Cane-Assisted Gait Biomechanics and Electromyography After Total Hip Arthroplasty

1966 Cane-Assisted Gait Biomechanics and Electromyography After Total Hip Arthroplasty Stan Ajemian, MSc, Deb Thon, BSc PT, Peter Clare, BSR(PT), MA, Lauri Kaul, BSc PT, Ronald F. Zernicke, PhD, Barbara Loitz-Ramage, PhD, PT ABSTRACT. Ajemian S, Thon D, Clare P, Kaul L, Zernicke RF, Loitz-Ramage B. Cane-assisted gait biomechanics and electromyography after total hip arthroplasty. Arch Phys Med Rehabil 2004;85:1966-71. Objective: To quantify the effects of cane use during walking on hip joint kinematics, kinetics, and muscle activity patterns after unilateral total hip arthroplasty (THA). Design: Nonrandomized experimental design. Setting: Urban inpatient hospital. Participants: Adults (n 9 men, 2 women) with no history of orthopedic or neuromuscular disease who underwent elective unilateral THA. Intervention: Gait was assessed preoperatively and 4 and 8 months postoperatively. Main Outcome Measures: Three-dimensional hip joint motion and moments and electromyographic patterns of gluteus medius, tensor fascia latae, lateral hamstring, and vastus lateralis were measured during level walking, with and without use of a straight cane. Results: When a cane was held in the contralateral hand, the abduction moment of the affected hip decreased by 26%, whereas that of the contralateral hip increased by 28%. Use of a cane in THA rehabilitation is important because it reduces the load on the operative hip so that bone and soft tissues can heal. Our results suggest that load reduction was successful on the operative side, but the loads on the contralateral side were increased. Conclusions: After unilateral arthroplasty, subjects using a cane had increased hip abduction moments on the nonoperative hip and decreased hip abduction moments on the operative hip. Clinicians should be mindful of the effects of cane use on the contralateral hip. Key Words: Arthroplasty, replacement, hip; Cane; gait; Electromyography; Rehabilitation. 2004 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation From the Faculties of Kinesiology (Ajemian, Zernicke) and Medicine and Engineering (Zernicke) and McCaig Centre for Joint Injury and Arthritis Research (Zernicke, Loitz-Ramage), University of Calgary; and Department of Physical Therapy, Foothills Hospital (Thon, Clare, Kaul), Calgary, AB, Canada. Supported by the Arthritis Society of Canada and Hip-Hip Hurray (Canadian Orthopaedic Foundation). No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the author(s) or upon any organization with which the author(s) is/are associated. Correspondence to Barbara Loitz-Ramage, PhD, PT, McCaig Centre for Joint Injury and Arthritis Research, 3330 Hospital Dr NW, Calgary, AB T2N 4N1, Canada, e-mail: ramage@ucalgary.ca. Reprints are not available from the authors. 0003-9993/04/8512-8849$30.00/0 doi:10.1016/j.apmr.2004.04.037 ASSISTIVE DEVICES are often used in the management of lower-extremity pain or weakness, and in some cases they provide enough assistance to make a difference between functional and nonfunctional ambulation. Canes have also been shown to reduce fall frequency among the elderly. 1 Among patients awaiting total hip arthroplasty (THA), Mont et al 2 reported that 50% of the patients they studied used 1 or 2 canes or crutches. In a similar population, McBeath et al 3 noted that only 29% of patients in their study could walk without an assistive device. While the preoperative use of assistive devices is common, postoperative use presents a dilemma to health care practitioners because most patients strive for unassisted ambulation. People with lower-extremity instability typically use a cane on the ipsilateral side, whereas persons with pain or weakness use the cane in the side opposite (contralateral) the affected extremity. 1,4-7 Contralateral use encourages a normal reciprocal gait pattern and reduces the hip contact forces more effectively than ipsilateral use. 4,6,7 The reduction in hip contact forces results primarily from the decreased hip abductor muscle force required to balance the pelvis during unilateral stance. Brand and Crowninshield 8 calculated that when patients with unilateral hip disease used a cane in the contralateral hand, hip contact forces decreased by 56% when compared with unassisted walking. Although some of the difference may be accounted for by a decreased walking velocity in the caneassisted trials, it is clear that contralateral cane use dramatically decreases joint loads. That was supported by Krebs et