The authors state that they have no conflicts of interest.
We experimentally studied the distribution of hip fracture types at different structural mechanical strength. Femoral neck fractures were dominant at the lowest structural strength levels, whereas trochanteric fractures were more common at high failure loads. The best predictor of fracture type across all failure loads and in both sexes was the neck-shaft angle.
Introduction: Bone geometry has been shown to be a potential risk factor for osteoporotic fractures. Risk factors have been shown to differ between cervical and trochanteric hip fractures. However, the determinants of cervical and trochanteric fractures at different levels of structural mechanical strength are currently unknown. In addition, it is not known if the distribution of fracture types differs between sexes. The aim of this experimental study on excised femora was to investigate whether there exist differences in the distribution of cervical and trochanteric fractures between different structural mechanical strength levels and different sexes and to identify the geometric determinants that predict a fracture type.
Materials and Methods: The sample was comprised of 140 cadavers (77 females: mean age, 81.7 years; 63 males: mean age, 79.1 years) from whom the left femora were excised for analysis. The bones were radiographed, and geometrical parameters were determined from the digitized X-rays. The femora were mechanically tested in a side impact configuration, simulating a sideways fall. After the mechanical test, the fracture patterns were classified into cervical and trochanteric.
Results: The overall proportion of cervical fractures was higher in females (74%) than in males (49%) (p = 0.002). The fracture type distribution differed significantly across load quartiles in females (p = 0.025), but not in males (p = 0.205). At the lowest load quartiles, 94.7% of fractures in female and 62.5% in males were femoral neck fractures. At the highest quartiles, in contrast, only 52.6% of fractures in females and 33.3% in males were cervical fractures. Among geometric variables, the neck-shaft angle was the best predictor of fracture type, with higher values in subjects with cervical fractures. This finding was made in females (p < 0.001) and males (p = 0.02) and was consistent across all failure load quartiles.
Conclusions: Femoral neck fractures predominate at the lowest structural mechanical strength levels, whereas trochanteric fractures are more common at high failure loads. Females are more susceptible to femoral neck fractures than males. The best predictor of fracture type across all structural strength levels and both sexes was the neck-shaft angle.
Osteoporotic fractures are an increasing health problem. A fracture is usually related to reduced bone strength (osteoporosis) and falling, alone or more often in combination. The clinical assessment of osteoporosis and fracture risk is currently based primarily on measurements of areal BMD using DXA. However, DXA-based BMD has been shown to lack high predictive ability of the individual risk of fracture.(1) In a prospective, population-based cohort study, most fractures occurred in individuals without osteoporosis as defined by DXA (T score below −2.5).(2) Robbins et al.(3) concluded that there exist distinct factors that are associated with hip fractures in women with high BMD. Several other studies have also shown that the possibility of osteoporotic fracture increases, when low BMD and other risk factors coexist in a patient.(4–10) Therefore, there exists a need for developing more accurate tools for the assessment of fracture risk, using not only BMD, but also other predictive parameters.(2)
Bone geometry has been previously studied as a potential risk factor and has been shown to improve the prediction of fracture risk.(4–6,9,10) Risk factors have been shown to differ between cervical and trochanteric fractures of the proximal femur.(5,6,10,11) It has also been shown that there is a significant difference in the hip fracture risk predictor variables between females and males.(12) Thus, the assessment of fracture risk may require separate approaches for the femoral neck and trochanteric fractures and for both sexes. However, it is not known what the specific determinants of cervical and trochanteric fractures are at different levels of structural mechanical strength (failure load), and whether these risk factors differ between male and female subjects. In addition, there is currently no knowledge on the distribution of fracture types at different structural mechanical strength levels and only little information on the difference in the distribution of fracture types between sexes.
For better understanding of the pathomechanism of fracture, it might be useful to elucidate whether the both types of fracture are equally probable at low and high energy impacts. Epidemiologic studies show that trochanteric fractures are an increasing problem compared with cervical fractures because their relative number increases progressively with age in women after the age of 60 years and because their incidence has been shown to increase in both sexes and all age groups during the recent decades.(13,14) However, epidemiological studies can not clarify the physical mechanisms that lead to the fracture. By standardizing the measurements in an experimental study design, the mechanisms behind the phenomena (fracture) might possibly be further elucidated.
