Effect of heel lifts on ground reaction force patterns in subjects with structural leg-length discrepancies.

Effect of Heel Lifts on Ground Reaction Force Patterns in Subjects with Structural Leg-Length Discrepancies

The purpose of this study was to determine the effect of heel lifts on ground reaction force patterns in subjects with structural leg-length discrepancies (LLDs). Eighteen subjects with LLDs ranging from 4.8 to 22.2 mm participated in this study. Subject age range was from 20 to 63 years. A force platform was used to obtain ground reaction force data for four conditions. Data were collected prior to fitting of the heel lift and after a three-week break-in period. Data were analyzed by use of a two-factor within-subject analysis of variance for repeated measures. Before heel-lift fitting, maximum lateral force was greater in the short leg than in the long leg. After heel-lift fitting, maximum vertical force was greater within both legs, and maximum medial force was greater in the long leg than in the short leg. The results suggest that although heel lifts are used to achieve pelvic levelness, the use of heel lifts also resulted in increased ground reaction forces, which may cause increased joint stresses within the lower extremities. [Schuit D, Adrian M, Pidcoe P: Effect of heel lifts on ground reaction force patterns in subjects with structural leg-length discrepancies. Phys Ther 69:663-670, 1989] Key Words: Ground reaction forces; Heel lifts; Kinesiology/biomechanics, gait analysis; Leg-length discrepancy.

Leg-length discrepancies (LLDs) are common in the general population and have been related to postural anomalies in standing and walking.[1-10] Most likely as a result of postural changes, LLDs have been implicated in the occurrence of pain in the knee, hip, and most often the low back.[2,5-9,11-15] A high incidence of asymptomatic sacroiliac joint malalignment in young women who exhibited LLDs has also been reported.[16] Increased wear patterns have been observed on the sole of the shoe of the short leg, indicative of a possible alteration in the ground reaction force resultant during walking.[5] Despite the clinical evidence, actual research documentation of the effects of a LLD on the typical ground reaction force pattern is limited. To our knowledge, only two studies have examined ground reaction forces in each leg in subjects with LLDs.[17,18] In the first study, 18 patients were fitted with two small force transducers on each foot--one at the center of the heel and the other underneath the metatarsal heads. The patients then walked on a treadmill before and after treatment for their LLD. Treatment was defined as fitting of a heel lift if the patient had a structural LLD and arch taping to minimize pronation if the patient had a functional LLD. Peak force magnitude was determined before and after treatment by the electrical activity of each piezoelectric force transducer. The authors reported force magnitudes in terms of "relative units." The force data were not calibrated in terms of pounds per square inch. A statistically significant difference was found between heel forces on the same heel before and after treatment. When comparing short-leg and long-leg forces prior to treatment, a majority of cases (78%) showed greater forces on the heel of the longer leg. All differences between short-leg and long-leg heel forces became more normalized after treatment. No statement regarding forces from the forefoot transducers was made. The second study examined 17 patients with structural LLDs. A series of small (0.5 cm) force transducers were fitted to the subjects' feet at the following locations: medial and lateral plantar calcaneal tubercles; metatarsal heads 1, 2, and 5; and interphalangeal joint of the hallux. Each subject walked a standard distance (9.14 m) on a concrete floor for all recordings both before and after addition of a heel lift. All heel lifts were 0.64 cm thick, and all subjects wore standardized running shoes. The authors concluded that the addition of the heel lift resulted in an equalization of cadence and compensatory mechanisms between both feet and in a reduction of pronatory forces within the longer limb and supinatory forces within the shorter limb. These studies have primarily been limited to the examination of the distribution of vertical forces at selected anatomical sites on the feet. Actual magnitudes of ground reaction forces as well as the components of these forces (fore-aft, medial-lateral, and vertical) have not been examined. The purpose of this investigation was to determine the effect of heel lifts on each of the components of the ground reaction force pattern during the stance phase of walking for individuals with structural LLDs. On the basis of the cited literature, we expected to find differences in the ground reaction force components between the short leg and the long leg prior to heel lift fitting. After fitting, we expected a reduction or normalization of the differences between the short leg and the long leg.