al, 9 who reported that, in a subject with an instrumented acetabular component, contralateral cane use reduced the peak acetabular contact pressure and gluteus medius electromyographic activity when compared with unassisted walking. Relative changes in muscle activity and kinetics during cane-assisted gait have been reported. Neumann 10 found a 31.1% decrease in gluteus medius activity during contralateral cane use and a 42.3% decrease in muscle activity when subjects were instructed to push as hard as possible on the cane in the contralateral hand. The data suggested that contralateral cane use effectively reduced demand on the hip abductors, subsequently reducing joint compression forces related to muscle contraction. Despite those findings, no report has described the muscle activity or joint kinetics on the side ipsilateral to the cane. Bilateral disease is common in people with hip osteoarthritis. If the decreased loading of 1 hip comes at the expense of the contralateral joint, use of a unilateral cane may need to be reconsidered. In this study, we describe the effect of unilateral cane use on bilateral joint kinematics, joint moments, and hip abductor muscle activity in adults preoperatively and 4 and 8 months after THA. METHODS Participants We recruited 14 patients scheduled for unilateral THA. Subjects were excluded if they had previous surgery on either hip; recent trauma that precipitated the scheduled total hip replacement surgery; pain or abnormality of the knee or ankle of the operative limb; pain or abnormality of the nonoperative leg; a

HIP MECHANICS AFTER JOINT ARTHROPLASTY, Ajemian 1967 cardiopulmonary disease that placed them at risk during data collection; or a neurologic disease manifesting as lack of balance, decreased sensation, or decreased strength. All subjects were ambulatory with and without an assistive device preoperatively. Recruitment and experimental procedures were approved by the institutional committee for ethical use of human subjects in research. The subjects were selected because they represented an active, healthy population presenting only with unilateral hip degeneration. By limiting the sample to these subjects, other confounding variables that could influence gait, such as increased age or frailty, were avoided. Subjects were examined from 0 to 22 days before surgery, and at 4 and 8 months postoperatively. Each testing session included a clinical assessment (standard manual muscle test, range of motion [ROM] measurement, interview) and gait analysis. Gait Analysis Subjects walked the length of a 6-m walkway wearing their own shoes and at a self-selected speed. Three types of data were collected during each trial: high-speed video, ground reaction forces, and electromyographic activity of selected lower-extremity muscles. To avoid encumbrance caused by the electromyography electrodes, trials from the affected and contralateral extremities were collected separately, with all the trials from 1 side completed before trials began with the other side. Six trials were collected from each side, with 3 being cane-assisted and 3 being independent. All subjects used the same standard straight cane, adjusted such that the handgrip was level with the wrist joint when the subject stood with his/her arm in a pendant position. The cane was grasped in the hand contralateral to the affected hip. A trial was accepted if a subject s foot struck entirely within the forceplate without the subject modifying his/her gait and the cane, if used, did not touch the plate. Kinematics High-speed video data were collected with a 2-camera video digitizing system a that captured images of reflective markers placed on the subject. Cameras were placed anterolateral to the subject, with a 45 angle between the cameras. Images from the cameras were synchronized and digitized at 200Hz. Three reflective markers were placed on each lower-extremity segment to record the 3-dimensional limb motion: greater trochanter, anterior mid-thigh, lateral femoral epicondyle, fibular head, anterior mid-tibia, lateral malleolus, superior navicular, lateral calcaneus, and the head of the fifth metatarsal. Subjects wore a pelvic belt to which 3 markers were attached. The belt fit snugly and was positioned with the markers overlying the anterior superior, posterior superior, and posterior inferior iliac spines. From an image of the subject standing and measured distances from the palpated joint center to each marker, marker position with respect to the joint coordinate system axes was computed. b