The Influence of High Heeled Shoes on Kinematics, Kinetics, and Muscle EMG of Normal Female Gait

JOURNAL OF APPLIED BIOMECHANICS, 2000, 16, 309-319 2000 by Human Kinetics Publishers, Inc. The Influence of High Heeled Shoes on Kinematics, Kinetics, and Muscle EMG of Normal Female Gait Darren J. Stefanyshyn, Benno M. Nigg, Veronica Fisher, Barry O Flynn, and Wen Liu The purpose of this investigation was to determine whether a graded response in gait kinematics, kinetics, and EMG occurs as shoe heel height increases. Four different shoes, including one flat shoe and three shoes with high heels, were tested in this investigation. The average heel heights of the four shoes were 1.4 cm, 3.7 cm, 5.4 cm, and 8.5 cm. Kinematics, kinetics, and muscle EMG were measured during the stance phase of gait on 13 healthy female subjects while wearing each of these 4 shoes. Systematic increases in the active vertical, propulsive, and braking forces were found as shoe height increased. Ankle and knee flexion and soleus and rectus femoris activity showed a graded response as heel height increased. One surprising result was the behavior of the maximal vertical impact force peak and the maximal loading rate during heel impact. The vertical impact force peaks and the maximal vertical loading rates were highest for the shoe with 3.7 cm heel height and lowest for the flat shoe and the shoe with heel height of 8.5 cm. Key Words: high heeled shoe, gait, joint motion, joint moment, muscle activity. Introduction High heeled shoes have a variety of features, which distinguish them from other shoes that are ordinarily worn by females. While typical shoes have a heel elevation of approximately 1 to 2 cm, high heeled shoes have a heel elevation of up to or even greater than 10 cm. Other differences include a narrow toe box, a rigid heel cap which protrudes anteriorly (Stephens, 1992), and an excessive plantar curvature (Schwartz & Heath, 1959 In some instances, these features have been associated with the development of various foot deformities. The narrow toe box has been speculated to cause Hallux valgus due to the increased pressure on the medial side of the forefoot (Frey et al., 1993 The rigid heel cap with the anterior protrusion is speculated to cause Haglund s deformity (a protrusion on the back of the calcaneus) due to increased pressure on the posterior aspect of the calcaneus. Additionally, high heeled shoes are often associated with low back pain, plantar foot pain, discomfort, and muscle fatigue (Adrian & Karpovich, 1966; Opila-Correia, 1990 The authors are with the Human Performance Laboratory in the Faculty of Kinesiology at the University of Calgary, Calgary, Alberta, Canada, T2N 1N4. 309

310 Stefanyshyn et al. When walking in high heeled shoes as compared to flat shoes, differences exist in stride parameters, kinematics, kinetics, muscle activity, energy consumption, and plantar foot pressure (Gehlsen et al., 1986; Joseph, 1968; Mathews & Wooten, 1963; Murray et al., 1970; Soames & Evans, 1987; Snow & Williams, 1990 Undoubtedly, many of these differences lead to the above mentioned pain and discomfort. However, it is unclear how these variables change as heel height systematically increases. With only a few exceptions (Loy & Voloshin, 1987; Ebbeling et al., 1994), previous studies on high-heeled gait have concentrated on comparing a flat shoe to a single high-heeled shoe. It was hypothesized that a graded response in biomechanical variables exists as heel height increases. Thus, the purpose of this investigation was to determine whether systematic increases in shoe heel height result in systematic changes in kinematic, kinetic, and EMG data during gait. Methods Four different shoes (Figure 1) were used in this investigation. The first shoe, the Enzo Angiolini Liberty, was a flat shoe and had an average heel height of 1.4 cm (0.55 in. The other three shoe models were all high heels with varying heel heights. The Amalfi Gina was a low heeled shoe with an average heel height of 3.7 cm (1.46 in. The Amalfi Dahlia was a medium heeled shoe with an average heel height of 5.4 cm (2.14 in.), and the Caressa Haute was a high heeled shoe with an average heel height of 8.5 cm (3.34 in. Throughout the report, the Enzo Angiolini Liberty, the Amalfi Gina, the Amalfi Dahlia, and the Caressa Haute will be referred to as flat, 3.7, 5.4, and 8.5, respectively, to represent the approximate heel heights of the different shoes. All shoes were commercially available and were chosen primarily due to their similarity of construction so that the main difference between the shoe models was the height of the heel. Thirteen female subjects were recruited for this investigation. Their mean age was 40.6 ± 8.3 years, mean height 164.1 ± 5.6 cm, and mean mass 67.7 ± 12.3 kg. Their shoe Figure 1 Photograph of the four different shoes used in this investigation. From left to right: Enzo Angiolini Liberty (flat), Amalfi Gina (3.7 cm), Amalfi Dahlia (5.4 cm), and Caressa Haute (8.5 cm

