

,

Johnson

,

C. L.

,

Reed

,

R. B.

,

Roche

,

A. F.

 &

Moore

,

W. Thousand.

 (

1979

)

Concrete growth: National Center for Health Statistics percentiles

.

Am. J. Clin. Nutr.

32

:

607

–

629

.

Kanis

,

J. A.

,

World Health Organization Report Group

(

1994

)

Assessment of fracture and its application to screening for postmenopausal osteoporosis: synopsis of a WHO report

.

Osteoporos. Int.

iv

:

368

–

381

.

Koo

,

W.West.K.

(

2000

)

Body composition measurements during infancy

.

Ann. N.Y. Acad. Sci.

904

:

383

–

392

.

Koo

,

W.West.Grand.

,

Bush

,

A. J.

,

Walters

,

J.

 &

Carlson

,

S. Eastward.

 (

1998

)

Postnatal development of bone mineral condition during infancy

.

J. Am. Coll. Nutr.

17

:

65

–

70

.

Koo

,

West.West.1000.

,

Massom

,

L. R.

&

Walters

,

J.

 (

1995a

)

Validation of accuracy and precision of dual free energy X-ray absorptiometry for infants

.

J. Os Miner. Res.

10

:

1111

–

1115

.

Koo

,

Westward.Westward.Chiliad.

&

Steichen

,

J. J.

(

1998

)

Osteopenia and rickets of prematurity

.

Polin

,

R.

Fox

,

West.

 eds.

Fetal and Neonatal Physiology

2nd ed. :

2335

–

2349

W. B. Saunders Philadelphia

,

PA

.

Koo

,

Westward.W.G.

,

Walters

,

J.

&

Bush

,

A. J.

 (

1995b

)

Technical considerations of dual energy 10-ray absorptiometry-based bone mineral measurements for pediatric studies

.

J. Bone Miner. Res.

ten

:

1998

–

2004

.

Koo

,

Due west.W.M.

,

Walters

,

J.

,

Bush

,

A. J.

,

Chesney

,

R. W.

 &

Carlson

,

Due south. E.

 (

1996

)

Dual energy Ten-ray absorptiometry studies of bone mineral status in newborn infants

.

J. Bone Miner. Res.

xi

:

997

–

1002

.

Lapillonne

,

A.

,

Braillon

,

P.

,

Claris

,

O.

,

Chatelain

,

P. G.

,

Delmas

,

P. D.

 &

Salle

,

B. L.

 (

1997

)

Body composition in advisable and in pocket-size for gestational age infants

.

Acta Paediatr

86

:

196

–

200

.

Lazarus

,

R.

,

Colditz

,

Chiliad.

,

Berkey

,

C. Southward.

 &

Speizer

,

F. E.

 (

1997

)

Furnishings of body fat on ventilatory function in children and adolescents: cross-exclusive findings from a random population sample of school children

.

Pediatr. Pulmonol.

24

:

187

–

194

.

Mehta

,

K. C.

,

Specker

,

B. L.

,

Bartholmey

,

Due south.

,

Giddens

,

J.

 &

Ho

,

M. L.

 (

1998

)

Trial on timing of introduction to solids and nutrient type on infant growth

.

Pediatrics

102

:

569

–

573

.

Namgung

,

R.

,

Tsang

,

R. C.

,

Specker

,

B. L.

,

Sierra

,

R. I.

 &

Ho

,

M. Fifty.

 (

1994

)

Low bone mineral content and high serum osteocalcin and 1,25-dihydroxyvitamin D in summertime- versus winter-born newborn infants: an early fetal event?

.

J. Pediatr. Gastroenterol. Nutr.

19

:

220

–

227

.

Nelson

,

D. A.

&

Koo

,

W. W. K.

(

1999

)

Estimation of absorptiometric bone mass measurements in the growing skeleton: issues and limitations

.

Calcif. Tissue Int.

65

:

i

–

3

.

Ortiz

,

O.

,

Russell

,

G.

,

Daley

,

T. L.

,

Baumgartner

,

R. N.

,

Waki

,

M.

,

Lichtman

,

S.

,

Wang

,

J.

,

Pierson

,

R. N.

,

Jr & Heymsfield

,

Due south. B.

 (

1992

)

Differences in skeletal muscle and os mineral mass between blackness and white females and their relevance to estimates of body composition

.

