body composition and its components in preterm and term newborns: a cross-sectional, multimodal...

Original Research Article

Body Composition and Its Components in Preterm and Term Newborns:A Cross-Sectional, Multimodal Investigation

IRFAN AHMAD,1 DAN NEMET,1,2 ALON ELIAKIM,1,2 ROBIN KOEPPEL,3 DONNA GROCHOW,3 MARIA COUSSENS,3

SUSAN GALLITTO,3 JULIA RICH,1 ANDRIA PONTELLO,1 SZU-YUN LEU,1 DAN M. COOPER,1 AND FEIZALWAFFARN1

1Department of Pediatrics, College of Health Sciences, University of California, Irvine, California2Department of Pediatrics, Meir Medical Center, Sackler Faculty of Medicine, Tel Aviv University, Israel3Program in Nursing Science, College of Health Sciences, University of California, Irvine, California

ABSTRACT A prospective, cross-sectional, observational study in preterm and term infants was performed to com-pare multimodal measurements of body composition, namely, limb ultrasound, bone quantitative ultrasound, and dualX-ray absorptiometry (DXA). One hundred and two preterm and term infants appropriate for gestational age were en-rolled from the newborn nursery and neonatal intensive care unit. Infants were included when they were medically sta-ble, in an open crib, on full enteral feeds and within 1 week of anticipated discharge. Correlations among the variousmeasurements of body composition were performed using standard techniques. A comparison between preterm infant(born at 28–32 weeks) reaching term to term-born infants was performed. Limb ultrasound estimates of cross-sectionalareas of lean and fat tissue in a region of tissue (i.e., the leg) were remarkably correlated with regional and whole-bodyestimates of fat-free mass and fat obtained from DXA suggesting the potential usefulness of muscle ultrasound as aninvestigative tool for studying aspects of body composition in this fragile population. There was a weak but significantcorrelation between quantitative ultrasound measurements of bone strength and DXA-derived bone mineral density(BMD). Preterm infants reaching term had significantly lower body weight, length, head circumference, muscle and fatcross-sectional area, bone speed of sound, whole-body and regional lean body mass, fat mass, and BMD compared toterm-born infants. Current postnatal care and nutritional support in preterm infants is still unable to match the in-utero environment for optimal growth and bone development. The use of relatively simple bedside, noninvasive bodycomposition measurements may assist in understanding how changes in different components of body composition earlyin life affect later growth and development. Am. J. Hum. Biol. 22:69–75, 2010. ' 2009 Wiley-Liss, Inc.

In healthy newborns, lean (predominately muscle), fat,and bone tissues, the key components of body mass,increase rapidly. The pattern of this increase is profoundlyaltered by premature birth, and disruptions in the ontog-eny [or as suggested by Wells et al. (2007), ‘‘program-ming’’] of body composition are associated with seeminglydisparate lifelong conditions such as obesity and increasedrisk of cardiovascular disease (Hofman et al., 2006; Jogle-kar et al., 2007) on the one hand and growth retardation(Farooqi et al., 2006) and bone disease of prematurity(Sharp, 2007) on the other. Although little is known aboutthe mechanisms that control the early-in-life developmentof muscle, fat, and bone, there is mounting evidence thatthe factors regulating body composition among these com-partments are distinct leading to differing rates of growth.In the case of muscle and fat tissues, for example, growthmay be antagonistic (Quinn, 2007): Shibata et al. (2006)recently noted in fetal animal models that muscle-derivedmyostatin may play a role in intramuscular fat bydecreasing myogenesis and increasing adipogenesis. Incontrast, in the case of muscle and bone, growth of the twotissues may be synergistic (Thomopoulos et al., 2007).

Body composition is difficult to study in prematurebabies largely because of the limited tools appropriate forthis fragile population. We used a multimodal approachby simultaneously employing three distinct techniquespreviously used in premature babies, albeit in relativelysmall samples: (1) dual X-ray absorptiometry (DXA),which measures lean, fat, and bone tissues and can beused for these measurements both for the whole organismand for specific anatomic regions; (2) limb ultrasound

(LUS), which quantifies regional muscle and fat mass(FM) (Pereira-da-Silva et al., 1999; Weiss and Clark,1985); and (3) bone quantitative ultrasound (QUS), whichis related to bone strength (Littner et al., 2003; Peredaet al., 2003).We chose a cross-sectional approach specifically so that

we could readily compare results obtained from the threedifferent methods of assessing body composition compo-nents over a wide range of body masses. This approachwould help determine the feasibility of these measure-ments ‘‘at the bedside.’’ Finally, the study would, webelieve, begin to demonstrate how these techniques couldbe optimally used to assess the development of body com-position in newborns in longitudinal studies.We also wanted to study the differences in body compo-

sition measurements in premature infants versus terminfants at the time they approach discharge to understandthe impact of premature birth and neonatal intensive careunit (NICU) stay on infant growth. We hypothesized that(1) regional muscle, fat, and bone measurements obtained

Contract grant sponsor: NIH; Contract grant numbers: R01NR009070,MO1-RR00827.

