Objective: The aim was to investigate a link between insulin sensitivity in childhood and 3D facial morphology in adolescence.
Methods: Facial laser scans of 400 children, 15½ years old, were chosen randomly from the data base of Cardiff Dental School(UK)and processed. Good quality scans of 273 children (124 males and 149 females) were included. The information on fasting insulin, glucose, height, weight and total fatness was obtained from the medical records at the age of 9. Body mass index (BMI), fat mass index (FMI) and quantitative insulin sensitivity check index (QUICKI) were calculated. Twenty one soft tissue landmarks were identified and the position of each landmark (X, Y and Z coordinates) collected. Generalised Procrustes Analysis was performed to register faces. Average faces for groups of subjects with high/low BMI, FMI and QUICKI indices were created and superimposed. The differences between groups of average faces were visualised by absolute and signed colour maps.
Results: Principal component analysis identified two major groups of landmarks responsible for facial shape variation. The distance between average points of two groups (centroids distance) was determined for each face. Significant correlation was found between centroids distance and height (0.52, and 0.35, p<0.01), in males and females with high QUICKI, respectively. Facial morphology of children with high and low QUICKI differs mostly in the size of the nose.
Conclusion: Three-dimensional facial morphology of adolescents is partly influenced by insulin sensitivity in childhood. Influence of different medical conditions on facial shape can be explored using described methodology.