Presented in part at the 28th Annual Meeting of the American Society for Bone and Mineral Research, Philadelphia, PA, September 15–19, 2006.
Version of Record online: 12 MAR 2007
Copyright © 2007 ASBMR
Journal of Bone and Mineral Research
Volume 22, Issue 6, pages 825–831, June 2007
How to Cite
Bouxsein, M. L., Szulc, P., Munoz, F., Thrall, E., Sornay-Rendu, E. and Delmas, P. D. (2007), Contribution of Trochanteric Soft Tissues to Fall Force Estimates, the Factor of Risk, and Prediction of Hip Fracture Risk*. J Bone Miner Res, 22: 825–831. doi: 10.1359/jbmr.070309
Published online on March 12, 2007;
The authors state that they have no conflicts of interest.
- Issue online: 4 DEC 2009
- Version of Record online: 12 MAR 2007
- Manuscript Accepted: 5 MAR 2007
- Manuscript Revised: 1 MAR 2007
- Manuscript Received: 25 SEP 2006
- hip fracture;
- trochanteric soft tissue;
- factor of risk;
- femoral strength;
We compared trochanteric soft tissue thickness, femoral aBMD, and the ratio of fall force to femoral strength (i.e., factor of risk) in 21 postmenopausal women with incident hip fracture and 42 age-matched controls. Reduced trochanteric soft tissue thickness, low femoral aBMD, and increased ratio of fall force to femoral strength (i.e., factor of risk) were associated with increased risk of hip fracture.
Introduction: The contribution of trochanteric soft tissue thickness to hip fracture risk is incompletely understood. A biomechanical approach to assessing hip fracture risk that compares forces applied to the hip during a sideways fall to femoral strength may by improved by incorporating the force-attenuating effects of trochanteric soft tissues.
Materials and Methods: We determined the relationship between femoral areal BMD (aBMD) and femoral failure load in 49 human cadaveric specimens, 53–99 yr of age. We compared femoral aBMD, trochanteric soft tissue thickness, and the ratio of fall forces to bone strength (i.e., the factor of risk for hip fracture, Φ), before and after accounting for the force-attenuating properties of trochanteric soft tissue in 21 postmenopausal women with incident hip fracture and 42 age-matched controls.
Results: Femoral aBMD correlated strongly with femoral failure load (r2 = 0.73–0.83). Age, height, and weight did not differ; however, women with hip fracture had lower total femur aBMD (OR = 2.06; 95% CI, 1.19–3.56) and trochanteric soft tissue thickness (OR = 1.82; 95% CI, 1.01, 3.31). Incorporation of trochanteric soft tissue thickness measurements reduced the estimates of fall forces by ∼50%. After accounting for force-attenuating properties of trochanteric soft tissue, the ratio of fall forces to femoral strength was 50% higher in cases than controls (0.92 ± 0.44 versus 0.65 ± 0.50, respectively; p = 0.04).
Conclusions: It is possible to compute a biomechanically based estimate of hip fracture risk by combining estimates of femoral strength based on an empirical relationship between femoral aBMD and bone strength in cadaveric femora, along with estimates of loads applied to the hip during a sideways fall that account for thickness of trochanteric soft tissues. Our findings suggest that trochanteric soft tissue thickness may influence hip fracture risk by attenuating forces applied to the femur during a sideways fall and provide rationale for developing improved measurements of trochanteric soft tissue and for studying a larger cohort to determine whether trochanteric soft tissue thickness contributes to hip fracture risk independently of aBMD.