*Correspondence to: Dan M. Cooper, Bldg 25, 2nd Floor, UC Irvine Medi-cal Center, 101 The City Drive, Orange, CA 92868, USA.E-mail: [email protected]

Received 6 November 2008; Revision received 24 February 2009;Accepted 17 April 2009

DOI 10.1002/ajhb.20955

Published online 16 June 2009 in Wiley InterScience (www.interscience.wiley.com).

AMERICAN JOURNAL OF HUMAN BIOLOGY 22:69–75 (2010)

VVC 2009 Wiley-Liss, Inc.

by ultrasound (US) at the bedside would correlate withsimultaneously obtained whole-body DXA measurementsof body composition; and (2) body composition in prema-ture infants approaching time of discharge would be dif-ferent from newborn infants of similar postmenstrual age(PMA).

PARTICIPANTS AND METHODS

Participants

A total of 102 preterm and term infants (Table 1), appro-priate for gestational age, hospitalized in the neonatalintensive care unit or newborn nursery at UC IrvineMedical Center were enrolled in this study. Infants wererecruited when they were medically stable, in an opencrib, on full enteral feeds, and approaching discharge. Theinfants were averaging over 120–130 kcal/kg/day anddemonstrated progressive weight gain. Exclusion criteriaincluded infants with evidence of chromosomal or othermajor genetic abnormality, suspected neuromuscular dis-orders, and significant chronic lung disease. Infants bornto mothers with diabetes, substance abuse, and other en-docrine abnormalities were also excluded. The study wasapproved by the UC Irvine Institutional Review Board(IRB) and written parental consent was obtained fromboth parents in accordance with IRB requirements. Allmeasurements were performed prior to discharge fromthe NICU or newborn nursery. The PMA of the infants atthe time of the measurement was calculated from themother’s last menstrual period.

Weight, length, and head circumference

Standard electronic calibrated scales were used to mea-sure weight. Length was measured on a standard length-measuring board by two trained NICU nurses. Head cir-cumference was measured by a disposable paper-mea-suring tape. All measurements were obtained on the dayof US and DXA measurements.

Dual X-ray absorptiometry

All participating infants were transported in a standardneonatal transport incubator to the DXA site located inthe same building floor as the NICU. Hologic QDR Discov-ery-A (Hologic, Bedford, MA) machine was used for DXAmeasurements. The infants were swaddled in a cotton

sheet and a pacifier with sucrose was used to minimizemovement artifacts. A heat lamp was used to maintain acomfortable body temperature. Pediatric software in theDXA scanner was used to analyze whole-body and re-gional (right tibia) bone area, bone mineral content (BMC),and bone mineral density (BMD). Apparent BMD was calcu-lated by dividing the BMD to body length (m).

Limb ultrasound

Clear visualization of muscle boundaries is possible,since epimysium surrounding the muscles is highly reflec-tive. LUS has been used to assess muscle mass (muscularhypertrophy) in an infant with myostatin deficiency(Schuelke et al., 2004), and it has also been used to assesscardiac and skeletal muscle mass in small animals (dePaiva et al., 2003; Dietz et al., 1999). Besides measure-ment of muscle mass, FM can also be estimated with thistechnique (Scholten et al., 2003). All studies were per-formed by a trained single ultrasonographer from the UCIrvine Department of Radiology using a Philips ATLmachine (model) with 12-MHz probe. Scans were obtainedof right calf, one-third the distance from the poplitial fossato the heel. Images were stored digitally as bitmap files.Total leg cross-sectional area and muscle cross-sectionalarea (inclusive of tibia and fibula) on these US scans weremeasured using Scion software (obtained from the NIHweb site, http://rsb.info.nih.gov/nih-image/about.html).The difference between the two was recorded as subcuta-neous fat cross-sectional area.

Bone quantitative ultrasound

Speed of sound (SOS) measurements were made usingthe Sunlight Premiere QUS machine by a trained studyteam nurse. A standardized procedure was followed andthe probe was placed on the left tibia at half the measureddistance between the apex of the heel and the distal patel-lar apex. After calibrating the machine with a standardphantom, three measurements were obtained from thesame site and the mean value was calculated.All technicians performing DXA, LUS, or QUS were

blinded to the prior medical history of the examinedinfants.

