Despite intensive research efforts over the last decade, female athletes continue to experience non-contact anterior cruciate ligament (ACL) injuries at a greater frequency than their male counterparts in the comparable sports of basketball, soccer and volleyball (Agel et al., 2005). Perhaps one reason female ACL injuries have persisted is that the bulk of research over the last decade has primarily focused on identifying sex differences during functional activities (Boden et al., 2000; Ford et al., 2003; Lephart et al., 2002; Malnizak et al., 2001; Myer et al., 2005; Padua et al., 2005; Shultz et al., 2001; Uhorchak et al., 2003; Wojtys et al., 1998). Little work, however, has examined the effectiveness of intervention programs (Hewett et al., 1999; Heidt et al., 2000; Mandelbaum et al., 2005; Myklebust et al., 2003; Petersen et al., 2005; Soderman et al., 2000).

Intervention programs that have been documented to reduce injury (Hewett et al., 1999; Mandelbaum et al., 2005; Myklebust et al., 2003; Petersen et al., 2005) have incorporated some level of: 1) instruction related to technique, 2) plyometric, 3) balance or 4) strength training exercise. While these programs have demonstrated an acute reduction in injury incidence, it remains unclear whether injury rates remain suppressed or return to pre-intervention levels over the long term. Arguably, after cessation of the intervention, without continued training and conditioning, key neuromuscular physiologic characteristics (i.e. strength, muscle response time), which ultimately may have been responsible for the findings, will return to or near baseline levels (Hakkinen et al., 2000). The ramifications of this ‘detraining’ could potentially negate the protective effect of the program. To effectively reduce injury rates over the long term, the effects of an intervention should not only be easily sustainable or permanent but also focused on modifiable risk factors.

A modifiable risk factor that has been identified, though received little attention, is that of excessive foot pronation (Hertel et al., 2004). Biomechanically, increased foot pronation has been associated with increased medial tibial rotation (Coplan, 1989). Given that medial rotation of the tibia on the femur increases ACL strain (Markolf et al., 1990), it is reasonable to assume that increased medial tibial rotation further increases ACL strain thereby heightening injury risk. Supporting this logic, several investigators have observed increased medial tibial rotation at or near the time of injury (McNair and Marshall, 1990; Olsen et al., 2004). Others have reported an increased incidence of non-contact ACL injury in athletes with pes planus (Beckett et al., 1992; Hertel et al., 2004; Loudon et al., 1996; Woodford-Rogers et al., 1994).

Reducing foot pronation and coupled medial tibial rotation would, therefore, seem to be desired. In fact, it is possible to reduce foot pronation and medial tibial rotation with a foot orthotic (McPoil and Cornwall, 2000; Mundermann et al., 2003). Recent work shows that a medially posted foot orthotic alters lower extremity kinematics during running (Mundermann et al., 2003) and walking (McPoil and Cornwall, 2000). Specifically, a medially posted foot orthotic demonstrated a decrease in both the rate and degree of medial tibial rotation during walking (McPoil and Cornwall, 2000). Furthermore, given that a foot orthotic provides a mechanical barrier to pronation, the effect of a foot orthotic on lower extremity kinematic patterns should be permanent since the intervention effect is maintained. Whether a medially posted foot orthotic is capable of altering lower extremity kinematics during activities that are more representative of a mechanism that simulates non-contact ACL injury (i.e. landing) however is unknown. Therefore, the purpose of our study was to examine the effect of a medially posted foot orthotic on transverse plane lower extremity kinematics in female athletes with pes planus during landing. We hypothesized that the foot orthotic would decrease both the rate and excursion of tibial rotation during landing.

An a- priori power analysis using results from a pilot study (Houglum and Carcia, 2004) conducted in our laboratory was performed. Results from this analysis suggested that inclusion of 20 subjects would yield a power of greater than 80%. To be included in the study, subjects had to be a university athlete participating in the sports of basketball, soccer or volleyball, have a navicular drop score of at least 8 millimeters (mm) and be able to perform simple landing tasks without pain. Subjects were excluded if they had a history of prior surgery on the dominant lower extremity or if they reported acute injury to the dominant lower extremity within the last six months. An acute injury was defined as one that necessitated the use of an assistive device for greater than a day. Twenty females (height = 1.69 ± 0.10 m; weight = 69.7 ± 9.7 kg; age = 20.1 ± 1. 0 yrs) volunteered to participate. Before participation, all subjects read and signed an informed consent form which was approved by the University’s Institutional Review Board.

