To determine whether there are differences in femoral skeletal geometry in fracture-prone subjects when size, positioning diagnosis, and age are controlled, we compared femoral measurements made from the uninvolved hip on 119 plane anteroposterior pelvis radiographs of women without fracture to those of the contralateral hip in a group of 43 female patients with hip fractures (neck, 23; intertrochanteric, 20). The hip was imaged in a standardized position of rotation and adduction. Race, age, and musculoskeletal diagnosis were known. Subjects were grouped by diagnosis of the opposite hip condition (rheumatoid arthritis, osteoarthritis, and normal) and compared. Measurements were also analyzed as ratios to head diameter (HD), neck diameter (ND), and pelvic width. Femoral neck length (NL) was measured from skeletal preparations and imaged in controlled positions of abduction and external rotation. No differences were found between the neck and intertrochanteric fracture groups. The differences between the fracture group and the controls were a thinner femoral cortex (measured at a point one head radius below the lesser trochanter) a larger femoral head, and a larger femoral ND in the fracture group (p < 0.025). The difference in cortical thickness was still significant when scaled by size, but the ratio of HD to ND was equivalent in fractures and controls. No difference in femoral NL could be demonstrated. The experimental measurements showed that apparent NL is significantly position sensitive and this may explain previously reported differences in fracture

No abstract