Exercise and gait effects on in vivo hip contact pressures.

Exercise is an integral part of preventive and rehabilitative programs to ameliorate disability from hip osteoarthritis, trauma, and various surgical procedures. Prescription decisions regarding specific exercises in such programs are currently based on criterial such as patient tolerance and potential for enhancing functional performance, but not for their effects on articular cartilage or endoprosthetic components. [1] Cartilage nourishment and mechanical integrity presumably benefit from intermittent loading, resulting in exudation and imbibition of joint fluid during exercise. [2,3] The changing acetabular pressures probably stimulatte normal cartilage development and hypertrophy in osteoarthritis. [2] Despite the prevalence of hip pathology and surgely [5]--and rehabilitation regimens designed to exercise these diseased hips--no experimental evidence has been reported that examines the effect of exercises on the cartilage of the living (in vivo) human hip. We will report on measurements of in vivo human hip contact pressures collected via a unique, instrumented Austin-Moore-type endoprosthesis. Because the endoprosthetic femoral head was carefully fitted to the natural femoral head to be identical in size and shape, and because the subjects acetabulum was clinically normal, we except the contact pressures recorded to be close to those that would occur in a normal hip joint.

Several groups have measured in vivo hip forces. Rydell, [6] in 1966, first reported forces of three times body weight during free-speed gait, probably attributable to the combined contribution of superincumbent body mass, acceleration, and muscle activity. English and Kilvingtonh, [7] in 1979, and Davy et al, [8] in 1988, reported similar peak hip forces during gait. Although force measurements during straight-leg-raising exercises were reported, no details of angular velocity or of resisted rehabilitation exercises were offered. Straigh-leg-raising contract forces were reported to be about half those that are generated during gait. [8]

We have developed and implemented a system that uses an instrumented Austin-Moore-type prosthesis to measure contact pressures within the human hip joint in vivo. [9] Following the implant in June 1984, the subject has been tested at least twice per year. The effects of assisted and unassisted gait and various other activities of daily living (ADL) since the time of the implant have been reported elsewhere. [10] A large volume of data have been obtained from this subject; we will report only examples of exercise activities of interest to the rehabilitation community.

Exercise is typically believed to be less stressful than functional activities and is frequently used as a preparatory activity in anticipation of functional activities. We therefore hypothesized (1)) that exercise peak acetabular pressures would be less than gait pressures and (2) that peak pressures and pressure rise rates would vary directly with load imposition. A single-subject research design with replications within and across experimental sessions was used.

Method

Subject

A 77-year-old white woman, who was 1.7 m in height and 68 kg in weight, participated in the experiments. She had sustained a traumatic right femoral neck fracture 4 years prior to data collection, which was treated by unipolar femoral head replacement with the Austin-Moore-like device described in this report. Her rehabilitation course was unremarkable except for the acetabular contact measurements taken during surgery and recovery. The subject was otherwise healthy, without any evidence of arthristis or other orthopedic or neurologic impairments. Her daily activities included 0.5. to 1 hour per day of either bicycle riding or swimming, on alternate days. Strenght, range of motion (ROM), gait, and ADL capacity appeared to be unimpaired. Manual muscle test (MMT) results were Good for right hip flexion, adduction, and internal and externasl rotation; Good plus for right hip extension and abduction; and Good plus for knee flexion and extension. Ankle and left-side MMT measurements were not obtained, but the patient could walk >10 m on her toes or heels and rise from a 35-cm-high stool without assistant. Furthermore, isokinetic test results revealed the subject could generate substantial force with the right hip muscle. Goniometry revealed right hip ROM to be 135 degrees for flexion and extension, 75 degrees for internal and external rotation, and 60 degrees for abduction and adduction. The subject walked and jogged without a limp or clinically apparent gait deviations. Informed consent was obtained for both the implant and subsequent tests.