Statistical analysis

Data are presented descriptively with mean and 95%confidence interval (CI) and frequency and percentage.Pearson’s product–moment correlations were used todescribe the relationships between different measure-ments. Since all the measurements are also correlatedwith age, Pearson’s partial correlations were also obtainedto adjust for the effect of PMA.A linear model was also fitted for US- and DXA-meas-

ured lean mass (LM), FM, and bone measurements withPMA as a predictor. All measurements were standardizedso that the slopes against age could be compatible. Themean standardized slope and 95% CI against PMA andgestational age are presented separately in bar chart.

RESULTS

Sample population

Descriptive data of the infants enrolled in the study areshown in Table 1. The mean gestational age at birth for

TABLE 1. Participant characteristics (N 5 102)

Mean 95% CI

Gestational age at birth (weeks) 33.7 (32.6, 34.7)Chronological age (days) 28.5 (21.7, 35.2)Post menstrual age (weeks) 37.8 (37.3, 38.4)Birth weight (g) 2,296.0 (2,076.1, 2,516.0)Birth length (cm) 44.2 (42.9, 45.5)Birth head circumference (cm) 30.7 (29.8, 31.5)

Frequency Percentage

Male 51 50.0Hispanic 49 48.0Caucasian 35 34.3African American 8 7.8Asian 8 7.8Others 2 2.0

70 I. AHMAD ET AL.

American Journal of Human Biology

the group was 33.7 6 0.5 weeks. Sixty-three infants wereborn premature (gestational age 30.3 6 0.3) and 39 wereborn at term (>37 weeks gestation, gestational age 39.4 60.2). Because of heterogeneity of the participants, infantswere divided into four groups according to their gesta-tional age [term: >37 weeks (n 5 39); preterm: 23–28weeks (n 5 20), 29–32 weeks (n 5 26), and 33–36 weeks (n5 17)]. By group, birth characteristics, physical measure-ments, and body composition (DXA and US) are presentedin Table 2. No gender differences were found in all meas-urements. DXA measurements of bone and muscle wereobtained in all infants. Because of incomplete visualiza-tion of muscle or subcutaneous fat on MUS, cross-sec-tional muscle area measurements were made in 100infants (98%) and fat area measurements were obtainedin 92 infants (90%). At discharge, only the group of 23- to28-week premature infants reached PMA of term; there-fore, we only compared this age group to the results of theinfants born at term. Body weight, length, head circumfer-ence, muscle and fat cross-sectional area, bone SOS,whole-body and regional lean body mass, FM, and BMDwere significantly lower in the preterm infants reachingterm. No significant difference was found in percent bodyfat and in the ratio of LM to FM. Interestingly, fat percent-age in the most preterm infants (23–28 weeks) reachingterm was significantly greater than in both other pretermgroups (29–32 and 33–36 weeks).

Correlations of bone measurements

Whole-body BMD was highly correlated with whole-bodylean tissue (see Fig. 1). Bone SOS had weak but significantcorrelation with BMD (r5 0.48 for whole-body and r5 0.45for regional DXA) (see Fig. 2). After adjusting for PMA,these correlations slightly improved (r 5 0.56 and r 5 0.49for whole-body and regional DXA, respectively).

Correlations among measurements of lean/muscle,fat, and body weight

There was a very strong correlation (r 5 0.99) betweenDEXA total mass and measured weight. There were sub-stantial correlations between muscle cross-sectional areadetermined by LUS with DXA-derived whole-body (r 50.82) and regional LM (r 5 0.71) (see Fig. 3). After adjust-ing to PMA, the correlations between LUS muscle cross-section and DXA-derived LM decreased (r 5 0.63 forwhole-body and r5 0.51 for regional LM).

TABLE 2. Participant characteristics by group

23–28 weeks (n5 20) 29–32 weeks (n 5 26) 33–36 weeks (n5 17) 37–42 weeks (n5 39)

Gestational age (weeks) 25.9 (25.1, 26.6) 30.8 (30.4, 31.3) 34.2 (33.6, 34.7) 39.4 (39.0, 39.8)Chronological age (days) 89.3 (77.7, 100.8) 31.2 (27.4, 34.9) 9.6 (6.3, 12.9) 3.8 (2.1, 5.4)Postmenstrual age (weeks) 38.6 (37.2, 40.0) 35.3 (34.9, 35.7) 35.6 (35.0, 36.2) 40.1 (39.7, 40.5)Birth weight (g) 864.8 (738.2, 991.3) 1,609.2 (1,498.1, 1,720.3) 2,225.0 (2,015.0, 2,435.0) 3,518.9 (3,360.5, 3,677.2)Birth length (cm) 34.0 (32.1, 36.0) 41.6 (40.5, 42.7) 45.2 (44.2, 46.2) 50.7 (50.0, 51.4)Birth head circumference (cm) 24.1 (22.5, 25.7) 29.0 (28.3, 29.7) 31.4 (30.6, 32.2) 34.8 (34.3, 35.3)Physical measurements