Method

Subjects

The subjects for this study were 18 volunteers (4 male, 14 female) who exhibited a structural LLD of between 4.8 and 22.2 mm (Tab. 1). Subjects were between the ages of 20 and 63 years (X[bar] = 26.0, s = 9.6). All subjects were screened by one of the investigators (DS) by direct questioning to determine history of pathology or trauma to the lumbar spine, pelvis, or lower extremities within the last six months and by clinical observation to determine the presence of skeletal deformities. The presence of a functional LLD was determined by palpation of the height of the iliac crests in the standing position with the bilateral subtalar joints placed in the neutral and then in a relaxed position.[17,19] The presence of an innominate rotation was determined by palpation of pelvic landmarks (bilateral anterior superior iliac spines [ASISs] and posterior superior iliac spines [PSISs]) and by performance of the supine-to-long-sitting test.[5] The presence of an innominate rotation or a change in iliac crest height from the bilateral subtalar joint neutral position to the bilateral subtalar joint relaxed position indicated the presence of a functional LLD. Volunteers exhibiting a functional LLD were excluded from this study. Measurement for structural LLD was accomplished by fitting thin metal plates of various thicknesses under the foot of the short leg until the pelvis was level. Levelness of the pelvis was determined by an apparatus consisting of a carpenter's level affixed to an adjustable rod with movable arms that were placed appropriately on the iliac crests. The superior precision of this measurement method as compared with tape measurements of the distance from the ASIS to the medial malleolus, ASIS to the lateral malleolus, and umbilicus to the medial malleolus has been described elsewhere[20] and recommended by several authors.[5,17,21] The standardized position of the feet of each subject prior to measurement was equidistant from a midheel line with the feet in a comfortable, relaxed position relative to angle and base.[22,23] Prior to testing, all subjects read and signed an informed consent form approved by the University of Illinois at Chicago's Institutional Review Board.

Instrumentation

A Kistler piezoelectric force platform(*) positioned in a 10.4-m walkway was used to sample force data. Force data were sampled at a frequency of 1,000 Hz for three seconds. A Hewlett-Packard analog-to-digital converter([dagger]) interfaced with a Hewlett-Packard Model 9217-A personal computer([dagger]) was used to collect amplified force data. To begin data collection, the force platform and analog-to-digital converter were triggered synchronously by interruption of a light beam to a photocell on the walkway. A Gould strip chart recorder([double dagger]) was used for immediate verification of all force data. All data were stored on 8.9-cm floppy disks.

Procedure

Each subject wore comfortable gym clothing and his or her own recreational or running shoes for all trials. Each subject was allowed as many practice walking trials across the force platform as necessary to establish the proper starting position to ensure complete contact with the force platform. Each subject was instructed to walk as naturally as possible and to avoid any excessive "pushing" on the force platform as he or she walked over it. All subjects were timed with a stopwatch prior to the experimental trials as they walked across the platform at their self-selected walking speed. The time was measured as the subjects walked between two fixed markers on the walkway. For use in the experimental trials, subject cadence was determined by the following formula:

Cadence = number of steps/time

(in seconds) x 60 seconds/minute

= steps/minute For all trials, a metronome was set to the cadence of each subject so that a consistent cadence for each subject occurred for each trial. Data from three walking trials for each leg, for a total of six trials, were collected for each subject during the first testing session. Each subject was then fitted with a heel lift equal to the amount of LLD up to a maximum thickness of 12.7 mm. For those subjects with LLDs greater than 6.4 mm, lifting was performed in gradual increments as recommended by previous authors.[5,6] After wearing the heel lift for 21 days, each subject returned for additional testing. The data-collection process for the second testing session was identical to that of the first session, except that all subjects wore their heel lifts during all trials of the second session. Metronome settings used during the second testing session for each subject were identical to those settings used in the first testing session. Brief visual postural inspection along with iliac crest height determination at the pelvic level was performed on each subject prior to the heel lift wearing period to assess pelvic symmetry while the subject was wearing the heel lift. Fourteen subjects whose LLDs were 12.7 mm or less achieved level iliac crest height when wearing the heel lift. The remaining 4 subjects with LLDs greater than 12.7 mm did not achieve a totally level pelvis, but the amount of asymmetry in iliac crest height was reduced because of the thickness of the heel lift.

