The Role of Foot Pronation in the Development of Femoral and Tibial Stress Fractures: A Prospective Biomechanical Study

ORIGINAL RESEARCH The Role of Foot Pronation in the Development of Femoral and Tibial Stress Fractures: A Prospective Biomechanical Study Iftach Hetsroni, MD,* Aharon Finestone, MD, Charles Milgrom, MD, David Ben-Sira, PhD, Meir Nyska, MD,* Gideon Mann, MD,* Sivan Almosnino, and Moshe Ayalon, PhD Objective: To examine whether dynamic parameters of foot pronation are risk factors for the development of stress fractures of the femur and tibia. Design: Observational prospective study. Setting: Infantry basic training course. Participants: 473 recruits evaluated for stress fractures of the femur and tibia every 2 weeks during 14 weeks of infantry basic training. The final analysis included 405 recruits. Assessment of Risk Factors: Two weeks before commencement of training, the recruits were evaluated during treadmill walking for their subtalar joint kinematics. Five independent variables were measured bilaterally: maximal pronation angle during the stance, pronation range of motion, time from heel strike to maximum pronation, pronation mean angular velocity, and time to maximum pronation as a percent of the total stance time. Main Outcome Measurements: Stress fractures of the femur and tibia. These were considered positive only when proven by imaging. Results: Ten percent of the participants were diagnosed with stress fractures of the femur and tibia. Recruits with longer duration of foot pronation had reduced odds ratio to develop this injury. Conclusions and Clinical Relevance: Longer duration of foot pronation may have a protective effect from stress fractures of the femur and tibia. This finding may promote the understanding of stress fracture pathomechanism, assist in the identification of subjects with increased risk who need augmented monitoring throughout training, and assist in future planning of impact reducing aids. Submitted for publication May 1, 2007; accepted October 12, 2007. From the *Orthopaedic Department, Meir University Hospital, Sapir Medical Center, Kfar Saba, Israel and the Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Foot and Ankle Unit, Assaf HaRofeh Medical Center, Zerrifin, Israel; Hadassah Medical Organization and Hebrew University Medical School, Jerusalem, Israel; and Biomechanics Laboratory, The Zinman College of Physical Education & Sport Sciences, Wingate Institute, Netanya, Israel. The authors state that they have no financial interest in the products mentioned within this article. Reprints: Iftach Hetsroni, MD, Meir Hospital, Sapir Medical Center, Tsharnichovski Street 59, Kfar Saba 44281, Israel (e-mail: hetsrony@ netvision.net.il). Copyright Ó 2008 by Lippincott Williams & Wilkins Key Words: subtalar joint, stance phase, stress fracture, tibial rotation (Clin J Sport Med 2008;18:18 23) INTRODUCTION The bony structures of the lower extremity endure ground impact forces as well as forces applied by the muscles during the initial phase of foot support. In a number of retrospective studies, increased ground reaction impact forces and loading rates have been related to the occurrence of stress fractures of the lower extremity. 1 4 Foot pronation, occurring primarily at the subtalar joint, has been hypothesized to play a role in the development of this injury as well. 5 9 As limited pronation was demonstrated to increase the magnitude of the impact loading experienced during locomotion, 10,11 it has been assumed that pronation through the subtalar joint contributes to the attenuation of ground reaction impact forces encountered by the bones of the lower limb. Pronation of the foot is also responsible for the unlocking of the transverse tarsal joint, increasing foot flexibility and again contributing to shock absorption. 12 Pronation through the subtalar joint is also coupled with internal rotation of the tibia during the first half of stance. 13,14 Thus, abnormal pronation may result in abnormal internal tibial torsion, affecting stresses exerted along the entire lower limb. 7,14,15 Despite these apparent rational assumptions, the literature is contradictory regarding the role of foot pronation in stress fracture development. Some cross-sectional retrospective studies 5,7,9 have reported that injured subjects who suffered stress reactions exhibited more pronation and had greater pronation velocity than noninjured subjects. Conversely, other studies 6,8 have correlated low arch feet, which are considered susceptible to a large degree of pronation, 16 with a protective effect from stress reactions and stress fractures. Some studies failed to support any association between pronation velocity 3 and subtalar range of motion 17,18 and the risk of developing stress fractures. Because of the inconclusive evidence, the justification for research and practice, based on the association between characteristics of foot pronation and the risk of developing stress fractures, is unresolved. The present study was designed to prospectively assess the risk for stress fractures of the tibia and the femur, associated with kinematical characteristics of foot pronation. It has been hypothesized that maximal pronation angle, pronation angular 18 Clin J Sport Med Volume 18, Number 1, January 2008

