Familial relationships of normal spine shape

Abstract

Familial correlations in the spinal shape of children are explored using statistical shape analysis. Measurements of the spine on the surface of the back were taken using an optical device, and recorded for several hundred children from the Leeds area of the U.K. A portion of the spine from the distal thoracic to proximal lumbar spine was used for the analysis, and measures of the shape of the spine line were determined using Procrustes analysis and principal components (PCs) analysis. Analysis was carried out on scans repeated in triplicate and scans taken six months apart, and good repeatability was demonstrated. Four groups of children were considered in the main study: monozygotic twins, dizygotic twins, same sex siblings and different sex siblings. Intra-familial correlations were calculated for the shape measures, and also for age and sex matched unrelated controls. The second PC score (which measures possible lordosis) had significant positive correlation in the family groups. It was observed that the familial correlations were higher for the genetically closer groups. Also, the same sex siblings had higher correlation than the different sex siblings. All sibling groups showed greater correlation of sagittal profile for the second PC score than unrelated controls. A significant correlation was observed in same sex pairings for the control data. This work suggests that some elements of spinal profile may be familial but also shows correlation with sex. Both of these observations may be important in the aetiology of idiopathic scoliosis.