Materials and Instrumentation

Prosthesis. Identical in external size, shape, and mechanical function to an Austin-Moore femoral prosthesis, the instrumented prosthesis used in this study (Fig. 1) has been described in detail elsewhere by the prosthesis engineer-designers and fabricators. [11,12] The device functions like a conventional endoprosthesis in all respects except for its data-generating capabilities. A 47.5-mm-diameter femoral head was instrumented with 3-mm-diameter pressure-sensing transducers. Each transducer diaphragm is flush to, and integral with, the surface of the prosthesis. Diaphragm pressure produces deflection of a silicon cantilever, with a straingauge bridge changing voltage proportional to applied pressures at the acetabulum-femoral head interface. The prosthesis is hermetically sealed by electron beam welding and polished repeatedly to create an ultrasmooth prosthetic femoral head.

Contact pressure measured by the device was defined as the pressure (force per unit of surface area) on a surface resulting from contact with another object. In the set of experiments described in this article, the contact pressures exerted on the surface of a prosthetic femoral head resulting from contact with the natural acetabulum were reported in 10 locations (Fig. 1). Applying Newtonian mechanics, [13] acetabular cartilage experiences equal and opposite pressures to those sustained by the endoprosthesis. Peak pressure was defined as the greatest contact pressure measured during a given experiment. Peak rate of pressure rise was defined as the maximum slope of the pressure/time curve (ie, [Delta]P/[Delta]; see Fig. 2, arrows).

Each transducer was sampled at 253 Hz via a telemetric FM transmitter built into the prosthesis. The electronics are powered by a wire induction loop worn around the thigh as a garter, which avoids the need for internal batteries or other power supplies. The device is energized, and transmits data, only during experimental sessions. Reliable and valid data have been obtained for the past 5 years; extensive testing of the device prior to implantation revealed linear calibration curves for 10 of the transducers. [10] Data from the other 4 transducers are not reported.

Kinematic data. During gait trials only, two optoelectric Selspot[TM] II cameras (*) and the TRACK[C] computer program (+) on a PDP 11/60 computer (++) were used to determine the position and orientation of the segments of ths subject's lower limb and thigh relative to the pelvis in three-dimensional, 6-degrees-of-freedom space. Thus, the locations of the femoral transducers on the acetabular cartilage could be determined to an accuracy of <1 degree of rotation and 1 mm of position. [14]

Kinetic data. Floor reaction force vector data were acquired from two Kistler[R] platforms (+++) processed on the same computer. The NEWTON[C] (+) program was used to solve, through inverse Newtonian mechanics, [13] the net moments and forces experienced by the patient, given the kinematic data on joint centers and segment mass properties and the force-plate data on the ground reaction vectors. Kinematic and kinetic data were sampled at 153 Hz and digitally filtered at 15 Hz.

Electromyographic data. Surface electromyographic (EMG) electrodes were applied, as described by Basmajian and Blumenstein, [15] to the skin above the bellies of the gluteus maximus, gluteus medius, rectus femoris, biceps femoris, medical hamstring, adductor longus, vastus lateralis, and vastus medialis muscles. Only the gluteal muscle EMG data are presented in this report. The 16-mm-diameter electrodes were separated by 1 cm and oriented parallel to the direction of the muscle fibers.

The EMG data were analogue sampled, rectified, and high-pass filtered at 16 Hz by a preprocessor. (+) The EMG data were then digitally sampled at 153 Hz, smoothed off-line on the laboratory's computer with a recursive zero-lag 6-Hz digital filter, and normalized to a percentage of maximal voluntary isometric contraction (MVIC) values to permit estimation of relative muscle activity during the exercise sessions. The signals were monitored in real time and off-line to ensure no electromagnetic interference-created EMG artifacts or interference with prosthesis pressure transmissions. Following EMG electrode application, the prosthesis power and transceiver cuff was donned and a series of mechanical and electrical checks of the system were made.