Clin J Sport Med Volume 18, Number 1, January 2008 Foot Pronation range of motion, time from heel strike to maximum pronation, mean pronation angular velocity, and relative time to maximum pronation with respect to total stance time, are risk factors for the development of stress fractures in these bones and that identified risk factors will be manifested bilaterally. METHODS This study was IRB approved, and all subjects signed informed consent. It included 473 infantry male recruits who were followed during a 4-month basic training. All subjects wore the same standard infantry boots. Throughout basic training, the subjects were examined by a single orthopedic surgeon for the presence of lower limb injuries every 2 weeks. Stress fractures were diagnosed and managed according to the IDF Stress Fracture Protocol. 19 Only tibial and femoral stress fractures confirmed by imaging provided the outcome measure. Two weeks before the commencement of the 4-month basic training, the recruits were examined clinically and biomechanically. The clinical examination confirmed that none had musculoskeletal abnormality, which could preclude him from participating in an infantry military service. For the biomechanical evaluation, the participants were filmed with a digital video camera at 60 Hz from a rear view, walking barefoot on a treadmill at 5 km/hr. This velocity was chosen because it represents the mean velocity employed during long-distance marching, which has previously been shown to be a major external causative factor in the development of stress fractures. 20 Four reflective markers were placed on the posterior aspect of the leg and foot by a single investigator (Figure 1), according to a protocol previously described. 21 Two inferior markers, representing the rearfoot segment, were placed on the bisector line of the calcaneus distal to the subtalar joint center of rotation: the distal marker at the tuber calcani and the other marker 1 cm distal to the joint center of rotation. Two superior markers representing the leg segment were thereafter placed on the bisector line of the posterior shank and Achilles, 2 and 8 cm proximal to the subtalar joint center of rotation. The angle formed between the rearfoot and leg segments represented the subtalar joint position (ie, the foot pronation angle). Using the computerized Ariel Performance Analysis System (Ariel Dynamics Inc., Trabuco Canyon, CA), 5 dynamic parameters of foot pronation were bilaterally measured during the stance phase: maximal pronation angle (degrees), subtalar pronation range of motion from heel strike to maximum pronation (degrees), time from heel strike to maximum pronation (seconds), pronation mean angular velocity (degree/seconds), and time to maximum pronation as percent of the total stance time (%). For each subject, the mean value of each parameter during 4 consecutive walking cycles was selected for analysis. Right and left extremities were evaluated separately in order to crossvalidate possible risk factors. Statistical Analyses Subjects were divided into quartiles to assess the risk of stress fracture associated with each of the 5 pronation parameters. The first group (Q1) was defined as the lower quartile (ie, all subjects with the lower 25% values of the measured FIGURE 1. Reflective markers placement. parameter). The middle-group (IQR) included subjects with values within the interquartile range, and the third group (Q4) included those with values in the upper quartile. We used odds ratios (OR) to assess the association between each of the 5 parameters and the risk to develop stress fracture. An increased risk for injury, or conversely a protective effect from injury, was expected to be manifested in a significant OR of injury incidence between an extreme quartile and the middlegroup (ie, Q4 versus IQR), or between the 2 extreme quartiles (ie, Q4 versus Q1). RESULTS The final analysis included the data of 405 subjects after excluding recruits who dropped