Data Analysis

Maximum force values and impulse values were calculated for each of the force-time curves (fore-aft, medial-lateral, vertical) measured.[24] For vertical force, maximum vertical loading rate (MVLR) was also calculated. The MVLR is the point of greatest slope as determined by a five-point moving average of all data points on the initial slope of the vertical force-time curve.[25] Mean data values for each set of three trials for each leg were analyzed by use of a two-factor within-subject analysis of variance for repeated measures. Tukey's Honestly Significant Difference (HSD) test post hoc comparisons were then performed to detect significant differences between all means for each experimental condition: 1) short leg, no heel lift (SLNL); 2) long leg, no heel lift (LLNL); 3) short leg, heel lift (SLL); and 4) long leg, heel lift (LLL). To determine the reliability of the force platform contact periods for each subject during both testing sessions (before heel-lift fitting and with heel lift), we calculated intraclass correlation coefficients (ICC, Type 3), according to the formula of Shrout and Fleiss,[26] for the force-plate contact times for all experimental trials over both testing days for each subject.

Results Three maximum force variables--maximum lateral force (MAX LAT), maximum vertical force (MAX VERT), and maximum medial force (MAX MED)--exhibited statistically significant differences between the experimental conditions. No significant differences were found for any of the impulse variable values or for the MVLR variable. The ICC value for force-plate contact times for all 18 subjects was .75. Maximum lateral force was significantly greater (p [is less than] .05) in the SLNL condition than in the LLNL condition (Tabs. 2,3). Without the heel lift, MAX LAT was 6.6 N greater in the foot of the short leg than in that of the long leg. With the heel lift, differences in MAX LAT between the short leg and the long leg were not significantly different. Maximum vertical force was significantly greater (p [is less than] .05) in the SLL condition compared with the SLNL condition and in the LLL condition compared with the LLNL condition (Tabs. 4, 5). Without the heel lift, there was a minimal difference in MAX VERT between the short leg and the long leg. With the heel lift, MAX VERT increased by 19.6 N within the short leg and by 36.0 N within the long leg. The MAX VERT variable also demonstrated a significant interaction (p [is less than] .05) (Fig. 1). Maximum medial force was significantly greater (p [is less than] .05) in the LLL condition compared with the SLL condition (Tabs. 6, 7). Without the heel lift, there was no significant difference in MAX MED between the short leg and the long leg. With the heel lift, MAX MED was 6.7 N greater in the foot of the long leg than in that of the short leg.