High Heeled Gait 311 size ranged from 6 to 9.5 (average, 7.7 All subjects were free from recent lower extremity injury or pain. Although not on a daily basis, all subjects regularly wore high heeled shoes during their work day. The research was approved by the University of Calgary ethics committee, and informed written consent was obtained from all subjects. Kinetic data was collected with a Kistler force platform sampling at 1,000 Hz. Kinematic data was collected simultaneously with the kinetic data using a Motion Analysis four video camera system at 200 Hz. Three spherical reflective markers were placed on each of the thigh, shank, and shoe for kinematic data collection. For each subject, the order of the shoes was randomly assigned, and four walking trials (1.4 ± 0.2 m/s) were collected for each shoe condition. The walking speed of the subjects was monitored with photocells placed immediately before and after the force plate. A standing neutral trial prior to each set of walking trials was used to define the joint centers. The ankle joint center in the vertical and anterior-posterior directions was defined by an additional marker placed on the lateral malleolus. The location of the ankle joint center in the medial-lateral direction was defined by a superior navicular marker that was placed on the midline of the foot. The knee joint center in the vertical and anterior-posterior directions was defined by the marker placed on the lateral epicondyle and in the mediallateral direction by an additional marker placed in the middle of the patella. Prior to calculation of any variables, a second-order low-pass Butterworth filter was used to filter the kinematic data (cut-off frequency of 10 Hz) and the kinetic data (cut-off frequency of 100 Hz The kinematic and kinetic variables analyzed in this investigation are listed in Table 1 and Table 2, respectively. Electromyographical activity of the gastrocnemius, soleus, peroneus longus, tibialis anterior, rectus femoris, semitendinosus, biceps femoris, and vastus medialis muscles were collected using a surface electrode EMG system. Skin preparation consisted of shaving the appropriate areas and treating with alcohol wipes to ensure dead skin removal. Circular Ag/AgCl electrodes of 1-cm diameter were placed approximately 2 cm apart in the middle of the muscle bellies in an attempt to minimize cross-talk and remained in the same placement for data collection of all shoes. The signals were amplified with a bandwidth frequency from 10 Hz to 1 khz (Common mode rejection ratio, CMMR = 120 db, gain = 1,000) and were sampled at 2.4 khz. Analysis of the EMG data consisted of calculating the root mean square for the activity of each muscle during ground contact. A one-way repeated measures MANOVA using SPSS software was used to compare the biomechanical variables (kinematic, kinetic, and EMG) between the different shoes. When a significant difference was found, a Tukey post hoc analysis was performed. The level of significance was chosen as p <.01. Results High heels resulted in higher ground reaction forces both in the anterior and posterior directions (Figure 2 The increased anterior-posterior forces corresponded to increases in the peak deceleration and acceleration forces in the vertical direction, Fz a1 and Fz a2, respectively (Figure 2 The maximal vertical impact force and the maximal vertical impact loading rate were significantly higher for both the 3.7 and the 5.4 shoes than either the flat or 8.5 shoes (Figure 3 The 8.5 shoe had the lowest values of the maximal vertical impact force and the maximal vertical impact loading rate. An increase in knee flexion during stance was found when wearing high heels (Figure 4 The rectus femoris muscle is more highly activated to control the increased knee flexion, as shown in the increased muscle EMG signals. The increase in extensor