out from the training and those who were diagnosed with other lower extremity injuries. This ascertained that risk factors for stress fractures were assessed in relation to healthy status. Of the 405 subjects, 42 (10%) subjects sustained 71 diagnosed stress fractures of the femur and tibia. Thirty-six fractures were related to the femur and 35 to the tibia. Twelve femur and 9 tibia stress fractures were bilateral, and 9 subjects were diagnosed with stress fractures in both the femur and the tibia. Altogether there were 20 subjects with single, 16 subjects with dual, 5 subjects with triple, and 1 subject with quadruple stress fractures. Injured and noninjured subjects did not differ significantly in q 2008 Lippincott Williams & Wilkins 19

Hetsroni et al Clin J Sport Med Volume 18, Number 1, January 2008 body mass index (BMI; means = 21.8 6 3.5 and 22.2 6 2.7, respectively, P = 0.52). Descriptive statistics of the pronation parameters are presented in Table 1. The 25th and 75th percentiles define the cutoff values between Q1 and IQR and between IQR and Q4, respectively. Median values demonstrate the difference in central tendencies among the 3 subgroups in distribution of each variable. The descriptive statistics indicate that the bilateral characteristics of the 3 groups are quite similar. As illustrated in Table 2, the classification of subjects bilaterally to the same lower or upper quartiles was only fractional, ranging from 39% to 50% in Q1 and 38% to 59% in Q4. Nevertheless, most subjects that were classified to another group on the contralateral limb were classified into the IQR middle group as evident from the low proportion of cases who were bilaterally classified on opposite tails of the distribution (5 to 13%). Intraclass correlation coefficients for these parameters over 4 cycles ranged between 0.74 and 0.96, indicating acceptable reliability over repetitive cycles. 22 The incidence of stress fractures in the 3 groups is presented as a percentage in Table 3. The ORs are presented in Table 4. None of the ORs were significant for either pronation angle or pronation range of motion. However, longer pronation time (ie, Q4) was associated with reduced risk for stress fracture in both lower extremities. Although the OR of 0.47 between Q4 and Q1 on the left side, indicating risk reduction, did not reach statistical significance because the upper limit at 95% confidence interval exceeded the value of 1, the same trend of risk reduction for stress fracture was found in the other ORs of pronation time, and its size indicates a practically significant reduction in risk. These observations were supported by the ORs for the relative pronation time as a percent of the stance time (Table 4). The ORs to develop stress fracture as well as the 95% confidence interval limits were all below 1.0 and statistically significant, implying that patients with longer pronation time as percent of the total stance time had lower ORs to develop stress fracture. As for pronation velocity on the left foot, significantly higher OR to develop stress fracture was demonstrated in Q4 compared to Q1 and borderline significant OR in Q4 compared to the IQR groups. However, these trends were not supported by the data of the right foot. DISCUSSION Previous studies that examined the relation between foot pronation and the development of stress reactions and stress fractures have usually been either cross-sectional or composed of small sample groups, without control, or evaluated only 1 or 2 parameters of the subtalar joint motion. 3,5 9,16 18 The conclusions they provided were contradictory. The present study is prospective and was based on a large cohort of healthy subjects with homogenuous characteristics (ie, sex, age group, BMI, and daily timetable) exposed to a structured controlled training program that involved a high level of mechanical stresses to the skeletal system. An incidence rate of 10% of stress fractures implies that a significant proportion of the healthy population with seemingly normal orthopedic baseline characteristics is at risk of developing stress fractures if exposed to intensive training routines. Identification of risk factors in such a design provides valuable information regarding prediction of stress fractures and in designing preventive measures intended for risk reduction. Although previously reported retrospective crosssectional designs provide evidence for differences in pronation characteristics among subjects affected by stress reactions and those who were not affected, 5,7,9 such differences cannot be considered predictive risk factors because they may be consequent to