Discussion The differences in MAX LAT values for each experimental condition (Tab. 3) are consistent with information previously described in the literature. A major compensatory mechanism for LLD is increased pronation of the foot of the long leg and increased supination of the foot of the short leg.[5,9,18] Increased pronation results in a relative shortening of the long leg, whereas increased supination results in a relative lengthening of the short leg.[27] The pronatory force occurs in the medial direction, and the supinatory force occurs in the lateral direction.[18] In this study, a greater lateral force was occurring in the foot of the short leg during walking without the presence of a heel lift. After the addition of the heel lift, the MAX LAT values decreased slightly in the short leg and increased slightly in the long leg. This normalization of MAX LAT or supinatory force value after the introduction of a heel lift has been documented previously.[18] Values for MAX VERT increased significantly in both the short leg and the long leg after the installation of the heel lift (Tab. 5). In addition, the increase in MAX VERT during the LLL condition was much greater than the increase during the SLL condition. This finding accounts for the significant interaction shown in Figure 1. On the basis of previous studies, these results were surprising. Prior to the heel-lift fitting, we expected greater MAX VERT values within the short leg than within the long leg. After the heel-lift fitting, we expected normalization or equaling of the MAX VERT values within both legs. The results show a very small difference in MAX VERT values between both legs prior to heel-lift fitting. After heel-lift fitting, the difference was greater between the short leg and the long leg, but the difference was not statistically significant. There was, however, a statistically significant increase in MAX VERT values within each leg after heel-lift fitting. Clinical observations of increased shoe wear on the short limb have been described previously, and these increases were reported to be due to the greater impact forces occurring at the heel of the short leg.[5] In contrast, greater heel forces within the long leg during the beginning of the stance phase of walking have also been reported, with normalization of these vertical forces between the short leg and the long leg after heel-lift fitting.[17] The results of this study do not support any of the findings previously reported. A possible explanation for the increase in MAX VERT values within the short leg after the addition of the heel lift may be based on the material used to manufacture the heel lifts. Each heel lift was made of a thermoplastic cork-latex material that has a density of 0.54 g/ [cm.sup.3]. This density classifies the heel-lift material as being rigid in nature.[28] We found no literature regarding the effect of different shoe sole materials on ground reaction forces during walking. In running, however, an increase in density or hardness of the shoe heel and sole has been shown to result in an increase in impact (heel) forces.[25] On the vertical force-time curve depicted in Figure 2, the first peak is an indication of the impact force at and immediately following heel-strike as the center of gravity of the body is elevated.[24] In this investigation, the MAX VERT value in the short leg occurred at the first peak of the two vertical force-time curve peaks in 70.4% of the SLNL condition trials and in 74.1% of the SLL condition trials (Tab. 8) These values are nearly identical. Because the stance phase duration times for the SLNL and the SLL conditions also are nearly identical (Tab. 9), we theorized that the increase in MAX VERT values within the short leg was due to the only variable that changed within the short leg--the addition of the heel lift. In this study, although the subjects were walking and although the heel lift was added inside the shoe, the introduction of a rigid material beneath the foot could have effectively increased the overall or combined density of all material (heel lift and shoe heel) between the foot and the ground. This increased density could account for the increased MAX VERT value during the SLL condition. The increased MAX VERT value in the LLL condition is likely based on the stance phase duration time of the long leg. Increased speed of walking has been shown to result in increased MAX VERT values, with all other variables being constant.[29] As previously described, we attempted to control the walking velocity of each subject by the use of a metronome. Subsequently, the reliability of the force-plate contact times for each subject for all trials in both testing sessions was determined to be .75. Some variations in the durations of the force-plate contact times (stance phase durations), therefore, were present. Note that the smallest stance phase duration time occurred in the LLL condition (Tab. 9). The longest stance phase duration time also occurred within the same limb, in the LLNL condition. The faster the walking speed, the smaller the stance phase duration time. The large, but not statistically significant, difference in stance phase duration between the LLNL and LLL conditions (based on the Tukey's HSD test results) could account for the increase in MAX VERT values within the long leg after installation of the heel lift. In the presence of a LLD, the long leg has been reported to have a greater stance time during walking than the short leg.[30] A decrease in stance time of the long leg after fitting of a heel lift in the shoe of the short leg has also been reported.[18] The pattern of stance phase durations in the LLNL and LLL conditions of this investigation are consistent with those studies previously reported. Medial force has been described as being directly related to the anatomical foot function of pronation.[31] A compensatory mechanism of increased pronation in the foot of the long leg prior to heel-lift fitting has also been described.[18] The data from this study do not support the presence of this compensatory mechanism because the MAX MED values (pronation) for the short leg and the long leg prior to heel-lift fitting are very similar. After the heel lift was installed, the MAX MED in the foot of the short leg decreased slightly, whereas the same force increased significantly in the foot of the long leg. Maximum medial force has been shown to be velocity dependent; a decreased stance phase time will result in an increase in MAX MED.[29] The LLL condition was also described earlier as having the smallest stance phase duration of all the experimental conditions (Tab. 9). The smaller stance phase duration in the LLL condition is the most likely reason for the greater MAX MED values for that condition. This result also contradicts a previous study that reported a reduction in medial (pronatory) force in the foot of the long leg after heel-lift fitting in the foot of the short leg.[18] Differences among the results of this investigation and those of previous studies cited may be due to the different experimental methods used. This study is the first to use a fixed force platform for the measurement of ground reaction forces in subjects with structural LLDs. Previously cited studies used small electronic sensors placed on the plantar surface of subjects' feet. Increased wear on the heel and sole of the shoe of the short leg has been documented, along with the theory that this increased wear was due to greater impact forces at heel-strike.[5] The results of this study suggest that increased lateral forces in the foot of the short leg may be responsible for the increased shoe wear because the MAX VERT values in both legs before heel-lift fitting were essentially equal. Further research in this area is warranted. The significant differences in MAX LAT AND MAX MED values between the short leg and the long leg are indicative of differences in supination and pronation that are occurring within the feet of the short leg and the long leg. Because the installation of a heel lift places the forefoot and the midfoot in a plantar-flexed position relative to the hindfoot, additional investigations should be undertaken to determine the effect of such positioning on force distribution within the foot. Center-of-pressure measurements may provide more detailed information in this regard. Perhaps optimum correction of a structural LLD will require functional orthoses in the shoes of both the short leg and the long leg, with the added leg-length correction built into the orthosis of the foot of the short leg. Finally, this study used a within-subject design, thereby permitting each subject to serve as his or her own control. The primary reason for this design was that each subject wore his or her own recreational or running shoes throughout the investigation. Additional research using similar subjects with standardized recreational or running shoes would permit between-subject comparisons of all force data obtained. Use of this design would also allow comparisons of varying leg-length discrepancy values with the resulting ground reaction force values.

