Acute phase.

Rehabilitation procedures for patients with painful hips secondary to arthritis or surgical implantations appear to be largely based on local custom. To examine hip contact pressures during the acute and postacute phases of rehabilitation, we conducted a two-part study of a single subject with a pressure-instrumented femoral prosthesis. This report compares in vivo acetabular contact pressures generated during selected "inpatient" rehabilitation activities and their relationship to pain, range of motion, and other clinical indicators.

Instrumented Hip Prostheses

Rydell(1) first reported in vivo hip force measurements from an instrumented femoral prosthesis. Strain gauges were applied to the surface of a "femoral neck" made overly long to ensure adequate instrument response. Rydell's measured forces thus may have differed from the forces experienced by a normal hip. The results were from a test session conducted 6 months after surgery and therefore only loosely pertain to rehabilitation of patients in acute pain. According to Rydell's data, peak forces during activities of daily living were 0.7 to 0.9 times body weight for a non-weight-bearing static standing position. The higher force occurred when both the hip and the knee were flexed, and the lower force occurred when the hip was neutral and the knee was flexed. Peak forces were 1.0 to 2.0 times body weight during straight-leg-raising exercises and 1.8 to 3.3 times body weight during level walking. Isometric hip abduction was not studied, but active hip abduction with the subject positioned supine resulted in a peak force of 0.7 times body weight.

English and Kilvington(2) and Davy et al(3) obtained in vivo force measurements from instrumented femoral and total hip prostheses during the first 4 to 6 postoperative weeks. Measured hip forces were lower during partial weight-bearing (PWB) ambulation than during full weight-bearing (FWB) ambulation. Forces measured in the postacute period were greater than forces measured for the same activity during the acute phase.

Hodge and colleagues(4-6) first reported in vivo acetabular contact pressures from the hip, obtained from the same subject who participated in this study. Contact pressures during the first 36 postoperative months were reported from various locations on the femoral head. "Contact pressure" occurs as a result of loading between the prosthetic head and the acetabular cartilage. Pressure is defined as the force divided by the area on which the force is applied.(6) Transducer pressure, as expressed by the telemetered signal, was established during preimplantation calibration, with the hemisphere exposed to known pressures. "Peak pressure" refers to the highest reading from the referenced transducer for a given activity. Non-weight-bearing ambulation generated a peak acetabular contact pressure of 2.4 MPa (1 MPa= 144 psi), PWB ambulation generated 3.5 MPa, resisted isometric hip abduction generated 4.2 MPa, and unsupported FWB ambulation generated 5.5 MPa.(6) The detailed relationship between in vivo hip contact pressures and total forces is unknown, but because force is the geometric integration of pressure, higher pressures infer higher local forces, as shown in vitro.

Relationship of Peak Pressures to Cartilage Behavior

The use of a femoral endoprosthesis is often the preferred surgical treatment for femoral head or neck fractures.(7) Migration of the endoprosthesis through the acetabular cartilage is a complication in at least 15% of patients followed 1 year or more after surgery.(8) Anecdotal reports suggest that early FWB ambulation ([equal to or less than] 3 months postsurgery) led to a 50% incidence of femoral endoprosthesis migration, whereas delaying FWB ambulation until [is greater than] 3 months postsurgery was associated with only 18% migration incidence.(9)

Though optimal in vivo articular cartilage stresses have yet to be determined, the mechanical and metabolic responses of normal articular cartilage apparently require the imbibition and exudation of fluid from intermittent contact pressures.(10,11) In vitro studies of articular cartilage have demonstrated that low levels of hydrostatic pressure (0.5-2.1 MPa [75-300 psi]) result in a 50% decrease in synthetic activity, whereas a higher pressure (2.6 MPa [375 psi]) results in a 10% to 15% increase in synthetic activity.(12) Hall et al(13) found articular cartilage synthetic activity was time and pressure dependent in bovine articular cartilage. Short-term pressures ([is not greater than]5 minutes) at 15 MPa stimulated synthetic activity, and long-term, high pressures (2 hours at 50 MPa) inhibited synthesis. Further studies are needed to determine the optimal range and temporal variation for hip contact pressures before specific recommendations can be made for human articular cartilage.