Am. J. Clin. Nutr.

55

:

viii

–

thirteen

.

Pettersson

,

U.

,

Nordstrom

,

P.

&

Lorentzon

,

R.

 (

1999

)

A comparison of bone mineral density and musculus forcefulness in young male person adults with dissimilar practise level

.

Calcif. Tissue Int.

64

:

490

–

498

.

Picaud

,

J.C.

,

Nyamugabo

,

K.

,

Braillon

,

P.

,

Lapillone

,

A.

,

Claris

,

O.

,

Delmas

,

P.

,

Meunier

,

P.

,

Salle

,

B. 50.

 &

Rigo

,

J.

 (

1999

)

Dual energy X-ray absorptiometry in small-scale subjects: influence of dual energy X-ray equipment on assessment of mineralization and body composition in newborn piglets

.

Pediatr. Res.

46

:

772

–

777

.

Prentice

,

A.

,

Parsons

,

T. J.

&

Cole

,

T. J.

 (

1994

)

Uncritical use of os mineral density in absorptiometry may lead to size-related artifacts in the identification of bone mineral determinants

.

Am. J. Clin. Nutr.

60

:

837

–

842

.

Ravn

,

P.

,

Cizza

,

G.

,

Bjarnason

,

N. H.

,

Thompson

,

D.

,

Daley

,

M.

,

Wasnich

,

R. D.

,

McClung

,

One thousand.

,

Hosking

,

D.

,

Yates

,

A. J.

 &

Christiansen

,

C.

 (

1999

)

Low body mass index is an important chance gene for low bone mass and increased os loss in early postmenopausal women. Early postmenopausal intervention cohort (EPIC) study

.

J. Bone Miner. Res.

xiv

:

1622

–

1627

.

Rigo

,

J.

,

Nyamugabo

,

K.

,

Picaud

,

J. C.

,

Gerard

,

P.

,

Pieltain

,

C.

 &

De Curtis

,

M.

 (

1998

)

Reference values of body limerick obtained past dual free energy X-ray absorptiometry in preterm and term neonates

.

J. Pediatr. Gastroenterol. Nutr.

27

:

184

–

190

.

Shetty

,

P. S.

(

1999

)

Adaptation to low energy intakes: the responses and limits to low intakes in infants, children and adults

.

Eur. J. Clin. Nutr.

53

:

S14

–

S33

.

Taylor

,

R. Due west.

,

Gold

,

East.

,

Manning

,

P.

 &

Goulding

,

A.

 (

1997

)

Gender differences in body fat content are nowadays well before puberty

.

Int. J. Obes. Relat. Metab. Disord.

21

:

1082

–

1084

.

Trotter

,

M.

&

Hixon

,

B. B.

(

1974

)

Sequential changes in weight, density, and pct ash weight of human being skeletons from an early fetal menstruation through onetime age

.

Anat. Rec.

179

:

1

–

18

.

Valdimarsson

,

O.

,

Kristinsson

,

J. O.

,

Stefansson

,

Due south. O.

,

Valdimarsson

,

S.

 &

Sigurdsson

,

K.

 (

1999

)

Lean mass and physical activity every bit predictors of os mineral density in sixteen–20 year onetime women

.

J. Intern. Med.

245

:

489

–

496

.

Widdowson

,

Due east. M.

(

1975

)

Changes in body proportions and composition during growth and limerick during growth

.

Davis

,

J. A.

Dobbing

,

J.

 eds.

Scientific Foundations of Pediatrics

1st ed. :

153

–

163

William Heineman Medical Books London

,

UK

.

Ziegler

,

E. Due east.

,

O'Donnel

,

A. Yard.

,

Nelson

,

S. E.

 &

Fomon

,

Southward. J.

 (

1976

)

Body composition of the reference fetus

.

Growth

46

:

537

–

556

.

Abbreviations

  • Area

  • BC

  • BMC

  • BMD

  • DXA

    dual energy Ten-ray absorptiometry

  • FM

  • %FM

    fatty mass as a per centum of body weight

  • LM

  • SEE

    standard mistake of estimate

  • TB

Footnotes

one

Supported by a University of Tennessee Medical Research Grant and by The University of Tennessee-Memphis Clinical Research Center, USPHS grant RR 00211–29.

© 2000 The American Society for Nutritional Sciences