Effect of shoe insert construction on foot and leg movement

Abstract

Purpose: 

The purpose of this study was to quantify changes in foot eversion and tibial rotation during running resulting from systematic changes of material composition of five shoe inserts of the same shape.

Methods: 

Tests were performed with 12 subjects. The inserts had a bilayer design using two different materials at the top and bottom of the insert. The functional kinematic variables examined in this study were the foot-leg in-eversion angle, β, and the leg-foot tibial rotation, ρ. Additionally, the subject characteristics of arch height, relative arch deformation, and active range of motion were quantified. The statistical analysis used was a two way repeated measures MANOVA (within trials and inserts).

Results: 

The average group changes resulting from the studied inserts in total shoe eversion, total foot eversion, and total internal tibial rotation were typically smaller than 1° when compared with the no-insert condition and were statistically not significant. The measured ranges of total foot eversion for all subjects were smallest for the softest and about twice as large for the hardest insert construction. Thus, the soft insert construction was more restrictive, forcing all feet into a similar movement pattern, whereas the harder combinations allowed for more individual variation of foot and leg movement and did not force the foot into a preset movement pattern. The individual results showed substantial differences between subjects and a trend: Subjects who generally showed a reduction of tibial rotation with all tested inserts typically had a flexible foot. However, subjects who generally showed an increase of tibial rotation typically had a stiff foot.

Conclusions: 

The results of this study suggest that subject specific factors such as static, dynamic, and neuro-physiological characteristics of foot and leg are important to match specific feet and shoe inserts optimally.

Athletes participating in running are susceptible to injuries in the lower extremities. Epidemiological studies suggested that excessive foot eversion is an important factor in the development of general running injuries and specifically knee injuries^{(4,5,13,17,23,24)}. It has been shown that part of the inversion/eversion movement of the foot is transferred through the ankle joint complex into external/internal rotation of the tibia ⁽¹²⁾. Excessive foot eversion coupled with excessive internal tibial rotation has been proposed to be one major reason for the development of knee injuries ⁽¹⁷⁾ since high relative rotation between the tibia and femur may alter the tracking of the patella on the femur. Consequently, internal/external rotation of the tibia relative to the femur would be the primary kinematic variable of interest to investigate these potential injury mechanisms. However, tibial rotation relative to the femur derived from skin markers cannot be used with an acceptable accuracy ⁽²⁰⁾. Thus, rotation of the tibia with respect to the foot or a laboratory coordinate system has been proposed as a substitute to study the development of movement related knee injuries⁽¹⁷⁾.

Orthotic shoes, specific shoe construction features, and/or shoe inserts of various materials and shapes are commonly assumed to improve comfort, performance and/or to reduce the frequency of overuse injuries^{(7,10,14,16,19,23,25-28)}. Shoe orthotics/inserts have often been constructed to control and reduce foot eversion and/or internal tibial rotation^{(3,4,7,8,10,13-15,18,19,21)}. Such shoe orthotics/inserts typically reduced the mean maximal foot eversion and tibial rotation during ground contact, but the reported kinematic effects were small and sometimes not significant^(18,29,30). Some research projects assessed the effect of specific orthotics/inserts on the injury frequency, the possibility to return to activities, the general acceptance and/or the subjective assessment of the patients. However, the authors are not aware of any research projects attempting to study the possible correlation between systematically changed shoe inserts and excessive foot eversion and specifically tibial rotation. Furthermore, a systematic approach addressing the relationship among anatomical, functional, and sensitivity related foot and leg characteristics and corresponding insert characteristics is missing.

The purposes of this investigation were (a) to quantify the effect of subject specific anthropometric characteristics on foot eversion and tibial rotation, (b) to quantify the effect of systematic changes in material composition of one shoe insert specifically constructed to reduce foot eversion and tibial rotation on these variables during running, and (c) to quantify the possible influence of individual foot characteristics on changes in foot eversion and tibial rotation resulting from the inserts.

METHODS

Twelve male subjects (age: 32.6 ± 9.6 yr; mass: 76.7 ± 8.0 kg; height: 1.78 ± 0.05 m) were recruited for the study. The subjects were physically active and at the time of the experiments involved in sports activities at least three times per week. Each subject gave informed written consent. All subjects suffered no pain or discomfort while running. Three-dimensional kinematic data of foot and leg and ankle joint range of motion (ROM) for all three rotations and arch stiffness were quantified.

