Osteoporos Int (2006) 17: 593–599 DOI 10.1007/s00198-005-0019-4
ORIGINA L ARTI CLE
Femur strength index predicts hip fracture independent of bone density and hip axis length K. G. Faulkner . W. K. Wacker . H. S. Barden . C. Simonelli . P. K. Burke . S. Ragi . L. Del Rio
Received: 27 July 2005 / Accepted: 30 September 2005 / Published online: 31 December 2005 # International Osteoporosis Foundation and National Osteoporosis Foundation 2005
Abstract Introduction: Proximal femoral bone strength is not only a function of femoral bone mineral density (BMD), but also a function of the spatial distribution of bone mass intrinsic in structural geometric properties such as diameter, area, length, and angle of the femoral neck. Recent advancements in bone density measurement include software that can automatically calculate a variety of femoral structural variables that may be related to hip fracture risk. The purpose of this study was to compare femoral bone density, structure, and strength assessments obtained from dual-energy X-ray absorbtiometry (DXA) measurements in a group of women with and without hip fracture. Methods: DXA measurements of the proximal femur were obtained from 2,506 women 50 years of age or older, 365 with prior hip fracture and 2,141 controls. In addition to the conventional densitometry measurements, structural variables were determined using the Hip Strength Analysis program, including hip axis length (HAL), cross-sectional moment of inertia (CSMI), and the femur strength index (FSI) calculated as the ratio of estimated compressive yield strength of the femoral neck to K. G. Faulkner (*) . W. K. Wacker . H. S. Barden GE Healthcare, 726 Heartland Trail, Madison, WI 53717, USA e-mail:
[email protected] Tel.: +1-608-8267681 C. Simonelli HealthEast Clinics, Woodbury, MN, USA P. K. Burke Osteoporosis Diagnostic and Treatment Center, Richmond, VA, USA S. Ragi Centro de Diagnõstico e Pesquisa da Osteoporose do Espirito Santo, Vitoria, Brazil L. Del Rio CETIR Centre Medic, Barcelona, Spain
the expected compressive stress of a fall on the greater trochanter. Results: Femoral neck BMD was significantly lower and HAL significantly higher in the fracture group compared with controls. Mean CSMI was not significantly different between fracture patients and controls after adjustment for BMD and HAL. FSI, after adjustment for T score and HAL, was significantly lower in the fracture group, consistent with a reduced capacity to withstand a fall without fracturing a hip. Conclusion: We conclude that BMD, HAL, and FSI are significant independent predictors of hip fracture. Keywords Bone density . Femur strength index . Hip axis length . Hip fracture
Introduction Osteoporosis is one of the world’s most significant health concerns, affecting 200 million individuals worldwide [1]. Hip fracture is recognized as the most serious osteoporotic fracture because of its association with high medical costs [2–5] and its profound influence on patient morbidity, functional capacity, and mortality [6–10]. A recent attempt to quantify the global burden of hip fracture in the year 1990 estimated more than 1.3 million incident hip fractures and 4.5 million prevalent hip fractures with disability were associated with 740,000 deaths and 1.75 million disabilityadjusted life years lost [8]. These numbers will increase significantly over the next 50 years due to population growth and extended life expectancy [11]. Mortality from hip fracture as high as 20% with 50% permanent loss in function has been reported [12]. Measurement of bone mineral density (BMD) with dualenergy X-ray absorptiometry (DXA) is currently the most frequently used technique for diagnosing osteoporosis and estimating fracture risk [13, 14]. Fractures occur when the mechanical stresses on bone exceed local material strength. Bone mass correlates highly with bone strength, accounting for 80–90% of the variance in ultimate compressive strength of compact and trabecular bone when tested in
594
vitro [15–22]. Reliance on BMD alone, however, may not provide the best predictive ability to detect individuals with the greatest risk for hip fracture. Proximal femoral bone strength is not only a function of femoral BMD but also a function of the spatial distribution of bone mass intrinsic in structural geometric properties, such as diameter, area, length, and angle of the femoral neck [23–25]. A wellrecognized example of a structural change affecting bone strength is periosteal apposition of femoral bone that results in increased femoral neck diameter and area and increased modulus of elasticity and strength [26]. This structural change in femoral neck bone distribution, more apparent in men than women, partially offsets the negative effects on strength of femoral bone loss with aging [24, 27]. Several recent in vitro studies of femora tested to failure in a loading configuration designed to simulate a fall impacting the greater trochanter reported a high correlation between femoral neck strength and DXA-derived femoral neck BMD (r=0.88–0.96) [28, 29]. Recent commercial advancements in bone density measurement include software that can automatically calculate a variety of femoral structural variables that may be related to hip fracture risk. Pioneering work by Martin and Burr showed that the cross-sectional bone mineral absorption curves generated by single photon absorptiometry (SPA) could be utilized to determine structural variables useful for assessing femoral strength indices [30]. This approach has been modified for use with DXA, and methods have been Fig. 1 Femur variables used to measure fall index (FI) and hip axis length (HAL)
developed to assess femoral size, shape, and strength [23, 31, 32]. For example, hip strength analysis (HSA) programs are now commercially available that can automatically asses structural variables such as the femoral neck crosssectional moment of inertia (CMSI), cross-sectional area (CSA), femoral neck shaft angle, and hip axis length (HAL). In addition, models have been proposed that combine these structural parameters with age, height, and weight to calculate the femur strength index (FSI), a measure of the ability of a hip to withstand a fall on the greater trochanter [31]. The purpose of this study was to compare femoral bone density, structure, and strength assessments obtained from DXA measurements in a group of women with and without hip fracture.