Isokinetic data. Isokinetic abduction-adduction and flexion-extension exercises were conducted using a LIDO[R] Active device[unkeyable character] set to concentric contraction mode. The device was calibrated and the subject was positioned according to the manufacturer's instructions. Torques and angular displacement data were collected simultaneously with the free-standing microcomputer that is part of the LIDO[R] system and by connecting the isokinetic device to our laboratory's PDP 11/60 computer. Infrared light-emitting diodes (LEDs) were placed on the resistance arm of the LIDO[R] device, and the positions of these LEDs were sampled at 153 Hz. These independent checks of the LIDO[R] system's speed-controlling and data-processing mechanisms revealed that the device provided constant angular velocity. We were thus able to precisely time-synchronize the isokinetic torque, position, and in vivo pressures generated. We did not, however, attempt to assess the reliability of torque measurements obtained with the LIDO[R] system for this patient.

Procedure

We used a relatively simple exercise protocol that was first implemented 3.5 years postimplantation. Exercises were chosen to be a similar as possible to those used in clinical physical therapy practice. Each acitivity was performed at least twice within a given experimental session, and each experimental session was conducted twice at 6-month intervals. During the first session, isokinetic exercises were performed first; in the second session, isokinetic exercises were performed last. All results reported were to within [+ or -] 0.25 MPa (1 MPa = 144 psi [[lb.in.sup.-2]]) within and across sessions, indicating that contact pressure measurement reliability was high for these activities.

Pressure, torque, force, kinematic, and EMG data were collected simultaneously. Compliance with the instructions and experimental maneuvers was observed by the first authors (DEK and LE), both of whom are physical therapists with more than 10 years each of clinical experience. Rest periods of 5 to 7 minutes between maneuvers were provided, and the subject denied experiencing any fatigue or muscle soreness during the course of data collection. During all exercise sessions, except during isometric exercises, at least two practice repetitions were provided. Data were collected during the third and fourth repetitions of the repeated contractions to assess "steady-state" exercise levels. During isometric exercises, data were collected throughout the contraction period.

Gait. Free-speed gait trials were conducted with the subject walking barefoot along a 10-m pathway with the force places mounted unobtrusively under the carpeted floor in the middle of the walkway. The subject was instructed, "Walk at your normal pace, as though you are walking in the park." The subject's average gait velocity was 1.2 m/s.

Isokinetic exercise. For the isokinetic abduction-adduction exercises, the subject was positioned side-lying on her left hip; for the flexion-extension exercises, the subject was positioned supine. Special care was taken to align the axis of the resistance arm of the LIDO [R] device with the anatomical axis of the subject's right hip. A computer-assisted maneuver for gravity compensation was conducted according to the manufacturer's instructions. The subject was instructed to move her leg "as fast and hard as possible" during the concentric abduction-adduction and flexion-extension exercise sessions, which were conducted at 30[degrees] and 90[degrees]/s angular velocities over the 45-degree ROM of hip adbudction-adduction and flexion-extension.

Isometric exercise. Maximal voluntary isometric contraction hip abduction pressures were measured during a grdually ascending contraction effort. The subject was placed supine on the floor, with the hip at 0 degrees of rotation, flexion, and abduction so that an area just proximal to the lateral malleolus was in contact with a padded, vertical rigid bar attached to the force plate, from which the transverse force measurements were collected. Isometric abduction force measurements were thus acquired from the force plate, while the subject's body was stabilized by two researchers. Although the LIDO[R] dynamometer can be modified to function isometrically, the subject was more easily positioned and the data were more easily time-synchronized using the available laboratory equipment. A metronome was set to 60 bpm, and the subject began an isometric contraction at one beat and gradually intensified her effort to a maximally comfortable level by the second beta. Verbal reinforcement to "push, push, harder, harder, as hard as you can" was also offered during the 3-second ascent to MVIC.

Gravity-resisted hip abduction. The subject was positioned side-lying on her left side, and using standard goniometric techniques, the position of her right foot at 0 and 30 degrees of hip abduction was determined. One of the experimenters' hands were placed at each position to define target end points to the exercise. No attempt was made to control the slight (estimated visually as [is less than] 5%) hip rotation and flexion that typically accompanied this type of exercise. The maneuver was repeated with a 2.2-kg cuff weight wrapped around the ankle.