312 Stefanyshyn et al. Table 1 The Kinematic Variables Analyzed in This Investigation Variable Definition min max min max minlcs maxlcs lcs min max o max max Minimal shoe eversion with respect to the leg value of minimum eversion of the shoe during the initial 50% of stance. Maximal shoe eversion with respect to the leg value of maximum eversion of the shoe during stance. Maximal difference in shoe eversion with respect to the leg the difference between the maximal eversion of the shoe during stance and the minimal eversion of the shoe ( max min Minimal tibial rotation with respect to the foot value of minimum tibial rotation during the initial 50% of stance. Maximal tibial rotation with respect to the foot value of maximum tibial rotation during stance. Maximal difference in tibial rotation with respect to the foot the difference between the maximal (internal) tibial rotation during stance and the minimal tibial rotation during stance ( max min Maximal tibial rotation with respect to the lab coordinate system value of minimum tibial rotation during the initial 50% of stance in the lab coordinate system. Maximal tibial rotation with respect to the lab coordinate system value of maximum tibial rotation during stance in the lab coordinate system. Maximal difference in tibial rotation with respect to the lab the difference between the maximal (internal) tibial rotation during stance and the minimal tibial rotation during stance ( max min Minimal foot dorsi-plantarflexion with respect to the leg value of minimum dorsi-plantarflexion angle of the ankle joint during the initial 50% of stance. Maximal foot dorsi-plantarflexion with respect to the leg value of maximum dorsi-plantarflexion angle of the ankle joint during stance. Maximal difference in foot dorsi-plantarflexion with respect to the leg the difference between the maximal dorsi-plantarflexion angle of the ankle joint and the minimal dorsi-plantarflexion angle of the ankle joint during stance ( max min Flexion-extension angle between the thigh and lower leg at touchdown value of the flexion-extension angle of the knee joint at touchdown. Maximal flexion-extension angle between the thigh and lower leg value of maximum flexion-extension angle of the knee joint during stance. Maximal difference in the flexion-extension angle between the thigh and lower leg the difference between the maximal flexion-extension angle of the knee joint and the flexion-extension angle of the knee joint at touchdown ( max o Maximal ab-adduction position of the foot in the lab coordinate system value of maximum ab-adduction position of the foot with respect to the lab during the first 50% of stance.

High Heeled Gait 313 Table 2 The Kinetic Variables Analyzed in This Investigation Variable Definition Fz i Gz i Fa 1 Fa 2 Fy min Fy max KM f-emin KM f-emax AM f-emin AM f-emax Maximal vertical impact force local maximum of the vertical ground reaction force during the first 50 ms after heelstrike. Maximal vertical loading rate maximal rate of change in the vertical ground reaction force between the time of heel strike and the time of the maximum vertical impact force (Fz i First maximal vertical active force local maximum of the vertical ground reaction force during the initial 50% of the stance phase. Second maximal vertical active force local maximum of the vertical ground reaction force during the final 50% of the stance phase. Minimal anterior-posterior force minimum of the anterior-posterior force during stance, represents the maximal braking force. Maximal anterior posterior force maximum of the anterior-posterior force during stance, represents the maximal propulsive force. Minimal flexion-extension knee moment minimal flexion-extension moment at the knee during stance, represents the maximum flexor moment. Maximal flexion-extension knee moment maximal flexion-extension moment at the knee during stance, represents the maximum extensor moment. Minimal dorsi-plantarflexion ankle moment minimal dorsi-plantarflexion moment at the ankle joint during stance, represents the maximum plantarflexor moment. Maximal dorsi-plantarflexion ankle moment maximal dorsi-plantarflexion moment at the ankle joint during stance, represents the maximum dorsiflexor moment. activity lead to increases in the knee extensor moments. There were no significant differences in the activity of the vastus medialis, biceps femoris, or semitendinosus muscles as heel height increased. An increase in ankle plantarflexion while wearing high heels was found in this study (Figure 5 Decreases were found in plantarflexor moments (Figure 6 There was no difference between flat or high heeled shoes in the dorsiflexor moments, as they were relatively small. Despite smaller plantarflexor moments, the activity of the soleus muscle increased (Figure 6) as heel height increased. There were no significant differences in the activity of the gastronemius or tibialis anterior muscles as heel height increased. There was no significant change in the amount of shoe eversion as heel height increased; however, there was a trend toward decreased eversion (pronation; Figure 7 As well as increases in the activity of the rectus femoris and soleus muscles when wearing high heels, there was also an increase in the activity of the peroneus longus. There was a significant increase in the amount of shoe adduction as heel height increased, which would correspond to the more supinated foot position. No significant difference was found between the different heel heights in the amount of tibial rotation. However, there was a trend of decreased tibial rotation as heel height increased.