All procedures were completed in the Kristen McMaster Motion Analysis Laboratory in a single forty-five minute session. After recording height and weight, lower extremity dominance was identified using a self-selection procedure. Each subject was asked to perform a single leg landing from a 20cm high wooden box. The lower extremity the subject chose to land on for two out of three trials was defined as the dominant lower extremity. All subsequent testing was performed using this extremity. Navicular drop was then measured using a modified Brody technique (Brody, 1982). Specifically, the subject stood in single leg stance adjacent to an examination table. She placed a hand lightly on the table top to facilitate a steady, static posture. Next, the examiner marked the skin over the most prominent aspect of the navicular with a pen. Once instructed, the subject supinated and pronated their foot while the examiner palpated the medial and lateral aspects of the talar head between his thumb and index finger. When the medial and lateral aspects of the talar head were congruent, the subject was asked to actively hold this position. The distance from the navicular mark to the floor was then measured in millimeters with a clear plastic ruler. The subject was then instructed to relax their foot and the distance between the pen mark and the floor was once again measured. Navicular drop was defined as the difference between the two measures. This process was repeated three times and an average of the scores was used for data analysis. All navicular drop measures were performed by the primary investigator (CRC) who had established a day to day intra-rater reliability (ICC_2,1) of 0.90 ± 0.5 mm before the study’s commencement.

Subjects then performed two to three practice trials of the selected tasks. A single-leg forward hop was performed from a distance equal to 45% of the subject’s height onto the center of a force plate (Bertec; Columbus, OH) (Figure 1). For the second task, subjects performed a drop landing onto the center of the force plate from a 20cm-high wooden box positioned 11cm from the edge of the force plate. The hop and land activities were chosen as they are reasonable laboratory simulations of non-contact ACL injury mechanisms. Additionally, similar methodology has been used by other investigators (Lephart et al., 2002) thereby facilitating the comparison of our findings to published work. To minimize the influence of upper extremity position on lower extremity kinematics, the subject’s hands were placed on their iliac crests for all trials. Neither the position of the subject’s head nor direction in which the subject was looking during the tasks was standardized. Once the subject and principal investigator were comfortable with the subject’s performance during the practice trials, the subject was prepared for kinematic motion analysis. Three electromagnetic sensors (Ascension Technology; Burlington, VT) coupled with The Motion Monitor Software (Innovative Sports Training; Chicago, IL) measured three-dimensional lower extremity kinematics while the force plate identified ground contact. Prior work has established the reliability of the motion analysis system (Bull et al., 1998; McQuade et al., 2002; Milne et al., 1996). The motion analysis system and force plate were synchronized and collected data at 100 and 1000 Hertz respectively. Sensors were placed over the L5-S1 junction, mid-lateral thigh and laterally just distal to the fibular head using prefabricated neoprene cuffs. To prevent undesired movement of the sensors, each neoprene cuff was further secured by circumferentially wrapping three-inch pre-wrap (Mueller^®; Praire Du Sac, WI) followed by 1½ inch athletic tape (Johnson & Johnson; New Brunswick, NJ) over each cuff. The proximal and distal aspects of the segments were then digitized using standard procedures with the subject standing in the anatomical position.

Using a repeated measures counterbalanced design, subjects completed three trials of the hopping and landing task with and without an orthotic device (Interpod^®; St Kilda, Australia) wearing standardized footwear (New Balance; Boston, MA). Approximately 15 seconds elapsed between trials. A repeated measures design has been recommended by prior investigators to account for the between subject variability associated with transverse tibial rotation (McPoil and Cornwall, 2000). Additionally, the day to day reliability of tibial rotation excursion (ICC = 0.82) and time to peak angle (ICC = 0.97) during a landing activity have been reported (Stevens and Schmitz, 2003) using identical hardware and software. The prefabricated foot orthotic was rigid and manufactured with an extrinsic 6° rear-foot medial (varus) post (Figure 2). Prefabricated orthoses are less expensive than custom orthoses, are readily available over the counter and have been shown to be just as effective as other types of orthoses for controlling tibial rotation during functional tasks (Brown et al., 1995; McPoil and Cornwall, 2000). Furthermore, a rear-foot varus post of 6° is not only a common clinical prescription, but with pilot testing (Houglum and Carcia, 2004) it was capable of limiting medial tibial rotation during forward hopping in a convenience sample of female subjects. The selected sneaker was a cross-trainer that had neither a supination nor pronation bias.