Rehabilitation Programs

Specific hip rehabilitation regimens are often reported only incidentally in the literature. Therapists and surgeons, therefore, have few empirically tested approaches to implement for patients who have undergone hip arthroplasty. In addition, most research has been done on total hip prostheses.

Thielen and Mueller(14) suggested a protocol for the management of the patient with a total hip replacement, recommending abductor exercises to prepare for holding the pelvis in equilibrium during gait. This emphasis is consistent with the decrease in maximum abductor torque following hip replacement surgery that has been reported by several authors.(15-17) The quadriceps femoris muscles have also been noted to be significantly weaker in patients who have undergone hip replacement. Other studies(16-18) attribute early femoral loosening to muscle weakness.

Most hip rehabilitation programs include isometric hip and knee extension, active hip abduction and flexion, weight shifting, and transfers to standing. These exercises are intended to isolate specific muscle groups necessary to accomplish higher-level functional activities. For example, isometric hip extension (gluteal muscle sets) was chosen because extension is important in the "sit-to-stand" activity. Additionally, the abductors and the quadriceps femoris muscles are essential in the activities of standing and gait.

The purposes of this portion of the two-pact study were (1) to compare hip pressures generated during selected acute phase "inpatient" rehabilitation activities and (2) to examine the relationship of hip contact pressures to pain, range of motion (ROM), and other commonly used clinical indicators of patient tolerance of rehabilitation programs. Previous reports(4-6) did not isolate the inpatient data from the long-term results. In addition, no prior attempt has been made to explore the relationships between clinical data (eg, ROM, manual muscle test [MMT] grades, independence in activities of daily living) and pressure data. Based on the usual clinical recovery pattern and acetabular contact pressure during functional activities noted by Hodge and colleagues,(4-6) we hypothesized that these acute recovery phase acetabular contact pressures will increase over time and that the contact pressures will correspond directly with clinical improvement as measured by goniometry, inversely with the amount of pain medication received, and directly with independence in gait as measured by the Functional Ambulation Scale.(19) We hypothesized the rank order of the selected activities, from the lowest to the highest expected peak pressures, to be as follows: (1) isometric knee extension (quadriceps femoris muscle setting); (2) isometric hip extension (gluteal muscle setting); (3) supine, active hip abduction; (4) supine, active hip flexion with knee flexed; (5) step with operated leg, using an assistive device; (6) step with nonoperated leg, using an assistive device; and (7) sit to stand, with assistance.

Method

Subject

A 73-year-old healthy white woman with a weight of 68 kg and a height of 1.68 m underwent hip arthroplasty for a Garden 111 fracture of the right femoral neck. Instead of the standard femoral head replacement, she received an instrumented right femoral head prosthesis. The patient provided written consent for participation in both the study and the surgical procedure. The size of the patient's femoral head (47.5 mm) was confirmed to be appropriate for the endoprosthesis, ensuring the optimal acetabularprosthetic fit. In addition, the acetabular cartilage was inspected upon surgical exposure to confirm that it was normal in appearance.

Instrumentation and Materials

Prosthesis data. The instrumented endoprosthesis was designed to quantify contact pressures at the acetabular cartilage. The prosthesis instrumentation is described in detail elsewhere.(20,21) Fourteen pressure transducers, each 3.96 mm in diameter, were distributed in two concentric circles on the hemisphere of a 47.5-mm-diameter femoral head (Fig. 1). The pressure-sensing diaphragm of each transducer was contiguous with the smooth, polished outer surface of the endoprosthesis, but deflected slightly ([equal to or less than] 0.0002 mm) by the pressure of the opposing acetabular cartilage. The diaphragm center point of movement was measured by a strain-gauged, single silicon crystal cantilever beam. An FM telemetry system hermetically sealed within the prosthetic head transmitted the output of each transducer as the magnitude of a pulse, with each frame of data consisting of the 14 pressures and the calibration signals. The pressure data were transmitted to a PDP 11/03 minicomputer(*) at a rate of 30 frames per second. To avoid the use of internal batteries, the electronics were powered by an inductive transmitting loop worn around the thigh as a garter, with the prosthesis energized only during the test periods. Prosthesis-acetabular contact pressures were reported from 10 transducers, because data from the other 4 transducers did not meet calibration standards.(20,21) Force and kinematic data were not recorded during this phase of the study.