Hip fractures are disabling events for the elderly, leading to significant social and financial burdens for societies worldwide. Among the strongest risk factors for hip fracture include increased age, low BMD, previous history of fracture, and low body weight or body mass index (BMI).(1–3) The vast majority of hip fractures occur secondary to a fall, and thus characteristics of the fall itself, as well as risk factors related to falls, are associated hip fracture risk.(4–7) Forces applied to the hip during a fall depend on the fall characteristics, impact surface, self-protective responses, and thickness of soft tissues overlying the hip. Therefore, it has been suggested that the relationship between low BMI and hip fracture may by attributable either to lower bone strength because of reduced mechanical loading and/or to decreased soft tissue thickness overlying the hip.(8) Greenspan et al.(5) reported that, among elderly fallers, independent predictors of hip fracture include low femoral areal BMD (aBMD), falling to the side, increased fall energy, and low BMI. Their observation that low BMI is associated with hip fracture independent of femoral aBMD supports the view that a low BMI may be a risk factor for fracture because of decreased trochanteric soft tissue thickness, and hence, reduced energy absorption during a fall impact. A recent meta-analysis confirmed that low BMI is a risk factor for hip fracture after adjustment for femoral aBMD.(2)
Further support for a key role of trochanteric tissues in hip fracture is provided by studies of external trochanteric padding systems. Laboratory studies indicate that trochanteric padding markedly reduces the loads applied to the hip during a sideways fall.(9–11) Initial randomized controlled studies of hip protectors showed significant prevention of hip fracture.(12,13) Subsequent studies with different randomization schemes have not confirmed the efficacy of hip protectors.(14–16) Thus, although the efficacy of trochanteric pads remains controversial, the equivocal results may be in part caused by poor compliance.(17–20) Indeed, the majority of fractures that occurred in the hip protector arm of randomized studies occurred when the hip protector was not being worn or was worn but incorrectly positioned.(12,17)
To better understand the impact of trochanteric soft tissue on hip fracture risk, it may be useful to use a biomechanical approach to assess hip fracture risk that compares forces applied to the hip during a sideways fall to the strength of the proximal femur. The ratio of applied force to bone strength is termed the factor-of-risk.(21) Recent studies have shown that age- and sex-specific differences in the factor-of-risk reflect the observed incidence of wrist, hip, and vertebral fractures more closely than do differences in BMD.(22,23) However, studies that have computed the factor-of-risk for hip fracture have not yet accounted for force attenuation caused by soft tissue overlying the hip.(22,24)
Thus, to further explore the impact of trochanteric soft tissue on hip fracture, the objectives of this study were to (1) compute the factor-of-risk for hip fracture, defined as the ratio of the force applied to the hip during a sideways fall to the strength of the proximal femur, both with and without accounting for the force-attenuating effects of trochanteric thickness; (2) determine whether trochanteric soft tissue thickness, aBMD, and the factor-of-risk for hip fracture differ between women with incident hip fracture and age-matched controls; and (3) examine the correlations between body habitus, aBMD, trochanteric soft tissue thickness, and forces applied to the hip during a sideways fall.
MATERIALS AND METHODS
Relationship between femoral aBMD and femoral strength in vitro
To estimate the strength of the proximal femur in a sideways fall configuration in vivo, we first determined the relationship between femoral aBMD and failure load of 49 human cadaveric femora tested to failure in a sideways fall configuration. Cadaveric specimens included 29 women and 20 men, with a mean age of 80.5 ± 10.5 yr (range, 53–99 yr). Before enrollment into the study, radiographs of each specimen were acquired and used to screen for abnormalities, such as metastatic or benign bone tumors, Paget's disease, or prior fracture. Specimens were harvested fresh and maintained frozen at −20°C until testing. Total femur, femoral neck, and trochanteric aBMD were assessed using DXA (QDR2000+; Hologic, Bedford, MA, USA), with femora submerged in a water bath to simulate soft tissue. Femoral strength was assessed using high rate loading designed to simulate impact to the greater trochanter experienced during a sideway fall, as previously reported.(25) Failure load was taken as the maximal force from the load-displacement curve.
Hip fracture subjects and controls
Subjects were selected from the OFELY study, a prospective cohort of women from France recruited between February 1992 and December 1993. The general description of this cohort is provided in detail elsewhere.(26–28) The sample for this study included 21 of 23 postmenopausal women who suffered a hip fracture in the 13-year follow-up period (11 femoral neck and 10 intertrochanteric fractures), and 42 age-matched controls who did not suffer a fracture. Two hip fracture cases were not included because their whole body scans could not be read for technical reasons. To select the control subjects, all women in the OFELY cohort with no prior history of hip fracture and who were within ±1 yr of age of each hip fracture case were identified. From these, two controls were randomly chosen for each hip fracture case. The mean age and years postmenopause of subjects in this study at baseline was 74.5 ± 7.9 and 23 ± 8 yr, respectively. Previous history of fragility fracture and history of falls in previous 12 mo (yes/no) were collected at the baseline exam and annually thereafter. Hip fractures occurred on average 60.3 ± 34.1 mo (range, 1–109 mo) after the baseline visit. Only nontraumatic hip fractures were included. All incident fractures were confirmed by radiologic review and/or surgical report. Weight and height were measured at baseline and annually thereafter and used to calculate BMI (kg/m2).