the injury itself. The hypotheses that maximal pronation angle, pronation range of motion, or mean pronation velocity are risk factors for tibial or femoral stress fractures could not be supported by our results. They concur with findings of others who failed to demonstrate any correlation between parameters of foot pronation such as subtalar joint range of motion and pronation TABLE 1. Descriptive Statistics of the Pronation Parameters Pronation Parameter Side Mean SD P 25 P 75 Q1 IQR Q4 Medians Maximum pronation angle (degree) L 6.9 4.2 4.6 9.6 2.0 7.0 11.3 R 8.3 4.4 5.5 10.9 3.7 7.9 13.7 Pronation range of motion (degree) L 7.8 2.6 6.0 9.1 5.1 7.4 10.8 R 7.8 2.5 6.0 9.2 5.3 7.4 10.7 Time to maximum pronation (sec) L 0.17 0.06 0.13 0.20 0.12 0.16 0.25 R 0.17 0.06 0.13 0.20 0.12 0.16 0.25 Pronation velocity (deg/sec) L 49.7 21.6 33.4 62.0 26.8 46.1 75.8 R 49.4 20.1 33.5 62.7 26.2 47.0 73.8 Time to maximum pronation/stance time (%) L 29.5 10.2 21.8 34.4 19.3 26.9 42.9 R 29.8 10.4 22.2 34.1 20.2 27.6 41.9 Maximum pronation angle, angle at the peak of pronation during the stance phase; Pronation range of motion, the change in the measured angle from heel strike to maximal pronation angle; Time to maximum pronation, the time from heel strike to maximal pronation angle; Pronation velocity, the average velocity from heel strike to maximal pronation angle, computed as pronation range of motion divided by time to maximum pronation; Time to maximum pronation/ stance time, the relative duration of the pronation phase from heel strike to maximum pronation, with respect to the overall duration of the support phase. Q1, Lower quartile; IQR, Interquartile range (ie, the 2 central quartiles); Q4, Higher quartile. 20 q 2008 Lippincott Williams & Wilkins

Clin J Sport Med Volume 18, Number 1, January 2008 Foot Pronation TABLE 2. Bilateral Distribution of Subjects Classification into Groups Variable Bilateral Q1 Bilateral Q4 Q1 R & Q4 L Q1 L & Q4 R Maximum pronation angle (degree) 43.6% 37.6% 8.9% 5.0% Pronation range of motion (degree) 46.5% 39.6% 12.9% 10.9% Time to maximum pronation (sec) 38.9% 59.2% 8.2% 5.6% Pronation velocity (deg/sec) 49.5% 46.5% 6.9% 10.9% Time to maximum pronation/stance time (%) 42.6% 56.4% 7.9% 6.9% Bilateral Q1, % of subjects classified as Q1 on both sides; Bilateral Q4, % of subjects classified as Q4 on both sides; Q1 R and Q4 L, % of subjects classified as Q1 on the right and Q4 on the left; Q1 L and Q4 R, % of subjects classified as Q1 on the left and Q4 on the right. velocity and the occurrence of stress reactions and stress fractures. 3,17,18 One mechanism by which impact force may be reduced is increased movement range of motion, which enables generation of lower average forces to accomplish a given amount of mechanical work. In this case, a reduced risk is expected to be associated with increase in maximal angle of pronation and pronation range of motion. On the other hand, increased maximum pronation angle may lead to excessive tension on the medial part of the subtalar joint, compression on the lateral aspect of the joint and torsion of the tibia, thus affecting the stress distribution along the long bones of the lower extremity. According to our results, neither maximum pronation angle nor pronation range of motion were significantly associated with stress fractures, suggesting that even if such mechanisms exist they do not transform into a protective effect. Furthermore, the trend of the ORs, being all greater than 1.0 (range, 1.10 to 1.97), is contradictory to the expected diminution in risk. Evidence from other prospective studies may elucidate this trend. While the spatial variables were rejected as potential independent risk factors for femoral and tibial stress fractures, both temporal parameters proved to provide protective effect with increase in both absolute and relative time of pronation. Subjects who were characterized by a longer time to maximum pronation, as well as time to maximum pronation as a percent of the total stance time (Q4), had lower ORs to develop a stress fracture than subjects in either the lower quartile or the IQR. The only OR that did not reach statistical significance was between Q4 and Q1 in time to maximum pronation, but its trend is consistent with all the other significant ORs and its magnitude (0.47) reveal a practically significant reduction in risk. Altogether, belonging to the group with the longer time to maximum pronation reduces the odds of experiencing a stress fracture to between 11% and 47% of the odds among those in the other 3 quartiles, typified by a shorter time to maximum pronation. Likewise, subjects