All running trials used one standard running shoe (Adidas T-Advance C). Six conditions were analyzed in this study, one condition in which the movement was performed with the test shoe without any insert and five conditions with specific inserts provided by Schering-Plough Inc. (Memphis, TN). The inserts were constructed to reduce foot eversion and tibial rotation and had a bilayer design using two different materials at the top and bottom of the insert. The materials and their specifications are shown in Table 1.

TABLE 1: The materials and specifications for the inserts.

The conditions tested were: O for no insert, A for U2/L1, B for U1/L2, C for U1/L3, D for U3/L2, and E for U3/L3 with A and B being the two softest and E being the hardest inserts.

The subjects ran on a 30-m indoor runway at a running speed of 4.0 ± 0.2 m·s^-1 and the movement was quantified for one foot fall in the middle of the runway. Seven running trials were collected for each condition (O, A, B, C, D, and E) with the sequence of the tested conditions selected at random. All subjects wore the running shoe without socks to minimize foot movement inside the shoe and to allow markers to be placed directly on the foot. The Adidas (Portland, OR) insoles provided with the shoes were removed and replaced for all conditions with contourless Poron(Portland, OR) insoles.

The spatial positions of the lower leg, foot, and shoe were defined using six spherical reflective markers attached to the subject's leg and shoe and two flat round reflective markers attached to the subjects foot with adhesive tape (Fig. 1). Marker positions were chosen to minimize the effects of soft tissue movement over the bones and to ensure unobstructed camera views of the markers during running. The markers were placed as follows: M1: anterior tibial tuberosity; M2: lower anterior tibia; M3: superior lateral malleolus; M4: proximal heel; M5: distal heel; M6: inferior lateral malleolus; M7: proximal shoe; M8: distal shoe.

Figure 1-Schematic of the marker attachment at the lower extremities: .

A hole was cut in the back of the left shoe to view the two flat circular markers glued onto the heel which allowed the quantification of movement of the posterior aspect of the heel in the shoe. The three-dimensional positions of all markers while running were recorded with four electronically shuttered, high speed video cameras (NAC MOS-TV, V14B, Japan) with 12.5-75 mm zoom lenses(Cosmicar, Japan) and a VP310 video processor (Motion Analysis Corporation, Santa Rosa, CA) at a nominal sampling frequency of 200 Hz. A two-pass fourth-order Butterworth filter was used to smooth the three-dimensional spatial coordinates of each marker with a 15 Hz cut-off frequency.

Three types of coordinate systems were defined in this study: the lab co-ordinate system, LCS; the marker co-ordinate system, MCS; and the segment co-ordinate system, SCS. The spatial data were calculated from relative marker positions in a standard neutral body position. Before the running trials for each insole, the subject was videotaped for 2 s while standing motionless on the force platform with knees straight and the feet shoulder width apart and positioned in a zero ad-abduction angle. The left foot was adjusted so that the second metatarsal and the calcaneus were centered along the anterior-posterior axis of the laboratory coordinate system as defined by the camera calibration.

Three steps were involved in calculating the three-dimensional body segment orientations: transformation of each MCS to the SCS, calculation of the rotation matrix for each SCS relative to the LCS during dynamic trials, and the use of three joint coordinate systems, JCS^(6,9) to calculate orientations between body segments.

Relative movement occurs between the heel of the foot and the shoe during running ⁽²²⁾. Therefore, two inversion/eversion variables were calculated in this study, one representing movement of the foot relative to the leg and one representing movement of the shoe relative to the leg. For the ankle joint, the inversion/eversion motion of the foot with respect to the leg was described in the FOOT-LEG coordinate system. This system was defined with the first body fixed axis (ê₁) as the medio-lateral axis of the leg SCS and the second body fixed axis(ê₃) as the antero-posterior axis of the foot SCS. The third axis was calculated as the vector cross productEquation

The inversion/eversion motion of the shoe with respect to the leg was described in a SHOE-LEG coordinate system. This coordinate system was similar to the FOOT-LEG coordinate system with the antero-posterior axis of the shoe replacing that of the foot as ê₃. To ensure that movement of the lower leg in the frontal plane did not affect the measurement of total foot inversion/eversion, motion of the foot relative to a lab-coordinate system was described using a FOOT-LAB coordinate system. This coordinate system was similar to the previous FOOT-LEG coordinate system except that the medio-lateral axis of the laboratory coordinate system replaced that of the leg as ê₁. Tibial rotation, the axial rotation of the leg with respect to the foot, was calculated using a LEG-FOOT coordinate system with ê₁ as the medio-lateral axis of the foot SCS and ê₃ as the longitudinal axis of the lower leg.