Assessment of BMD, body composition, and trochanteric thickness
DXA (QDR2000; Hologic) was used to assess aBMD (g/cm2) of the proximal femur, as previously described.(26) Whole body scans were used to assess lean body mass and fat mass as percentages of total body weight. The average thickness of soft tissue overlying the greater trochanter was assessed manually from the whole body scans by measuring the lateral distance between the greater trochanter and the air-soft tissue boundary in the whole body analysis algorithm (“Enhanced array whole body”, V5.54A; Hologic). The soft tissue thickness was assessed by measuring the distance between the most lateral aspect of the greater trochanter and the lateral aspect of the skin-air boundary. Visualization of the air-soft tissue boundary was optimized by manually adjusting the luminosity and contrast. Both right and left sides were measured from the whole body scan, and the average soft tissue thickness used for subsequent analyses. Repeated analysis of 20 whole body scans on three separate occasions indicated that reproducibility of trochanteric soft tissue measurements, assessed by the mean CV for repeat measurements, was 3.9% (range. 1.7–7.1%). Height and weight were measured and used to compute BMI.
Estimation of femoral strength, loads applied to the hip during a sideways fall, and the factor-of-risk for hip fracture
Femoral strength in a sideways fall configuration was estimated from the linear regression between trochanteric aBMD and femoral failure load, as determined from the mechanical testing in cadaveric specimens.
The peak force applied to the hip during a sideways fall was estimated from previously published studies describing the kinematics of sideways falls(29,30) and femoral impact forces in falls on the hip(31):
where g is the gravitational constant (9.81 m/s2), h is the height of an individual's center of gravity for a fall from standing height, taken as 0.51 × height (m), m is the effective mass (kg), and k is the stiffness constant (N/m).(31) Because the peak force applied to the hip is attenuated by trochanteric soft tissues, we also computed an “attenuated force” using the observation from Robinovitch et al.(8) that the peak force applied to the hip was reduced 71 N/mm of trochanteric soft tissue.
We computed the factor-of-risk for hip fracture (Φ) as the ratio of the applied force to estimated femoral strength in a sideways fall.(21) The factor of risk was calculated using both the peak force (Φpeak) and the attenuated force (Φatten) estimates.
Linear regression analysis was used to determine the association between femoral aBMD and femoral failure load for the in vitro testing. For the clinical study, we compared mean values of body composition, trochanteric thickness, femoral BMD, and the factor-of-risk in the cases and controls using unpaired Students t-test. Predictors of hip fracture were evaluated by performing univariate logistic regression to compute ORs (and 95% CIs), initially without adjustment for total femoral aBMD and then with adjustment for total femoral aBMD. In addition, multiple logistic regression was used to test whether femoral strength and the force applied to the hip during a sideways fall were independently associated with hip fracture. Relationships among body composition measures, aBMD, and trochanteric thickness were determined using Pearson correlations.
Relationship between aBMD and femoral strength in vitro
Total hip aBMD in the cadaveric femora averaged 0.64 ± 0.27 g/cm2 (range, 0.11–1.12 g/cm2). Using the NHANES database, T-scores for the total femur ranged from −6.8 to 1.3 and from −4.3 to 1.0 in women and men, respectively. The average femoral failure load was 3353 ± 1809 (SD) N and ranged from 709 to 8841 N. Regression analysis indicated that femoral strength was strongly associated with aBMD of the femoral neck (r2 = 0.73, p < 0.01), total femur (r2 = 0.79, p < 0.01), and trochanteric regions (r2 = 0.82, p < 0.01; Fig. 1). Because the trochanteric aBMD was slightly (although not significantly) better correlated with femoral failure load than the other femoral regions, we used trochanteric aBMD to derived estimates of femoral strength in the hip fracture subjects and controls:
Predictors of hip fracture risk
Subjects who suffered a hip fracture were similar in age, height, and weight to controls at baseline; however, they had significantly lower femoral aBMD and therefore lower estimated femoral failure load (Table 1). Women with fractures had decreased trochanteric soft tissue thickness compared with controls (p = 0.04) and tended to have lower BMI (p = 0.07). The estimated peak force applied to the hip during a sideways fall did not differ between hip fracture cases and controls; however, the estimated force applied to the hip after adjusting for the attenuation caused by soft tissues tended to be higher in hip fracture cases versus controls (Table 1; p = 0.07).
Logistic regression confirmed that women with decreased femoral aBMD had an increased risk of subsequent hip fracture, with an OR of 2.1 (p < 0.001) for each SD decrease in aBMD (Table 2). Each SD decrease in trochanteric soft tissue thickness was associated with a 1.8-fold increased hip fracture risk (p < 0.05; Table 2). Increased force applied to the hip after adjustment by trochanteric soft tissue thickness tended to be associated with increased risk of hip fracture; however, this did not reach statistical significance. Similarly, there was a trend for a higher factor-of-risk for hip fracture to be associated with increased risk of hip fracture (Table 2). After adjustment for femoral aBMD, none of the factors remained significantly associated with hip fracture.