Weight (g) 2,679.4* (2,428.8, 2,929.9) 2,215.5 (2,096.5, 2,334.5) 2,256.4 (2,060.1, 2,452.6) 3,453.3 (3,296.5, 3,610.0)Length (cm) 46.0* (44.6, 47.4) 45.1 (44.3, 45.9) 45.7 (44.6, 46.8) 50.5 (49.7, 51.2)HC (cm) 33.4* (32.6, 34.1) 32.1 (31.5, 32.6) 31.9 (31.2, 32.7) 35.1 (34.6, 35.7)

Ultrasound measurementsMuscle area (cm2) 5.4* (4.9, 5.9) 4.8 (4.6, 5.1) 2.7 (2.2, 3.2) 6.6 (6.2, 7.0)Fat area (cm2) 3.7* (3.2, 4.3) 2.8 (2.5, 3) 3,068.0 (2,997.3, 3,138.6) 4.3 (4.1, 4.6)Bone SOS (m/s) 2,797.4* (2,720.4, 2,874.4) 3,003.9 (2,949.8, 3,058) 2,470.3 (2,267.2, 2,673.4) 3,168.4 (3,129.0, 3,207.9)

DEXA measurements—Whole bodyLean mass (g) 2,556.9* (2,374.2, 2,739.6) 2,251.0 (2,152.8, 2,349.1) 2,250.4 (2,098.7, 2,402.1) 3,162.8 (3,033.7, 3,291.8)Fat mass (g) 345.6* (240.3, 450.9) 142.0 (114.9, 169.1) 177.3 (112.1, 242.5) 417.6 (357, 478.2)BMD (g/cm3) 0.145* (0.135, 0.154) 0.139 (0.133, 0.145) 0.161 (0.152, 0.170) 0.204 (0.199, 0.210)Apparent BMD 0.314* (0.298, 0.330) 0.308 (0.297, 0.320) 0.352 (0.339, 0.367) 0.405 (0.396, 0.414)BMC (g/cm) 43.0* (37.2, 48.8) 35.3 (32.7, 38.0) 42.6 (37.6, 47.6) 72.5 (68.3, 76.7)% fat 11.1 (8.8, 13.5) 5.7 (4.8, 6.6) 6.8 (4.8, 8.7) 11.2 (9.9, 12.5)

DEXA measurements—RegionalLean mass (g) 52.2* (45.7, 58.7) 47.1 (43.7, 50.5) 52.2 (45.3, 59) 74.4 (70.3, 78.4)Fat mass (g) 13.9* (9.6, 18.2) 6.9 (5.4, 8.4) 8.1 (5.2, 10.9) 19.0 (16.0, 22.0)BMD (g/cm3) 0.106* (0.098, 0.115) 0.113 (0.103, 0.123) 0.128 (0.119, 0.138) 0.172 (0.164, 0.180)Apparent BMD 0.231* (0.216, 0.246) 0.249 (0.230, 0.269) 0.280 (0.262, 0.299) 0.340 (0.325, 0.355)BMC (g/cm) 1.2* (1.0, 1.4) 1.2 (1.0, 1.3) 1.5 (1.3, 1.7) 2.7 (2.5, 2.8)% fat 19.0 (15.7, 22.4) 12.1 (10.0, 14.3) 12.1 (9.1, 15.1) 19.2 (16.8, 21.7)

*Significant difference between term born and preterm infants born at 28–32 weeks reaching term.

Fig. 1. Bone mineral density as a function of DXA-derived meas-urements of whole-body lean tissue by group. BMD was significantlycorrelated to lean tissue (r 5 0.80).

71BODY COMPOSITION IN TERM AND PRETERM BABIES


There were also strong correlations between fat cross-sectional area by LUS and FM measured by both whole-body DXA (r 5 0.82) and regional DXA (r 5 0.81) (see Fig.4). After adjusting to PMA, the correlations slightlydecreased [whole-body DXA (r 5 0.67) and regional DXA(r5 0.67)].Mean slopes and 95% CI for measurements of body com-

position adjusted for PMA in the linear regression modelare presented in Figure 5. All slopes were significantly dif-ferent from zero.