314 Stefanyshyn et al. Figure 2 Comparison of the minimal anterior-posterior force, Fy min, and the maximal anteriorposterior force, Fy max (a), and the first, Fz a1, and second, Fz a2, maximal vertical active forces (b) (mean and standard error) for the different heel heights. Significant differences from the flat, 3.7, 5.4, and 8.5 shoes are indicated by #, *, **, and ***, respectively. Discussion Similar to the data presented by Ebbeling et al. (1994), it was found that vertical and anterior-posterior ground reaction forces increase with increasing heel heights. The larger braking forces caused a greater deceleration of the center of mass, which had to be counteracted by an increase in the peak propulsive force to accelerate the center of mass again at take-off. Thus, as heel height increased, the subjects gait pattern became less fluent, characterized by larger accelerations and decelerations. As heel height increases, the center of mass moves more and more forward due to the increased plantarflexion of the ankle joint. Thus, knee flexion continually increases as a counteractive measure to offset this forward movement of the center of mass. Therefore, the rectus femoris muscle becomes more active to control the increased knee flexion.

High Heeled Gait 315 Figure 3 Comparison of the maximal vertical impact force, Fz i, and the maximal vertical loading rate, Gz i, (mean and standard error) for the different heel heights. Significant differences from the flat, 3.7, 5.4, and 8.5 shoes are indicated by #, *, **, and ***, respectively. As a result of the increased plantarflexion when walking in high heels, smaller propulsive moments were required at take-off, which was evidenced by the decrease in plantarflexor moments. The increase in activity of the soleus muscle with high heels corresponds to data previously reported by Joseph (1968 This may appear to be a confusing result at first; however, it can be explained mechanically by the geometry at the ankle joint. As the ankle becomes more plantarflexed, as is the case for high heels, the length of the moment arm of the Achilles tendon decreases. Therefore, despite an increase in force in the Achilles tendon due to increased soleus muscle activity, there is a decrease in the moment that the force produces due to a decrease in moment arm length. Loy and Voloshin (1987) used skin mounted accelerometers to measure tibial acceleration at heel strike when walking in different heel heights. They found an increase in peak accelerations of the tibia as heel height increased from 0.5 in. (1.1 cm) to approximately 3 in. (7.6 cm This corresponds well with the increased impact force and loading rate found in this study, as heel height increased up to 5.4 cm. Unfortunately Loy and Voloshin (1987) did not include any shoes with heel heights greater than approximately 3 in. Therefore, the result that the impact force and loading rate for the 8.5 shoes decreases from the other high heeled shoes to levels equivalent to the flat shoes was not substantiated by their study. The rather unexpected impact force results for the 8.5 shoe in this study may be due to various reasons: (a) The foot can not be placed normally when the heel height passes a certain threshold. Consequently, impact forces decrease after passing this threshold heel height and landing occurs with less impact. (b) The shoe models used for the 8.5 shoe were different in their construction and/or composition than the other shoes. (c) It is possible that a certain threshold exists above which the musculoskeletal system acts in a manner to decrease the impact force and loading rate to prevent possible injury. There was a trend toward decreased eversion (pronation) as heel height increased. In addition, a trend of decreased tibial rotation as heel height increased may be a result of the coupling mechanism between the foot and lower leg (Inman, 1976 The peroneus longus is a foot everter, and the increased activity of this muscle when wearing high heels