We performed a multiple logistic regression to assess whether femoral strength and the attenuated force applied to the hip were independently related to hip fracture. In this regression model, each SD increase in femoral strength was associated with a 60% reduction in hip fracture (OR = 0.41; 95% CI: 0.21, 0.81; p = 0.01), and each SD increase in force applied to the hip was associated with a 1.5-fold increased risk for fracture, although this did not reach statistical significance (OR = 1.56; 95% CI: 0.82, 2.95; p = 0.18).
Trochanteric soft tissue thickness was moderately positively correlated with body weight (r = 0.64, p < 0.001), percent body fat (r = 0.65, p < 0.001), and BMI (r = 0.75, p < 0.001), but was weakly inversely correlated with height (r = −0.26, p = 0.04). Trochanteric soft tissue thickness was unrelated to the ratio of waist-to-hip circumference. Total femur aBMD was moderately positively correlated with body weight (r = 0.47, p < 0.001), BMI (r = 0.39, p = 0.02), and trochanteric soft tissue thickness (r = 0.31, p = 0.02).
In this study, we explored the impact of trochanteric soft tissue thickness on hip fracture risk. We did this by computing the factor-of-risk for hip fracture before and after adjusting for force-attenuating capacity of trochanteric soft tissue and by comparing femoral BMD, trochanteric soft tissue thickness, and factor-of-risk in postmenopausal women with and without incident hip fracture. We found that each SD decrease in trochanteric soft tissue thickness was associated with a 1.8-fold increased risk of hip fracture. Our finding that reduced trochanteric thickness increases risk for hip fracture is consistent with previous assertions that trochanteric soft tissues influence hip fracture risk by reducing the forces applied to the femur during a sideways fall.(32,33) Moreover, our observation is consistent with the view that a primary reason that reduced BMI is associated with increased hip fracture risk is through an association with reduced trochanteric soft tissue thickness, because the correlation between BMI and trochanteric thickness was stronger (r = 0.76) than the correlation between BMI and total femoral aBMD (r = 0.34). Our findings, along with the results of Greenspan et al.,(5) who showed that aBMD and BMI were independent predictors of hip fracture among elderly fallers, support further studies of tissue thickness as a predominant mechanism for the influence of BMI on hip fracture risk.
After adjustment for total femoral BMD, the decrease in trochanteric soft tissue thickness was no longer a significant predictor of hip fracture, despite a relatively weak association between total femur BMD and trochanteric soft tissue thickness (r = 0.31). This finding supports a predominant role for femoral BMD in determining hip fracture risk. However, the OR for trochanteric thickness after adjustment for femoral BMD (OR = 1.44), although not significant, does infer that there is some residual effect of trochanteric thickness on hip fracture risk after adjustment for BMD. Clearly, one must recognize that our relatively small sample size gave limited power to detect multiple independent predictors of hip fracture, and a larger sample size is necessary to confirm these inferences. Moreover, BMD measurements were likely to be more accurate than our trochanteric soft tissue measurements, and therefore, additional studies are needed to better define the relative predictive value of aBMD, BMI, and trochanteric soft tissue thickness to hip fracture.
We also showed that estimates of the forces applied to the hip depend greatly on inclusion of variations in soft tissue thickness. After accounting for the energy attenuating properties of trochanteric soft tissue, forces applied to the hip were reduced on average by 50% (2869 N) and 61% (3537 N) in hip fracture cases and controls, respectively. Whereas the peak fall force did not differ, the estimated fall force after adjustment for trochanteric soft tissue thickness tended to be higher in hip fracture cases than controls (Table 1). Also, when using the peak force estimates, the mean factor-of-risk for hip fracture exceeded the biomechanical “fracture threshold” of 1.0 in both cases and controls. However, after correcting for the force attenuating properties of trochanteric soft tissue, the factor-of-risk was substantially <1.0 in the controls and was ∼1.0 in the fracture cases (Table 1). These data suggest that accounting for trochanteric soft tissue is likely to improve estimates of the forces applied to the hip in a sideways fall and thereby improve biomechanical estimates of fracture risk. One may have expected the factor of risk to be even higher in the fracture group (i.e., much greater than 1.0). We recognize that in this initial attempt to use a biomechanical assessment of hip fracture risk, several factors may eventually be revised to provide more precise estimates of both loading and femoral strength. For example, the whole body DXA-based measurement may overestimate trochanteric soft tissue thickness because of lateral spreading of soft tissues in the supine positioning. This would lead to an underestimation of loads applied to the hip, and therefore the factor-of-risk would be underestimated. Additional research and advanced imaging techniques may improve the accuracy of both the estimates of the applied force as well as femoral strength.