DISCUSSION

The aim of the present study was to examine the feasi-bility of noninvasive bedside US measurement of bodycomposition in infants, compared to DXA measurements.We found significant correlations between LUS measure-ments of muscle and fat, with regional and whole-bodyDXA. These correlations remained significant whenadjusted to PMA. Weaker correlations were foundbetween bone strength, measured by bone QUS, andBMD, BMC, and apparent BMD, measured by DEXA. Ourdata support the idea that relatively simple bedside meas-urements could be used to assess body composition in thisfragile population.

Consistent with our second hypothesis, we found thatpreterm infants reaching term had significantly lowerbody weight, length, head circumference, muscle and fatcross-sectional area, bone SOS, whole-body and regionallean body mass, FM, and BMD compared to term-borninfants. These results demonstrate that the optimal carein the NICU (advancement of treatment and nutritionalsupport) is still unable to equal intrauterine conditions.DXA is now widely accepted as a reliable and accurate

measure of body composition in infants (Venkataramanand Ahluwalia, 1992). In the past, the accuracy of DXA insmall animals and human infants was questioned. Inpoorly mineralized bone, BMC may be underestimated(Roubenoff et al., 1993), and in small subjects, FM may beoverestimated (Rigo et al., 1998). However, with correc-tional equations (Rigo et al., 1998) and with the develop-ment of the faster fan beam scans, such as those used inthe present study, DXA was found to be valid, reliable,and accurate in small animals. Reference data for pretermand term infants have been reported previously (Kooet al., 2004; Rigo et al., 1998). In the present study, totalmass measured by DXA is highly correlated with meas-ured weight (r 5 0.99). Our DXA measurements of LM

Fig. 2. Comparison of bone quantitative ultrasound with bone min-eral density measured by DXA for the whole body (upper panel) andregional (lower panel), by group. There were modest but significantcorrelations between the two modalities for both (whole body, r 50.48; regional, r 5 0.45). Fig. 3. US leg muscle CSA as a function of DXA-measured regional

(leg) lean mass (upper panel) and whole-body lean mass (lower panel),by group. In both cases, the US measurement in a single leg washighly correlated with the DXA results (for regional lean mass, r 50.72; for whole-body lean mass, r 5 0.81).

72 I. AHMAD ET AL.


and FM are remarkably similar to chemical analysis ofstillborns (Widdowson and Spray, 1951), to body composi-tion measurements using whole-body water and whole-body potassium (Butte et al., 2000), and are also consist-ent with reference values for term newborns (Fomonet al., 1982). Our DXA body composition and BMD andBMC measurements in preterm infants (28–32 weeks) arecomparable to similar gestational age infants studied atdischarge (Cooke et al., 1999). Previous studies reportedhigher FM percentage for term infants and preterminfants approaching term, ranging from 15 to 21% (Venka-taraman and Ahluwalia, 1992). Those relatively high fatpercentages were reported using older versions of DXAmachines and software, and are now believed to be anoverestimation corrected by using appropriate equationsin the newer DXA machines. Percentage of body fat interm infants was also similar to previously reported chem-ical analysis values (Ziegler et al., 1976).

A positive correlation was found between LM andwhole-body BMD. There is mounting evidence to supportthe concept of the functional ‘‘bone-muscle unit’’ in whichmuscle activity can stimulate bone growth throughmechanoreceptors (Schoenau and Frost, 2002) andthrough the activity of hormones like IGF-I, which influ-ence both muscle and bone (Zofkova, 2008). Indeed, absent

or poor fetal movement is associated with impaired bonemineralization (Chen et al., 1995; Rodriguez et al., 1988).More recently, in fetal Myod-Myf5-deficient mice—con-genitally lacking striated muscle and having no function-ing ‘‘bone-muscle unit’’ or in-utero mechanical loadingthrough muscle activity—Gomez et al. (2007) found thatbone was profoundly abnormal and poorly mineralized.DXA measures BMD from the absorption of X-rays; the