The aim of this experimental study on excised femora was to investigate the association of geometric factors and structural mechanical strength with the distribution of cervical versus trochanteric hip fractures and to study whether these differ in males and females.
MATERIALS AND METHODS
The femora studied were obtained from a course of macroscopic anatomy from the Institute of Anatomy at the Ludwig Maximilians University of Munich (Germany). Sample preparation and storage methods have been previously described in detail.(15) The criterion for inclusion in the course was the testamentary decree to the institute several years before death and the sample should thus representative of the elderly population in southern Germany. The individuals belonged to a wide range of medical and social backgrounds, but no detailed medical or social history was available. To identify specimens with bone diseases other than osteoporosis or osteopenia, biopsy specimens were taken from the left iliac crest for histology. Inappropriate individuals (i.e., individuals with bone disease other than osteoporosis or osteopenia, such as osteosarcoma) were excluded from the study.
Eventually, the sample consisted of 171 cadavers (86 females: mean age, 81.3 years; 85 males: mean age, 78.2 years) whose left femora were used for the study (Fig. 1). Thirty-one samples had a shaft failure in mechanical testing, and they were excluded from the analysis. Thus, the sample size for final analyses was comprised of 140 cadavers (77 females: mean age, 81.7 years; 63 males: mean age, 79.1 years). In some study samples, DXA scans (N = 2) or X-rays (N = 1) were unavailable. In addition, part of the bone existed outside the image area in some radiographs, and the definition of some geometrical parameters in these samples was thus impossible (Fig. 1). The specific sample size is reported throughout the results section.
After the dissection course, the proximal femora were excised, cleaned of the surrounding soft tissue, and kept moist throughout the study. The bones were radiographed with a Faxitron X-ray system (Model 43885A; Faxitron, Hewlett Packard, McMinnville, OR, USA) at 40–85 kV, 2 mA, time = 120 s, using a 18 × 24-cm X-ray film (Agfa Structurix D7DW; Agfa, Leverkusen, Germany). X-rays were digitized together with a calibration scale using a scanner with a resolution of 900 dpi. A set of geometrical parameters was determined from the digitized X-rays using Image Tool software (version 3.00; University of Texas Health Science Center, San Antonio, TX, USA). The neck-shaft angle (NSA), femoral shaft cortex width (FSC), calcar femoral cortex width (CFC), femoral neck axis length (FNALa and FNALb), femoral head and neck diameters (HD and ND), trochanteric width (TW), and femoral shaft diameter (FSD) were measured (Fig. 2) as reported previously in clinical studies.(5,10) The root mean square CV (CVrms) was used as a measure of reproducibility, being 1.1%, 5.2%, 9.9%, 0.9%, 1.5%, 0.7%, 1.5%, 2.5%, and 2.5% for NSA, FSC, CFC, FNALa, FNALb, HD, ND, TW, and FSD, respectively.(5)
In vitro DXA scans of the femora were obtained using a standard narrow fan beam scanner with multiview image reconstruction (GE Lunar Prodigy; GE Lunar Corp., Madison, WI, USA) with the femoral specimens submerged in a water bath.(15) Standard positioning was used across all specimens. Total femoral BMD was evaluated with the software provided by the manufacturer.
Mechanical test of the proximal femur
The femora were tested in a side impact configuration, simulating a sideways fall on the greater trochanter, as previously presented.(15–18) Briefly, the femoral head and shaft faced downward and were able to move independently of one another on the support plates during loading of the trochanter. One half of a tennis ball with a lubricant was used to simulate cartilage contact with the femoral head. The load was applied to the greater trochanter through a pad, simulating a soft tissue cover. The shaft was positioned at 10° from horizontal, and the neck at 15° internal rotation. This standard position was used across all specimens. Loads were applied at a rate of 6.6 mm/s. The failure load was defined as the peak of the load-deformation curve. This test has been shown to display good reproducibility in previous tests on paired femora, the upper limit of the precision error being about 15%.(15)
After the mechanical test, the fracture patterns were classified from the broken bones to cervical and trochanteric by a single observer (VK), according to the standard AO classification. Subcapital and transcervical (basicervical and midcervical) fractures were classified to cervical fractures and pertrochanteric, intertrochanteric, and subtrochanteric fractures to trochanteric fractures. Femoral shaft failures were excluded from the analysis.