A metronome was set to 60 bpm, and beginning from a neutral position, the subject was instructed to move her leg "up ... and ... down ... and ... up ..." in time to the metronome. A practice trial was provided to ensure a reproducible movement. This cadence resulted in an average angular velocity of 30 [degrees]/s with a rest interval of approximately 1 second before and after each abduction effort. The exercise was then repeated, with the metronome set to yield a nominal average velocity of 90 [degrees]/s.

Straight leg raise. The patient was positioned supine, with the left knee bent. A nominal ROM of 0 to 30 degrees of flexion and a 30 [degrees]/s angular velocity were enforced, as described for gravity-resisted hip abduction. The maneuver was then repeated with a 2.2-kg cuff weight wrapped around the right ankle.

Manually resisted exercise. For this exercise, the patient was positioned supine on the plinth and the metronome was set to 60 bpm to time-synchronize the verbal commands. The command sequence of "and ... pull up and out ... now push down and in ..." was given to elicit a proprioceptive neuromuscular facilitation (PNF) pattern (ie, hip flexion-abduction-external rotation followed by extension-adduction-internal rotation) of the lower extremity. During the command "and" previous to the massed flexion, manual longitudinal traction was applied to the extremity via the dorsum and the calcaneus of the foot. A 'quick stretch" was similarly offered prior to the extension phase to encourage joint compression. During the flexion maneuver, resistance was applied to the dorsum of the foot and the anterior thigh; during the massed-extention component, manual resistance was applied to the plantar surface of the foot and the posterior thigh.

Powder-board abduction. A plastic low-surface-friction sliding board was placed under the right leg of the subject while she was lying supine. A nominal ROM of 0 to 30 degrees of abduction and a 30 [degrees]/s angular velocity were enforced, as described for gravity-resisted hip abduction.

Results

Results are summarized in the Table and in Figures 2 through 5. The maximum pressure observed during these experiments was 4.14 MPa (600 psi) during isometric hip abduction. The maximum rate of pressure rise was 12.18 MPa/s (1,754 psi/s) during manual (PNF) resistance exercise, closely followed by 90 [degrees]/s isokinetic hip abduction and gait. The minimum pressure of -0.48 MPa was recorded during manual hip traction during PNF exercise; pressures recorded during quiet supine resting are typically 0 MPa. In every case, maximum pressure was coincident with the hip muscles' maximum EMG activity for that trial, and these pressures were recorded from the transducers near the superior dome of the femoral head, along the calculated point of application of the force vector.

Gait

The peak contact pressure and the peak rate of pressure rise (Table) occurred between heel-strike and early mid-stance (Fig. 2). No gait defects such as right-left asymmetries, limps, or lurches were discernible in this subject, whose average self-selected velocity was 1.2 m/s.

Isokinetic Exercise

The peak contact pressure during isokinetic exercise was recorded during hip abduction-adduction at 90 [degrees]/s, and the peak rate of pressure rise occurred during hip flexion-extension at 90 [degrees]/s. Peak rates of pressure rise occurred simultaneously with maximal rates of torque development; peak pressures during isokinetic exercise occurred during peak torque output (Figs. 4, 5).

Isometric Exercise

Maximal isometric abduction produced the highest contact pressure recorded in any experiment (Table). Pressure magnitude and rate of pressure rise were closely associated with abduction force and EMG activity (Fig. 3) during this gradual ascent to MVIC. That is, voluntarily intensifying the contraction over 1 second controlled

[TABULAR DATA OMITTED]

the rate of both hip abductor force and contact pressure rise.

Isotonic Exercise

Gravity-resisted abduction. Isotonic exercise with a 2.2-kg cuff weight at a nominal angular velocity of 30 [degrees]/s generated a peak pressure during side-lying hip abduction of 10% less than the same exercise at 90 [degrees]/s (Table). Peak rate of pressure rise during the 30 [degrees]/s contraction was about one third of that attained by trebling the angular velocity the 90 [degrees]/s (Table).