Kinetics Ground reaction forces were measured using a piezoelectric forceplate c embedded in the middle of the walkway. The raw forceplate signal was amplified and digitized d at 1kHz. Vertical, anteroposterior (AP), and mediolateral (ML) force vectors were calculated from the raw force channels and filtered with a zero-lag low-pass Butterworth filter at a cutoff frequency of 200Hz. All forces were expressed in terms of the subject s body weight. Two peaks in the vertical ground reaction force were identified (fig 1). The first peak, or landing peak (Fz 1 ), represents Fig 1. Typical vertical ground reaction force curve. Labels represent landing peak (Fz 1 ), pushoff peak (Fz 2 ), minimum during midstance (Fz min ), yield range (FzR 1 ), and pushoff range (FzR 2 ). Abbreviation: BW, body weight. weight acceptance on the limb. The second peak, or pushoff peak (Fz 2 ), represents pushoff onto the opposite limb. A minimal vertical force (Fz min ) occurred between the 2 maximal peaks. The difference between Fz 1 and Fz min was identified as the yield range (FzR 1 ); change in force from Fz min to Fz 2 is the pushoff range (FzR 2 ). Internal joint moments were calculated using an inverse dynamics model of the lower limb that assumed the limb to be a set of 3 rigid ellipsoids linked by freely rotating joints. 11 Input to the model included ground reaction force magnitude and direction, position, velocity, and acceleration of the limb segments, joint centers of rotation, and estimates of the limb segment masses 12 and inertial properties. 13 Electromyography Surface electrodes recorded the activity, bilaterally, of the gluteus medius, tensor fascia lata, and erector spinae and the lateral hamstring and rectus femoris of the operative extremity. Before electrode placement, skin was shaved and cleansed with isopropyl alcohol. Surface electrodes were attached to a signal multiplexer contained within a belt pouch worn by the subject. The signal was then transmitted over an 11-m cable to an amplifier, e where it was amplified to a 5-V range. The amplified signal was digitized at 1kHz d and stored on a personal computer. Before further processing, movement artifacts were removed by passing the data through a high-pass filter at a cutoff frequency of 25Hz. For each subject, time-normalized, filtered data were averaged across the 3 trials for each condition. Electromyographic activity was considered on when the amplitude exceeded 15% of the maximal filtered signal for that session. To eliminate isolated spikes, the time to cross the threshold voltage was recorded only if the activity remained on or off for more than 10% of stance. 14 The total duration for which the muscle was considered on during the stance phase was calculated by dividing stance into 10% time intervals (bins) and then assigning a value indicating on to any bin during which muscle activity occurred. Because day-to-day and subject-to-subject differences in electrode placement precluded comparison of electromyographic amplitude across testing sessions, no attempt was made to interpret differences in electromyographic amplitude within or between subjects.

1968 HIP MECHANICS AFTER JOINT ARTHROPLASTY, Ajemian Limb and Test Session Table 1: Kinematic Variables for the Hip, Knee, and Ankle Joints During the No-Cane Trials Peak Hip Flexion Total Hip Frontal ROM Knee Flexion at Heel Strike Knee Extension Midstance Knee Flexion at Toe-Off Ankle Dorsiflexion at Heel Strike Ankle Plantarflexion at Toe-Off Affected Preoperative 23.5 2.2 7.7 1.0 3.6 1.6 6.5 1.7 52.1 1.4 10.1 1.2 30.7 2.2 4mo 27.8 2.5 10.0 1.8 2.8 2.8 10.1 3.4 54.2 4.1 10.1 1.3 31.9 2.3 8mo 26.7 2.1 8.5 2.0 0.4 1.8 7.5 2.2 54.4 3.4 11.0 1.4 32.9 2.3 Unaffected Preoperative 27.2 1.4 8.5 2.0 6.1 5.0 14.7 3.8 60.3 4.1 12.1 2.0 27.9 2.9 4mo 29.1 3.7 8.9 1.7 1.7 4.2 15.3 4.0 60.0 4.9 10.5 1.5 30.0 2.3 8mo 30.0 2.6 9.4 2.5 3.3 4.7 16.2 4.2 61.7 5.0 11.1 1.1 31.7 2.5 NOTE. Values are mean SD in degrees. Statistical Analyses Kinematic, ground reaction force, and joint moment data were averaged for the 3 trials within each condition. The average of the 3 trials for each subject was then used in all statistical tests. To assess the effect of the cane, a 3-way repeated-measures analysis of variance (ANOVA) was performed with session (preoperative, 4-wk postoperative, 8-wk postoperative), limb, and cane as the main effects. f Contrasts were performed to determine where differences lay within the session main effect, and simple effects were tested for variables in which significant interaction