The functional kinematic variables examined in this study were the foot-leg in-eversion angle, β, and the leg-foot tibial rotation, ρ. Specifically, these variables were determined for the time immediately before ground contact and for the maximal amplitude.

Arch height was quantified with the subjects standing in upright posture with the ankle joint in a neutral position on a raised platform⁽¹¹⁾. The highest point along the soft-tissue margin of the medial plantar curvature (marked with an ink pen) was measured with a modified Mitutoyo digital caliper. One arch height measurement, AH, was taken in a full weight bearing position of the left foot with the right foot resting lightly on the platform. A second measure of arch height, AHU, was taken in an unloaded position with the body weight transferred to the right foot and the left foot resting lightly on the platform. High arch heights were arbitrarily defined for this study for AHU values, which were above 34 mm, low arches for AHU values, which were less than 30 mm.

Relative arch deformation, RAD, has been used in this study as a measure of the stiffness of the arch of the foot. The relative arch deformation, RAD, was defined as Equation

RAD was normalized with respect to body weight (in Newtons) to account for inter-subject differences in body weight, which may influence arch compression. Stiff arches were arbitrarily defined for this study for RAD values which were less than 1.0/N, flexible arches for RAD values, which were above 2.0/N.

Active range of motion, ROM, of the left foot relative to the leg was evaluated using an apparatus that allowed independent measurement of three rotational degrees of freedom: (plantarflexion-dorsalextension, in-eversion and ab-adduction) and three translational degrees of freedom as described earlier ⁽¹⁾. Shortly, the subjects were seated in the apparatus with their knee bent at 90° flexion in an upright upper body posture. On the left foot they wore a standard running shoe without socks to reduce movement of the foot inside the shoe. The shod foot rested on a horizontal footplate and was attached to the footplate with self-stick straps around the heel, the midtarsal, and the metatarsal bones. The foot was positioned such that the second metatarsal and the center of the calcaneus were aligned with the in/eversion axis of the plate. The leg was secured in the apparatus with C-clamps just above the malleoli while the anterior aspect of the knee was pressed into a V-bar. The ROM was measured with a constant compressive load of 100 N applied along the tibial axis. The subject was asked to perform a maximal active motion starting from the neutral position for six types of movement (plantarflexion, dorsalextension, inversion, eversion, abduction, and adduction). The angular displacement was measured with calibrated potentiometers attached to the axes of the footplate. The between-day repeatability for 10 subjects has been reported as 0.60 ± 0.78° with a maximal individual difference of 1.55°⁽¹⁾. Three trials of each motion were taken and then averaged.

The statistical analysis selected was two way repeated measures MANOVA(within trials and inserts).

RESULTS

Results and discussion are organized in three sections. First, the effect of subject specific anthropometric characteristics is discussed independent of inserts. Second, general results are presented and discussed with respect to the tested inserts. Third, individual results are presented and discussed.

Effect of anthropometric characteristics. The correlation coefficients for all insert conditions (N = 72) are listed inTable 2. Relative arch deformation, RAD, showed a negative correlation with the total internal tibial rotation movement during the first 50% of ground contact, Δρ = ρ_max -ρ_o. Stiff arches showed more tibial rotation than flexible arches. The ROM from maximal dorsal-extension to maximal plantar-flexion, ROM-DP, showed negative correlations with most of the kinematic variables analyzed. Low ROM-DP was associated with high foot and shoe eversion and high internal tibial rotation. The ROM from maximal eversion to maximal inversion, ROM-EI, was negatively correlated with internal tibial rotation. However, ROM-EI did not show any correlation for the actual eversion movement during stance in running. The ROM from maximal adduction to maximal abduction, ROM-AA, showed a negative correlation for the maximal foot and shoe eversion during actual running, i.e., subjects with low ab-adduction ROM showed high foot and shoe eversion.

