Journal Logo


Predicting Diaphyseal Cortical Bone Status Using Measures of Muscle Force Capacity


Author Information
Medicine & Science in Sports & Exercise: July 2018 - Volume 50 - Issue 7 - p 1433-1441
doi: 10.1249/MSS.0000000000001581
  • Free


Muscle and bone tissues originate from similar cellular origins and “the muscle–bone unit” is used to describe both the mechanical and biochemical coupling of the tissues throughout the life-span (1–3). Much research has focused on the mechanical aspect of the muscle–bone unit as muscle force accounts for many of bones adaptive responses, owing to the fact that internal muscle forces are the highest experienced by the skeleton, trauma aside (2). In the context of physical activity (PA), increasing muscle force may be an avenue for the optimization of bone mineral accrual in youth and slowing of bone loss in adulthood (4).

When assessing the effects of PA on bone strength, researchers often use muscle mass or size as surrogates for the forces applied to bones by muscle because of the link between muscle cross-sectional area (MCSA) and force production (5). However, the reasoning that muscle size and force output are interchangeable in their association with the forces applied to bone is flawed (6), with one of several reasons being that force capacity in two muscles of the same size is dependent on several factors such as fiber-type distribution, with fast-twitch isoforms producing 40% to 70% greater isometric forces than slow-twitch isoforms (7). Furthermore, the accurate measurement of muscle mass has its own inherent drawbacks. Researchers must decide among expensive clinical methodologies, which are typically more accurate, such as magnetic resonance imaging or to a lesser extent dual-energy x-ray absorptiometry (DXA) or peripheral quantitative computed tomography (pQCT), or less expensive but generally less accurate methodologies such as quantitative ultrasound. Furthermore, commonly used methodologies for the assessment of bone accrual and osteoporosis risk such as DXA and pQCT also expose participants to x-ray radiation, with risks in children being higher than adults because of their greater susceptibility to the risk of radiation-induced biological effects (8).

Thus, there is a need to compare the utility of different muscle force capacity measures against currently used surrogates to identify noninvasive, feasible measures that might be used within the clinical and research settings in place of often costly or radiative methods of assessing muscle mass (9). In this context, the primary aims of this study were twofold: 1) to determine whether clinical laboratory–based methodologies for measuring muscular force capacity such as the Nottingham leg extensor power rig (peak lower body power) and the Biodex dynamometer (peak joint torque) account for a similar portion of variance in diaphyseal cortical bone status (density, structure, and estimated bending strength) as a commonly used muscular force surrogate, MCSA, at the midtibia in young men and women, and 2) to determine whether a noninvasive, field measure of muscle force capacity, the vertical jump (peak anaerobic power), accounts for a similar portion of variance in cortical bone status as corresponding clinical laboratory–based methodologies (Biodex dynamometer and Nottingham leg extensor power rig). We hypothesized that (a) muscle force capacity derived from clinical laboratory–based methodologies would predict at least as much variance in cortical bone status as the corresponding site-specific MCSA, and (b) muscle force capacity derived from vertical jump scores would predict at least as much variance in cortical bone status as the clinical laboratory–based methodologies.


Recruitment and participant characteristics

This cross-sectional study included participants recruited between fall 2016 and spring 2017 at a major university in the southeast United States. Recruitment consisted of mass e-mails to university-provided accounts, posted flyers, and word-of-mouth advertising. All participants (n = 146; n = 77 female) were white, non-Hispanic, or Latino, and age was restricted to 18 to 21 yr to ensure a similar skeletal maturation (>90% bone accrual). Inclusion criteria required participants to (a) be a current University of Georgia student; (b) be free of orthopedic limitations that precluded PA; (c) not be current smokers (past 6 months); (d) not be pregnant or planning to become so for the duration of their participation; (e) not take medications known to affect bone metabolism (i.e., glucocorticoids), dietary intake, or PA; (f) be free of medical conditions know to affect bone metabolism (i.e., Crohn’s disease); (g) not currently be diagnosed with an eating disorder; and (h) not have undergone recent weight loss surgery. All participants provided written informed consent before participation. The institutional review board of the university approved all aspects of the protocol.

Anthropometric measures

Body mass and standing height were measured via digital scale (Seca Bella 840, Columbia, MD) and stadiometer (Novel Products Inc., Rockton, IL) to the nearest 0.1 kg and 0.1 cm, respectively. Tibia length was measured using a sliding anthropometer to the nearest 0.1 cm as the distance from the distal medial malleolus to the tibial plateau.