Conclusions Ground reaction force patterns were recorded for 18 male and female subjects with structural LLDs. Prior to heel-lift fitting, MAX LAT (supination) was greater in the foot of the short leg than in the foot of the long leg. After heel-lift fitting, MAX VERT increased in both legs, and MAX MED (pronation) was greater in the foot of the long leg than in the foot of the short leg. The results of this study are important to any clinician who treats patients with structural leg-length discrepancies. The primary reason for the use of a heel lift is to level the pelvis and the sacral base. This is a desired outcome because normal pelvic and spinal alignment reduces the erosion of these structures that results from asymmetry. The results of this study confirm that normal or near-normal pelvic levelness was achieved by the use of heel lifts. The clinician should be aware that the joints of the lower extremities may undergo additional wear because of the increased ground reaction forces described in this study. The results of this study also raise questions as to whether the use of a heel lift results in an alteration of the normal biomechanical function of the foot under which it is placed. In patients with structural LLDs, perhaps functional interventions in addition to heel lifting may be required for proper foot function to occur. Further research in this regard is needed. [Tabular Data Omitted] [Figure 1 to 2 Omitted]

(*)Kistler Instrument Corp, 75 John Glenn Dr, Amherst, NY 14120. ([dagger])Hewlett-Packard Corp, PO Box 3640, Sunnyvale, CA 94088-3640. ([double dagger])Gould Inc, Instrument Systems Div, 3631 Perkins Ave, Cleveland, OH 44114.