χ2 test was used to test the overall difference in the proportion of fracture types between sexes. The sample was divided into failure load quartiles of females and males for studying the fracture type distribution at different levels of structural mechanical strength. The difference in the distribution of fracture types across the failure load quartiles was studied by Kruskal-Wallis test.
The normality in the distribution of geometrical parameters was tested by the Kolmogorov-Smirnov test with the correction of Lilliefors. Statistical evaluation for the normally distributed variables was performed with the independent samples Student's t-test or the Mann-Whitney test for non-normally distributed variables. Levene's test was used to test the equality of variance. The sex and bone strength interaction with geometric variables was studied with ANCOVA, using sex and load quartiles as factors and age and total BMD as covariates. Similarly, the fracture type and bone strength interaction with geometric variables was clarified by using fracture type and load quartiles as factors and age and total BMD as covariates. Tukey's posthoc test was used for the comparison of differences between quartiles.
Spearman's correlation analysis was used to test the relationship between measured parameters and fracture type and failure load. All statistical analyses were performed using the SPSS for Windows software, version 12.0.1 (SPSS, Chicago, IL, USA).
The overall proportion of cervical fractures was higher in females (74%) than in males (49%) (p = 0.002). The fracture type distribution differed significantly across load quartiles (Table 1) in females (p = 0.025) but not in males (p = 0.205; Figs. 3A and 3B). However, a similar trend was found also in men, with neck fractures predominating at lower load levels and trochanteric fractures at higher load levels. At the lowest load quartiles, 94.7% of fractures in female and 62.5% in male were femoral neck fractures, whereas at the highest quartile, only 52.6% of fractures in females and 33.3% in males were femoral neck fractures. The lowest quartile for failure load in male specimens was equivalent to the third quartile for failure load in female specimens (Table 1).
Table Table 1.. Failure Load Quartiles for Both Females and Males and for the Entire Study Population
Differences in geometrical parameters between cervical and trochanteric fractures are shown in Table 2. In females, the NSA was significantly higher (p = 0.001) and the FNAL significantly longer (p = 0.004) in the cervical fracture cases, whereas in males, only the NSA differed significantly (p = 0.02) between the fracture types. These differences also remained after adjustment for age and total BMD.
Table Table 2.. Differences in Geometrical Parameters Between Fracture Types by Sex Before and After Adjustment for Age and Total BMD
With the exception of NSA, all geometrical parameters differed remarkably between sexes in both cervical and trochanteric fracture cases (Table 3). After adjustment for age and total BMD, the differences in CFC and FSC disappeared in the cervical fracture cases, and the difference in CFC disappeared in trochanteric fractures.
Table Table 3.. Values of Statistical Significance for the Geometrical Parameters Between Females and Males by Fracture Types Before and After Adjustment for Age and Total BMD
In comparison between load quartiles, CFC and FSC were significantly lower in the first load quartile compared with the other quartiles in the cervical fracture cases of females (p < 0.001–0.003). In the trochanteric fractures of males, CFC was significantly lower in the first and second quartiles compared with the highest quartile (p < 0.01). However, these differences disappeared after adjusting for age and total BMD. No other sex, fracture type, or failure load quartile interactions with geometrical parameters were observed.
Within sex-specific failure load quartiles, differences in geometrical parameters between cervical and trochanteric fractures were not statistically significant because of the limited sample size in subgroups of female and male quartiles. Because there was no sex-specific difference in NSA (Table 3), females and males were pooled together for a more detailed analysis of NSA within failure load quartiles. The NSA was higher in the cervical cases in all failure load quartiles (p < 0.01–0.05) except in the quartile of lowest failure load, where the difference did not reach statistical significance (p = 0.07) because of the small number of specimens (n = 7) in the trochanteric group. The NSA did not differ between load quartiles within each fracture type.
The fracture type was significantly related to NSA (r = 0.393, p < 0.001 in females and r = 0.293, p = 0.02 in males). In females, fracture type correlated also with FNALb (r = 0.308, p = 0.007). There was also a significant relationship between fracture type and failure load in both sexes (r = 0.348, p = 0.002 and r = 0.250, p = 0.047 in females and males, respectively). However, neither NSA nor FNAL was related to failure load directly.
Previous clinical studies have indicated that certain predictive variables differ between cervical and trochanteric hip fractures.(5,6,10,11) Here we have analyzed, for the first time, the association of different levels of structural mechanical strength and geometric factors with the occurrence of cervical versus trochanteric hip fractures in a controlled experimental setup. The analysis was performed in a relatively large study sample, permitting us to perform sex-specific analyses as well as to look at fracture behavior at different levels of mechanical failure load.
This study shows that, in an experimental setup, hip fractures in femora with low mechanical strength predominantly occur at the femoral neck, in particular in females, whereas trochanteric fractures seem to be more common in femora with higher failure loads, in particular in males. This suggests that the strength properties of the femoral neck dominate at loads lower than the breaking force of the trochanter (i.e., femora with low bone strength are particularly vulnerable at the level of the femoral neck), whereas those with higher bone strength preferably fail at the trochanter. Considering that bone strength (failure load level) is likely to decrease with age, this seems to be somewhat controversial to epidemiological findings. In the review of 15 published reports, Baudoin et al.(19) presented that, in women, the ratio of cervical to trochanteric fractures (C/T) evolves in three periods: (1) before the age of 50 years, the incidence of cervical fractures is close to that of trochanteric fractures; (2) between 50 and 60 years, cervical fractures increase markedly, and the C/T ratio is well above unity at an age when the fracture incidence is still very low; and (3) this imbalance progressively diminishes to reach unity in the very old, as the result of a progressive increase in trochanteric fractures. In men, cervical fractures are progressively more common with increasing age, and the C/T ratio exceeds unity after 70 years of age. This controversy might be explained by the standardized testing configuration, mimicking a sideways fall on the greater trochanter, using a fixed low loading rate. Numerous comprehensive studies on risk factors of fractures among elderly people(20–22) have shown that falling is the strongest single risk factor for a fracture, and the type and severity of falling are crucial in determining whether a fracture occurs.(21,23,24) Compared with BMD, the relative risk of hip fracture for a sideways fall is about 5, whereas if such a fall results in an impact to the greater trochanter of the proximal femur, the corresponding risk ratio is ∼30.(25,26) Furthermore, it was already shown in 1996 by Pinilla et al.(27) that the impact direction associated primarily with a fall is a critical determinant of hip fracture risk that is both independent of BMD and associated with fall biomechanics. Using an analysis of covariance to adjust for total hip BMD, they showed that failure load decreased by 24% as the loading angle changed from 0° to 30°, which reduction in failure load is comparable with that associated with about 25 years of age-related bone loss after the age of 65. Instead of fall biomechanics, this study gives information on the hip fracture mechanisms independent on the falling. The fracture mechanism has been shown to be different for cervical and trochanteric fractures,(5,6,10,11) which suggests fracture type to be taken into account in the evaluation of pathomechanism of fracture. It is indisputable that most fractures occur through a fall and that falling mechanism plays a crucial role in determining whether a fracture occurs. However, the type of falling as a strong determinant for the occurrence of a fracture may lack the structure-related information behind the fracture mechanisms. In the experimental study design, the type of falling can be eliminated as the well-known factor for the fracture susceptibility, and the mechanisms behind the phenomena (fracture) can be further elucidated. Our finding might thus be clinically important, because it suggests focusing on the identification of risk factors associated with low mechanical strength.
The study also shows that the distribution of fracture types differs by sex, with females being more susceptible to encounter a femoral neck fracture. This is in contrast to an experimental study of Cheng et al.,(28) who found no significant sex difference in the distribution of fracture types in a smaller sample (28 females and 36 males). The finding is also in controversy to existing epidemiological data presented above. Here again, this may be caused by the ignorance of the falling mechanism in the experimental study design. It is, however, well accepted that males generally have bigger and stronger bones and a lower hip fracture incidence than females. The primary finding of the study of Cody et al.(12) was a significant difference between male and female hip fracture risk predictor variables. They concluded that effective hip fracture prevention strategies may require separate approaches for men and women. Our finding further supports this and suggests focusing more on cervical fractures of females, which seems to be the most prominent fracture risk group in the sole perspective of mechanical strength.
We found the neck-shaft angle to be a significant discriminator between cervical and trochanteric fractures, with the NSA being substantially higher in cervical fractures. This was evident for all loading levels, for both sexes and also before and after adjustment for age and BMD. Alonso et al.(4) reported in a cross-sectional clinical study that an increase of 1 SD in NSA was associated with an OR of hip fracture of 2.45 in men and 3.48 in women, but they did not report a difference in NSA between the fracture types. Gnudi et al.(6) and us(10) have reported that NSA was a better predictor of femoral neck fractures than BMD, and the importance of NSA is also supported by this experimental study. These findings might be explained by the high bending moment at the neck during a fall on the greater trochanter in femora with high NSA. The importance of NSA might also explain our controversial findings with epidemiological data concerning the distribution of fracture types. Theoretically, the rotation angle of a leg at the moment of the floor contact may actually have a profound influence on the type of the hip fracture. This aspect is supported by the finding of Nevitt and Cummings,(25) where the nature of fall determines the type of fracture. In a falling situation, the rotation angle of a leg might compensate the effect of high NSA and a trochanteric fracture occurs.
Interestingly, the NSA was not found to differ between the low and high failure load fractures within each fracture type, suggesting that it is not a good predictor of failure load. In general, clinical results concerning the importance of NSA in predicting fracture risk have been conflicting: whereas some studies did not find a significant impact,(29,30) others found an important relationship between NSA and fracture risk.(4–6,9) Kukla et al.(31) have reported that NSA showed a slight positive correlation with load-to-failure in an experimental study with the loading direction parallel to the shaft (vertical loading of the femur). In contrast, no significant correlation was reported between mechanical strength and NSA in the experimental study of Cheng et al.(28) with the mechanical test simulating a fall on the greater trochanter. Based on this experimental study, it seems that NSA is the best discriminator for the fracture type, albeit NSA is not associated with the magnitude of failure load in a side impact load configuration. The findings suggest that NSA may be used as a clinical tool for identifying subjects specifically with increased risk of cervical hip fracture.
We also observed that the differences in FNAL between fracture types were statistically significant in females but not in males. The significant difference remained also after adjustment for age and total BMD, indicating that FNAL is independent of the changes in material properties affected by age, and it is also independent of the bone mass defined by areal BMD. Mechanically, the FNAL, or hip axis length (HAL, which also includes acetabular structures), might be one potential risk factor for femoral neck fracture, but the results are still partly inconsistent. Some studies have shown HAL not to be directly associated with an increased risk of hip fracture either in women or in men,(4,5,29,30) whereas in other studies, HAL was shown to be an important risk factor.(6,9,11,32) In the experimental study of Cheng et al.,(28) FNAL correlated significantly (r2 = 0.24) with femoral strength, and the authors found a significant difference between sexes, values being significantly longer in males. In this study, FNAL also differed between sexes, being longer in males. However, it was not remarkably different between load quartiles, indicating that FNAL does not predict fracture load.
In addition to FNAL, we found that most geometrical parameters differed between sexes, even after adjustment for age and total BMD. This is consistent with previous studies, where it has been shown that the material properties, structural and geometrical properties of femur,(12,18,33,34) and even the predictive variables of hip fracture risk(12) differed between females and males. In this study, however, NSA was not dependent on sex. It seems therefore that for the prediction of fracture type, a similar risk model based on NSA can be formulated for both sexes, whereas the models for failure load have to be sex-specific.
Cortical thickness differed here between sexes and between load levels, the cortex being thinner in female and at the lowest load levels. However, the differences disappeared after adjustment for age and BMD. This indicates that cortical thickness is not independent of BMD and supports the previous findings, where it has been shown that cortical thickness correlates with DXA-based BMD(10,35) (i.e., they convey the same message). BMD measurement by DXA, which is a projectional method, is biased by the thickness of the cortical shell (i.e., the thicker the cortex, the higher the values for DXA-based BMD).
In conclusion, femoral neck fractures predominated in femora with low mechanical strength, whereas trochanteric fractures were more common at the highest failure loads. The distribution of fracture types differed by sex, with females being more susceptible to femoral neck fractures. The best independent predictor of fracture type across all failure load levels and for both sexes seemed to be neck-shaft angle.
The authors thank Risto Bloigu for statistical consultation. The study was financially supported by the National Technology Agency of Finland and the Finnish Foundation for Advancement of Technology.