Objectively measured PA was assessed over a period of ≥7 d using a triaxial ActiGraph GT3X+ accelerometer (Firmware v3.2.1.). Participants wore the accelerometer on the mid axillary line of their right hip during all waking hours, except during water-based activities. Data were considered valid if the participant had accrued at least 10 h·d−1 of data on at least two weekdays and one weekend day, determined via the wear time macro developed by the National Cancer Institute (Centers for Disease Control and Prevention/National Center for Health Statistics) from the Troiano algorithm (10). Periods with consecutive raw activity values of zero (with a 2-min spike tolerance of ≤100 counts) for 60 min or longer were interpreted as “nonwear” and excluded from this analysis (11). Data analyses were performed using ActiGraph software (v6.10.1; ActiGraph, LLC, Fort Walton Beach, FL), with 15-s epochs and the VM3 vector magnitude approach (12). As such, the cut point used to classify minutes per day spent in moderate-to-vigorous PA (MVPA) was ≥2691 counts per minute. A weighted average PA time [(weekday average × 5) + (weekend average × 2)/7] was used to represent mean weekly activity variables.

Bone assessments by pQCT

pQCT (XCT-3000; Stratec Medizintechnic GmbH, Pforzheim, Germany) scans were taken at sites 38% and 66% of the length of the nondominant tibia using the distal metaphysis as the anatomical marker. Scans used a 0.4-mm voxel and slice thickness of 2.4 mm. A software-automated scout view was used to position the measurements, using the medial end plate as an anatomic reference. The following variables were derived from the 38% site scans: cortical volumetric bone mineral density (Ct.vBMD; mg·mm−3), cortical area (Ct.Ar; mm2), cortical thickness (Ct.Th; mm), periosteal circumference (PC; mm), endosteal circumference (EC; mm), and polar strength-strain index (pSSI; mm3). The 66% site scans were taken to acquire MCSA (mm2), which was determined via automated analyses, which used edge detection, threshold techniques, and image filters to separate tissues. Cortical parameters were assessed using cort mode 2 and the default threshold of 710 mg·cm−3, except for pSSI, which used a threshold of 480 mg·cm−3. All pQCT measures were performed and analyzed by one operator who was trained for acquisition and analysis following guidelines provided by Bone Diagnostic (Spring Branch, TX), using Stratec software (version 6.20). The manufacturer phantom was scanned daily to maintain quality assurance (Stratec Medizintechnik GmbH, Pforzheim, Germany).

Muscle force capacity

First, the Nottingham Leg Extensor Power Rig (Medical Engineering Unit, University of Nottingham Medical School, Nottingham, UK) was used as a clinical measure of peak lower body power (13). Participants pushed out on a pedal with the nondominant leg as hard and fast as possible. Up to 10 trials were performed until a plateau in force output was achieved. Second, a Biodex System Pro 4 dynamometer (Biodex Medical Systems, Inc., Shirley, NY) was used as the clinical measure of peak isokinetic torque at the knee and ankle joints. Participants were positioned per the manufacturer’s guidelines and performed five maximal effort voluntary contractions at 60°·s−1 to assess peak torque (N·m) of their nondominant side. Ankle dorsiflexion and knee extension via dynamometry have been found to be highly reliable in young men and women (14), and were chosen because they reflect muscle–bone units with direct force transfer onto the tibia. Finally, the Vertec vertical jump system (Vertec, USA) was used as a field measure of peak anaerobic power. Participants jumped as high as possible over five individual attempts, tapping the rotating flags on the Vertec to register jump height. Peak anaerobic power was obtained using the Sayer’s equation (15), peak anaerobic power (W) = 60.7 × (jump height [cm]) + 45.3 × (body mass [kg]) − 2055. Jump height from Vertec testing is strongly correlated (r = 0.91) with criterion methodologies such as three-camera motion analysis (16). A single set of familiarization trials were performed for each methodology before testing.

Dietary calcium and vitamin D

To assess the primary micronutrients known to influence bone status, dietary calcium (mg·d−1) and vitamin D3 (IU·d−1) intakes, a 3-d diet record required participants to log all foods consumed over two weekdays and one weekend day. Trained interviewers using food models to aid in estimation of portion sizing interviewed participants before receipt of the complete log. Dietary data were analyzed using the Nutrition Data Systems for Research software (University of Minnesota, Minneapolis, MN), and dietary supplements were included within all estimates based on responses to a simple questionnaire. All dietary data were checked by another trained interviewer for quality control.

