| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Paper |
Departments of Pediatrics, Obstetrics and Gynecology, and Preventive Medicine, and Division of Biostatistics and Epidemiology (A.B.), The University of Tennessee, Memphis, Memphis
Address reprint requests to: Winston W.K. Koo, MBBS, FACN, Hutzel Hospital, Department of Pediatrics, 4707 St. Antoine Blvd, Detroit, MI 48201.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Methods: Cross-section observational study of total body bone mineral content (TB BMC) and density (TB BMD) of 130 healthy infants (71 male and 59 female with 63 white and 67 African American) between 1 and 391 days. Whole body dual energy X ray absorptiometry (DXA) scans were performed on unsedated infants using a bone densitometer with pediatric platform. Scan analyses were performed with software version V5.64P. The ability of study independent variables to explain variance in bone mineral status was determined by multiple linear regression analysis.
Results: During infancy, average TB BMC increased by 389% and TB BMD increased by 157%. The best determinant of bone mineral status is body weight which accounted for 97% of TB BMC, 98% of TB area and 86% of TB BMD variation. Postnatal age and body length jointly added only 1%, <1% and 2.5%, respectively, to the explained variation of these DXA measurements; race, gender and season all failed to reach statistical significance.
Conclusion: In healthy infants, body weight is the dominant predictor of bone mineral status. The percent increase in TB BMC differs from increase in TB BMD. Normative data generated from this study would be useful in the identification of abnormal bone mineral status in infants.
Key words: infant, dual energy X ray absorptiometry, bone mineral content, calcium
Abbreviations: PA=postnatal age • DXA=dual energy x-ray absorptiometry • TB BMC=total body bone mineral content • TB area=total body skeletal area • TB BMD=total body bone mineral density • TB Ca=total body calcium.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All anthropometric measurements: body weight, recumbent length and head circumference were measured by one of the two investigators (JW or WK) and another assistant according to standard procedures as described previously [2,7]. Anthropometric measurements obtained by each of the investigators were highly significantly correlated (r=0.993 to 0.999, p<0.000). Body weight was determined separately from weights of the materials (cotton blanket, and diaper if present) covering the infants during scan acquisition.
DXA scan acquisition with a whole body scanner (Hologic QDR 1000/W densitometer, Hologic Inc., Waltham, MA) and the use of pediatric platform were described previously [2,7,9]. With our densitometer, the typical entry radiation exposure during a pediatric whole body scan showed a maximum dose of 3 uSv (1 uSv=0.1 mrem) with the use of a pediatric platform interposed between the x-ray source and the infant. The radiation scatter at 90 cm from the scanner was <0.03 uSv from 10 minutes of measurement. Long-term (> 3 years) coefficient of variation (CV) for the determination of bone mineral content (BMC), area, and bone mineral density (BMD), using an anthropometric spine phantom is < 0.31% for all parameters. The average annual rate of change for each of these measurements is not significantly different from zero. The in vivo replication of DXA measurements in 50 infants was highly significantly correlated (r=0.99, p<0.000). The standard deviation of differences [10] between paired DXA measurements in these infants for total body (TB) BMC was 3.8% at a mean of 93 g, TB area was 1% at a mean of 371 cm2, and TB BMD was 2.8% at a mean of 0.228 g/cm2 respectively.
All infants were scanned without sedation or additional restraint. Only scans without significant movement artifact [9] were analyzed using the software developed in conjunction with the manufacturer (Version V5.64P). In addition to analysis of the whole body scan, analyses of different regions: head, trunk and each of the four extremities were also performed using the same software if the position of the infant allowed adequate delineation of separate regions. Body calcium content was calculated based on the assumption that calcium accounts for 34% of osseous minerals and that the skeleton contains 99% of the body’s calcium [11].
Multiple regression analyses were used to determine the predictive effect of six independent variables that may affect growth and bone mineralization: race, gender, birth weight, body weight, length, and postnatal age (PA) at study, on each of the three dependent variables (BMC, area, and BMD). For infants older than 1 week, the type of milk intake at study was used as an additional independent variable. Regression models were formed to identify the most powerful individual and joint explanations for the variance of each of the study’s overall bone mineral status variables. Once the most powerful explanatory variable was identified, other independent variables were systematically incorporated to determine their unique effects. Identical procedures were followed to determine the best explanation in regional (head, trunk, upper limbs, lower limbs) BMC, area, and BMD variations. In the case of head region, head circumference was also taken into account as an additional independent variable.