in Q4 of relative time to maximum pronation (relative to total stance time) had significantly lower odds of experiencing a stress fracture in the order of 9% to 29%, compared to those with shorter relative pronation time (Q1 and IQR). These results may imply that temporal values that characterize the upper quartile (time to maximum pronation.0.25 seconds; relative pronation time.42%) have a major protective effect from stress fracture injuries. These cutoff values may be tentatively used for identification of individuals in high risk for the purpose of either implementing a more stringent monitoring or for the application of preventive treatments. The practical value of this information depends on our ability to modify the duration of foot pronation. Theoretically, this can be accomplished by specific training programs, modification of gait pattern, or modification of footwear attributes. However, before such alternatives are adopted into preventive or rehabilitative practices, they should be supported by appropriate evidence from appropriately designed clinical trials. The reduced risk by extended time of foot pronation can be straightforwardly explained by simple mechanical principles. Reduction in the body momentum during the stance phase depends on the integration of force and time. A longer time may require reduced magnitude of average forces and provide for more gradual attenuation of the momentum, inflicting less damage on the skeletal system. The considerable protective effect of relative pronation time signifies that gait pattern during foot support, particularly temporal distribution between deceleration and propulsion, have long-term consequences with regard to the risk of developing stress fractures, and significantly lower risk is associated with relative time of 42% of the support phase or longer. However, an exact mechanism by which extended duration of pronation is related to reduction in mechanical stresses is not deducible from the current data. Pronation velocity was not found to be an apparent risk factor for stress fractures. Yet, the significant OR between Q4 TABLE 3. Stress Fracture Incidence as Percentage of the Subgroups Variable Side Q1 IQR Q4 Maximum pronation angle (degree) L 11.0% 11.3% 14.9% R 12.0% 17.2% 14.9% Pronation range of motion (degree) L 12.9% 10.2% 15.3% R 11.4% 12.1% 13.1% Time to maximum pronation (sec) L 18.5% 13.7% 2.3% R 10.7% 16.6% 5.4% Pronation velocity (deg/sec) L 5.7% 11.4% 20.5% R 4.3% 17.1% 10.8% Time to maximum pronation/stance time (%) L 20.7% 12.9% 2.3% R 13.3% 16.0% 4.3% q 2008 Lippincott Williams & Wilkins 21

Hetsroni et al Clin J Sport Med Volume 18, Number 1, January 2008 TABLE 4. Odds Ratios (OR) Between Groups and 95% Confidence Interval for Stress Fractures Pronation Parameter Side Groups OR Maximum pronation angle (degree) Pronation range of motion (degree) Time to maximum pronation (sec) Pronation velocity (deg/sec) Time to maximum pronation/stance time (%) *Significant OR. Lower Limit Upper Limit L Q4 vs. Q1 1.38 0.65 2.92 Q4 vs. IQR 1.42 0.57 3.54 R Q4 vs. Q1 1.53 0.66 3.55 Q4 vs. IQR 1.97 0.92 4.20 L Q4 vs. Q1 1.21 0.51 2.89 Q4 vs. IQR 1.58 0.74 3.41 R Q4 vs. Q1 1.18 0.47 2.93 Q4 vs. IQR 1.10 0.50 2.40 L Q4 vs. Q1 0.11* 0.02 0.47 Q4 vs. IQR 0.15* 0.03 0.65 R Q4 vs. Q1 0.47 0.15 1.47 Q4 vs. IQR 0.29* 0.11 0.77 L Q4 vs. Q1 4.22* 1.48 12.05 Q4 vs. IQR 2.01 0.99 4.08 R Q4 vs. Q1 2.71 0.80 9.14 Q4 vs. IQR 0.59 0.27 1.31 L Q4 vs. Q1 0.09* 0.02 0.40 Q4 vs. IQR 0.16* 0.04 0.69 R Q4 vs. Q1 0.29* 0.09 0.95 Q4 vs. IQR 0.23* 0.08 0.69 and Q1 (OR = 4.22) and a borderline significant OR between Q4 and IQR (OR = 2.01) on the left foot are in line with the assumption that higher mean pronation velocity is a risk factor for stress fractures. However, these trends were not supported by the right foot data. Given that pronation velocity is derived from both spatial and temporal variables, it is possible that higher pronation velocity is not by itself a risk factor for stress fracture, and it merely reflects the combined effects of both a significant and a nonsignificant risk factor, masking any unique epidemiological or mechanical explanations. It seems that stride duration dominates the inverse trend for the increased risk associated with higher pronation velocities, and that the reason it is not as obvious as in the temporal variables is because of the indefinite association between stress fractures and pronation range of motion. Another possible explanation for the inconsistency in the pronation velocity results is that the inherent measurement error in mean pronation velocity, being calculated as the ratio between 2 