Dependent variables included initial contact angle, peak angle, excursion and time-to-peak angle for both the tibia and femur. Excursion was defined as the difference between peak and initial contact joint angles for each segment. Time-to-peak angle was the time in milliseconds it took for the respective segment to rotate from initial contact to peak angle. Initial contact was defined as the time at which the vertical ground reaction force exceeded 50 Newtons.

A mean of three trials for both tasks (hop and land) and each condition (with and without orthotic) was used for data analysis. A separate paired t-test was used to determine the effect of the orthotic on the dependent variables for the hop and land tasks. Alpha levels for all analyses were set a-priori at p < 0.05.

Descriptive statistics by sport for navicular drop are provided in Table 1. Means and standard deviations for kinematic data for each task with and without the orthotic device are detailed in Tables 2 and 3">3. Composite tracings of each task and condition are represented in Figures 3, 4">4, 5">5 and 6">6.

The primary finding of our study was that the prefabricated foot orthotic did not alter lower extremity transverse plane kinematics in female athletes with large navicular drop scores during either the hop or the land task. Several reasons, including the limitation of a static measure to predict dynamic movement, large inter-subject variability, as well as the characteristics of the orthotic likely account for the findings.

Creating a link between static, clinical anatomical measures and an increased injury incidence has been the thrust of several studies (Beckett et al., 1992; Hertel et al., 2004; Loudon et al., 1996; Woodford-Rogers et al., 1994). A link between certain static measures (single or multiple), which are different from established norms, and increased injury risk is an appealing model to both investigators and clinicians. A model of this nature provides a means for clinicians to assess an athlete’s injury risk via pre-participation screening. Once an athlete is identified as having a greater injury risk, specific intervention programs may then be prescribed and implemented. Ultimately, it is the intent of this process to reduce injury frequency. Creating a link between static anatomical measures and injury incidence, however, neglects the influence of the neuromuscular system. As clinicians, we are well aware that a positive static test which suggests mechanical instability does not necessarily correlate with functional instability. Exemplifying this, Eastlack et al demonstrated that a sub-group of ACL deficient subjects with increased anterior tibial translation was not related to functional knee instability (Eastlack et al., 1999). The authors note that with sufficient neuromuscular control despite a deficiency in the static stabilizers, it is possible to keep the joint stable during functional activity. Similarly, it is plausible that subjects with large navicular drop scores and adequate neuromuscular control do not exhibit increased foot pronation and tibial rotation. Without excessive motion, a foot orthotic which is designed to block excessive pronation and subsequent segment rotation is not warranted and therefore unlikely to produce an effect. As we did not measure dynamic foot pronation or any aspect of neuromuscular control, the validity of this hypothesis is unknown. Our pilot work (Houglum and Carcia, 2004), however, indirectly lends credence to this hypothesis. Tibial rotation excursion during an identical hopping task was 3.5° greater in the female subjects of our pilot study when compared to the athletes in this study. Given that the activity level of the females in the pilot study was less than that of the athletes in this study, it is likely they did not possess the level of neuromuscular control compared to the present competitive, athletic population. These findings suggest that there may be an interaction between neuromuscular control and lower extremity kinematic patterns. Further study is necessary to clarify the influence of neuromuscular control on foot pronation and lower extremity rotation in subjects with large navicular drop scores.

As demonstrated by the large standard deviations, there was substantial variability observed among the participants for all dependent measures during both tasks.

This variability is similar in magnitude when compared to the transverse plane values (both excursion and time to peak) reported by Lephart et al., 2002 using similar procedures in a comparable population. Likewise, the standard deviations were within a degree of the tibial transverse plane values reported by McPoil and Cornwall, 2000 during gait.

The variability may be partially explained by the fact that human movement is diverse. Exemplifying this, during the hop task after ground contact, nine subjects demonstrated a medial rotation kinematic pattern of the tibia while the remaining 11 subjects demonstrated a lateral rotation kinematic pattern. Unlike tibial kinematics and despite large standard deviations, femoral movement occurred in a similar fashion among all subjects. In summary, our work reinforces substantial variability is present in transverse plane lower extremity kinematics during functional activities. This variability makes differences between test conditions challenging to identify.