latter is correlated with the degree of bone mineralization.QUS is used to derive an index of ‘‘bone strength’’ basedon the SOS wave propagation through the heterogeneousbone tissue. The technique has been shown to correlatewith BMD measured by DXA in adults (Kang and Speller,1998; Tromp et al., 1999), but there are conflicting reportsin the pediatric population (Gianni et al., 2008; Mughalet al., 1996). Although QUS has been used to study bonestrength in premature babies (Nemet et al., 2001), thereare few, if any, studies, in which the two modalities wereused simultaneously in premature babies. We found mod-est but significant correlations between bone QUS andboth whole-body and regional DXA-derived BMD (see Fig.2). Clearly, the two modalities are measuring somewhatdifferent but related properties of bone in the newborn.While DXA measures quantitative aspects of BMD, boneQUS is also influenced by qualitative factors that contrib-ute to bone strength such as bone elasticity, microarchitec-ture, and fatigue damage (Greenfield et al., 1981). Thus,although the two modalities are not interchangeable, acomplementary use of both methods may better reflect‘‘true’’ bone strength. Both bone strength and BMD byDXA were found to be significantly lower in the preterminfants, even in those reaching term. It is not known yetwhether this osteopenia of prematurity is a self-resolvingcondition or a disease with long-term consequences. Chanet al. (2008) recently demonstrated reduced bone mineral-ization in 7-year-old children who were born prematurely.Even if bone mineralization improves spontaneously, theprolonged period of demineralization in early childhoodmay lead to overt rickets, bone fractures, and potentiallyreduce peak bone mass.

Fig. 4. US-determined leg fat CSA as a function of DXA-measuredregional (leg) fat mass (upper panel) and whole-body fat mass (lowerpanel), by group. In both cases, the US measurement in a single legwas highly correlated with the DXA results (for leg fat mass, r 5 0.81;for whole-body fat mass, r 5 0.82).

Fig. 5. Mean slopes and 95% CI adjusted for PMA in the linearregression model of ultrasound and DXA measurements of bone SOS,regional BMD, whole-body BMD, muscle and fat CSA, regional andwhole-body lean mass (LM) by DXA, regional and whole-body fatmass (FM) by DXA. All slopes except for bone SOS were significantlydifferent from zero.



Most remarkable was the observation of the substantialcorrelation between the measure of muscle and fat CSA ina single limb by US with DXA-derived measurements oflean and fat tissues both at the regional and whole-bodylevel (Figs. 3 and 4). While reference values of muscle sizeusing US are available for the adult population, normativedata for the premature infant are scarce and mostlyfocuses on muscle thickness rather than cross-sectionalarea or volume (Heckmatt et al., 1988; Scholten et al.,2003). In a subgroup of infants (data not shown), weobtained US measurements of the calf and measured mus-cle depth from the epimysium to the tibia similar to thetechnique described by Scholten. We found no significantcorrelation between muscle thickness and DXA LM. Incontrast, our data indicate that cross-sectional areas ofthe muscle and subcutaneous fat in a single limb obtainedby US were highly correlated with DXA measurements oflean and fat tissue both regionally and for the whole body.On a technical note, we found that it is important for thesemeasurements to be obtained at precise locations withconsistent technique to obtain valid results.Using all three techniques, we studied the body compo-

sition of preterm infants approaching time of discharge.For this purpose, we used the PMA to compare appropri-ate for gestational infants at different degrees of maturity.Only the most premature group (born at 23–28 weeks)had PMA equivalent to term at the time of the measure-ments. We found that this group of preterm infants reach-ing term had significantly lower body weight, length, headcircumference, muscle and fat cross-sectional area, boneSOS, whole-body and regional lean body mass, FM, andBMD compared to term-born infants. However, similar toprevious observation (Cooke et al., 1999), these infantsmaintained LM to FM proportions similar to term infants.Interestingly, in the two other preterm groups (born at28–32 and 33–36 weeks), who were discharged earlierfrom the NICU and did not reach term, fat percentagewas lower and thus the ratio of LM to FM was higher.This is consistent with the fact that during fetal life pre-mature infants appear to follow a similar pattern with anearlier accretion of LM followed by accumulation of FM.Therefore, preterm infants who did not reach term gainedmainly LM, while those reaching term before dischargegained both LM and FM.The relationship between the magnitude and rate of

change of the FM early in life with conditions such as obe-sity and its accompanying insulin resistance andincreased cardiovascular disease risk in adolescence andadulthood has spurred renewed interest into the mecha-nisms that control fat accretion in the fetus and newborn,in particular, on the hormones and mediators that consti-tute the ‘‘fat-brain axis’’ even in fetuses (McMillen et al.,2006). A disproportionately higher rate in the recovery ofbody fat compared to lean tissue in preterm infants duringearly postnatal life (i.e. preferential ‘‘catch-up fat’’), asrecently suggested by Dulloo (2008) may be a centralevent in growth trajectories to obesity and to diseases thatcluster into the insulin resistance (metabolic) syndromeseen in this population (Casey, 2008).