measured variables (ie, range of motion and time of pronation) is larger than for each of the single variables, thus masking a possible stronger true relationship between pronation velocity and incidence of injury. We classified the cohort into quartiles according to the distinctive distribution of each of the 10 predictive variables to assess the independent predictive potential of each of the 5 parameters and to ascertain possible risk factors by crossvalidation with the results on the contralateral limb. The classification of subject to the same bilateral quartile of 39% to 50% in Q1 and 38% to 59% in Q4 is well above random expectation of 25%, but it still reflects only partial correspondence between the 2 limbs in the classification of subjects to quartile groups. The explanation for mismatches may be measurement error or a mere bilateral asymmetry. On the other hand, only a small fraction of the sample was classified into 2 opposite quartiles, attesting that extreme cases of asymmetry are relatively rare in our population. Despite the partial correspondence between the bilateral samples, both temporal variables exhibited similar bilateral trends as significant risk factors. This finding provides support for the validity of the temporal risk factors being independent of any incidental subclassification of the sample. In a normal population, the higher and lower quartiles contain values that do not necessarily deviate considerably from mean, with P 25 and P 75 located at 60.67 standard deviations from the mean. The fact that higher risks may be associated with moderate deviations of kinematical values from the population mean suggests that predisposition to injury does not necessarily require very extreme values. An interaction between relatively moderate deviations and other factors may trigger such predisposition. An interesting course of investigation might have been to assess the potential risk factors with respect to specific subclassifications of stress fracture (ie, tibial versus femoral, isolateral versus contra-lateral prediction, or single versus multiple fractures). We avoided such detailed analyses for several reasons. From a practical point of view, the debilitating effects of the different outcomes are similar, namely a long-term disablement of the injured subject. Furthermore, assessment of the associations between particular subgroups of stress fractures and potential risk factors is of limited practical value, especially when considering the frequency distribution of the 42 diagnosed subjects. From a biomechanical point of view, it is assumed that stress fractures are consequent of repetitive local overload of bone tissue at the lower extremity that in turn depends both on the multitude of forces acting along the tibial and femoral shafts as well as on the structure and local alignment of the lower extremity. As a result, the currently available diagnostic tools are not able to identify the exact mechanism of each unique episode of either tibial or femoral stress fracture. From a statistical point of view, the incidence of confirmed stress fractures was too low to grant sufficient statistical power in order to investigate the risk factor for particular subsets of stress fractures. Apparently, a larger sample would have resulted in better statistical power, thus providing more adequate evidence for analyses of hypotheses regarding subsets of stress fractures. The study was delimited to a male population only. Female recruits have been reported to sustain up to 4 times the incidence of stress fractures as compared to males 23,24 ; therefore, kinematic parameters of the lower limb may be differently associated with stress fractures in females as compared to males. The measurements of subtalar pronation were performed barefoot, as we assumed that any difference in foot pronation among the subjects while walking with their shoes during the training period would be related mostly to inherent foot characteristics, as their shoe design was similar. In addition, we contemplated that placing markers on the shoes would not reflect the true foot motion inside the shoes. This approach has some limitation, for there is uncertainty with regard to the extent to which shod walking reflects spatial and temporal 22 q 2008 Lippincott Williams & Wilkins