Straight leg raise. Peak pressures with or without an ankle cuff weight were similar; straight-leg-raising (SLR) peak rates of pressure rise were also similar with or without the 2.2-kg cuff weight (Table). Both peak pressures and rates of pressure rise occurred immediately after the flexion phase was initiated.

Manually Resisted Exercise

During the manually resisted PNF maneuver, peak pressures reached 2.75 MPa, but the highest rate of pressure rise of all experiments (ie, 12.18 MPa/s) was recorded. This rapid pressure rise occurred immediately after manual hip joint traction and a quick hip flexor muscle stretch were applied. Interestingly, the application of a manual traction force resulted in transient negative pressures (minimum, -0.48 MPa) on the femoral head.

Powder-Board Abduction

Gravity-eliminated hip abduction at a nomnal angular velocity of 30 [degrees]/s generated the lowest peak pressure during any experimental trial (ie, 1.69 MPa), which occurred just after the initiation of abduction and coincident with the maximal EMG activity of the gluteus medius muscle during the maneuver. The peak rate of pressure rise (ie, 1.87 MPa/s) was also minimal during the powder-board exercises (Table).

Discussion

Our results clearly show that, in this patient, acetabular cartilage pressure magnitudes and rates of pressure development during commonly used exercises may exceed those developed during free-speed, full weight-bearing, unassisted gait. These hip pressures are apparently independent of exercise type, but appear to be highly dependent on the hip loading force and rate of force development, two factors that can be titrated in exercise prescription.

Contact pressures within the hip joint are important for at least two reasons. First, the number of individuals who undergo partial hip replacement surgery in the United States exceeds 100,000 per year, [16] and about 50% of these hip replacements fail after 5 years from excessive mechanical stresses on the acetabular cartilage and the endoprosthetic components. [17] Second, cartilage longevity is probably influenced by mechanical stresses. Our data indicate that, for this patient, at least one type of mechanical stress (ie, contact pressure) may be controlled by careful choice of exercise type and intensity.

Gait

Acetabular contact pressures during free-speed gait show a highly consistent pattern closely linked to the events of the gait cycle. The maximum rate of pressure rise and the peak pressure occurred between hell-strike and mid-stance; peak vertical ground reaction force occurred during push-off (Fig. 2). Abductor muscle activity, as indicated by the EMG results, rose co-linearly with hip pressure; both EMG activity and hip pressure rose prior to foot-floor contact. Acetabular pressures during gait were therefore related to muscle contraction as well as to ground reaction forces from body mass and acceleration. [18] That muscle activity alone can increase hip contact pressure during gait is clear from the data prior to heel-strike: no ground reaction forces can contribute to hip contact pressures during the swing phase. Thereafter, hip pressures and abduction torques varied directly. Congruent with prior reports of force-instrumented hip prosthesis data, peak hip stresses occurring during stance are apparently determined by the combined contribution of muscular co-contractions required to increase lower extremity joint impedances sufficiently to provide propulsion and balance during the stance phase, as well as by the ground reaction forces. [19,20] During the swing phase of gait, pressures decrease but are always greater than zero (Fig. 2).

Exercise

Exercise has long been a part of the rehabilitation process. As Hippocrates noted 2,500 years ago:

All parts of the body which have a function, if used in moderation and exercised, become thereby healthy and well-developed and age slowly, but if unused and left idle they become liable to disease, defective and age quickly.

No prior reports have addressed the effect of individual resistive exercises on joint contact pressures.

Maximal exercise generated greater acetabular pressure than forces from the body mass combined with muscular activity required for unaided, free-speed gait. Obviously, maximal muscle activity is not required for normal gait, and this patient's ability to generate flexor and abductor force and moments during exercise (Figs. 3-5) at 4 years postimplantation would be expected to exceed that of debilitated, acutely postoperative patients. Therefore, one should not generalize directly from these data to the acute postoperative hip. In the patient with hip arthritis or other joint impairments, strengthening exercises are frequently recommended. Recently, isokinetic exercises have been encouraged as a means of slowing atrophy in patients with arthritis. [21] Prior to our report, however, no quantitative data were available to compare the effects of isokinetic and other excercise modalities on cartilage pressures.