CONCLUSION

In summary, the results of this cross-sectional multimo-dal analysis suggest the potential usefulness of muscle USas an investigative tool for studying aspects of body com-

position in this fragile population. There was also a weakbut significant correlation between QUS measurements ofbone strength and DXA-derived BMD, but, when appliedto investigating the rate of bone mineralization, the tech-niques proved not to be interchangeable. Current post-natal care and nutritional support in preterm infants isstill unable to match the in-utero environment for optimalgrowth and bone development. Understanding how differ-ent components of body composition change early in lifemay shed light on the mechanisms that link these earlychanges in fat, muscle, and bone tissue with obesity andosteopenia that occur in later periods of growth anddevelopment.

LITERATURE CITED

Butte NF, Hopkinson JM, Wong WW, Smith EO, Ellis KJ. 2000. Body com-position during the first 2 years of life: an updated reference. PediatrRes 47:578–585.

Casey PH. 2008. Growth of low birth weight preterm children. Semin Peri-natol 32:20–27.

Chan GM, Armstrong C, Moyer-Mileur L, Hoff C. 2008. Growth and bonemineralization in children born prematurely. J Perinatol 28:619–623.

Chen H, Blackburn WR, Wertelecki W. 1995. Fetal akinesia and multipleperinatal fractures. Am J Med Genet 55:472–477.

Cooke RJ, Rawlings DJ, McCormick K, Griffin IJ, Faulkner K, Wells JC,Smith JS, Robinson SJ. 1999. Body composition of preterm infants dur-ing infancy. Arch Dis Child Fetal Neonatal Ed 80:F188–F191.

de Paiva SAR, Zornoff LAM, Okoshi MP, Okoshi K, Matsubara LS, Matsu-bara BB, Cicogna AC, Campana AO. 2003. Ventricular remodelinginduced by retinoic acid supplementation in adult rats. Am J PhysiolHeart Circ Physiol 284:H2242–H2246.

Dietz MW, Dekinga A, Piersma T, Verhulst S. 1999. Estimating organ sizein small migrating shorebirds with ultrasonography: an intercalibrationexercise. Physiol Biochem Zool 72:28–37.

Dulloo AG. 2008. Thrifty energy metabolism in catch-up growth trajecto-ries to insulin and leptin resistance. Best Pract Res Clin EndocrinolMetab 22:155–171.

Farooqi A, Hagglof B, Sedin G, Gothefors L, Serenius F. 2006. Growth in10- to 12-year-old children born at 23 to 25 weeks’ gestation in the1990s: a Swedish national prospective follow-up study. Pediatrics118:e1452–e1465.

Fomon SJ, Haschke F, Ziegler EE, Nelson SE. 1982. Body composition ofreference children from birth to age 10 years. Am J Clin Nutr 35(5Suppl):1169–1175.

Gianni ML, Mora S, Roggero P, Mosca F. 2008. Quantitative ultrasoundand dual-energy X-ray absorptiometry in bone status assessment of ex-preterm infants. Arch Dis Child Fetal Neonatal Ed 93:F146–F147.

Gomez C, David V, Peet NM, Vico L, Chenu C, Malaval L, Skerry TM.2007. Absence of mechanical loading in utero influences bone mass andarchitecture but not innervation in Myod-Myf5-deficient mice. J Anat210:259–271.

Greenfield MA, Craven JD, Huddleston A, Kehrer ML, Wishko D, Stern R.1981. Measurement of the velocity of ultrasound in human cortical bonein vivo. Estimation of its potential value in the diagnosis of osteoporosisand metabolic bone disease. Radiology 138:701–710.

Heckmatt JZ, Pier N, Dubowitz V. 1988. Measurement of quadriceps mus-cle thickness and subcutaneous tissue thickness in normal children byreal-time ultrasound imaging. J Clin Ultrasound 16:171–176.

Hofman PL, Regan F, Cutfield WS. 2006. Prematurity–another example ofperinatal metabolic programming? Horm Res 66:33–39.

Joglekar CV, Fall CH, Deshpande VU, Joshi N, Bhalerao A, Solat V, Deo-kar TM, Chougule SD, Leary SD, Osmond C, Yajnik CS. 2007. Newbornsize, infant and childhood growth, and body composition and cardiovas-cular disease risk factors at the age of 6 years: the Pune Maternal Nutri-tion Study. Int J Obes (Lond) 31:1534–1544.

Kang C, Speller R. 1998. Comparison of ultrasound and dual energyX-ray absorptiometry measurements in the calcaneus. Br J Radiol71:861–867.

Koo WW, Walters JC, Hockman EM. 2004. Body composition in neo-nates: relationship between measured and derived anthropometrywith dual-energy X-ray absorptiometry measurements. Pediatr Res56:694–700.