existed. An level of.05 was used to determine significance for all ANOVA tests. Given the small sample size, a procedure that would account for multiple comparisons, such as a multivariate analysis of variance, was not used. RESULTS Of the original 14 subjects, 11 (9 men, 2 women) completed the 3 sessions of data collection. Two subjects withdrew before the second data collection session and a third subject withdrew before the third session. All 3 withdrew for reasons unrelated to their gait or surgery. The mean age standard deviation (SD) of the subjects who completed data collection was 62.6 8.6 years. Kinematics The free speed of walking increased over time, with a significant difference noted between the preoperative (.95.16m/s) and 8-month postoperative (1.08 0.09m/s) trials. When the subjects walked with a cane, velocity decreased 3% compared with when the same subjects walked independently (P.07). During cane-assisted walking, hip sagittal plane ROM showed both a significant limb effect (affected less than the contralateral) and a session effect (preoperative less than at either 4- or 8-mo postoperative). ROM of the contralateral hip during cane-assisted walking did not change during the study, whereas ROM of the affected hip increased significantly from the preoperative to 4-month postoperative sessions. At all time points studied, ROM of the contralateral hip significantly exceeded that of the affected hip during cane-assisted walking (table 1). When walking without a cane, subjects demonstrated no differences in peak hip flexion angle between the affected and contralateral limbs. Peak hip flexion did not differ among the preoperative and postoperative trials. However, at all sessions, the contralateral hip achieved a significantly greater extension angle than the affected hip (fig 2). Kinetics Vertical ground reaction forces. Neither the landing (Fz 1 ) nor the pushoff peak (Fz 2 ) differed between testing sessions, but there was a significant main effect of limb and cane. Fz 1 and Fz 2 for the contralateral limb significantly exceeded Fz 1 and Fz 2 of the affected limb in both the cane and no-cane trials. For the affected limb, use of a cane reduced Fz 1 by 7% and Fz 2 by 9%, compared with the no-cane trials; the cane had no effect on the vertical forces for the contralateral limb. Fz min was similar between the affected and contralateral limbs when a cane was not used. Use of the cane significantly decreased Fz min for the affected limb, but not for the contralateral limb. For both limbs, Fz min decreased at the 4- and 8-month postoperative trials compared with the preoperative trials. Yield range (FzR 1 ) and pushoff range (FzR 2 ) were significantly less for the affected side than for the contralateral side. Both ranges increased slightly after surgery. A significant session-cane interaction occurred for FzR 1, but there were no main effects. FzR 2 showed a significant interaction between the limb and cane. The pushoff range on the affected limb decreased when the cane was used (P.07) and was significantly less than FzR 2 for the contralateral limb for cane and no-cane trials. AP ground reaction forces. The anterior and posterior shear components of the ground reaction forces were significantly smaller on the affected limb than on the contralateral limb. There was a significant interaction for the anterior shear between the limb and session and the limb and cane effects. Simple effects showed a significant increase in anterior shear for the affected limb over the 8-month protocol and no change in the anterior shear on the contralateral side. The affected limb Fig 2. Maximal hip extension (mean SD) during gait. The hip extension angle was significantly smaller (P<.05) for the affected limb than the contralateral limb at each time point.

HIP MECHANICS AFTER JOINT ARTHROPLASTY, Ajemian 1969 Limb and Test Session Table 2: Peak Internal Moments for the Hip, Knee, and Ankle Joints for the No-Cane Trials Hip Flexion Hip Extension Knee Flexion Knee Extension Knee Varus Ankle Plantarflexion Affected Preoperative 13.7 2.1 11.9 2.4 5.9 1.0 11.5 2.6 5.1 1.4 5.6 1.7 4mo 12.4 2.0 13.9 1.5 8.5 1.0 10.0 3.0 3.9 1.7 7.2 0.8 8mo 14.7 1.9 11.6 2.7 6.5 1.5 11.6 1.4 3.7 0.8 5.1 1.2 Unaffected Preoperative 12.6 1.8 15.3 1.2 7.6 1.0 8.0 1.5 6.5 1.4 8.7 1.3 4mo 12.9 2.7 18.4 2.5 10.0 1.2 9.9 3.9 4.1 1.2 10.8 1.8 8mo ND 17.9 3.5 9.5 1.9 8.4 2.4 4.6 1.1 10.3 2.5 NOTE. Values are mean SD in Nm BW 10 2. Abbreviation: ND, no data. had a significantly smaller anterior shear than did the contralateral limb