Statistical analysis

Analyses were performed using SPSS for Windows (SPSS 22.0, Chicago, IL) with an α level of P < 0.05. Assumptions were examined across all combinations of variables, and multicollinearity among predictors were assessed via variance inflation factor, with values >10 indicating multicollinearity (17); no variables met these criteria. Participants (n = 4) were excluded from analysis after being deemed influential multivariate outliers on the basis of Mahalonobis distance critical values for a chi-squared distribution with k degrees of freedom (k = number of predictors), leaving the final analyzed sample at n = 142.

Means and SDs were calculated for all participant characteristics and primary outcome variables. Independent t-tests identified any differences between sexes for descriptive purposes (Table 1), and semipartial correlations assessed the relationship between proposed predictors and cortical bone outcome variables, controlling for the effect of sex and age on each predictor (Table 2). Potential covariates were then entered into a regression model in a sequential manner to assess their relationship with cortical bone status. Calcium (mg·d−1) and vitamin D3 (IU·d−1) were both identified as having no relationship with any bone status variables in any model and were thus not included in final analyses. The following predictors of cortical bone status were assessed using simultaneous entry into a multivariate regression model and squared semipartial correlations (sr2): age (years), sex (female, 0; male, 1), tibia length (cm), body mass (kg), MVPA (min·d−1), muscular force variables, and MCSA (mm2). Squared semipartial correlations represent the percent of variance in the dependent variable uniquely explained by each predictor, offering readers the ability to also calculate the amount of variance accounted for by the combination of all predictors by summing sr2 and subtracting from the overall model R2. All results are presented as standardized beta coefficients (β; Tables 3 and 4) to aid in comparison among predictors and 95% confidence intervals highlight significant effects. All data are expressed as mean ± SD, unless otherwise indicated.

Descriptive characteristics (n = 142).
Semipartial correlations between predictor variables and cortical bone outcomes (n = 142).
Multiple regressions comparing the predictive ability of descriptive variables, MCSA, and Biodex knee extension peak torque on cortical bone outcomes (n = 142).
Multiple regressions comparing the predictive ability of descriptive variables, MCSA, and Vertec peak anaerobic power on cortical bone outcomes (n = 142).

Adequate power to detect an effect was ensured through the use of G-power to run an a priori power analysis examining an R2 increase in a fixed linear multiple regression model. Using a seven-predictor model with a power of 0.8 and α = 0.05 to test the effect of adding one predictor to the model with an expected effect size of f2 = 0.15 (moderate), a sample size of n = 55 is required. Previous research suggests that partial correlations between muscle torque and cortical outcomes, controlling for stature and maturity, range between r = 0.56 and 0.62 (f2 = 0.45–0.61); (9) thus, we believed that this was a conservative estimate, providing adequate power for the analysis.



A total of 452 individuals completed the online screening survey between September 2016 and February 2017, of which 108 were deemed ineligible based on the following inclusion/exclusion criteria: older than 20 yr (n = 45), did not complete the screening survey in enough detail to contact (n = 30), taking medications known to affect bone status (n = 8), current smoker (n = 7), diagnosed with a disease known to affect bone status (n = 4), not willing to wear a PA monitor (n = 4), were currently injured and unable to complete testing (n = 4), not willing to meet the time commitments of the testing (n = 4), or had an eating disorder (n = 2). Thus, 344 potential participants were contacted via e-mail, of which 184 were scheduled for testing. Upon scheduling or following the first of two testing visits, 24 participants dropped out from the study, with 6 additional participants excluded because of not having worn their PA monitors for the required time, leaving a total useable sample of 154. Because of the known moderating effect of race on bone outcomes, only white non-Hispanic or Latino participants were included (n = 146), with a further 4 participants excluded as multivariate outliers, leaving the final analyzed sample at 142.

Preliminary analysis

Combined and sex-specific descriptive data are reported in Table 1. MVPA was similar between sexes, with both groups accruing large amounts of activity. Male participants were greater in stature and mass than female participants, with cortical bone status mirroring whole-body differences. Bone outer diameter was 12.8% larger in male participants as reflected by PC, and although their medullary cavity was 11.0% wider and Ct.vBMD 2.4% lower than their female counterparts, Ct.Th and pSSI suggested that their bones were 14.7% thicker and 40.2% stronger, respectively (all P < 0.05). Direct measures of muscular force capacity and MCSA followed a similar pattern, with male participants having 18.2% greater midtibial MCSA, and 50.6%, 58.0%, 47.9%, and 64.1% greater peak lower body power, peak knee extension torque, peak ankle dorsiflexion torque, and peak anaerobic power, respectively, compared with female participants (all P < 0.05).

