Open Access Research article

Assessing newborn body composition using principal components analysis: differences in the determinants of fat and skeletal size

Beverley M Shields1*, Bridget A Knight12, Roy J Powell3, Andrew T Hattersley1 and David E Wright4

Author Affiliations

1 Peninsula Medical School, Barrack Road, Exeter, UK

2 Heavitree Hospital, Royal Devon and Exeter NHS Foundation Trust, Exeter, UK

3 Research and Development Support Unit, Royal Devon and Exeter NHS Foundation Trust Exeter, UK

4 School of Mathematics and Statistics, University of Plymouth, Plymouth, UK

For all author emails, please log on.

BMC Pediatrics 2006, 6:24  doi:10.1186/1471-2431-6-24

Published: 17 August 2006



Birth weight is a composite of skeletal size and soft tissue. These components are likely to have different growth patterns. The aim of this paper is to investigate the association between established determinants of birth weight and these separate components.


Weight, length, crown-rump, knee-heel, head circumference, arm circumference, and skinfold thicknesses were measured at birth in 699 healthy, term, UK babies recruited as part of the Exeter Family Study of Childhood Health. Corresponding measurements were taken on both parents. Principal components analysis with varimax rotation was used to reduce these measurements to two independent components each for mother, father and baby: one highly correlated with measures of fat, the other with skeletal size.


Gestational age was significantly related to skeletal size, in both boys and girls (r = 0.41 and 0.52), but not fat. Skeletal size at birth was also associated with parental skeletal size (maternal: r = 0.24 (boys), r = 0.39 (girls) ; paternal: r = 0.16 (boys), r = 0.25 (girls)), and maternal smoking (0.4 SD reduction in boys, 0.6 SD reduction in girls). Fat was associated with parity (first borns smaller by 0.45 SD in boys; 0.31 SD in girls), maternal glucose (r = 0.18 (boys); r = 0.27 (girls)) and maternal fat (r = 0.16 (boys); r = 0.36 (girls)).


Principal components analysis with varimax rotation provides a useful method for reducing birth weight to two more meaningful components: skeletal size and fat. These components have different associations with known determinants of birth weight, suggesting fat and skeletal size may have different regulatory mechanisms, which would be important to consider when studying the associations of birth weight with later adult disease.