One characteristic not accounted for in our current biomechanical approach to fracture assessment is the probability of falling, which if included, may further improve biomechanically based estimates of fracture risk. Thus, assessment of fracture risk in an individual would ideally include the probability of suffering a sideways fall, along with the probability of fracture were a sideways fall to occur by comparing the forces applied to the hip to estimated femoral strength. As such, there is evidence that incorporating age, sex, fall history, and simple gait and balance assessments into the model might allow reasonable assessment of the probability of falling,(34,35) although there is little information on key risk factors specifically for sideways falls.
Although use of whole body DXA scans is not optimal for assessment of trochanteric soft tissue thickness because of limited resolution, we found similar values of soft tissue thickness as Maitland et al.,(36) who used ultrasound to directly measure trochanteric soft tissue thickness. The measurement of trochanteric soft tissue with the subject in a supine position likely overestimates the soft tissue thickness and therefore the force-attenuating capacity of the tissue. However, Maitland et al.(36) showed a strong associated between direct ultrasound measurements of trochanteric thickness in subjects while standing and DXA-derived measurements from subjects in supine position. Moreover, use of the fan-beam mode may lead to accuracy errors in linear measurements.(37,38) In this study, the measurement of trochanteric soft tissue from whole body scans was also limited by the scan resolution; in our case, a 2.04-mm point size and 1.3-mm line spacing. More sophisticated techniques or higher-resolution DXA scans could readily be applied to improve the measurement of trochanteric soft tissues. If future studies with a larger number of fractures show that decreased trochanteric soft tissue thickness is associated with hip fracture independently of BMD, it may be possible to derive two independent risk factors for hip fracture from a single DXA exam. Alternatively, 3D methods of bone assessment, such as QCT, may also be used to assess trochanteric thickness along with volumetric bone density and geometry.
Our study had several shortcomings. Among them was the relatively small number of hip fractures (n = 21), which limited our ability to detect independent risk factors for hip fractures. Thus, in multiple logistic regression analyses, femoral BMD was the only significant risk factor. In addition, our estimates of femoral strength in vivo were determined from an empirical relationship between femoral BMD and failure load. Although we used elderly cadaveric specimens and tested them in a sideways fall configuration at high loading rate, it is not known whether this relationship, derived from 49 cadaveric specimens, is valid across all clinical populations. Moreover, several reports indicate that changes in femoral BMD do not fully account for bone strength changes seen after therapeutic intervention,(39,40) and therefore, it is likely that the femoral strength estimates will be improved by incorporating 3D imaging techniques and/or finite element analyses.(41–44) Also, our estimate of the force attenuation provided by soft tissues was derived from a single study in a small number of cadavers,(8) which did not differentiate between trochanteric thickness provided by fat versus muscle.
These limitations notwithstanding, in this study, we showed that it is possible to compute a factor of risk for hip fracture by combining estimates of femoral strength, based on an empirical relationship between femoral aBMD and bone strength in cadaveric femora, along with estimates of loads applied to the hip during a sideways fall that account for thickness of trochanteric soft tissues. This approach, when combined with estimates of fall risk, may enhance predictions of hip fracture risk and may help to explain the fracture benefits of therapeutic interventions. Our findings suggest that variation in trochanteric soft tissue thickness may contribute to hip fracture risk by attenuating forces applied to the hip during a sideways fall. These results provide strong rationale for conducting a larger clinical study, developing improved measurements of trochanteric soft tissue, and determining whether bone strength, trochanteric soft tissue thickness and fall impact forces contribute independently contribution to hip fracture risk.
We kindly thank the participants in the OFELY trial. This study was funded in part by a grant from the National Institute of Health (CA41295), INSERM, and an unrestricted research grant from Eli Lilly.
- 172000 Hip protectors for preventing hip fractures in the elderly. Cochrane Database Syst Rev 4: CD001255., ,
- 241994 Factor of risk is associated with frequency of hip fracture in a case-control study. Transactions of the Orthopaedic Research Society., , ,
- 321988 Biomechanical aspects of fractures. RiggsBL, MeltonLJIII (eds.) Osteoporosis: Etiology, Diagnosis and Management. Raven Press, New York, NY, USA, 111–132., ,