Semipartial correlations controlling for age and sex are shown in Table 2. All body stature and muscle force predictors were significantly related to Ct.vBMD (sr range = −0.14 to −0.22, all P < 0.05); however, minutes per day of MVPA and MCSA were not. Contrary to the negative relationships observed with volumetric density, bone structural variables were positively related to both body stature and muscle force variables, including MCSA. Tibia length and body mass were positively related to all bone structural outcomes (r range = 0.22–0.42, all P < 0.05), with the exception of the relationship between tibia length and Ct.Th, which did not reach significance (P = 0.057). Furthermore, muscle force estimates and MCSA were positively related to all cortical bone structural measures except EC, with the strongest of these relationships being between peak anaerobic power and Ct.Th (sr = 0.46, P < 0.001). Finally, MVPA was positively related to Ct.Th (sr = 0.18, P = 0.013), but was unrelated to any other cortical bone outcome.

Primary analysis—MCSA, peak torque, and peak power

Multivariate regression models assessing the predictive utility of MCSA, Biodex knee extension peak torque, and Vertec peak anaerobic power on cortical bone outcomes of interest are reported in Tables 3 and 4, respectively. Models accounted for 30.8% (EC) to 78.4% (pSSI) of the variance in cortical bone outcomes; however, when peak lower body power and ankle dorsiflexion peak torque were entered into regression models, neither emerged as a significant predictor of cortical bone outcomes with 95% confidence intervals spanning zero (data not shown). Furthermore, none of the measures of muscle force predicted Ct.vBMD either individually or when combined with MCSA.

Biodex knee extension peak torque

Knee extension peak torque emerged a significant predictor of all cortical structural measures except EC (β = 0.21–0.44, all P < 0.05), independent of MCSA (see Table 3). The inclusion of knee extension peak torque and MCSA accounted for 78.4% of the variance in bending strength (pSSI) and 53.5%–78.3% of the variance in bone structure. The individual contribution of each muscular variable to cortical bone outcomes ranged from 2.6% to 7.5% and 1.0% to 4.5% for MCSA and knee extension peak torque, respectively (all P < 0.05).

Vertec peak anaerobic power

Peak anaerobic power was a positive predictor of Ct.Ar, Ct.Th, and pSSI, and a negative predictor of EC, independent of MCSA (see Table 4). Furthermore, the magnitude of variance in Ct.Th accounted for by peak anaerobic power was greater than that of MCSA (5.5% vs 4.7%, respectively), and MCSA did not significantly predict EC (β = −0.07, P = 0.549). In contrast, MCSA accounted for a greater proportion of variance in Ct.Ar and pSSI than peak anaerobic power (4.0% and 3.0% vs 2.9% and 0.8%, respectively), and was a significant predictor of PC (β = 0.24, P = 0.001), whereas peak anaerobic power was not.


We hypothesized that a) muscle force capacity derived from clinical laboratory–based methodologies would predict at least as much variance in cortical bone status as the corresponding site-specific MCSA, and b) muscle force capacity derived from vertical jump would predict at least as much variance in cortical bone status as the clinical laboratory–based methodologies. Our primary hypothesis was rejected because none of the laboratory-based measures of muscle force capacity predicted as much variance in cortical bone structure or estimated bending strength as MCSA. However, the measure of knee extension peak torque did predict variance in cortical bone structure and estimated bending strength, which was shared with and independent of MCSA. Furthermore, in support of our secondary hypothesis, a field estimate of peak anaerobic power derived from vertical jump height predicted cortical bone structure and estimated bending strength to an equal to and/or greater extent than both knee extension peak torque and MCSA, depending on the cortical bone outcome assessed. These data are novel because they provide a basis for the vertical jump to be used as a field-based measure of muscular power in the prediction of cortical bone structure and estimated bending strength in young adults, without the need for expensive and/or radiative methodologies for attaining muscle mass. The results add to previous research by comparing the predictive utility of multiple commonly used and validated measures of muscle force capacity, both muscle power and peak torque, in a highly active population who are not affected by maturational or major physical function differences.