As hypothesized, hip pressures and imposed loads varied directly during all exercises. Both the external moments from the isokinetic device and the EMG data varied directly with hip pressure. Taken together, these findings suggest that the therapist's role in modifying hip stress and stress application rates can best be implemented by controlling muscle activity required for hip exercise. Neumann and colleagues' [22] EMG and hip abductor torque data from young healthy subjects also appear to support this conclusion. Our data indicate that trebling the angular velocity commensurately increased the rate of pressure rise, and gradually ascending to MVIC revealed a similarly gradual rise in contact pressures. Thus, controlling the contraction rate apparently controls acetabular contact pressure rise rates.

The data suggest that hip contact pressures during slow, "gravity-eliminated" exercise are minimal. For patients with acute postoperative cementless hip prostheses or those whose cartilage should not be stressed, powder-board exercises appear to be low-stress activities, whereas antigravity abduction stresses the acetabulum and hip muscles to a much greater extent. Peak hip pressures generated during SLR exercises were among the lowest developed during any exercise session. This finding calls into question the commonly held clinical concept that SLR is particularly stressful to the human hip. Indeed, SLR peak pressures were half those of free-speed gait and only 10% greater than those of powder-board hip abduction (Table).

Compressive stresses. Minimizing compressive stress is important when designing preventive and rehabilitative programs for individuals with osteoarthritis or other degenerative diseases of the hip joint. Certainly, the inorganic materials used in prosthesis fabrication degenerate more rapidly with greater contact pressures. [23] Exercises inducing high contact pressures, such as maximal isokinetic or isometric contractions, should be avoided for patients whose cartilage is already fibrillated or whose cartilage is physiologically at risk for deterioration. Although the physiological processes that result in the predisposition of the cartilage to degeneration are not yet fully elucidated, an important component of the degeneration may be mechanical stresses on the cartilage. [24]

Rates of pressure rise. Our data suggest that submaximal or slow maximal exercises control rates of pressure rise effectively (Table, Fig. 3). Cartilage loading rate, defined as the speed at which pressure develops, may be a useful exercise prescription factor. Cartilage, like all poroviscoelastic materials, has a complex, rate-dependent reaction to mechanical stresses. [25] Mow and colleagues stated,

In a rapid loading and unloading situation ... (for example during jumping), [cartilage] will behave more or less like an elastic, single-phase material, deforming instantaneously upon loading and recovering instantaneously upon unloading. However, if the load is applied slowly or kept constant on the tissue, as for example during prolonged standing, the tissue deformation will continue to increase in time as the fluid is squeezed out. [3](p70)

Therefore, it may be desirable to minimize loading rates as well as absolute magnitudes od contact pressures in patients with prosthetic hips or arthropathy, but not necessarily in those patients with normal cartilage. [4]

Exercise and Gait Effects

on Normal Cartilage

Normal cartilage is nourished by pressure-mediated fluid exudation and imbibition during normal movement. [25,26] Hall and Urban [27] recently reported that brief ( [is less than] 5 minutes) static pressures of [is less than] 15 MPa stimulate normal bovine articular cartilage synthesis, but prolonged high pressures (2 hours at 50 MPa) inhibit cartilage growth in vitro. Our in vivo data suggest that exercise of typical durations and intensities may generate pressures within the levels reported by Hall and Urban to encourage cartilage synthesis. Normal cartilage probably regenerates; it "does not thin simply as a result of physiological joint activity," [TABULAR DATA OMITTED]

but rather chondrocyte activity in the appropriate mechanical and cell-matrix environment stimulates regrowth. [28] The extent to which contact pressures during exercise and ADL inhibit or enhance this process is not known, but our data indicate that similar mechanical stresses may be engendered by maximum force exercise and gait. Submaximal exercise, including SLR and powder-board abduction-adduction exercises, generates lower acetabular stresses than gait, but the rates of pressure development during rapid submaximal exercise can nonetheless be similar to those of maximal exercise (Table).