Pressure data were low-pass filtered at 10 Hz to attenuate high frequency transducer noise, which ranged between 0.2 and 0.35 MPa in unfiltered pressure data.(22) After filtering, the transducer noise was well below 0.2 MPa, so pressure differences of 0.2 MPa or more were considered significant.

Clinical data. Clinical data included measurements of ROM, pain, and gait and MMT grades. These data were collected and summarized using a retrospective review of the patient's medical record. A standard goniometer was used to measure hip flexion and abduction ROM during both physical therapy intervention and testing periods. Physical therapy was discontinued on postoperative day 3 because the patient was noted to be coughing blood-streaked sputum. Physical therapy sessions were resumed on postoperative day 8 after an angiogram ruled out pulmonary embolus.

Gait performance was documented in both the physical therapy and nursing notes. The patient began ambulation on postoperative day 8 and was discharged walking independently with crutches on day 17. Conversion to the Functional Ambulation Classification(19) was performed independently by two researchers, who attained 100% classification agreement. The physical therapy notes revealed documentation of MMT grades only for the quadriceps femoris and hip abductor muscles, and these data were documented only on postoperative day 12. The patient's pain intensity was not specifically assessed; therefore, the actual pain medication dosage she requested and received during hospitalization was used to estimate pain levels. Three different analgesics--Tylenol[R][daggar] (650 rag), codeine (60 mg), and Darvon[R] [double daggar] (65 mg)--were given. The dosages given for these different medications were standardized to the strength of Tylenol[R] (650 mg= 1) (Tab. 1).(23)

Procedure

Data were collected during 11 separate sessions during the patient's 17-day hospital stay. Each activity was tested on at least 2 Separate days. The number of activities varied from one session to another, dependent upon subject tolerance and pain experienced during testing. On days 6 and 7, most selected maneuvers were tested within the same session.

Isometric exercise sets. Maximal isometric contraction of both the right knee extensors and hip extensors were performed while the patient was positioned supine on the hospital bed. On postoperative days 1 and 2, peak pressures were obtained during isometric exercise with the right limb in balanced suspension. On test days 4, 5, 7, and 12, however, the hip was positioned in neutral flexion, abduction, and rotation without suspension of the limb.

Active exercises. Pressure data for active hip flexion were recorded with the patient positioned supine in bed. Active hip abduction tests were performed with the patient in a supine position and the hip in neutral flexion, abduction, and rotation. The patient was then asked to move the limb horizontally toward abduction.

Sit to stand. Pressure data during sit-to-stand activities were first acquired on the sixth postoperative day. The patient was instructed to arise from a seated position using the armrests of the chair, as necessary. The chair height was not documented; however, because hospital procedures at the time were to prevent hip flexion [is greater than] 90 degrees, it is probable that the chair was at least 110% of knee height.

Step right/step left. Contact pressure data during stepping with the right leg first (ie, ipsilateral stance) and stepping with the left leg first (ie, ipsilateral swing phase), both with the use of crutches, were reported on only 2 postoperative days.

Data Analysis

Acetabular pressures were tabulated separately for each activity. The peak pressure for each activity was noted. The activities were then ranked by peak pressure level. Peak pressure data were plotted graphically against the clinical indicators (goniometry, pain medication, and gait performance).