TABLE 2: Summary of correlation coefficients for foot eversion,Δβ_foot, shoe inversion, Δβ_shoe, and internal tibial rotation, Δρ, on one side and arch height, AH, relative arch deformation, RAD, range of motion for dorsi-plantarflexion, ROMDP, in-eversion, ROMEI, and ab-adduction, ROMAA.

Effect of inserts on group results. The average results of the total shoe eversion movement, Δβ_shoe, and the total foot eversion movement, Δβ_foot, were not systematically and significantly changed by the various inserts used in this project. Differences in the mean values for the foot or shoe eversion movement between the different insert conditions were typically small. The average maximal everted foot positions showed a nonsignificant trend. All maximal everted foot positions with inserts were about 1-2° higher than the maximal everted foot positions for the no-insert condition. Thus, the inserts may have slightly shifted the foot toward a more everted position.

The results for tibial rotation showed a similar trend as the results for foot/shoe eversion. There was a trend that total internal tibial rotation movement was slightly (but not significantly) smaller for the insert conditions when compared with the no-insert condition.

Effect of inserts on individual results. The effects of the tested inserts on foot eversion and internal tibial rotation for four different test subjects are illustrated in Figure 2.Figure 2 shows the results for foot eversion on the left side boxes and for the results for internal tibial rotation on the right side boxes. The left side of each box shows the mean result for the variable of interest for the no-insert condition, the right side the mean results for the specific insert conditions. To indicate the change for one condition, the results for each insert condition are connected through a solid line with the results for the no-insert condition. Additionally, each subject is characterized with the relative arch deformation, RAD, and the unloaded arch height, AHU.

Figure 2-The influence of different shoe inserts on foot eversion and internal tibial rotation. The four boxes on the left side show the results for foot eversion, the boxes on the right side the results for internal tibial rotation. The point on the left side of each box represents the subject specific result for the no-insert condition. The points on the right side of each box represent the results for the specific insert conditions. The information under a set of two boxes describes the foot stiffness (relative arch deformation, RAD) on the left side and the unloaded arch height on the right side: .

Subject 1 showed a small reduction of internal tibial rotation but no consistent results for foot eversion for all insert conditions. Some inserts reduced foot eversion slightly while others increased them slightly. Overall, the inserts produced only minimal changes of foot or leg movement(<2.5°) for subject 1.

Subject 2 showed a reduction of foot eversion (up to 4°) because of the inserts. However, there was no systematic trend for internal tibial rotation. The combination subject 2 and insert A (a soft insert), for instance, showed a reduction of foot eversion of about 3° but (contrary to expectation) an increase of tibial rotation of about 4°. The combination of subject 2 and insert E (the hardest insert) showed a reduction of foot eversion of about 2° and (as expected) a concurrent reduction of internal tibial rotation of about 3°.

Subject 4 showed a reduction of foot eversion and an increase of internal tibial rotation for all insert conditions.

Subject 7 was the only subject who showed a reduction of foot eversion and internal tibial rotation for all but one insert condition, the result which was expected for most subjects.

DISCUSSION

Effect of anthropometric characteristics. The finding that the relative arch deformation showed a correlation with tibial rotation but no correlation with foot or shoe eversion indicates that arch stiffness influenced the movement coupling between foot and leg. The connection between foot and leg was stronger for the stiff arches. Consequently, the coupling between foot and leg is high for stiff arches and low for flexible arches. The transfer of foot eversion into tibial rotation was higher for the stiff than for the flexible arch group, a result that is in agreement with findings of earlier studies ⁽¹⁷⁾. Therefore, excessive foot eversion during locomotion will be increasingly critical for subjects with stiff foot arches.

The effect of inserts on group results. The indicated trend in the changes of tibial rotation because of the inserts was in the right direction to potentially reduce knee injuries. However, the changes were so minimal that they are assumed to be physiologically not relevant.

The inserts, which were developed to reduce foot eversion and internal tibial rotation, did not, as expected, reduce foot eversion and internal tibial rotation in a systematic way for the whole group tested. However, some results were significantly different for the whole group of subjects.Table 3 lists the ranges of the measured foot eversion,Δβ_foot, and internal tibial rotation, Δρ, for all tests conditions. The ranges for insert B were about 7° for total foot eversion and about 5° for total internal tibial rotation. In comparison, insert C allowed a range of about 14° for total foot eversion and close to 9° for total tibial rotation, about twice the range allowed for both variables by insert B. Insert B, therefore, seemed to be more restrictive, forcing all different feet into a similar movement pattern, while insert C seemed to allow/produce more variation in movement patterns for the different test subjects and, consequently, may have allowed other factors such as foot anatomy or arch stiffness to have more influence on the results of the movement patterns.