Conclusion

It has previously been assumed that active exercise serves as "preparation" for gait training and may be safely instituted well before weight-bearing is allowed. This notion may be valid for submaximal exercise, but it is not supported by our articular cartilage contact pressure findings for maximal exercise. Although these de novo data stem from a single subject, and generalizations are thereby limited, this is the first report of in vivo acetabular contact pressures during rehabilitation exercises. Consistent with prior reports of force-instrumented prostheses and in vitro experimentation, [11] joint contact pressures vary directly with force imposition and can be limited by factors under patients' and clinicians' control. For example, to limit the magnitude and rate of acetabular pressure development in the patient described in this report, one could request low-velocity, supine hip abduction exercise on a powder board. Antigravity isotonic, isokinetic, and isometric exercises apparently increased peak hip joint contact pressures in proportion to estimates of hip force exerted. Transducers near the femoral head's superior apex experienced the greatest pressures, indicating nonuniform distribution of contact pressures throughout the joint surfaces. Such highly localized pressures are to be expected from exercise trials according to prior in vitro studies. Further studies on a larger series of subjects with instrumented hip prostheses are needed before firm conclusions can be drawn.

(*1) Selective Electronics. Partille, Sweden.

(*2) Developed at the Massachusetts Institute of Technology, Cambridge, Mass.

(*3) Digital Equipment Corp, 146 Main St, Maynard, MA 01754.

(*4) Type 9281A, Kistler Instruments AG, Winterthur, Switzerland.

(*5) Loredan Biomedical Inc, 1632 Da Vinci Ct, Davis, CA 95616.

References

[1] Moncur C. Attacking the sacred cows. Arthritis Care Research. 1988;1:116-211.

[2] Radin EL, Paul IL. A consolidated concept of joint lubrication. J Bone Joint Surg [Am]. 1972;54:607-613.

[3] Mow VC, Roth V, Armstrong CG. Biomechanics of joint cartillage. In: Frankel VH, Nordin M, eds. Basic Biomechanics of the Skeletal System. Philadelphia, Pa: Lea & Febiger; 1980: chap 2.

[4] Mankin HJ, Lippiello L. Biochemical and metabolic abnormalities in articular cartilage from osteo-arthritic human hips. J Bone Joint Surg [Am]. 1970;52:424-434.

[5] Wilcock GK. The epidemiology of diseases of the hip: a review of the literature. Int J Epidemiol. 1979;8:247-250.

[6] Rydell NW. Forces acting on the femoral head-prosthesis: a study on strain gauge supplied prostheses in living persons. Acta Orthop Scand. 1966;37(suppl 88):1-132.

[7] English TA, Kilvington M. In vivo records of hip loads using a femoral implant with telemetric output: a preliminary report. J Biomed Eng. 1979;1:111-115.

[8] Davy DT, Kotzar GM, Brown RH, et al. Telemetric force measurements across the hip after total arthroplasty. J Bone Joint Surg [Am]. 1988;70:45-50.

[9] Hodge WA, Fijan RS, Carlson KL, et al. Contact pressures in the human hip joint measured in vivo. Proc Natl Acad Sci USA. 1986;83:2879-2883.

[10] Hodge WA, Carlson KL, Fijan RS, et al. Contact pressures from an instrumented hip endoprosthesis. J Bone Joint Surg [Am]. 1989;71:1378-1386.

[11] Rushfeldt PD, Mann Rw, Harris WH. Influence of cartilage geometry on the pressure distribution in the human hip joint. Science. 1979;204:413-415.

[12] Carlson CE, Mann Rw, Harris WH. A radio telemetry device for monitoring cartilage surface pressure in the human hip joint. IEEE Trans Biomed Eng. 1974;21:257-264.

[13] Bresler B, Frankel JP. The forces and moments in the leg during level walking. Trans ASME. 1950;72:1851-1859.

[14] Antonsson EK, Mann RW. Automatic 6-dof kinematic trajectory acquisition and analysis. Journal of Dynamic Systems, Measurement, and Control. 1989;111:31-39.