The data analyzed were obtained from the transducer that most frequency recorded the maximum peak pressure for all test sessions and for all activities. We report only values from transducer 5, which is located posterior and superior on the femoral head of the endoprosthesis when the hip is in neutral alignment (Fig. 1).(6) Similar trends over time and across activities occurred in transducers 2, 3, and 7. The designated transducer peak pressures were then compared for each activity across sessions. Hence, care must be exercised in comparing data reported in this article with data reported in other publications describing this subject, because prior reports did not confine their analyses to a single transducer.

Results

Peak Pressures Peak acetabular contact pressure results are summarized in Figures 2 and 3. Among the activities examined during the acute phase recovery period, active hip flexion on day 7 produced the highest pressure (ie, 4.79 MPa). As hypothesized, isometric knee extension exhibited the lowest peak pressure (ie, 3.44 MPa). The pressures generated by the other activities did not rank as predicted. Active hip flexion and isometric hip extension produced the greatest peak pressures of all activities (4.79 and 4.65 MPa, respectively). Because the maximum pressures during these activities were within the 0.2-MPa significance criterion, they should be considered equivalent. Step right and left (right swing and stance phase, respectively) exhibited pressures nearly equal to active abduction ([is not greater than]0.2 MPa). Figure 3 illustrates the differences in peak pressure for the various activities.

As hypothesized, contact pressures during step right/step left exhibited an upward trend with increasing time after surgery (Fig. 2). The exercise activities did not follow this same pattern. Instead, the pressures peaked on approximately day 7 and declined rapidly thereafter.

Reliability can be inferred from the consistency of measurements on consecutive days. No reliability studies were performed, however, in part because of the limited data available. In addition, as the prosthetic device had been extensively calibrated before implantation, the patient's day-to-day changes most likely reveal real physiological variations. Variability between individual test measurements within activities on consecutive days ranged from a low of 0.01 MPa for isometric quadriceps femoris muscle exercises to a high of 1.2 MPa for isometric gluteal muscle sets (Fig. 2).

Clinical Data

Although gait training was begun on day 8, independence on stairs with crutches (Functional Ambulation Scale score=5) was not achieved until discharge on postoperative day 17 (Fig. 4). Figure 5 illustrates the patient's gradual progression of active hip flexion.

An overall decline in use of pain medication is noted in Figure 6. As expected, a sharp peak was seen on postoperative day 2. The increase in pain medications on days 6 and 7 occurred concurrently with increases in contact pressures during the rehabilitation activities.

The patient's activity levels varied inversely with pain medication dosage (Figs. 6 and 7). For example, on day 9, more rehabilitation activities were tested than on any other day and the subject took no pain medication. An exception to this relationship is noted on days 3, 8, and 16. On days 3 to 8, pulmonary embolism was suspected and physical therapy was suspended. Day 16 was just prior to discharge, when no tests were performed to permit the patient to be maximally functional on the day of discharge.

Discussion

This study is the first to document acetabular contact pressure data for these common rehabilitation maneuvers during the acute postoperative phase. Current therapeutic regimens have been based primarily on clinical observations and to a lesser extent on research. The joint pressure data from this subject indicate that isometric and active exercises could have negative effects. In those patients with compromised acetabular cartilage, careful consideration should be given to both the rationale for and the effectiveness of performing maximal isometric gluteal muscle exercises. Ambulation activities appeared to produce less femoro-acetabular stress than did isometric exercise in this patient during the acute postoperative phase.

Peak Pressures

Exercises are frequently prescribed preparatory to ambulation because unloaded joints in non-weight-bearing activities are believed to experience less force and therefore reduce contact pressures on joint surfaces. Our data, however, do not support this practice if the purpose of the exercise is to move with minimal pressure on the hip joint. Except for isometric knee extension, peak pressures among the selected maneuvers did not follow the hypothesized rank order, or therefore traditional practice. Indeed, active hip flexion created the highest acetabular contact pressure (Fig. 3). Hip contact pressures during exercise are highly effort dependent,(21) whereas functional activities are less demanding of patient efforts. These data suggest that therapists should focus on controlling the vigor, rather than the type, of exercise prescribed for the patient during acute phase recovery following hip surgery.