Analysis of variance was performed to determine the presence of seasonal variability in anthropometric and DXA measurements according to whether the birth month falls within a 3 sequential calendar month period beginning in January. We tested for interactive effect of PA, gender and race on the DXA and anthropometric measurements. In these analyses, we categorized PA into five age intervals: newborn, and thereafter at four intervals with approximately equal number of subjects. In addition, general linear models were employed to explore for interactive effects for season with gender and race on these same dependent variables. All tests were performed with SPSS Windows Version 7.5.2 at an adopted significance level of 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
The continuous independent variables: body weight, length, and PA were strongly interrelated with r values of 0.860 to 0.999 (p<0.0001 for all comparisons). Birth weight had much lower correlation to other independent variables with r values of 0.051 to 0.211. DXA derived total weight was significantly (p<0.001) correlated with study bare weight (r=0.98) and total study weight including all materials covering the infant (r=0.98).
Study bare weight consistently proved to be the single best predictor of TB BMC, TB area, and TB BMD, with r2 values of 0.97, 0.98, and 0.86, respectively. Distribution of TB BMC according to study bare weight is shown in Fig 1. For modeling TB BMC and TB BMD, PA and study length were additional predictors that resulted in a statistically significant (p<0.001) but minimal increase in the model’s r2 beyond that achieved by study bare weight alone. Specifically, PA explained an additional 0.7% and 1.7% of the variability of TB BMC and TB BMD respectively, while study length explained an additional 0.2% and 0.8% of the variability of the same measures beyond that of the models detailed below. For modeling TB area, study length was the only statistically significant (p<0.001) predictor other than study bare weight, but its contribution to the additional explanatory power of the model was <0.5%.
|
TB BMC g=77.24 (25.85)+24.94 (1.96) Bare Weight kg+0.21 (0.03) PA days-1.889 (0.62) Length cm.
TB Area cm2=-94.28 (36.75)+49.95 (3.74) Bare Weight kg+4.94 (0.98) Length cm.
TB BMD g/cm2=0.295 (0.04)+0.016 (0.003) Bare Weight kg+0.0002 (0.0001) PA days-0.003 (0.001) Length cm.
The final r2 values, including all statistically significant predictors, were 0.976, 0.985 and 0.858 for TB BMC, TB area and TB BMD, respectively. Incorporating any other independent variable including type of milk intake (10 infants were fed human milk, nine infants were fed homogenized whole cow milk and the others were fed infant formulas) concurrent with DXA assessment failed to improve prediction.
Actual distributions of TB BMC, TB area, and TB BMD are shown in Tables 1 and 2 respectively. TB calcium (Ca) based on TB BMC also are shown in the same Tables. Neither gender nor seasonal effects were present for any of the study’s dependent variables. When considered in isolation, race proved to be a statistically significant predictor of our DXA measurements with a maximum r2 of 0.046. But when included in a model containing study weight, the effect disappeared from the models for all three DXA variables.
|
Study bare weight remains the best predictor for all regional BMC (r2=0.83 to 0.93) and area (r2=0.88 to 0.96) although head circumference was a significant predictor for head area. Study bare weight’s predictive ability decreases for regional BMD (r2=0.54 to 0.86) and particularly for lower limb BMD (r2=0.06), although it remains statistically significant for all regions. With increasing body weight from 2.5 kg to 13.5 kg, BMC of head as a proportion of TB BMC increased by 27% while BMC of trunk, upper limbs and lower limbs decreased by 34%, 19%, and 23%, respectively (p<0.05 for all comparisons). The DXA measured area of lower limbs as a proportion of TB area increased by 54% while the area of upper limbs decreased by 14% (p<0.005 for all comparisons), and the area of head or trunk as a proportion of TB area did not change significantly.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
DXA determined mass measurements of the total weight of the subject being scanned i.e., including all materials covering the infant. Total study weight is subject to variabilities in the amount and type of diaper, clothing, blanket, etc., and it is important to record the type and amount of material covering the infant during scan acquisition since they may constitute a significant proportion of the total mass particularly for subjects with low body mass. However, the consistency of our technique, including the use of consistent type and quantity of covering appears to be the major factor accounting for the high collinearity among study bare weight, total study weight and DXA determined weight. However, the use of study bare weight as a predictor of DXA measurements has the advantage of ease of reproducibility without concern for variability of covering materials and allows for comparison among different studies.