Muscle force surrogate—MCSA

Since the discovery that bone mass and structure respond directly to the forces placed upon it (18), it has become commonplace for researchers to use mass- or size-based surrogates of muscle force capacity (5,19–22). Surrogates such as MCSA or total lean mass strongly predict bone strength; however, this widespread use of muscle size– and mass-based surrogates of muscle force oversimplifies the muscle–bone relationship with regard to loading. This oversimplification is eloquently highlighted in a study comparing the differences between playing and nonplaying arms in female tennis players (21). In this cohort, PA-induced differences between forearm bone status as measured by magnetic resonance imaging were of the magnitude of 6.0%–13.0%; however, only 11.8%–15.9% of the variance was independently accounted for by differences in muscle area despite significant relationships (r = 0.36–0.40, P < 0.05). This example illustrates how, especially in an active cohort, muscular factors including but not limited to fiber-type distribution and pennation angle, motor unit activation, and tendon length may also contribute to absolute force and/or power production and transfer onto bone surfaces (6). Thus, when characterizing the mechanical effects of muscle on bone, the potential drawbacks such as a lack of specificity, the expense to researchers and/or patients, or the potential for radiation exposure (<1 to 27 μSv, depending on the methodology used, which equals 2.0% to 54.0% of the recommended yearly dose [8,23–25]) that usually accompanies muscle force surrogates, provides a convincing argument against their use when specific measures of force or power are available.

Muscular force capacity

Multiple methodologies have been used to examine the predictive capacity of muscle force and or/power on bone status including the Nottingham Leg Extensor Power Rig (23,26), the Biodex dynamometer (16,24), and vertical jump assessments (9,25,27,28), among others (24,27,29–31). However, few studies have directly compared measures, with those attempting to do so reporting conflicting results in regard to their predictive utility for bone status (23,26–29,31,32). Some reports suggest a difference in the predictive utility of muscle force compared with muscle power (28,30), with estimates of muscle power potentially having stronger relationships with bone outcomes. These articles used jump mechanography and ground reaction forces during a two-legged counter movement jump to assess power and force, respectively. This is important to note because peak muscular force derived from ground reaction force is not equal to dynamometry-derived measures of peak isokinetic torque, the criterion methodology for quantifying internal muscle forces.

Clinical methodologies—Biodex dynamometer (peak torque)

Peak torque derived from dynamometry has been assessed as a predictor of both DXA-based and pQCT-derived bone status across the life-span (16,19,29,31–33). However, comparison of these results is challenging because of the differences in sample characteristics, variability in the joint assessed and muscle actions used to acquire peak torque, and the paucity of studies using pQCT-based bone status measures.

Our data suggesting that knee extension peak torque is a predictor of cortical bone geometry and estimated bending strength (Ct.Ar, Ct.Th, PC, and pSSI), independent of MCSA, agrees with one prior study, albeit in a very different population. In that study, knee extension torque was the strongest predictor of pQCT-derived distal tibia bone strength index (β = 0.30, P < 0.001) controlling for height, weight, and age in postmenopausal women (n = 139; 62 ± 4 yr old) (31). These results mirror the effect sizes seen in our bone strength predictive models (pSSI; β = 0.31, P = 0.001). The previous study, however, did not account for 1) lean mass or limb length, both of which are important predictors of cortical bone status; 2) reported only results at the distal tibia; and 3) derived knee extension torque from isometric contractions, inhibiting direct comparison with isokinetic torque.

Unlike the current study, Wetzsteon et al. (16) found that a model including ankle dorsiflexion isometric peak torque predicted up to 92% of the variance in both strength and structure of cortical bone, independent of MCSA. It may be that their heterogeneous sample of n = 321 male and female participants 5 to 35 yr old may have introduced greater variance to outcome measures, highlighting relationships not seen in our narrow age range. Alternatively, in such a highly active sample as the present study, forces arising from the quadriceps may be applied to the proximal tibia at a greater magnitude and frequency than those from the ankle dorsiflexors. With the reasoning behind this being that the knee extensors are involved to a greater extent in eccentric muscle actions during high-force movements due to their crossing of both the hip and knee.