TABLE 3: Maximal and minimal total foot eversion,Δβ_foot, and tibial rotation, Δρ, and their corresponding ranges for the condition without inserts, O, and each insert condition, A, B, C, D and E for all subjects.

Insert B was composed of a soft upper and a soft lower layer, insert C of a soft upper layer and a hard lower layer. It is surprising and was certainly not expected that the softer insert would provide more control of foot eversion and internal tibial rotation than the harder insert. Whether the material properties and the specific shape of insert B are advantageous or disadvantageous with respect to the reduction of pain and/or injury remains open for discussion. However, the trend that soft inserts produced a more uniform movement response than hard inserts is interesting, and this initial result should be verified/rejected in further studies using similar but more pronounced material and shape combinations.

Effect of inserts on individual results. The individual results showed changes in foot/shoe eversion and internal tibial rotation. It should be noted, however, that for all subjects the changes resulting from the added insert were typically less than 4° for foot eversion and less than 5° for internal tibial rotation.

In a comparison of the insert with the no-insert condition, total tibial rotation decreased for all inserts for 6 of the 12 subjects but increased for all inserts for 3 of the 12 subjects (Table 4). A comparison of the relative arch deformation, RAD, for the two groups (those with a consistent increase and those with a consistent decrease) indicated that those subjects with a consistent reduction of total tibial rotation were on the average those with a flexible foot arch (average RAD = 1.8/N), while those subjects with a consistent increase of total tibial rotation with the inserts were on the average those with a stiff foot arch (average RAD = 0.57/N). It should be noted, however, that subject 7 (in the flexible group) did not fit this description.

TABLE 4: Summary of effects of inserts on total foot eversion,Δβ_foot, and total tibial rotation, Δρ, for the individual subjects. The numbers in the table represent the different subjects.

In general, the changes in foot and leg kinematics for the different subjects did not allow a clear conclusion about the effect of specific inserts for the variables quantified in this study. Some trends did appear. However, further studies are necessary to understand the effects of specific shoe inserts on lower leg kinematics for specific populations and/or leg-foot characteristics. Some of the results suggest that the leg-foot characteristics tested in this study should be complemented with other mechanical and structural variables. Furthermore, it is suggested that variables assessing the sensitivity of the foot and leg to low and high frequency vibrations should be added to characterize the individual feet and legs.

The results of this study indicate that the tested inserts do not change the kinematics of the ankle joint complex significantly for the analyzed group. A power analysis conducted on the kinematic variables suggested that there was a 90% chance of detecting any differences in these variables between insert conditions that were greater than about 5°. Differences less than this, although not found to be statistically significant, may have existed. Therefore, the discussion of trends in this study was justified as they may have represented real effects. The trends that were observed were rather minimal, however, and it is suggested that these trends should only be used in guiding possible methodological changes for further developments. Furthermore, the changes in the material properties, which were used for the different inserts, had for the whole group a minimal effect on foot eversion and tibial rotation. This result is in agreement with previous studies^(18,29,30), indicating that group changes in kinematics and kinetics are only minimally influenced by changes in material. These previous studies indicate that significant and relevant changes in kinematic and kinetic variables are most effectively achieved by changing the geometry (shape) of the insert or the shoe. However, in some individual subjects specific material-shape combinations did have a substantial effect, changing the movement by more than 4°.

The results of this study suggest that criteria for mass production of shoe inserts may not be the same as criteria for individual insert fitting. The results of this study showed substantial differences in the individual reactions to the different tested inserts. The study was not designed to identify the optimal insert design for one specific foot-leg type and, consequently, conclusions about the “optimal” insert are not possible. However, the results suggest that further research may reveal specific construction features that are optimal for foot and leg types with a specific anatomy, morphology, functional behavior, and sensitivity to external signals.

This study has been supported by Schering-Plough (Dr. Scholl), Memphis, TN. They provided the inserts and financial help.

Address for correspondence: Dr. Benno M. Nigg, Human Performance Laboratory, The University of Calgary, Calgary, Alberta, Canada T2N 1N4.