While the variability in the data contributed to an inability to identify significant differences between test conditions, mean differences, regardless of the standard deviations, were small (<1° for excursion and <1 ms for time to peak). While no studies were identified that compared lower extremity kinematics between no orthotic and orthotic conditions during landing tasks, evidence regarding the ability of an orthotic device to influence lower extremity kinematics has been reported during walking (Nester et al., 2001; McPoil & Cornwall, 2000) and running (Mundermann et al., 2003). The lack of differences between conditions in this study may have been influenced by the physical characteristics of the foot orthotic. Baitch et al., 1991 reported for every 5° of rear-foot posting only a 1° change in static calcaneal position is evident. These findings imply that it is likely that even a greater amount of posting would have been necessary to alter the kinematics of the more proximal segments. Our pilot work (Houglum and Carcia, 2004), however, indicated the chosen orthotic did have the capability of limiting tibial rotation during the hopping task. As previously mentioned, the root of this phenomenon may be related to neuromuscular control. Interestingly, without solicitation, most athletes reported that they felt more comfortable and ‘controlled’ when using the orthotic during the tasks. It is possible that the orthotic device altered some other variable (i.e. tibiofemoral moments, time to stabilization) not quantified in this study. Beyond the size of the extrinsic posts, research suggests that molding methods are more dominant on altering lower extremity kinematics and kinetics when compared to traditional posting methods alone (Mundermann et al., 2003). Clearly, additional research is warranted to explore the effect of different types of orthotic devices on lower extremity kinematics during landing tasks.

Our study, like all research studies, had several limitations that should be acknowledged. Subjects were tested within minutes after the orthotic device was placed inside the sneaker. Whether a greater period of time between introducing the orthotic device into the sneaker and testing would have altered the manner in which the subject executed the tasks is unknown. Further, we did not formally track the number of subjects who wore foot orthoses on a regular basis for daily or sport activities. Information in this regard would have provided insight as to whether subjects had experienced symptoms at some juncture secondary to their foot posture and function. While the mean navicular drop of our healthy subjects (9.6mm) was greater than the navicular drop scores of the ACL injured subjects reported by Hertel et al., 2004 (8.4mm) and Woodford-Rogers et al., 1994 (5.0mm) it was less than the values reported by Beckett et al., 1992 (12.7mm). Additionally, our navicular drop scores were less than the ‘abnormal threshold’ suggested by Mueller et al., 1993 (10mm) and Brody, 1982 (15mm). If anything, we suspect methodological differences may have underestimated the navicular drop scores of our subjects. Specifically, the first measure of navicular height using the technique suggested by Brody, 1982 is quantified with the subject seated. Our first (and second) measure of navicular height was taken in a weight bearing posture. It is likely that body weight alone reduced the height of the navicular thereby making the difference score smaller than what we would have ascertained had the first measure been taken in sitting. Research is on-going to address this issue. Lastly and as previously mentioned, we did not measure dynamic foot pronation and, therefore, are uncertain as to whether our subjects actually did display excessive pronation during the land and hop tasks. Future studies should investigate the relationship between dynamic foot pronation and lower extremity transverse plane rotation with and without an orthotic device during landing tasks in female athletes.

A prefabricated , rigid foot orthosis with a six-degree extrinsic varus post does not alter the quantity or rate of transverse plane lower extremity kinematics in female athletes with navicular drop scores greater than or equal to eight millimeters when performing a forward hop or landing task in the laboratory. The findings suggest that other factors (i.e. neuromuscular control) may be responsible for controlling tibial rotation in female athletes with large navicular drop scores during landing activities. Future study should assess various aspects of this hypothesis to determine its validity. Additionally, before firm conclusions regarding the effectiveness of orthotic intervention on transverse plane kinematics in this population can be drawn, this line of study should continue to be pursued. An orthotic device which has the capability of favorably altering lower extremity movement and or neuromuscular recruitment patterns has great promise for reducing the incidence of non-contact ACL injury in female athletes. This type of intervention would be cost-effective and the impact should be immediate. Future work should examine the effectiveness of different types of orthoses (i.e inverted), as well as consider other dependent measures such as tibiofemoral kinetics and neuromuscular activity.