A larger body of research has examined the relationships between isokinetic measures of peak torque and DXA-derived measures of bone mass, structure, and strength (19,29,32,33). In a study where the results mirror our own, Daly et al. (32) reported that in prepubertal girls (n = 103, 7.8 ± 0.6 yr old) both leg lean mass and isokinetic knee extension torque were independently and similarly predictive of femoral neck diameter, cross-sectional area, and estimated strength, and DXA-derived leg and femoral neck bone mineral content. Despite bone status being measured in a nonvolumetric fashion by DXA, comparable additional variance of 2.0%–5.0% was accounted for by knee extension peak torque once leg lean tissue mass was included in the model concurrently, similar to results seen in our knee extension models of 1.0%–4.5%. Others have also reported that peak torque is related to bone geometric properties (33); however, most DXA-based studies only assess relationships between peak torque and measures of bone mass (19,29). Despite the methodological differences in assessing muscle force and bone status, dynamometry measures of peak torque have shown strong, consistent relationships with bone status across multiple populations, supporting its utility as a predictor of bone status, independent of muscle surrogates.

Field methodologies—Vertec vertical jump (peak anaerobic power)

Although clinical measures of muscle force output, both peak torque, and peak power show promise in the prediction of bone status, field-based measures are required for population-level application. Peak anaerobic power derived from a vertical jump was chosen for this study because of its ease of application, low cost, and the strong emerging research backing its validity as a predictor of bone status in multiple populations (9,25,30).

An important finding of the present study was that peak anaerobic power independently predicted cortical structure and strength, explaining more variance in Ct.Th than in MCSA when both were included in the model. These data support the use of vertical jump as a field-based predictor of bone status and agree with recent findings using a similar population and equation for estimating peak anaerobic power (22). Janz et al. (9) also predicted peak anaerobic power using the Sayer’s equation in adolescent boys and girls (n = 303, 17.5 ± 0.4 yr old), with mediation models suggesting that peak anaerobic power was a direct predictor of bone compressional strength, pSSI, and Ct.Ar at the 66% tibia in both boys and girls, independent of MCSA. Moreover, effect sizes were of a similar magnitude, with a 1-SD increase in peak anaerobic power predicting a 0.38- and 0.15-SD increase in Ct.Ar (in boys vs girls, respectively), compared with a combined sex increase of 0.39 SD in Ct.Ar in the current study; effect sizes for pSSI mirrored Ct.Ar.

The positive relationship between peak anaerobic power and EC in this population is a novel finding from the current study and highlights a potential mechanism whereby peak muscle force being applied to bone rapidly might be beneficial to bone strength. In a highly active sample, peak forces applied regularly to cortical structures may stimulate bone resorption on endosteal surfaces to support the migration of bone mineral to outer surfaces of bone, increasing overall diameter and strength (34). This agrees with findings in healthy prepubertal children in which muscle power, assessed via counter movement jump, independently accounted for 15% of the variance in pSSI at the midtibia, whereas muscle force and MCSA only accounted for 4% and 14%, respectively (28). Thus, muscle power may be of higher importance with regard to bone strength in youths when compared with less ballistic measures of muscle force such as peak torque. Importantly, researchers have used different predictive equations from those used in the current study and have reported similarly promising results. Baptista et al. (25) predicted peak anaerobic power from vertical jump height using sex-specific equations in n = 114 prepubertal children (55% male, 8.6 ± 0.4 yr old). Peak anaerobic power predicted all DXA-derived bone status outcomes with reasonable sensitivity and specificity after adjusting for skeletal age, accounting for 74.3%–77.0% of the variance in regression models (all P < 0.001). However, this study did not include important covariates such as muscle mass when examining effects.

Although studies support the use of vertical jump as a predictor of bone status, not all have demonstrated significant associations. Two studies used raw vertical jump height (cm) as the predictor of bone status in healthy prepubertal and early-pubertal children (22,32) and found that it did not significantly predict bone status at any site (measured by DXA and pQCT), whereas MCSA predicted most sites after adjustment for stature and developmental covariates. Thus, it follows that when using vertical jump as a field-based predictor of bone status, one should apply a validated prediction equation to convert raw scores into peak anaerobic power estimates.

