Purpose: Animal study to evaluate DXA and MRI/MRS as tools to assess body composition (BC) in newborn infants. MRI/MRS data are presented in a second abstract. Methods: 23 piglets (age: 1-37d, body weight(BW): 848-7550g). After weighing the animal (electronic scale), total body water was measured (deuterium dilution). The animal was sacrificed, BC was assessed using MR and DXA (QDR-1500 W, Hologic, Waltham, USA, infant software ver. 5.67, triplicate scans). The carcass was homogenized in 4 steps designed for complete recovery and homogenization: (1) grinding in a small butchery grinder, (2) rinsing with water, (3) lyophilization of the moeity, (4) grinding of dry matter using a smaller grinding device. CA of total ash (TA), calcium (Ca), phosphorus (P), total fat (TF), raw protein (RP), residual water(RW) was done in duplicate using standard methods. The quality of homogenization was assessed by measuring RP, TA, and RW in 10 aliquots from 1 piglet. Approvement by cantonal and federal commissions for animal studies.Results: Recovery was >99%. CV of RP, TA, and RW (1.7, 2.1 and 1.8% resp.) indicate perfect homogenization. BC by CA (%BW, mean±SD): body water 75.3±3.9, RP 13.9±8.8, TF 6.5±3.7 (range absolute terms 7.0-674.1 g). TA varied from 34.9-236.7 g, Ca accounted for 25.6±1.3 of TA, P for 17.0±0.7%. Ca/P ratio was 1.51±0.10. Hydration factor of fat-free mass was 80.4±1.1% decreasing with increasing BW (r2=0.419) independently from age. BW by DXA was highly correlated with BW by electronic scale (r2=0.999) but overestimated by 1.2%. DXA bone mineral content (BMC) was highly correlated with TA, Ca and P (r2=0.993, 0.992, and 0.992), Ca and P account for only 44.1±4.2 and 29.1±1.8% of BMC, TA exceeded BMC by a factor of 1.72±0.13. DXA fat mass (FM) was highly correlated with TF(r2=0.961) with a slope (S) of 0.763 and an intercept (I) of 52.4 g. Conversion equations are given:

BW = 0.9879 BWDXA+25.28; TF=1.311 FMDXA-68.75; TA=1.537 BMC+8.79; Ca=0.4015 BMC+1.65; P=0.2610 BMC + 1.54. Discussion: The high correlations between DXA and CA allow to develop conversion equations. With its use DXA gives accurate and precise estimates of neonatal BC and may become a valuable tool for the nonivasive assessment of neonatal growth.