Questions regarding conclusions reached in “Age dependence of femoral strength in white women and men”

The paper by Keaveny and colleagues entitled, ‘‘AgeDependence of Femoral Strength in White Women and Men,’’ provides interesting data and analysis related to the use of biomechanical computed tomography (BCT) for estimating femoral strength (FS). In the BCT technique, quantitative CT scans are used at a given anatomic site (in this article at the proximal femur) in conjunction with sophisticated computational techniques (finite-element analyses) to estimate the biomechanical properties associated with the femur. Based on the data and analysis described in the article, Keaveny and colleagues present two main conclusions. The first is that agerelated declines in femoral strength are much greater than suggested by age-related declines in femoral neck areal bone mineral density (aBMD). The second is that far more of the elderly may be at high risk of hip fracture because of low femoral strength than previously assumed based on the traditional classification of osteoporosis (ie, T-score< 2.5). Implicit in these conclusions is that BCT is superior to aBMD in terms of identifying individuals at increased risk of fracture. In the following it will be shown that the first conclusion does not adequately explain the data and that the second is a useful observation but can be incorporated into standard aBMD analyses. Regarding the first conclusion, the data demonstrate that agerelated declines in femoral strength (as estimated by BCT) are much larger than age-related declines in femoral neck aBMD. For example, in women, the decline in femoral strength over five decades of life was 55%, whereas the decline in aBMD was only 26%. The implication is that femoral strength, as estimated by BCT, would be a much better quantity to estimate fracture risk than aBMD. However, the percent decline of a parameter is not necessarily relevant in terms of its potential utility as a proxy for fracture risk. To see this, it is useful to plot femoral strength versus aBMD (Fig. 1). As may be seen, there is a strong linear relationship between femoral strength and aBMD, with R values equal to 0.98 for women and 0.96 for men. In addition, the equation relating FS to aBMD is FS1⁄4a aBMD – b, with a1⁄4 10,530 and b1⁄4 5680 for women and with a1⁄4 8956 and b1⁄4 4326 for men. While it is true that percentage changes in aBMD over a period of time are less than the percentage changes in FS over the same time period, the relevant comparison is to examine the relative change in femoral strength that results from a relative change in aBMD. This can be computed according to the following formula: