Journal Logo


Sex Differences in Parameters of Bone Strength in New Recruits

Beyond Bone Density


Author Information
Medicine & Science in Sports & Exercise: November 2008 - Volume 40 - Issue 11 - p S645-S653
doi: 10.1249/MSS.0b013e3181893cb7
  • Free


One of the most common and debilitating injuries observed during military recruit training is stress fracture (SF). SF occurs when bone is exposed to repetitive, high-volume, or novel exercise, particularly when a person significantly increases their level of activity over a short period. This injury is much more common in women (3,11,19), who are at two to six times greater risk for SF than men undergoing similar training (3,19). Previous research indicates that factors such as low bone density and smaller bone size may contribute to fracture susceptibility (2,12).

Studies relating low bone density in women to SF have relied heavily on two-dimensional imaging using dual-energy x-ray absorptiometry (DXA), which calculates mineral content within the total area scanned. DXA does not estimate porosity, marrow area, or bone shape but rather provides an areal measure of bone structural density (aBMD; mg·cm−2). Thus, studies using typical DXA images to assess sex differences may not accurately depict inherent anthropometric differences as there is no correction for bone and body size (20). In recent years, approaches to address this concern have included the calculation of bone mineral apparent density (8) and statistical adjustment for body size (14). This modified calculation may still not adequately account for individual differences in bone geometry related to age, chronic physical activity, calcium intake, or other genetic predeterminants.

A more direct approach to investigating sex differences in bone mineralization is available using cross-sectional imaging technologies such as peripheral quantitative computed tomography (pQCT). Calculation of volumetric bone mineral density (vBMD, mg·cm−3) can be made from isolated bone tissue, allowing for an assessment of tissue density independent of bone size. As small changes in bone tissue density can result in large changes in bone strength (5) and as research suggesting a relationship between aBMD and SF risk is equivocal, it is important in future studies to assess tissue density through measures of vBMD when attempting to determine fracture resiliency related to mineralization.

Smaller bone size, inherent in women, has also been related to SF susceptibility. Recent findings in a focused gender study that performed mechanical testing on human cadaver cores suggest that a smaller tibia may have reduced tissue ductility than larger tibia, independent of sex (26). Whether women are at greater risk for SF simply due to smaller bone size is not clear, but anthropometric data suggest that individuals with smaller, more slender tibia are at greater risk for SF (2,4,6,10,23). Given the contribution of bone shape and geometry to bone strength, researchers have attempted to estimate bone strength parameters in vivo using typical imaging methods. Radiographs were used to compute bone width and cross-sectional moments of inertia, which were related to SF risk in recruits (10). Programs have also been developed to derive geometric and strength parameters from DXA data (1). Using this method, Beck et al. (2) found significant sex differences in both structural density (aBMD) and geometry of the tibia and femur in military recruits. The fact remains, however, that these methods are estimates from two-dimensional images and may not accurately reflect bone tissue density and true geometry.

Peripheral QCT is a relatively new, noninvasive cross-sectional imaging tool used to assess parameters of bone quality of the extremities. In addition, to providing a measure of tissue density by calculating vBMD, differential assessment can be made between cortical and trabecular bone, which cannot be accomplished using DXA. The image additionally allows for more accurate determination of geometric measures and estimates of bone strength and quality at the scanned site. Nieves et al. (17) reported that pQCT images revealed significant sex differences in skeletal size and vBMD at one cortical site of the tibia in a physically active population of elite military cadets. Whether these same sex differences are evident in a cross section of age-matched military conscripts has not been determined.

Although pQCT provides an excellent method of determining parameters of whole-bone cross sections, analysis of regions within the cross section can provide additional information related to site-specific differences in geometry and tissue density of the tibia (13). Sector analysis related to sex differences in bone distribution is not available and may reveal additional insight into parameters of strength related to bone susceptibility to overuse injuries.

The purpose of this study was to assess sex differences in vBMD, geometry, and strength parameters of trabecular and cortical bone of the tibia in a cohort of military recruits. Within this article, we also present a novel pQCT image analysis method to analyze parameters in whole-bone and at defined anatomical sectors.



A total of 135 healthy women (n = 115) and men (n = 20) entering a gender-integrated recruit training program in the Israeli Defense Forces (IDF) volunteered to participate in this study. All volunteers were medically cleared by a physician before entering the training program and were eligible to participate in the study only upon providing written informed consent. Volunteers traveled to Heller Institute for baseline data collection on the first or second day after reporting to their unit for a 3-yr mandatory service period. This investigation was approved by the institutional review board of the Committee for Research on Human Subjects, Sheba Medical Center, Tel Hashomer, Israel.

Anthropometric Variables

Height (cm) was measured using a stadiometer, and weight (kg) was determined with a metric scale. Skinfold thickness was measured at four sites (biceps, triceps, suprailiac, and subscapular) with Lange skinfold calipers (Beta Technology, Santa Cruz, CA). The same investigator performed all skinfold measurements at all periods. Percent body fat was estimated from the skinfold measurements using previously established methods (9). Fat mass was determined by taking the subject's weight in kilograms and multiplying this by percent body fat determined by skinfold measurements. Lean mass was then calculated by subtracting the volunteer's fat mass from their total body weight. Tibial length was determined by measuring the distance from the distal aspect of the medial malleolus to the medial joint line.

Peripheral Quantitative Computed Tomography (pQCT)

Peripheral QCT (XCT 2000; Stratec Medizintechnik, Pforzheim, Germany) was used to measure bone characteristics of the tibia according to previously established methods (25). To assure measurement quality, a calibration check was performed on each data collection day by scanning a standard phantom with known densities of 168.5, 317.4, and 462.5 mg·cm−3. Volunteers were positioned on a chair with the nondominant leg extended through the scanning cylinder and were asked to maintain a convenient and stable position for the duration of the procedure (10-15 min). Initial scout scans were conducted at a scan speed of 40 mm·s−1 to identify the distal end plate of the tibia. After this, scans of the tibia (single axial slices of 2.2-mm thickness, voxel size 0.5 mm, measure diameter 140 mm) were taken at a translation speed of 20 mm·s−1 at 4%, 38%, and 66% of the approximated segment length proximal to the distal endplate of the tibia. These sites are typically used to analyze trabecular (4%) and cortical (38% and 66%) bone characteristics of the tibia.

Image Analysis Procedure

Image sets obtained using pQCT were analyzed using Matlab software (MathWorks, Natick, MA). To standardize positioning of the images across all scans, the image sets first underwent a rotation and registration procedure. Each 4% slice image was centered on the tibial perimeter. The 38% and 66% slices were centered on the intramedullary canal because the neutral axis seems better defined by the center of the canal than the center of the periosteal boundary. The crest of the tibia at 66% was then rotated to point in the anterior direction, and the 38% and the 4% slices were rotated accordingly by the same angle as at 66%. The images from two subjects who had images collected from their right legs were inverted so that they could be compared with the left tibias of the other subjects. Alignment was checked for each subject by overlaying plots of the periosteal boundaries for each subject.

After isolating the tibia from the fibula in each image, each voxel within the tibia was classified based on its vBMD value as being either trabecular (100-600 mg·cm−3), transitional (600-800 mg·cm−3), or cortical (800-1500 mg·cm−3). Histograms for each image indicated there were relatively few pixels in the transitional zone, which is in keeping with the known separation between cortical and trabecular vBMD values. For this reason, voxels in the transitional zone were assumed to be largely partial-volume artifacts and were not included in the final analysis.

Because the analysis in this article was not based on parameters calculated from the native, proprietary pQCT software, we present a detailed description of how each quantity was calculated using codes written for this purpose in Matlab.

Volumetric Measures of Bone Mineralization

Trabecular density (Tb.Dn)

After converting grayscale values to tissue densities using calibration constants, the average of all voxels falling within the trabecular thresholds (100-600 mg·cm−3) was calculated for the entire cross section and for each 60° polar sector at the 4% site.

Cortical density (Ct.Dn)

After converting grayscale values to tissue densities using calibration constants, the average of all voxels falling within the cortical threshold of 800 to 1500 mg·cm−3 was determined for the entire cross section and for each 60° polar sector at both the 38% and 66% sites.

Measures of Bone Geometry

Total cross-sectional area (CSA)

Total CSA of images taken at the 4%, 38%, and 66% sites was calculated by counting the number of pixels circumscribed by the periosteal boundary and then multiplying by the image resolution.

Tibial diameter (anterior-posterior [AP] and medial-lateral [ML])

Diameter of the tibia at the 4%, 38%, and 66% sites was calculated by measuring the distance (mm) from the locations of the voxels at the greatest directional extent in the AP and ML directions.

Trabecular area (Tb.Ar)

The CSA of trabecular bone at the 4% site (mm2) was obtained by calculating the areal sum of the voxels in the trabecular range. Although calculated at 38% and 66%, there were too few voxels in this range for a meaningful analysis.

Cortical area (Ct.Ar)

The areal sum of the voxels in the cortical range was used to determine Ct.Ar (mm2) at the 38% and 66% sites. Calculations were made for each polar sector (e.g., Ct.ArLat-Ant) and for the whole tibial cross section (Ct.ArTot).

Cortical thickness (Ct.Th)

This was calculated in 10° sector increments by calculating the average radial distance between the periosteal boundary and the endosteal boundary in each 10° sector. Values calculated for 10° sectors were averaged to determine values for 60° sectors.

Normalized canal radius (Ca.|Rd|)

The normalized Ca.|Rd| was calculated as the ratio of the endosteal radius to periosteal radius for each 10° and 60° sector. This gives an indication of the relative thickness of the cortical wall.

Measures of Bone Strength

CSA moments of inertia (IAP, IML, and J)

Moments of inertia (IAP and IML) were calculated about the AP and the ML axes, respectively, as a measure of bending strength independent of the degree of cortical mineralization. The polar moment of inertia (J = IAP + IML) provided a measure of torsional strength. Moments of inertia were calculated at the 38% and 66% sites using only those voxels in the cortical threshold range.

Bone strength index (BSI)

BSI, a measure of bending stiffness that takes into account bone mineralization, was calculated at the 38% and 66% sites. It is the product of average cortical tissue density (mg·mm−3) at the site and cross-sectional moment of inertia (mm4). Because the moment of inertia can be calculated as an ML, an AP, or a polar value, the BSI was calculated about each axis for both the 38% and 66% locations.

Slenderness index (SIAP and SIML)

SI (kg·mm−2) is a measure of bone mass distribution about the central axis that incorporates body weight and tibial length (25,28). SI values were calculated as the inverse ratio of the AP and the ML section moduli (J divided by AP/2) to the product of tibial length (L) and body weight (BW). For example, the SI about the AP axis would be

A higher SI is indicative of a more gracile bone, which is more likely to be composed of highly mineralized (and less ductile) cortical tissue (26).

Sector Analysis

In addition, to calculating parameters of bone mineralization, geometry, and strength for whole bone, images were further divided into six 60° polar sectors, using the positive x-axis as the 0° reference point and moving counterclockwise: lateral-anterior (Lat-Ant), anterior (Ant), medial-anterior (Med-Ant), medial-posterior (Med-Post), posterior (Post), and lateral-posterior (Lat-Post) (Fig. 1). Volumetric BMD, area, and cortical width and normalized Ca.|Rd| parameters were calculated for each sector for regional analysis. Additionally, these parameters were calculated for each of 36 10° sectors, and an image was generated to depict the mean bone "shape" for the entire cohort of men and women.

Representative male (A and C, left) and female (B and D, right) tibial cross-sectional images at the 4% site (top) and the 38% site (bottom). Images are divided into 60° sectors measured counterclockwise, with the positive x-axis defined as 0°: lateral-anterior (Lat-Ant), anterior (Ant), medial-anterior (Med-Ant), medial-posterior (Med-Post), posterior (Post), and lateral-posterior (Lat-Post) defined according to sector.

Statistical Analysis

Whole-bone and 60° sector parameters were analyzed using two-tailed t-tests (Statistical Package for the Social Sciences, version 15.0; SPSS Inc., Chicago, IL) to assess differences in bone measures between men and women. For whole-bone measures, an ANCOVA was used to calculate means adjusted for individual differences in body size (height and body weight).


The pQCT images were obtained from all 135 volunteers. Nine images were discarded due to noise within the image, most likely due to excessive movement during the scanning process. Included in our final analysis were images from 108 women and 20 men. Characteristics of 106 women and 20 men (two women refused to have anthropometric measures taken) are presented in Table 1. All but two volunteers were right leg dominant.

Volunteer characteristics (mean ± SD).

Bone mineralization

Values for trabecular and cortical vBMD of whole bone for men and women are presented in Table 2 as both unadjusted (left) and adjusted for height and weight (right). Trabecular vBMD (adjusted) at the distal tibia (4% site) was 10.2% greater (P < 0.001) in men than women. Sector analyses, depicted in Figure 2A, indicate that tissue density was 10.4 to 38.9 mg·cm−3 greater for men at each sector. These differences were statistically significant (P ≤ 0.05) with the exception of the postsector (P = 0.17).

Measurements (mean ± SD) of bone geometry, density, and strength values at three sites (4%, 38%, and 66% of the approximated segment length proximal to the distal end plate of the tibia) for men and women entering gender-integrated recruit training in the IDF.
Sector analysis of Tb.Dn (A) at the 4% site and Ct.Dn (B) at the 38% site (mg·cm−3) in men and women (mean ± SD, unadjusted for height and weight). * Significant sex difference for each sector at P < 0.001; ^ Significance at P < 0.05.

Conversely, cortical vBMD of whole bone, adjusted for height and weight at both the 38% and the 66% sites, was 2.7% and 2.0% greater (P = 0.006) in women than men, respectively (Table 2). Sector analysis revealed that cortical vBMD was 28.6 to 39.5 mg·cm−3 greater (P < 0.001) in women than men at the 38% site for all sectors (Fig. 2B). With the exception of the Lat-Ant sector, where no sex differences were observed, cortical vBMD at the 66% site was also 25.6 to 42.5 mg·cm−3 greater in women than men.

Bone geometry

Total CSA was 21.7% to 25.6% greater in men at all three scan sites (P < 0.001; Table 2). After adjusting for height and weight, differences were less (9.7-15.8% greater) but still statistically significant (P ≤ 0.002).

Tibial diameter was greater in men in both the AP and the ML directions (P < 0.0001) and remained significant after adjusting for height and weight at all measurement sites (Table 2).

Unadjusted values for trabecular and Ct.Ar in whole bone were 21.1% to 23.9% greater in men and remained 9.4% to 15.0% greater after adjusting for height and weight (Table 2). Values for CSA of trabecular (Fig. 3A) and cortical bone (Fig. 3B and C) at each 10° sector were plotted to depict data as the mean bone "shape" for men and women. Statistical analysis at each 60° sector revealed that values for men were greater than women (P < 0.001) by 30.5 to 45.0 mm2 for Tb.Ar and 5.4 to 17.6 mm2 for Ct.Ar (38% and 66% sites combined; Fig. 3).

Mean bone area (mm2) for men (•) and women (○) at each 10° sector at sites 4% (A), 38% (B), and 66% (C) from the distal aspect of the tibia. Asterisks at the 60° sector labels (lateral-anterior [Lat-Ant], anterior [Ant], medial-anterior [Med-Ant], medial-posterior [Med-Post], posterior [Post], and lateral-posterior [Lat-Post]) represent significant sex differences (unadjusted for height and weight, P < 0.001) for the area of that sector, derived by summing the areas for each 10° sector.

Ct.Th was 16.7 (P < 0.001) and 9.4% (P = 0.01) greater in men at the 38% and the 66% sites, respectively. These differences narrowed to 8.9 (P = 0.002) at the 38% site and became nonsignificant at the 66% site after adjusting for body size (Table 2). Sector analysis revealed that Ct.Th was from 0.5 to 1.0 mm greater (P < 0.001) at all sectors at the 38% site (Fig. 4A). At the 66% site, thickness was similar between men and women for all but the Post and Lat-Post sites (0.53 and 0.46 greater, respectively, P = 0.001; Fig. 4C). Ca.|Rd| did not differ significantly between men and women for whole bone (Table 2) or at regional sectors (Fig. 4B and D), with the exception of a small but significant increase in men at the 66% Med-Post site (P = 0.03).

Mean Ct.Th (A and C) and normalized Ca.|Rd| (B and D) at the 38% and 66% sites, respectively, at each 10° sector for men (•) and women (○). Asterisks at the 60° sector labels (lateral-anterior [Lat-Ant], anterior [Ant], medial-anterior [Med-Ant], medial-posterior [Med-Post], posterior [Post], and lateral-posterior [Lat-Post]) represent significant sex differences for that sector (unadjusted for height and weight; P ≤ 0.001); ^ Significance P < 0.05.

Bone strength

Measurements of area moments of inertia (IML, IAP, and J) determined at the 38% and 66% sites were 41.2% to 44.6% greater in men (P < 0.0001; Table 2). After adjusting for body size, the sex difference was still significant (P < 0.001) but had narrowed (27.4-31.2% greater in men). Bone strength and slenderness indices were not adjusted for body size because body mass and tibial length are adjusted for within the measure. The BSI for men was 40% greater at both the 38% and the 66% sites (Table 3). Women exhibited greater slenderness indices in the AP and the ML directions at the 38% site (P = 0.006 and 0.03, respectively) but were greater only in the AP direction at the 66% site (P = 0.01; Table 3).

Bone strength and slenderness indices (mean ± SD) measured at cortical sites 38% and 66% from distal aspect of tibia for men and women entering gender-integrated recruit training in the IDF.


Our study results indicate that women entering recruit combat training in the IDF possess disadvantages in bone geometry, strength, and tissue mineralization that may result in greater susceptibility to bone overuse injury relative to their male counterparts. Measures of tibial geometry and strength were significantly lower in women. Further, women had a higher tibial slenderness index (SI) than men, indicating a more "gracile" bone, particularly at the 38% site, a frequent site of SF. Trabecular vBMD in the distal epiphysis was significantly lower, whereas cortical vBMD of the diaphysis, a frequent site of SF in both men and women, was significantly higher. These differences remained significant even after adjusting for differences in height and weight and suggest that tibial bones of women may be less able, relative to men, to withstand loads imposed by repetitive mechanical loading.

Lower trabecular tissue density in the distal tibial epiphysis may result in greater strain on the tibial shaft under conditions of repetitive loading. In particular, lower Tb.Dn in the scaffold supporting the epiphyseal cortices of long bones, as observed in our study, may lower the force threshold for fracture by potentially placing increased load on the cortical structure (21). Given our additional observation that the tibial bone of women was more slender than that of men, the load to the bone shaft may be even greater, placing additional strain on the cortex of the tibial shaft during strenuous exercise.

In addition to having smaller, more slender bones, the women in our study exhibited significantly greater cortical vBMD of the tibia than men. Contrary to our findings, tibial vBMD assessed using pQCT was significantly greater in men than women in a group of elite military cadets (17). The apparent contradiction may simply be methodological. Tibial vBMD values previously reported by Nieves et al. (17) were 794.3 mg·cm−3 in women and 850.1 mg·cm−3 in men versus the values in this study of 1194 mg·cm−3 for women and 1159 mg·cm−3 for men. The disparity between studies may reflect that the vBMD reported by the resident pQCT software uses a cortical threshold of more than 710 mg·cm−3, whereas our study used a threshold of 800 mg·cm−3 to define cortical bone (17). It is also not clear whether the vBMD reported in their study represents an average density of the total bone area (which is significantly greater than the Ct.Ar) or density of the Ct.Ar. It is also possible that the values reported were taken at the proximal tibia, which may reflect bone with a lower average tissue density (17).

Our results support recent work to suggest that smaller, more slender bones in men and women are composed of tissue with higher mineralization (27). The higher tissue density in smaller bone is hypothesized to be a coadaptation to the lower relative amount of cortical bone within the tibial shaft (27). There is compelling evidence to support the idea that a compensatory increase in mineralization in smaller tibia may result in more damageable bone under conditions of high-volume mechanical loading in young adults (27,28). Researchers used standardized beams of bone prepared from cadaver tibia harvested from young adults to assess tissue-level mechanical properties that included stiffness, strength, ductility, toughness, and damageability. They found that, in male specimens, narrower tibiae were comprised of tissue that was more brittle and more prone to accumulating damage compared with tissue from wider tibias (28). Thus, although we acknowledge that small increases in tissue density confer greater overall strength of a whole bone, more mineralized cortical tissue of smaller bones in both men and women may be "stiffer" and more prone to microcracking under conditions of repetitive loading.

The notion that higher mineralization increases whole-bone strength in bending (5) while at the same time produces "brittle" cortices that are less able to withstand repeated loads is an interesting one, given the proclivity to relate fragility fractures with decreased structural density measured by DXA as typically seen with osteoporotic fracture. The literature, however, does support the concept that bone with higher ash content (i.e., higher tissue density) may be more susceptible to fracture because microfractures can more readily propagate through bone that is highly mineralized (28). Other researchers have shown that the majority of microdamage occurs in the extra-osteonal portion of bone (19,23), believed to be more mineralized because it is comprised of fragments of older osteons and primary circumferential lamellar bone. The majority of evidence supporting low aBMD as a key risk factor for SF in women is contradictory (16,17) and is based on DXA scans that do not take bone size into account (16). Given the evidence linking high mineral content to a degradation of toughness in bone (7), we hypothesize that whereas low structural density (aBMD) may increase risk for fragility fracture, high tissue density (vBMD) may predispose an individual to SF. This interpretation of BMD measurements, and their relationship to overall bone quality, might provide insight into the contradictory findings in the literature related to the relationship between bone mineralization and fracture risk in young men and women.

Measures of whole-bone geometry and strength were significantly lower in women; however, the Ca.|Rd| measures were not significantly different suggesting that, based on cross-sectional shape alone, women and men have similar cortical distributions away from the central axis. Because the predominating mechanical environment varies relative to the centroid of a given cross section (e.g., tensile in the anterior regions and compressive in the posterior regions), it is expected that bone quality will vary regionally as well. Our results suggest that regional differences in bone geometry and mineralization exist and follow similar patterns in men and women. Further analysis of adaptations related to varying anthropometric and lifestyle factors between men and women that may cause bone to be loaded differently, particularly specific regional adaptations, may provide further insight into this observation.

In conclusion, we observed that significant sex differences exist in parameters of bone tissue mineralization, geometry, and strength that remain evident after adjusting for body size. Our findings suggest that these sex differences may contribute to a decreased ability to withstand the demands imposed by novel, repetitive exercise in untrained individuals entering military recruit training. The contribution of inherent sex differences in bone strength to the higher incidence of SF fracture injury in women warrants further study.

The opinions and assertions in this article are those of the authors and do not necessarily represent official interpretation, policy, or views of the US Department of Defense or the Israeli Defense Forces.

We would like to thank Heath Isome, Daniel Catrambone, and Dan Schiferl for their technical expertise during conduct of this study.


1. Beck TJ, Ruff CB, Mourtada FA, et al. Dual-energy x-ray absorptiometry derived structural geometry for stress fracture prediction in male U.S. Marine Corps recruits. J Bone Miner Res. 1996;11:645-53.
2. Beck TJ, Ruff CB, Shaffer RA, Betsinger K, Trone DW, Brodine SK. Stress fracture in military recruits: gender differences in muscle and bone susceptibility factors. Bone. 2000;27:437-44.
3. Brudvig TJ, Gudger TD, Obermeyer L. Stress fractures in 295 trainees: a one-year study of incidence as related to age, sex, and race. Mil Med. 1983;148:666-7.
4. Brukner P, Bennell K. Stress fractures in female athletes. Diagnosis, management and rehabilitation. Sports Med. 1997;24:419-29.
5. Carter DR, Hayes WC. The compressive behavior of bone as a two-phase porous structure. J Bone Joint Surg Am. 1977;59:954-62.
6. Crossley K, Bennell KL, Wrigley T, Oakes BW. Ground reaction forces, bone characteristics, and tibial stress fracture in male runners. Med Sci Sports Exerc. 1999;31(8):1088-93.
7. Currey JD, Brear K, Zioupos P. The effects of ageing and changes in mineral content in degrading the toughness of human femora. J Biomech. 1996;29:257-60.
8. Cvijetic S, Korsic M. Apparent bone mineral density estimated from DXA in healthy men and women. Osteoporos Int. 2004;15:295-300.
9. Durnin JV, Rahaman MM. The assessment of the amount of fat in the human body from measurements of skinfold thickness. Br J Nutr. 1967;21:681-9.
10. Giladi M, Milgrom C, Simkin A, et al. Stress fractures and tibial bone width. A risk factor. J Bone Joint Surg Br. 1987;69:326-9.
11. Jones BH, Knapik JJ. Physical training and exercise-related injuries. Surveillance, research and injury prevention in military populations. Sports Med. 1999;27:111-25.
12. Jones BH, Thacker SB, Gilchrist J, Kimsey CD Jr, Sosin DM. Prevention of lower extremity stress fractures in athletes and soldiers: a systematic review. Epidemiol Rev. 2002;24:228-47.
13. Lai YM, Qin L, Hung VW, Chan KM. Regional differences in cortical bone mineral density in the weight-bearing long bone shaft-a pQCT study. Bone. 2005;36:465-71.
14. Looker AC, Beck TJ, Orwoll ES. Does body size account for gender differences in femur bone density and geometry? J Bone Miner Res. 2001;16:1291-9.
15. Myburgh KH, Hutchins J, Fataar AB, Hough SF, Noakes TD. Low bone density is an etiologic factor for stress fractures in athletes. Ann Intern Med. 1990;113:754-9.
16. Nattiv A. Stress fractures and bone health in track and field athletes. J Sci Med Sport. 2000;3:268-79.
17. Nieves JW, Formica C, Ruffing J, et al. Males have larger skeletal size and bone mass than females, despite comparable body size. J Bone Miner Res. 2005;20:529-35.
18. Norman TL, Wang Z. Microdamage of human cortical bone: incidence and morphology in long bones. Bone. 1997;20:375-9.
19. Pester S, Smith PC. Stress fractures in the lower extremities of soldiers in basic training. Orthop Rev. 1992;21:297-303.
20. Prentice A, Parsons TJ, Cole TJ. Uncritical use of bone mineral density in absorptiometry may lead to size-related artifacts in the identification of bone mineral determinants. Am J Clin Nutr. 1994;60:837-42.
21. Reich T, Gefen A. Effect of trabecular bone loss on cortical strain rate during impact in an in vitro model of avian femur. Biomed Eng Online. 2006;5:45.
22. Schaffler MB, Choi K, Milgrom C. Aging and matrix microdamage accumulation in human compact bone. Bone. 1995;17:521-5.
23. Seeman E, Duan Y, Fong C, Edmonds J. Fracture site-specific deficits in bone size and volumetric density in men with spine or hip fractures. J Bone Miner Res. 2001;16:120-7.
24. Selker F, Carter DR. Scaling of long bone fracture strength with animal mass. J Biomech. 1989;22:1175-83.
25. Sievanen H, Koskue V, Rauhio A, Kannus P, Heinonen A, Vuori I. Peripheral quantitative computed tomography in human long bones: evaluation of in vitro and in vivo precision. J Bone Miner Res. 1998;13:871-82.
26. Tommasini SM, Nasser P, Jepsen KJ. Sexual dimorphism affects tibia size and shape but not tissue-level mechanical properties. Bone. 2007;40:498-505.
27. Tommasini SM, Nasser P, Schaffler MB, Jepsen KJ. Relationship between bone morphology and bone quality in male tibias: implications for stress fracture risk. J Bone Miner Res. 2005;20:1372-80.
28. Zioupos P, Currey JD. The extent of microcracking and the morphology of microcracks in damaged bone. J Math Sci. 1994;29:978-86.


©2008The American College of Sports Medicine