References [1]Cathie AC: The influence of the lower extremities upon the structural integrity of the body. J Am Osteopath Assoc 49:443-446, 1950 [2]Jones CL: The damaging effects of a disaligned musculoskeletal system. J Am Podiatr Assoc 61:369-381, 1971 [3]Klein KK, Redler I, Lowman CL: Asymmetrics of growth in the pelvis and legs of children: A clinical and statistical study, 1964-1967. J Am Osteopath Assoc 68:153-161, 1968 [4]Pearson WM: A progressive structural study of school children. J Am Osteopath Assoc 51:155-166, 1951 [5]Porterfield JA: The sacroiliac joint. In Gould JA, Davies GJ (eds): Textbook of Physical Therapy: Orthopaedics and Sports. St Louis, MO, C V Mosby Co, 1985, pp 550-580 [6]Subotnick SI: The short leg syndrome. J Am Podiatr Assoc 66:720-723, 1976 [7]Bailey HW: Theoretical significance of postural imbalance, especially the "short leg." J Am Osteopath Assoc 77:452-455, 1978 [8]Beal MC: A review of the short leg problem. J Am Osteopath Assoc 50:109-121, 1950 [9]Botte RR: An interpretation of the pronation syndrome and foot types of patients with low back pain. J Am Podiatr Assoc 71:243-253, 1981 [10]Kendall HO, Kendall FP, Boynton DA: Posture and Pain. Baltimore, MD, Williams & Wilkins, 1952, pp 103-151 [11]Amstutz HC, Sakai D: Equalization of leg length (editorial comment). Clin Orthop 136:2-6, 1978 [12]Eggleston EE: Abnormalities of the lower extremities in relation to the unstable low back. J Am Osteopath Assoc 43:112-114, 1943 [13]Johnson JW: Sacroiliac strain. J Am Osteopath Assoc 63:1015-1029, 1964 [14]Pratt WA: The lumbopelvic torsion syndrome. J Am Osteopath Assoc 51:335-341, 1952 [15]Stoddard A: A Manual of Osteopathic Practice, ed 2. London, England, The Hutchinson Publishing Group Ltd, 1983, pp 108-109 [16]Schuit D, McPoil TG, Moe P: Sacroiliac joint malalignment in the presence of functional and structural leg length discrepancies. J Am Podiatr Med Assoc, to be published [17]Beekman S, Louis H, Rosich JM, et al: A preliminary study on asymmetrical forces at the foot to ground interphase. J Am Podiatr Med Assoc 75:349-354, 1985 [18]D'Amico JC, Dinowitz HD, Polchaninoff M: Limb length discrepancy: An electro-dynographic analysis. J Am Podiatr Assoc 75:639-643, 1975 [19]McPoil TG, Brocato R: The foot and ankle: Biomechanical evaluation and treatment. In Gould JA, Davies GJ (eds): Textbook of Physical Therapy: Orthopaedics and Sports. St Louis, MO, C V Mosby Co, 1985, pp 314-341 [20]Woerman AL, Binder-MacLeod SA: Leg length discrepancy assessment: Accuracy and precision in five clinical methods of evaluation. Journal of Orthopaedic and Sports Physical Therapy 5:230-239, 1984 [21]Bolz S, Davies GJ: Leg length differences and correlation with total leg strength. Journal of Orthopaedic and Sports Physical Therapy 6:123-129, 1984 [22]Denslow JS, Chase JA, Gutensohn OR, et al: Methods in taking and interpreting weight-bearing x-ray films. J Am Osteopath Assoc 54:663-670, 1955 [23]Bielke MC: Roentgenological spinal analysis and the technic for taking standing x-ray plates. J Am Osteopath Assoc 35:414-418, 1936 [24]Mann RA: Biomechanics. In Jahss MH (ed): Disorders of the Foot. Philadelphia, PA, W B Saunders Co, 1982, pp 37-67 [25]Nigg BM: Experimental techniques used in running shoe research. In Nigg BM (ed): Biomechanics of Running Shoes. Champaign, IL, Human Kinetics Publishers Inc, 1986, pp 27-61 [26]Shrout PE, Fleiss JL: Intraclass correlations: Uses in assessing rater reliability. Psychol Bull 86:420-428, 1979 [27]Okun SJ, Morgan JW, Burns MJ: Limb length discrepancy: A new method of measurement and its clinical significance. J Am Podiatr Assoc 72:595-599, 1982 [28]McPoil TG: The Effect of Foot Supports on Ground Reaction Force Patterns. Doctoral Dissertation. Urbana, IL, University of Illinois at Urbana-Champaign, 1987 [29]Andriacchi TP, Ogle JA, Gallante JO: Walking speed as a basis for normal and abnormal gait measurements. J Biomech 10:261-268, 1977 [30]Blustein SM, D'Amico JC: Limb length discrepancy: Identification, clinical significance, and management. J Am Podiatr Med Assoc 75:200-206, 1985 [31]Bates BT, Osternig LR, Sawhill JA, et al: An assessment of subject variability, subject-shoe interaction, and the evaluation of running shoes using ground reaction force data. J Biomech 16:181-191, 1983

D Schuit, PhD, PT, is Assistant Professor, Department of Physical Therapy, College of Associated Health Professions, University of Illinois at Chicago, 4th Floor, 1919 W Taylor St, Chicago, IL 60612 (USA). M Adrian, PED, is Professor of Physical Education and Director of the Biomechanics Research Laboratory, Department of Kinesiology, College of Applied Life Studies, University of Illinois at Urbana-Champaign, Urbana, IL 61801. P Pidcoe, MS, is Bioengineer and Research Associate, Department of Physical Therapy, College of Associated Health Professions, University of Illinois at Chicago.