Isometric exercise sets. As predicted, isometric contraction of the quadriceps femoris muscle generated the lowest contact pressures, probably because the contraction of this muscle primarily produces force at the knee. With the knee and hip in neutral positions, only the rectus femoris muscle can directly produce force at both the hip and knee joints. In contrast, the muscles involved in the hip exercises are prime movers of the hip.

The high peak pressure elicited during isometric hip extension may be due to the large cross-section of the gluteus maximus muscle. The amount of force a muscle generates depends on the number, size, and type of active motor units and on their firing frequencies. The larger the number of fibers active, the greater the force produced. In addition, greater force can be developed in an isometric contraction than in concentric contractions.(24)

Active exercise. The highest contact pressure occurred during active anti-gravity hip flexion. Active hip abduction was also performed in a supine position, thus minimizing the effect of gravity on the abductors. Contact pressure, therefore, resulted primarily from muscle activity rather than from muscle forces combined with superincumbent body weight. The MMT data demonstrated that the abductor strength was Poor during the test periods. This decrease in relative strength, and the minimal force required to abduct the hip while the patient was positioned supine, may have reduced the abductor force required and thus reduced the maximum peak acetabular pressure obtained for this activity.

Step left/step right. No significant difference was found in peak pressure between the stance and swing phases with upper-limb support. This "weight shifting" also resulted in less pressure than did active exercises (Tab. 2). During weight shifting, one foot must clear the floor during the swing phase. Contraction of the stance limb's hip abductors is required to keep the pelvis level, and the swing limb's flexors must also contract. Thus, during weight shifting, stresses to the hip were relatively low in this patient, suggesting that practice ambulation, with upper-limb support, may be as safe or safer than active exercise.

The similar peak pressures (Tab. 2) in step left (right swing phase) when compared with step right (right stance phase) may be due to the use of an assistive device that reduced hip abductor muscle activity and unloaded the stance limb during the acute recovery phase.

Sit to stand. The sit-to-stand task also revealed a high peak pressure. During this activity, hip flexion is followed by hip extension, which is essential to clear the buttocks from the chair.(25) In our study, both active hip flexion and isometric hip extension exercise generated high contact pressures, similar to the sit-to-stand results. In a related study of the same patient 1 year after surgery, the sit-to-stand activity produced the highest peak pressure among all tested maneuvers (18 MPa).(6) Performance of the sit-to-stand activity with the use of armrests or a high chair resulted in a decrease of forces acting on the hip and knee joint.(26) Therefore, use of an armchair or a high chair during the early postoperative period may limit the peak acetabular contact pressure.

Clinical Data

The decrease in analgesic intake over time was consistent with a normal postoperative pain pattern. The maximum peak pressure for all the maneuvers, except isometric hip extension, occurred on day 7. On day 7, the patient received a moderate dose of analgesics (Fig. 6), suggesting she was experiencing a rise in discomfort.

Hip flexion active range of motion increased over the postoperative period as expected, with the patient progressively improving in ambulation during the acute recovery phase. She began PWB ambulation using a walker. Prior to discharge, she had progressed to crutches on both level surfaces and stairs (Fig. 4). The decrease in acetabular contact pressures during the latter part of the patient's hospital stay may be due to enhanced movement efficiency from improved activities of daily living techniques (learning), which may involve decreased muscle co-contraction or other unidentified factors. Confirming such speculations is beyond the scope of this study, but the patient clearly experienced a clinically normal recovery from surgery.

Limitations of Study

The single-subject design precludes broad generalization of the results. The data we obtained may be idiosyncratic to this individual. No similar investigations, however, have been reported in the literature. Further, it is not clear what constitutes an important change in contact pressure. Bench tests revealed instrument error to be less than 0.2 MPa.(22) In addition, because increased velocity is thought to increase pressure, differences in angular velocity could have influenced the maximum pressures obtained.(21)| Some activities (eg, active hip flexion, abduction, sit to stand) would be more susceptible to variations in velocity.