preoperatively, but by 4 months postoperatively, the difference was no longer significant. The simple effects test for the limb versus cane interaction showed a significant reduction in the anterior shear on the affected side with use of a cane, and no effect on the contralateral side. The posterior shear component was 3% smaller with use of a cane than without (P.05). ML ground reaction forces. The medial shear force increased significantly on the contralateral limb between the preoperative and 8-month postoperative trials. No session effect was noted with the affected limb. In both postoperative trials, the medial shear force was greater for the contralateral limb compared with the affected limb. The lateral shear force decreased by 15% (P.05) during the cane trials, with no difference noted across time or between limbs. Joint moments. Sagittal-plane internal moments for the hip, knee, and ankle joints differed between the affected and contralateral limbs and across trials (table 2). Hip extension and ankle plantarflexion moments of the affected limb were significantly less than those of the contralateral limb, whereas the knee extension moment of the affected limb significantly exceeded that of the contralateral limb. Peak knee flexion moment tended to be smaller on the affected limb, although the difference was not significant. Use of the cane decreased the knee extension moment (P.02) and the hip flexion moment (P.06). Maximal hip abduction moment did not change over time, but there was a significant limb-cane interaction. Use of the cane significantly decreased the hip abduction moment on the affected limb by 26% and increased the abduction moment of the contralateral hip by 28% (fig 3). Without the cane, the hip abduction moment was larger on the affected limb (P.07). When the cane was used, hip abduction moment decreased on the affected limb to less than that of the contralateral limb. Muscle Activity Preoperatively, 79% of the subjects had some gluteus medius activity on the affected side from prestance to 60% of stance, with 14% of subjects having activity at 80% of stance. At 4 months postoperatively, 71% had activity from prestance to 60% of stance and 7% had activity continuing to 80% of stance. Eight months postoperatively, 82% of subjects had activity to 60% of stance, whereas in only 9% of subjects did the activity continue through 80% of stance. The tensor fascia latae on the affected limb demonstrated a similar timing pattern, with 86% of subjects showing activity from heel strike to 60% of stance and 73% demonstrating activity from 80% of stance to toe-off. Duration of gluteus medius and tensor fascia latae activity did not differ significantly over time or between limbs during the no-cane trials. The gluteus medius of the contralateral limb was active for 11% more of the stance cycle at 4 months postoperatively compared with the preoperative trial, whereas the duration of muscle activity on the affected limb remained unchanged across trials. At the 8-month session, the side-toside differences in gluteus medius duration remained the same. By 4 months postoperatively, tensor fascia latae duration decreased by 7% of stance on the contralateral side and increased by 6% of stance on the affected limb. By 8 months postoperatively, tensor fascia latae duration differed between the limbs by only 2% of stance. Gluteus medius duration of the affected hip decreased significantly when the cane was used during the 8-month postoperative trials. Prior to this testing session, no differences in the duration of muscle activity occurred between the cane and no-cane trials. Use of the cane also significantly reduced the duration of tensor fascia latae activity on the affected limb. DISCUSSION Canes are used in postarthroplasty rehabilitation to reduce the force on the prosthetic hip and the incised abductor muscles by reducing the abductor muscle force required for pelvic stability. A cane held in the contralateral hand may also prevent a lurching gait by reducing weight-bearing pain and assisting weak abductor muscles. Postoperatively, if patients are encouraged to walk without a cane before adequate hip abductor strength has been regained, the abnormal preoperative lurching pattern may reappear. The most important finding of this study was that when a cane was used in the contralateral hand, abduction moment of Fig 3. Maximal hip abduction moment (mean SD). *Differences between affected and contralateral limbs (P<.07). A cane significantly decreased the affected hip abduction moment.