Limitations and future directions

Despite the comprehensive comparison among predictors of cortical bone status, our study is not without limitations. The narrow age and racial/ethnic demographic of this study, although a strength in increasing statistical power, precludes generalization to other populations, especially those at the extremes of muscular fitness. Future testing in diverse samples is required to confirm the predictive ability and operational safety of vertical jump testing, especially in clinical populations. With regard to the testing of muscle force capacity, a participant’s motivation can influence the results of maximal effort tests. However, we are confident that our data were minimally affected as standardized verbal motivation was given to all participants. Moreover, sensitivity analyses (data not shown) suggested no effect of including participant motivation as a continuous variable (1–10 scale; derived from verbal questions immediately following each test) in the regression model. Finally, bone strength in young adulthood depends on the material and structural properties of bone acquired throughout youth, especially during the pubertal period. Thus, it is recognized that cross-sectional research has limitations due to assumptions that the current physiological and behavioral measures are indicative of similar habits and trajectories throughout growth. To avoid this assumption, future research must assess relationships between muscle force capacity and cortical bone status using prospective or intervention research designs.


In summary, our data suggest that surrogates of muscle force capacity such as MCSA are strong independent predictors of cortical bone structure and estimated bending strength. However, both clinical measures and field estimates of peak torque and peak anaerobic power, respectively, are promising alternatives that explain similar, and in some cases greater, variance than MCSA. Importantly, peak anaerobic power predicted from vertical jump height emerged as the strongest alternative predictor of midtibia Ct.Ar and Ct.Th, whereas isokinetic knee extension peak torque prevailed in the prediction of midtibia PC and pSSI. Neither MCSA nor any muscle force capacity measure predicted Ct.vBMD. Future research should apply these findings to other populations, especially those where undergoing radiative measures of muscle size and mass might be contraindicated or in situations where expensive clinical laboratory equipment is not available.

Data collection was performed by S. H., C. M. S., M. V., and J. G. A. Review of literature and statistical analyses were performed by S. H. Manuscript preparation and revision were performed by S. H., R. D. L., M. D. S., and E. M. E. Preparation of the final document for submission was performed by S. H.

No sources of funding were used in preparation of this manuscript. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

The authors declare no conflict of interest. The results of the present study do not constitute endorsement by the American College of Sports Medicine.