Documentation of the subject's clinical status in the physical therapy notes was neither frequent nor complete. There was minimal strength testing, and the MMT that was used lacked precision. The incomplete clinical data hindered comparisons of the pressure data with the patient's perceptions of pain. Although the dosage of analgesics provides indications of daily fluctuations in pain, analgesic dosage may not be a sufficiently precise and accurate method of measuring pain.

Finally, the timing of the activity-specific pressure tests was less than ideal. In some instances, as much as 9 days elapsed before testing was repeated for an activity. Isometric quadriceps femoris muscle exercise was not studied after day 9. The maneuvers tested, however, generally exhibited consistent data with little variability between testing sessions.

Conclusions

Quantitative contact pressure data during the acute phase of rehabilitation were collected from an implanted endoprosthesis. Because the data stem from a single subject, these results may not be generalizable. The data, however, suggest that maximal isometric gluteal exercises produce very high magnitude and highly variable acute postoperative acetabular contact pressures, probably because this exercise is highly effort dependent. Understanding the full implications of rehabilitation protocols for patients with femoral implants and other acutely painful hip conditions requires more information than can be obtained from a single subject. These data, however, allow clinicians to reconsider their approaches to acute rehabilitation following femoral head arthroplasty. In addition, this study demonstrates that exercises commonly used in the acute postoperative phase as preparation for gait activities may produce higher femoroacetabular pressures than those produced by weight shifting and sit-to-stand functional tasks.

Table 1.   Dosage and Duration of
Action of Prescribed Medications During
Inpatient Staya
                   Duration
          Dosage of Action Ratio to
Medication  (mg)      (h)     Tylenol(R)
Tylenol[R]   650       4       1 unit
Codeine       60     4-6       2 units
Darvon[R]     65     5-6       2 units

"This information was used to estimate the relative effectiveness of each medication standardized to the analgesic strength of Tylenol* (650 mg = 1 unit of medication).

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

(daggar) McNeil Consumer Products Co, Div of McNeil-PPC Inc, Fort Washington, PA 19034.

(double daggar) Eli Lilly & Co, Lilly Corporate Center, Indianapolis, IN 46285. TABULAR DATA OMITTED