1970 HIP MECHANICS AFTER JOINT ARTHROPLASTY, Ajemian the affected hip decreased by 26% while that of the contralateral hip increased by 28%. An important role of a cane in THA rehabilitation is to reduce the load on the hip and thereby permit healing of the bone and soft tissues. Our results suggest that the reduction was successful on the affected side, but the abductor hip moment on the contralateral side was greater. Although our subjects had only unilateral hip disease, the increased moments on the nonoperative hip may hasten degeneration of the heretofore asymptomatic joint. All subjects could walk without a cane or other walking aid at all testing sessions. Thus, differences between cane and no-cane trials may be less dramatic than in subjects who have difficulty walking without an assistive device. When using a cane, subjects walked 3% slower than when walking unassisted. Brand and Crowninshield 8 reported a 23% slower walking speed among THA patients before surgery. Their patients walked without a cane at.57m/s, in contrast to our patients, who walked at.95m/s preoperatively. The differences between the Brand and Crowninshield data and the our data may be accounted for by the degree of joint degeneration and concomitant pain in the respective study populations. At the 8-month interval, our subjects walked unassisted at an average speed of 1.08m/s, consistent with the 1.02m/s speed reported by Kyriazis and Riga 15 in patients 12 months after hip arthroplasty. Kleissen et al 16 reported a case study of a THA patient who walked 4% slower when using a cane. Thus, cane use appears to reduce walking speed. This is important to know when data are compared across studies, because walking speed influences gait kinetics and kinematics. Our subjects represented a population of active, fast-walking adults and our results must be considered in this light. A study quantifying cane-related changes in gait biomechanics in an older, less active group of adults would be appropriate. By using a cane, subjects decreased vertical and shear ground reaction forces on the affected limb. The peak landing (Fz 1 ) and pushoff (Fz 2 ) vertical ground reaction forces were reduced 7% and 9%, respectively, with use of a cane. Those decreases were expected because the cane transmitted load during the stance phase of the affected limb, thereby reducing the load on the limb. 5 Cane use had no effect on the contralateral limb because it was not in contact with the ground during contralateral stance. The cane also significantly reduced the braking shear (anterior) and propulsive shear (posterior) ground reaction forces for the affected limb, possibly by transmitting some of those forces as well. The lateral shear was 15% smaller for both limbs during the cane trials. The cane may have decreased the lateral oscillation of the center of mass, which could have reduced the lateral shear force. Other researchers have reported decreased limb loading during cane-assisted gait. Ely and Smidt 5 reported that for patients with a limp because of a hip disability,.15% body weight applied to a cane in the contralateral hand reduced the maximal vertical reaction forces from 100% to 89% body weight. Although speed remained the same for gait with or without a cane, cane-assisted gait had a longer stride length and decreased cadence. McGibbon et al 17 recorded in vivo pressures applied to a femoral endoprosthesis in 1 subject during independent and cane-assisted walking. Seven months postoperatively, a cane used on the unaffected side significantly reduced pressures on the endoprosthesis by as much as 35% (6.51MPa vs 3.88MPa), compared with unassisted walking, whereas a cane used on the affected side had no substantial effect on hip pressures. Given the technology used in their study, pressures on the nonoperative hip could not be measured. Thus, although their data support use of a cane in the contralateral hand to decrease loads on the operative hip, it provides no additional information on loads sustained by the nonoperative hip. In healthy adults, the gluteus medius contracts from ipsilateral heel strike to contralateral heel strike, which occurs in the first 10% to 15% of the stance phase. In our subjects, postoperative cane use decreased the duration of hip abductor contraction. Preoperatively, the gluteus medius duration was greater during cane trials, although the effect was not statistically significant. To produce a larger abduction moment about the contralateral hip than about the affected limb, contralateral hip abductor muscles were likely contracting with greater force; this can only be inferred from our data because of difficulties inherent in predicting muscle force from electromyography data collected from different sites on different days. 18 Long 19 and Kleissen 16 and colleagues have reported a 25% decrease in gluteus medius electromyographic amplitude during cane trials, consistent with the decrease in the internal hip abduction moment we found. Our findings and conclusions examine the effects of cane use in subjects after THA. The preoperative data are also clearly relevant to nonsurgical subjects with hip degeneration. Given that hip degeneration is often bilateral, the implications of single cane use need to be considered in any patient with hip pain. The clinician s challenge becomes one of what other assistive device to prescribe for pain relief during weight bearing without eliciting the loading asymmetry that we found. Ideally, the assistive appliance should promote symmetry, such as bilateral crutches or a walker. However, given the reluctance of most patients to accept an assistive device, this is likely an unacceptable clinical solution despite the data reported here. Limitations of any study need to be considered before drawing conclusions. In this study, the effects of the small sample size of relatively active adults and the lack of correction for multiple comparisons made with the statistical analysis must be considered. Limitations notwithstanding, the changes in electromyographic activity elicited by use of a cane were relatively robust, suggesting that a larger sample might strengthen the conclusions rather than negate them. It is unclear whether older, more sedentary adults would show similar changes in muscle moments because age-related changes in muscle force generation may result in compensations not seen in our sample. Clearly, the confounding effect of the subjects activity levels warrants further exploration. CONCLUSIONS The following conclusions were drawn from our data. First, activity patterns of the muscles controlling the affected and unaffected hip joints were modified during assisted walking. Second, those altered muscle activity patterns may account for the increased hip abduction internal moment on the contralateral hip. Finally, use of a cane on the unaffected side may detrimentally increase loads on the unaffected hip. Thus, there appears to be an optimal postoperative interval during which cane-assisted walking should be encouraged to promote postoperative healing, but sustained cane use after adequate healing should be monitored to avoid overloading the nonoperative hip. Acknowledgment: We thank posthumously Dr. Gary Hughes for his medical expertise. References 1. Dean E, Ross J. Relationship among cane fitting, function, and falls. Phys Ther 1993;73:494-504. 2. Mont MA, Maar DC, Krackow KA, Jacobs MA, Jones LC, Hungerford DS. Total hip replacement without cement for noninflammatory osteoarthrosis in patients who are less than fortyfive years old. J Bone Joint Surg Am 1993;75:740-51.