1. Frost HM, Schonau E. The “muscle–bone unit” in children and adolescents: a 2000 overview. J Pediatr Endocrinol Metab. 2000;13(6):571–90.
2. Robling AG. Is bone’s response to mechanical signals dominated by muscle forces? Med Sci Sports Exerc. 2009;41(11):2044–9.
3. Brotto M, Bonewald L. Bone and muscle: interactions beyond mechanical. Bone. 2015;80:109–14.
4. Weaver CM, Gordon CM, Janz KF, et al. The National Osteoporosis Foundation’s Position Statement on peak bone mass development and lifestyle factors: a systematic review and implementation recommendations. Osteoporos Int. 2016;27(4):1281–386.
5. Tan VP, Macdonald HM, Kim S, et al. Influence of physical activity on bone strength in children and adolescents: a systematic review and narrative synthesis. J Bone Miner Res. 2014;29(10):2161–81.
6. Anliker E, Toigo M. Functional assessment of the muscle–bone unit in the lower leg. J Musculoskelet Neuronal Interact. 2012;12(2):46–55.
7. Linari M, Bottinelli R, Pellegrino MA, Reconditi M, Reggiani C, Lombardi V. The mechanism of the force response to stretch in human skinned muscle fibres with different myosin isoforms. J Physiol. 2004;554(Pt 2):335–52.
8. Damilakis J, Adams JE, Guglielmi G, Link TM. Radiation exposure in x-ray–based imaging techniques used in osteoporosis. Eur Radiol. 2010;20(11):2707–14.
9. Janz KF, Letuchy EM, Burns TL, Francis SL, Levy SM. Muscle power predicts adolescent bone strength: Iowa Bone Development Study. Med Sci Sports Exerc. 2015;47(10):2201–6.
10. Troiano RP. Large-scale applications of accelerometers: new frontiers and new questions. Med Sci Sports Exerc. 2007;39(9):1501.
11. Choi L, Liu Z, Matthews CE, Buchowski MS. Validation of accelerometer wear and nonwear time classification algorithm. Med Sci Sports Exerc. 2011;43(2):357–64.
12. Sasaki JE, John D, Freedson PS. Validation and comparison of ActiGraph activity monitors. J Sci Med Sport. 2011;14(5):411–6.
13. Bassey EJ, Short AH. A new method for measuring power output in a single leg extension: feasibility, reliability and validity. Eur J Appl Physiol Occup Physiol. 1990;60(5):385–90.
14. Blew RM, Lee VR, Farr JN, Schiferl DJ, Going SB. Standardizing evaluation of pQCT image quality in the presence of subject movement: qualitative versus quantitative assessment. Calcif Tissue Int. 2014;94(2):202–11.
15. Sayers SP, Harackiewicz DV, Harman EA, Frykman PN, Rosenstein MT. Cross-validation of three jump power equations. Med Sci Sports Exerc. 1999;31(4):572–7.
16. Wetzsteon RJ, Zemel BS, Shults J, Howard KM, Kibe LW, Leonard MB. Mechanical loads and cortical bone geometry in healthy children and young adults. Bone. 2011;48(5):1103–8.
17. Marquardt DW. Generalized inverses, ridge regression, biased linear estimation, and nonlinear estimation. Technometrics. 1970;12(3):591–612.
18. Rubin CT, Lanyon LE. Kappa Delta Award Paper. Osteoregulatory nature of mechanical stimuli: function as a determinant for adaptive remodeling in bone. J Orthop Res. 1987;5(2):300–10.
19. Nordström P, Thorsen K, Nordström G, Bergström E, Lorentzon R. Bone mass, muscle strength, and different body constitutional parameters in adolescent boys with a low or moderate exercise level. Bone. 1995;17(4):351–6.
20. Heinonen A, McKay HA, Whittall KP, Forster BB, Khan KM. Muscle cross-sectional area is associated with specific site of bone in prepubertal girls: a quantitative magnetic resonance imaging study. Bone. 2001;29(4):388–92.
21. Daly RM, Saxon L, Turner CH, Robling AG, Bass SL. The relationship between muscle size and bone geometry during growth and in response to exercise. Bone. 2004;34(2):281–7.
22. Macdonald H, Kontulainen S, Petit M, Janssen P, McKay H. Bone strength and its determinants in pre- and early pubertal boys and girls. Bone. 2006;39(3):598–608.
23. Cousins JM, Petit MA, Paudel ML, et al. Muscle power and physical activity are associated with bone strength in older men: The Osteoporotic Fractures in Men Study. Bone. 2010;47(2):205–11.
24. Lorbergs AL, Farthing JP, Baxter-Jones AD, Kontulainen SA. Forearm muscle size, strength, force, and power in relation to pQCT-derived bone strength at the radius in adults. Appl Physiol Nutr Metab. 2011;36(5):618–25.
25. Baptista F, Mil-Homens P, Carita AI, Janz K, Sardinha LB. Peak vertical jump power as a marker of bone health in children. Int J Sports Med. 2016;37(8):653–8.
26. Arden NK, Spector TD. Genetic influences on muscle strength, lean body mass, and bone mineral density: a twin study. J Bone Miner Res. 1997;12(12):2076–81.
27. Yingling VR, Webb S, Inouye C, O J, Sherwood JJ. Muscle power predicts bone strength in Division II athletes. J Strength Cond Res. 2017; [Epub ahead of print]. doi:10.1519/JSC.0000000000002222.
28. Binkley TL, Specker BL. Muscle–bone relationships in the lower leg of healthy pre-pubertal females and males. J Musculoskelet Neuronal Interact. 2008;8(3):239–43.
29. Witzke KA, Snow CM. Lean body mass and leg power best predict bone mineral density in adolescent girls. Med Sci Sports Exerc. 1999;31(11):1558–63.
30. Hardcastle SA, Gregson CL, Rittweger J, Crabtree N, Ward K, Tobias JH. Jump power and force have distinct associations with cortical bone parameters: findings from a population enriched by individuals with high bone mass. J Clin Endocrinol Metab. 2014;99(1):266–75.
31. Munukka M, Waller B, Multanen J, et al. Relationship between lower limb neuromuscular performance and bone strength in postmenopausal women with mild knee osteoarthritis. J Musculoskelet Neuronal Interact. 2014;14(4):418–24.
32. Daly RM, Stenevi-Lundgren S, Linden C, Karlsson MK. Muscle determinants of bone mass, geometry and strength in prepubertal girls. Med Sci Sports Exerc. 2008;40(6):1135–41.
33. Dowthwaite JN, Rosenbaum PF, Sames CA, Scerpella TA. Muscle function, dynamic loading, and femoral neck structure in pediatric females. Med Sci Sports Exerc. 2014;46(5):911–9.
34. Bouxsein ML, Karasik D. Bone geometry and skeletal fragility. Curr Osteoporos Rep. 2006;4(2):49–56.


Copyright © 2018 by the American College of Sports Medicine