Correlation between cortical thickness and bowing of the femoral diaphysis in healthy elderly people

Abstract

Purpose: The cortical thickness of the femoral diaphysis is presumed to change with aging due to various mechanical factors, such as bony structure (e.g., femoral bowing, neck anteversion, and condyle shape), lower limb alignment, body weight, load bearing due to movement, and mechanical stimulation by muscles. Above all, the femoral bowing is assumed to be important as one of the mechanical factors correlating with the cortical thickness of the femoral diaphysis. The purpose of this study is to examine the correlation between cortical thickness and bony bowing of the femoral diaphysis, measured three-dimensionally (3D), with high accuracy from CT data of healthy elderly people. Methods: This study was performed based on a protocol approved by the institutional review board of our university. The lower extremities of 54 elderly Japanese volunteers (29 men and 25 women) were included in this study. The average ± SD (range) of age and body mass index (BMI) were 70 ± 5.9 years (60 to 83 years) and 21.6 ± 2.4 kg/m2 (17 to 27 kg/m2), respectively. The participants were not obese and had no history of trauma or other diseases that influence cortical thickness, such as osteometabolic diseases except for primary osteoporosis. They were classified into Kellgren-Lawrence grades 0-1, with no radiographic knee OA. The cortical thickness of the femoral diaphysis was calculated automatically in a 3D space by using Stradwin software (version 5.3; Medical Imaging Group, Machine Intelligence Laboratory, Cambridge Engineering Department, Cambridge, UK), which is available for free download and is a new tool with a demonstrated sub-voxel accuracy in assessing cortical bone properties using routine low-resolution CT [Figure 1]. As the accuracy, it was reported that cortical bone thickness estimates were accurate up to 0.3 mm. The number of measurement points per subject was 5000-8000, depending on femur length. An anatomical coordinate system for the femur was constructed using original software. Twelve assessment regions were created by combining three heights (proximal: 54%-71%; central: 37%-54%; distal: 20%-37%) and four areas of the axial plane (xy-plane) at 90° (medial, anterior, lateral, and posterior) [Figure 2]. The assessment parameter was the average of the cortical thickness data from 200-800 points in each region in the femoral diaphysis, divided by each height and area. When the data were analyzed, standardized values divided by the femoral length (cortical thickness / femoral length), not actual values, were applied. The other evaluation parameter was femoral bowing in the coronal plane and sagittal plane in a 3D space. To accurately define femoral bowing in 3D space, an axial center (AC) was automatically calculated for the 10 respective cross-sectional planes that divide the femoral diaphysis into 11 equal sections in the femoral bony coordinate system. Line A is defined as the line connecting the top of the AC with its bottom. Line B is a perpendicular line from the AC of the maximum bowing point to line A. Femoral bowing in the coronal plane (lateral bowing) is defined by the following equation: Length of line B/length of line A × 100 (%), projected on the XZ plane. Femoral bowing in the sagittal plane (anterior bowing) is defined by the following equation: Length of line B/length of line A × 100 (%), projected on the YZ plane [Figure 3]. To examine the correlation between cortical thickness and femoral bowing, Pearson's correlation coefficient was calculated. The probability level accepted for statistical significance was set at p < 0.05 (version 21; SPSS, Inc., Chicago, IL, USA). Results: The average values of anterior and lateral bowing were 3.19 ± 0.49% and 0.75 ± 0.45%, respectively. The anterior bowing showed a low positive correlation with the posterior area of the proximal diaphysis [Correlation coefficient (CC) = 0.269, p = 0.049], the anterior area of the central diaphysis (CC = 0.372, p = 0.006), the posterior area of the central diaphysis (CC = 0.343, p = 0.011), and the posterior area of the distal diaphysis (CC = 0.351, p = 0.009). The lateral bowing was negatively correlated with the lateral area of the central diaphysis (CC = -0.317, p = 0.019) and the lateral area of the distal diaphysis (CC = -0.342, p = 0.011). Conclusions: In healthy elderly people, the femoral bowing was predominantly confirmed in the sagittal plane. While the anterior and lateral bowing showed a correlation with femoral cortical thickness, the trends were reverse (anterior bowing - positive correlation; lateral bowing - negative correlation). In Curry ’s rule, the periosteum side and medullary cavity side are the compression side and tension side, respectively. The balance of remodeling between “bone formation by compression” and “bone resorption by extension” was different depending on the region, and may be related to cortical thickness and bowing of the femoral diaphysis.View Large Image Figure ViewerDownload Hi-res image Download (PPT)View Large Image Figure ViewerDownload Hi-res image Download (PPT)