316 Stefanyshyn et al. Figure 4 Comparison of the variables associated with flexion-extension of the knee. Muscle activation of the rectus femoris muscle and maximum knee extensor moment (mean and standard error) for the different heel heights. Positive values for O and max correspond to knee flexion. Significant differences from the flat, 3.7, 5.4, and 8.5 shoes are indicated by #, *, **, and ***, respectively. is possibly to control the increase in supination of the foot. The peroneus longus is also often considered an ankle stabilizer, and the increased activity may be required to stabilize the ankle joint when wearing high heels. In conclusion, there was a graded response in the ground reaction forces, ankle and knee kinematics, and activity of the rectus femoris, soleus, and peroneous longus muscles. The activity of the gastrocnemius, tibialis anterior, semitendinosus, biceps femoris, and vastus medialis did not show a graded response to increasing heel heights. The vertical impact forces and loading rates showed a graded response for heel heights up to 3.7 cm but then decreased for larger heel heights.

High Heeled Gait 317 Figure 5 Comparison of the variables associated with dorsi-plantarflexion of the ankle (mean and standard error) for the different heel heights. Values for min and max greater than 0 correspond to ankle dorsiflexion, while values less than 0 correspond to plantarflexion. Significant differences from the flat, 3.7, 5.4, and 8.5 shoes are indicated by #, *, **, and ***, respectively. Figure 6 Comparison of the plantarflexor moment and the muscle activation of the soleus muscle (mean and standard error) for the different heel heights. Positive values for AM f-emin correspond to dorsiflexor moments. Significant differences from the flat, 3.7, 5.4, and 8.5 shoes are indicated by #, *, **, and ***, respectively.

318 Stefanyshyn et al. Figure 7 Comparison of the maximum eversion of the shoe, muscle activation of the peroneus longus muscles, adduction of the shoe, and tibial rotation (mean and standard error) for the different heel heights. Significant differences from the flat, 3.7, 5.4, and 8.5 shoes are indicated by #, *, **, and ***, respectively. References Adrian, M.J., & Karpovich, P.V. (1966 Foot instability during walking in shoes with high heels. The Research Quarterly, 37, 168-175. Ebbeling, C.J., Hamill, J., & Crussemeyer, J.A. (1994 Lower extremity mechanics and energy cost of walking in high-heeled shoes. Journal of Orthopaedic and Sports Physical Therapy, 19, 190-196. Frey, C., Thompson, F., Smith, J., Sanders, M., & Horstman, H. (1993 American orthopaedic foot and ankle society women s shoe survey. Foot & Ankle, 14, 78-81. Gehlsen, G., Braatz, J.S., & Assmann, N. (1986 Effects of heel height on knee rotation and gait. Human Movement Science, 5, 149-155. Inman, V.T. (1976 The joints of the ankle. Baltimore: Williams & Wilkins.

High Heeled Gait 319 Joseph, J. (1968 The pattern of activity of some muscles in women walking on high heels. Annals of Physical Medicine, 9, 295-299. Loy, D.J., & Voloshin, A.S. (1987 Biomechanical aspects of high heel gait. Proceedings of the American Society of Biomechanics 11th Annual Meeting (pp. 135-136), Davis, CA. Mathews, D.K., & Wooten, E.P. (1963 Analysis of oxygen consumption of women while walking in different styles of shoes. Archives of Physical Medicine & Rehabilitation, 44, 569-570. Murray, M.P., Kory, R.C., and Sepic, S.B. (1970 Walking patterns of normal women. Archives of Physical Medicine & Rehabilitation, 51, 637-651. Opila-Correia, K.A. (1990 Kinematics of high-heeled gait. Archives of Physical Medicine & Rehabilitation, 71, 304-309. Schwartz, R.P., & Heath, A.L. (1959 Preliminary findings from a roentgenographic study of the influence of heel height and empirical shank curvature on osteo-articular relationships in the normal female foot. Journal of Bone and Joint Surgery, 41-A, 1065-1077. Snow, R.E., & Williams, K.R. (1990 Effects on gait, posture, and center of mass position in women wearing high heeled shoes. Medicine and Science in Sports and Exercise, 22, S23. Soames, R.W., & Evans, A.A. (1987 Female gait patterns: The influence of footwear. Ergonomics, 6, 893-900. Stephens, M.M. (1992 Heel pain. The Physician and Sports Medicine, 20, 87-95. Acknowledgment We would like to acknowledge the financial support of Schering-Plough HealthCare Products, Inc.