Females are not proportionally smaller males: Relationships between radius anthropometrics

Distal radius fracture reduction by internal fixation is most commonly achieved using implants/prosthetics such as volar locking plates (VLP). However, recent studies suggest that upwards of 27% of patients experience complications associated with the implanted hardware, leading to device removal. Radius implant design has traditionally been based on anthropometric population averages and the high complication rates may therefore be a result of designing implants without accounting for patient‐specific differences. Therefore, the purpose of this research was to: i) determine the relationships between clinically relevant radius anthropometrics; and ii) determine if proportional anthropometric differences exist between males and females. It was hypothesized that: i) a principal component analysis (PCA) would reveal overall relationships between radius anthropometrics; and ii) radius anthropometrics would change disproportionately between the sexes, with respect to changes in radius length. Three‐dimensional radius bone geometries were created from the CT scans of 17 male and 16 females age‐matched (12 left and 20 right, mean age = 74.9) forearm specimens in Mimics Medical Imaging software ( Materialise ). Twenty anthropometric measurements were taken from the reconstructions, including lengths (5), breadths (3), diameters (10), and angles (2). A PCA was performed to determine which variables were most significantly (0.722) correlated with each of the first three principal components (as determined from scree plots). Based on the outcome of the PCA, it was determined that principal component 1 (PC1) was a reflection of overall size and radius length was found to account for the largest proportion of variability in PC1. Therefore, a linear regression analysis was performed to normalize the other 19 measurements to radius length; a second PCA was performed on the linear regression residual values. The PCA correlation matrices were assessed, and separate scatter plots of PC1 vs. PC2 and PC1 vs. PC3 were developed for both the un‐normalized and length‐normalized eigenvalues. As described above, the initial PCA correlation matrix revealed that PC1 was representative of overall radius size, as all measurements related to size (i.e., length, breadth, diameter) were positively correlated with PC1. Males and females were initially separated by PC1 and all female specimens had negative PC1 eigenvalues in contrast to the positive PC1 eigenvalues for the male specimens ( Figure 1). Following normalization to radius length, males and females remained segregated by PC1 ( Figure 2). The results presented here, suggest that the anthropometrics of a female radius vary by different proportions compared to a male radius, as a function of radius length. Therefore, accounting for these individualized sex differences in future radius implant designs may reduce implant revision and improve overall patient outcomes. Support or Funding Information Partially funded by Fowler‐Kennedy Sport Medicine Clinic. 1A comparison between the PC scores for Principal Component 1 and 2. Scores represent variation in the raw measurements between each subject.2A comparison between the PC scores for Principal Component 1 and 2. Scores represent variation in residual values following regression to radius length between each subject.