[15] Basmajian JV, Blumenstein R. Electrode Placement in EMG Biofeedback. Baltimore, Md: Williams & Wilkins; 1982:79-82.

[16] Kelsey JL. Epidemiology and impact. In: NIH Consensus Development Conference: Total Hip Joint Replacement Program Abstracts. Bethesda, Md: National Institutes of Health; 1982:23-30.

[17] Sreide O, Skjaerven R, Alho A. The risk of acetabular protrusion following prosthetic replacement of the femoral head. Acta Orthop Scand. 1982;53:791-794.

[18] Catani F, Hodge WA, Mann RW. Hip dynamics in level walking, stair climbing and rising from a chair. In: ASME Winter Annual Meeting: Biomechanics of Normal and Prosthetic Gait. New York, NY: American Society of Mechanical Engineers; 1987;4:113-116.

[19] Frankel VH, Burstein AJ, Lygre L, Brown RH. The telltale nail. J Bone Joint Surg [Am]. 1971;53:1232.

[20] Rydell N. Biomechanics of the hip-joint. Clin Orthop. 1973;92:6-15.

[21] Danneskiold-Samsoe B, Grimby G. Isokinetic and isometric muscle strength in patients with rheumatoid arthritis. Clin Rheumatol. 1986;5:459--467.

[22] Neumann DA, Soderberg GL, Cook TM. electromyographic analysis of hop abductor musculature in healthy right-handed persons. Phys Ther. 1989;69:431-440.

[23] Charnley J, Halley DK. Rate of wear in total hip replacement. Clin Orthop. 1975;112:170-179.

[24] Radin EL, Paul IL. Response of joints to impact loading: in vitro wear. Arthritis Rheum. 1971;14:356-362.

[25] Mow VC, Torzilli PA. Fundamental fluid transport mechanisms through articular cartilage. Ann Rheum Dis. 1975;34(suppl):82.

[26] McCutchen CW. Mechanism of animal joints: sponge-hydrostatic and weeping bearings. Nature. 1959;184:1284-1285.

[27] Hall AC, Urban JPG. Responses of articular chondrocytes and cartilage to high hydrostatic pressure. In: Proceedings of the 35th annual meeting of the Orthopedic Research Society; February 6-9, 1989; Las Vegas, Nev. Page 49.

[28] Meachim G, Stockwell RA. The matrix. In: Freeman MAR. Adult Articular Cartilage. 2nd ed. Kent, England: Pitman Medical; 1979:1.

D Krebs, PhD, PT, is Associate Professor, Program in Physical Therapy, MGM Institute of Health Professions, 15 River St, Boston, MA 02108-3402 (USA), and Lecturer, Massachusetts Institute of Technology, Cambridge, MA 02138. Address all correspondence to Dr Krebs at the first address.

L

L. Elbaum, MM, PT, is Associate Professor, Florida International University, Miami, FL 33199. At the time of this study, he was a Nation Instituteof Disability and Rehabilitation Research Advanced Rehabilitation Fellow, MGH Biomotion Laboratory, Massachusetts General Hospital, Boston, MA, and Department of Mechanical Engineering, Massachusetts Institute of Technology.

R Riley, PhD, is Technical Director, MGH Biomotion Laboratory, Massachusetts General Hospital, Fruit St, Boston, MA 02114.

WA Hodge, MD, is Assistant in Orthopaedics, Massachusetts General Hospital, 5 Longfellow Pl, Ste 201, Boston, MA 02114.

R Mann, ScD, is Whitaker Professor of Biomedical Engineering Department of Mechanical Engineering, Massachusetts Institute of Technology.

This research was supported in part by the US Department of Education Grant H133P90005; the Veteran's Administration Medical Center, Grant REUVDHT9899; and Loredan Biomedical Inc, Davis, CA.

This study was approved by the Massachusetts Institute of Technology and Massachusetts General Hospital institutional review boards.

This article was submitted March 19, 1990, and was accepted October 30, 1990.