Clin J Sport Med Volume 18, Number 1, January 2008 Foot Pronation characteristics of pronation, as observed during barefoot walking, and to whether interindividual differences in adjusting to a new shoe may modify the locomotion pattern of the subtalar joint and thereby generate additional risk. Subtalar pronation is not the only potential risk factor for lower extremity stress fractures. Other factors, such as aerobic fitness, nutritional status, psychosocial factors and genetic factors, may prove as only a partial list of additional risk factors for stress fractures. Future research may address a wider range of potential risk factors, providing additional information to our understanding of the etiology of this complex multifactorial injury. In conclusion, prolonged pronation may play a role in reducing the risk to develop stress fractures of the femur and tibia during an extended period of training. Within the limitations of the current study, tentative values that define reduced risk are available for diagnostic purposes and for further development of risk reduction alternatives. Further studies are required to determine whether modification of temporal aspects of pronation by means of either footwear modification or specific training can assist in reducing the incidence of this type of injury. ACKNOWLEDGMENTS We thank Ganit Meyer for her important contribution to the conclusion of this study as a research assistant at the Biomechanics Laboratory, The Zinman College of Physical Education & Sport Sciences, Wingate Institute, Netanya, Israel. REFERENCES 1. Milner CE, Ferber R, Pollard CD, et al. Biomechanical factors associated with tibial stress fracture in female runners. Med Sci Sports Exerc. 2006; 38:323 328. 2. Grimston SK, Engsberg JR, Kloiber R, et al. Bone mass, external loads, and stress fracture in female runners. International Journal of Sports Biomechanics. 1991;7:293 302. 3. Hreljac A, Marshall RN, Hume PA. Evaluation of lower extremity overuse injury potential in runners. Med Sci Sports Exerc. 2000;32:1635 1641. 4. Zifchock RA, Davis I, Hamill J. Kinetic asymmetry in female runners with and without retrospective tibial stress fractures. J Biomech. 2006;39: 2792 2797. 5. DeLacerda FG. A study of anatomical factors involved in shin splints. J Orthop Sports Phys Ther. 1980;2:55 62. 6. Giladi M, Milgrom C, Stein M, et al. The low arch, a protective factor in stress fractures. A prospective study of 295 military recruits. Orthop Rev. 1985;14:709 712. 7. Messier SP, Pitala KA. Etiologic factors associated with selected running injuries. Med Sci Sports Exerc. 1988;20:501 505. 8. Simkin A, Leichter I, Giladi M, et al. Combined effect of foot arch structure and an orthotic device on stress fractures. Foot Ankle. 1989;10:25 29. 9. Viitasalo JT, Kvist M. Some biomechanical aspects of the foot and ankle in athletes with and without shin splints. Am J Sports Med. 1983;11: 125 130. 10. Freychat P, Belli A, Carret JP, et al. Relationship between rearfoot and forefoot orientation and ground reaction forces during running. Med Sci Sports Exerc. 1996;28:225 232. 11. Perry SD, Lafortune MA. Influences of inversion/eversion of the foot upon impact loading during locomotion. Clin Biomech. 1995;10: 253 257. 12. James SL, Bates BT, Osternig LR. Injuries to runners. Am J Sport Med. 1978;6:40 50. 13. Czerniecki JM. Foot and ankle biomechanics in walking and running. A review. Am J Phys Med Rehab. 1988;67:246 252. 14. Nigg BM, Cole GK, Nachbauer W. Effects of arch height of the foot on angular motion of the lower extremities in running. J Biomech. 1993;26: 909 916. 15. Gross MT. Lower quarter screening for skeletal malalignment suggestions for orthotics and shoewear. Foot Ankle Ther Res. 1995;21: 389 405. 16. Subotnik SI. The biomechanics of running. Sports Med. 1985;2: 144 153. 17. Kaufman KR, Brodine SK, Shaffer RA, et al. The effect of foot structure and range of motion on musculoskeletal overuse injuries. Am J Sports Med. 1999;27:585 593. 18. Winfield AC, Moore J, Bracker M, et al. Risk factors associated with stress reactions in female Marines. Mil Med. 1997;162:698 702. 19. Milgrom C, Finestone A, Shlamkovitch N, et al. Stress fracture treatment. Orthopaedics (Int Ed). 1995;3:363 367. 20. Jordaan G, Schwellnus MP. The incidence of overuse injuries in military recruits during basic training. Mil Med. 1994;159:421 426. 21. Mueller MJ, Norton BJ. Reliability of kinematic measurements of rearfoot motion. Phys Ther. 1992;72:731 737. 22. Shrout PE, Fleiss JL. Intraclass correlation: Uses in assessing rater reliability. Psychol Bull. 1979;86:420 428. 23. Hauret KG, Shippey DL, Knapik JJ. The physical training and rehabilitation program: duration and rehabilitation and final outcome of injuries in basic combat training. Mil Med. 2001;166:820 826. 24. Pester S, Smith PC. Stress fractures in the lower extremities of soldiers in basic training. Orthop Rev. 1992;21:297 303. q 2008 Lippincott Williams & Wilkins 23