Preliminary data have shown different relationships between bone mineral status and body mass for infants with abnormal versus normal bone mineralization [13,14]. For example, TB BMC is significantly lower in small preterm infants, and it is significantly higher in infants with congenital osteopetrosis, when compared to healthy infants with similar body mass. Thus specific measurement of bone mineral status is critical to the understanding of pathophysiologic mechanisms in normal and abnormal situations, and DXA measurements of bone mass corrected for body weight may be a more sensitive index than DXA BMC alone to detect deviations of bone mineral status from normal.
This study documented that race and gender did not play a significant role in the determination of bone mineral status throughout infancy after accounting for variations in body weight. This is consistent with the lack of racial and gender effect on bone mineral status based on DXA [7] or single photon absorptiometry [15] measurements in neonates and direct measurement of skeletal weight, density and percent ash in infants and children [16], although one report shows a higher bone mass in male infants [8].
Conflicting reports exist on the seasonal effect on bone mineral status. In adults, summer months are associated with an average of 3.6% higher BMD at distal radius [17] and 1.4% higher BMD at lumber spines [18] compared to fall and winter months, whereas, neonates born in summer were reported to have an 8% lower distal radial BMC compared to those born in winter [15]. Our data extends the previous observation that lack of seasonal variability in whole body BMC or BMD in neonates [7] persists throughout infancy. The reason for discrepancies reported for seasonal variability is not clear although it is known that regional measurement of bone mineralization correlates well but is not identical to whole body bone mineral status. Furthermore, changes in regional bone status is relatively small and may not be sufficient to significantly change the whole body bone mineral status.
The predictive value of body weight on DXA BMD throughout infancy is lower than that for BMC or area. This can be explained by the fact that DXA BMC and DXA area measurements increase at different rates leading to different rate of change in DXA BMD since the latter is an areal density calculated by BMC/area. For example, our data demonstrated that tripling the body weight is accompanied by about four-fold increase in TB BMC but only a 2.5-fold increase in TB area, thus resulting in 1.5-fold increase in TB BMD. Furthermore, DXA area measurement can be affected by the subject’s posture during scan acquisition. For example, it is conceivable that all extremities of an infant might be superimposed over the trunk leading to a lower measured DXA area and a disproportionately higher BMD. Therefore, the caution for using DXA BMD as an indicator of bone mineral status in older subjects [19] is also applicable to infants.
The ability to obtain regional measurements from whole body DXA scans offers the opportunity to determine the regional changes under normal and pathological conditions although one should be aware that these measurements are less precise than whole body measurements. The marked increase in DXA measured lower limb area versus a lesser increase in lower limb BMC as a proportion of total body DXA measurements in the growing infant exemplifies the caution needed when using DXA BMD as an indicator of bone mineral status. Thus, regional DXA data obtained from whole body scans are suitable primarily as a research tool [7,9].
Our DXA derived TB Ca is based on the observations by us [2] and by others [1,3,4] that DXA derived BMC correlates directly with carcass ash content of animals with body mass between 886 g to 35 kg, and the assumption that Ca is a constant proportion of bone ash [11]. Our TB Ca values cannot be compared directly with the widely used reference TB Ca data for infants [20] since the latter values were extrapolated from increase in body length (reflecting increase in skeletal mass) and fat-free femoral calcium content at birth, 4 months and 1 year of age. However, the TB Ca from DXA, when calculated as daily increment up to 5 months is 177 mg/day and is similar to the net retention of 164 to 210 mg/day from balance studies in infants fed human milk and infant formulas between 8 to 122 days; whereas, the increase in TB Ca from DXA increased between 5 and 12 months is about 150 mg which is lower than the average daily net retention of calcium at 238 to 259 mg from studies of infants fed whole cow milk, milk-based or soy-based formulas between 4 months and 12 months [20]. Limited dietary information on our subjects precludes any conclusion on the role of dietary intake on bone mineral status during infancy.
Our study demonstrated that DXA measurement of whole body bone mineral status may be a useful additional tool for determining the nutritional calcium requirement in human infants. Our data on the variation in DXA-derived TB Ca and potentially other bone minerals, if confirmed, could have significant impact on the current estimates of the developmental changes in elemental body composition and nutrient requirement of infants.
|
ACKNOWLEDGMENTS |
|---|
Received March 1, 1997.
Revised July 1, 1997.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