HIP MECHANICS AFTER JOINT ARTHROPLASTY, Ajemian 1971 3. McBeath AA, Bahrke MS, Balke B. Walking efficiency before and after total hip replacement as determined by oxygen consumption. J Bone Joint Surg Am 1980;62:807-10. 4. Edwards BG. Contralateral and ipsilateral cane usage by patients with total knee or hip replacement. Arch Phys Med Rehabil 1986;67:734-40. 5. Ely DD, Smidt GL. Effect of cane on variables of gait for patients with hip disorders. Phys Ther 1977;57:507-12. 6. Joyce BM, Kirby RL. Canes, crutches and walkers. Am Fam Physician 1991;43:535-42. 7. McConnel EA. Teaching a patient to use a cane correctly. Nursing 1991;21:83. 8. Brand RA, Crowninshield RD. The effect of cane use on hip contact force. Clin Orthop 1980;Mar-Apr(147):181-4. 9. Krebs DE, Robbins CE, Lavine L, Mann RW. Hip biomechanics during gait. J Orthop Sport Phys Ther 1998;28:51-9. 10. Neumann DA. Hip abductor muscle activity as subjects with hip prostheses walk with different methods of using a cane. Phys Ther 1998;78:490-501. 11. Bresler B, Frankel JP. The forces and moments in the leg during level walking. Trans Am Soc Mech Eng 1950;72:27-36. 12. Clauser CE, McConville JT, Young JW. Weight, volume, and center of mass of segments of the human body. Wright-Patterson Air Force Base (OH): Aerospace Medical Research Laboratories; 1969. AMRL Tech Rep 69-70. 13. Dempster WT. Space requirements of the seated operator. Wright- Patterson Air Force Base (OH): Aerospace Medical Research Laboratories; 1955. WADC Tech Rep 55-159. 14. Bogey RA, Barnes, LA, Perry J. Computer algorithms to characterize individual subject EMG profiles during gait. Arch Phys Med Rehabil 1992;73:835-41. 15. Kyriazis V, Rigas C. Temporal gait analysis of hip osteoarthritic patients operated with cementless hip replacement. Clin Biomech (Bristol, Avon) 2002;17:318-21. 16. Kleissen RF, Hermens HJ, den Exter T, de Kreek JA, Zilvold G. Simultaneous measurement of surface EMG and movements for clinical use. Med Biol Eng Comp 1989;27:291-7. 17. McGibbon CA, Krebs DE, Mann RW. In vivo hip pressures during cane and load-carrying gait. Arthritis Care Res 1997;10: 300-7. 18. Basmajian JV, DeLuca CJ. Muscles alive: their functions revealed by electromyography. 5th ed. Baltimore: Williams & Wilkins; 1985. 19. Long WT, Dorr LD, Healy B, Perry J. Functional recovery of noncemented total hip arthroplasty. Clin Orthop 1993;Mar(288): 73-7. Suppliers a. Motion Analysis Corp, 3617 Westwind Blvd, Santa Rosa, CA 95403. b. Matlab; The MathWorks Inc, 3 Apple Hill Dr, Natick, MA 01760-2098. c. Kistler Instruments Corp, 75 John Glenn Dr, Amherst, NY 14228-2171. d. RC Electronics Inc, 6464 Hollister Ave, Santa Barbara, CA 93117. e. Bertec Electronics, 1940 19th St NW, Calgary, AB, Canada. f. BMDP Statistical Software Inc, 1964 Westwood Blvd, Los Angeles, CA 90025.