Methods Study population DXA measurements of the proximal femur were obtained from 2,506 women 50 years of age or older, 365 with prior hip fracture and 2,141 controls, using the Lunar Prodigy DXA system (GE Healthcare). Femur scans were acquired at four densitometry centers located in the USA (Minnesota and Virginia), Spain (Barcelona), and Brazil (Vitoria). Qualified personnel reviewed each scan to assure accurate analysis. Control subjects consisted of women over 50 years of
mean SD skew
Fx 1.34 0.36 0.87
Cx 1.56 0.40 0.67
fracture
fracture control
control relative frequency
relative frequency
595
0.0
0.5
1.0
1.5
2.0
2.5
3.0
mean SD
0. 3
0 .5
0. 7
0 .9
1. 1
Fx 0.722 0.127
Cx 0.812 0.125
1 .3
1. 5
femural neck BMD (g/cm2)
Femur Strength Index
Fig. 2 Relative frequency distributions of femur strength index (FSI) values for fracture and control groups
Fig. 4 Relative frequency distributions of bone mineral density (BMD) values of fracture and control groups
age with no history of osteoporotic fracture. Nonfracture controls were otherwise unselected, except for the requirement that femur scans must have included the inner margin of the pelvic rim to allow measurement of HAL. Hip fractures were confirmed radiographically, and DXA measurements were performed on the nonfractured femur. DXA scans of hip fracture subjects were acquired after the fracture, but the time between scan and fracture was generally unknown or not reported. Scans were acquired and analyzed according to the manufacturer’s recommended protocol. All DXA systems were operating within the manufacturer’s performance specifications, so no attempt was made to cross-calibrate the densitometers used by the four centers.
turer’s HSA program that is commercially available. The HSA program automatically calculates a number of bone geometry variables from the scan image and bone distribution variables derived from information contained within DXA X-ray absorption curves [30, 31]. These variables included (Fig. 1):
Structural variables and BMD
mean SD skew
Fx Cx 0.387 0.474 0.243 0.241 -0.07 -0.12
These variables were used to calculate FSI, the ratio of estimated compressive yield strength of the femoral neck to
fracture control relative frequency
relative frequency
Conventional bone density values, including the BMD (g/cm2), BMC (g), and area (cm2) of the femoral neck, were determined using the manufacturer’s analysis software. Structural variables were determined using the manufac-
1. CSMI, mm4, of the section of minimum CSMI within the neck region of interest (ROI), a measure of the distribution of material around the neck axis necessary to calculate resistance to bending 2. CSA, mm2, of the minimum CSMI section within the neck ROI 3. Distance from the center of the femoral head to the minimum CSMI (d1, mm) 4. Distance from the center of mass (centroid) to the superior neck margin for the section of minimum CSMI (y, mm) and 5. Femoral neck angle (θ, degrees)
-0.5
0.0
0.5
1.0
mean SD
Fx 103.3 6.8
Cx 101.0 6.3
fracture control
1.5
adjusted ln(Femur Strength Index)
Fig. 3 Relative frequency distributions of log-transformed femur strenght index (FSI) values adjusted for bone mineral density (BMD) and hip axis length (HAL)
70
80
90
100
110
120
130
Hip Axis Length (mm)
Fig. 5 Relative frequency distributions of hip axis length (HAL) values for fracture and control groups
596 Table 1 Mean values of the fracture and control populations. Standard deviations are shown in parentheses Number Age (years) Height (cm) Weight (kg) Neck BMD g/cm2 HAL (mm) CSMI (cm4) CSA (mm2) FSI Fracture Controls
365 2141
71 (10) 66 (10)
156 (7) 156 (6)
63 (12) 64 (11)
0.72 (0.13)* 0.81 (0.12)
103 (7)* 101 (6)
0.82 (0.24) 0.83 (0.22)
109 (20) 120 (20)
1.34 (0.36)* 1.56 (0.40)
BMD bone mineral density, HAL hip axis length, CSMI cross-sectional moment of inertia , CSA cross-sectional area, FSI femur strength index *Significantly different from controls (p