Littner Y, Mandel D, Mimouni FB, Dollberg S. 2003. Bone ultrasound ve-locity curves of newly born term and preterm infants. J Pediatr Endocri-nol Metab 16:43–47.

74 I. AHMAD ET AL.


McMillen IC, Edwards LJ, Duffield J, Muhlhausler BS. 2006. Regulationof leptin synthesis and secretion before birth: implications for the earlyprogramming of adult obesity. Reproduction 131:415–427.

Mughal MZ, Langton CM, Utretch G, Morrison J, Specker BL. 1996. Com-parison between broad-band ultrasound attenuation of the calcaneumand total body bone mineral density in children. Acta Paediatr 85:663–665.

Nemet D, Dolfin T, Wolach B, Eliakim A. 2001. Quantitative ultrasoundmeasurements of bone speed of sound in premature infants. Eur JPediatr 160:736–740.

Pereda L, Ashmeade T, Zaritt J, Carver JD. 2003. The use of quantitativeultrasound in assessing bone status in newborn preterm infants.J Perinatol 23:655–659.

Pereira-da-Silva L, Veiga GJ, Clington A, Videira-Amaral JM, BustamanteSA. 1999. Upper arm measurements of healthy neonates comparing ultra-sonography and anthropometric methods. Early HumDev 54:117–128.

Quinn LS. 2007. Interleukin-15: a muscle-derived cytokine regulating fat-to-lean body composition. J Anim Sci 86(14 Suppl):E75–E83.

Rigo J, Nyamugabo K, Picaud JC, Gerard P, Pieltain C, De CM. 1998. Ref-erence values of body composition obtained by dual energy X-ray absorp-tiometry in preterm and term neonates. J Pediatr Gastroenterol Nutr27:184–190.

Rodriguez JI, Garcia-Alix A, Palacios J, Paniagua R. 1988. Changes in thelong bones due to fetal immobility caused by neuromuscular disease. Aradiographic and histological study. J Bone Joint Surg Am 70:1052–1060.

Roubenoff R, Kehayias JJ, Dawson-Hughes B, Heymsfield SB. 1993. Useof dual-energy X-ray absorptiometry in body-composition studies: notyet a ‘‘gold standard.’’ Am J Clin Nutr 58:589–591.

Schoenau E, Frost HM. 2002. The ‘‘muscle-bone unit’’ in children andadolescents. Calcif Tissue Int 70:405–407.

Scholten RR, Pillen S, Verrips A, Zwarts MJ. 2003. Quantitative ultraso-nography of skeletal muscles in children: normal values. Muscle Nerve27:693–698.

Schuelke M, Wagner KR, Stolz LE, Hubner C, Riebel T, Komen W, Braun T,Tobin JF, Lee SJ. 2004. Myostatin mutation associated with gross musclehypertrophy in a child. N Engl J Med 350:2682–2688.

Sharp M. 2007. Bone disease of prematurity. Early Hum Dev 83:653–658.Shibata M, Matsumoto K, Aikawa K, Muramoto T, Fujimura S, Kadowaki

M. 2006. Gene expression of myostatin during development and regener-ation of skeletal muscle in Japanese Black Cattle. J Anim Sci 84:2983–2989.

Thomopoulos S, Kim HM, Rothermich SY, Biederstadt C, Das R, Galatz LM.2007. Decreased muscle loading delays maturation of the tendon enthesisduring postnatal development. J Orthop Res 25:1154–1163.

Tromp AM, Smit JH, Deeg DJ, Lips P. 1999. Quantitative ultrasoundmeasurements of the tibia and calcaneus in comparison with DXA meas-urements at various skeletal sites. Osteoporos Int 9:230–235.

Venkataraman PS, Ahluwalia BW. 1992. Total bone mineral content andbody composition by X-ray densitometry in newborns. Pediatrics90:767–770.

Weiss LW, Clark FC. 1985. Ultrasonic protocols for separately measuringsubcutaneous fat and skeletal muscle thickness in the calf area. PhysTher 65:477–481.

Wells JC, Chomtho S, Fewtrell MS. 2007. Programming of body composi-tion by early growth and nutrition. Proc Nutr Soc 66:423–434.

Widdowson EM, Spray CM. 1951. Chemical development in utero. ArchDis Child 26:205–214.

Ziegler EE, O’Donnell AM, Nelson SE, Fomon SJ. 1976. Body compositionof the reference fetus. Growth 40:329–341.

Zofkova I. 2008. Hormonal aspects of the muscle-bone unit. Physiol Res 57Suppl 1:S153–S157.



body composition and its components in preterm and term newborns: a cross-sectional, multimodal...

Documents