References

  1 Rydell NW. Forces acting on the femoral
head prosthesis: a study on strain gauge sup
-plied prostheses in living persons. Acta Orthop
Scand. 1966;37(suppl 88):1-132.
  2 English TA, Kilvington M. In vivo records of
the hip loads using a femoral implant with
telemetric output: a preliminary report.
J Biomed Eng. 1979;1:111-115.
  3 Davy DT, Kotzar GM, Brown RH, et al. Tele
-metric force measurements across the hip
after total arthroplasty. J Bone Joint Surg ]Am].
1988;70:45-50.
  4 Hodge WA, Fijan RS, Carlson KL, et al. Con
-tact pressures in the human hip joint mea
-sured in vivo. Proc Nail Acad Sci USA. 1986;83:
2879-2883.
  5 Hodge WA, Fijan RS, Mann RW, Harris WH.
Preliminary in vivo pressure measurements in
a human acetabulum. In: Proceedings of the
31st annual meeting of the Orthopedic Re
-search Society; Las Vegas, Nev; 1985:284.
Abstract.
  6 Hodge WA, Carlson KL, Fijan RS et al. Con
-tact pressures from an instrumented hip en
-doprosthesis. J Bone Joint Surg ]Am]. 1989;71:
1378-1386.
  7 Lestrange NR. Bipolar arthroplasty for 496
hip fractures. Clin Orthop. 1980;251:7-19.
  8 Anderson LD, Hamsa WR, Waring TL.
Femoral head prosthesis: a review of 356 operations
and their results. J Bone Joint Surg [Am]. 1964;
46: 1049-1065.
 9 Harris WH. Sinking prosthesis. Surg Gynecol
Obstet. 1966;123:1297-1302.
 10 Tepic S, Macirowski T, Mann RW. Com
-puter simulation of interarticular fluid flow. In:
Perren SM, Schneider E, eds. Biomechanics:
Current Interdisciplinary Research. Boston,
Mass: Martinus Nijhoff Publishing; 1985:221-226.
  11 Mankin HJ, Brandt KD. Biochemistry and
metabolism of cartilage in osteoarthritis. In:
Moskowitz RW, Gelberman P,H, Mankin HJ,
Treadwell BV, eds. Osteoarthriis: Diagnosis
and Management. Philadelphia, Pa: WB Saun
-ders Co; 1984:43-44.
  12 Lippiello L, Kaye C, Numata T, et al. In vi
-tro metabolic response of articular cartilage
segments to low levels of hydrostatic pressure.
Connect Tissue Res. 1985;3:9%107.
  13 Hall AC, Urban JPG, Gehl KA. The effects of
hydrostatic pressure on matrix synthesis in
articular cartilage. J Orthop Res. 1991;9:1-10.
  14 Thielen PL, Mueller KH. Immediate postop
-erative management of patients with total hip
replacement. Phys Ther. 1973;53:949-954.
  15 Murray MP, Brewer BJ, Zuege RC. Kinesio-logic measurements of the functiona
l perfor
-mance pre and post total hip replacement.
J Bone Joint Surg [Am]. 1972;54:237-256.
  16 Stauffer RN, Smidt GL, Wadsworth JB. Clini
-cal and biochemical analysis of gait following
Charnley total hip replacement. Clin Orthop.
1982;99:70-77.
  17 Kelly IG, Brown TR, Hamblen BG, Paul JP.
A biomechanical analysis of gait following a
unilateral hip replacement. J Bone Joint Surg
[Br]. 1981;63:638.
  18 Perfin T, Dorr LD, Perry J, et al. Functional
evaluation of total hip arthroplasty with 5 to 10
year follow-up evaluation. Clin Orthop. 1985;
195:252-259.
  19 Holden MK, Gill KM, Magliozzi MR, et al.
Clinical gait assessment in the neurologically
impaired: reliability and meaningfulness. Phvs
Ther. 1984;64:35-40.
  20 Mann RW, Burgess RG. An instrumented
endoprosthesis for measuring pressure on ace
-tabular cartilage in vivo. In: Proceedings of the
workshop on implantable telemetry in ortho
-paedics; Berlin, Federal Republic of Germany;
April 1990:1-13.
  21 Krebs DE, Elbaurn L, Riley PO, et al. Exer
-cise and gait effects on in vivo hip contact
pressures. Phys Ther. 1991;71:301-309.
  22 Carlson K. Local Pressures in the Human
Hip Joint In vivo, Correlated with Motion
Kinematics and External Forces. Cambridge,
Mass: Massachusetts Institute of Technology;
1985. Thesis.
  23 Spring Health Drug Reference. Spring
Health, Pa: Spring Health Corp; 1988.
  24 Norkin CC, Levangie PK. Joint Structure
and Function: A Comprehensive Analysis. Phil
-adelphia, Pa: FA Davis Co; 1983:85-114.
  25 Nuzik S, Lamb R, VanSant A, Hirt S. Sit-to-stand movement pattern. Phys The
r. 1986;66:
1708-1713.
  26 Burdett RG, Habasevich R, Pisciotta J, Si
-mon SR. Biomechanical comparison of rising
from two types of chairs. Phys. Ther. 1985;65:
1177-1183.

EM Strickland, PT, is Staff Coordinator, Orthopedics, Department of Physical Therapy, Department of Veterans Affairs Medical Center, Boston, MA 02130. She was a master's degree candidate at the time of this study at the MGH Institute of Health Professions, Boston, MA.

M Fares, PT, is a doctoral candidate at Boston University. She was a master's degree candidate at the time of this study at the MGH Institute of. Health Professions.

DE Krebs, PhD, PT, is Associate Professor, Program in Physical Therapy, MGH Institute of Health Professions, 15 River St, Boston,/VIA 02108-3402 (USA), and Lecturer, Massachusetts Institute of Technology, Cambridge, MA 02139. Address all correspondence to Dr Krebs at the first address.

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

DL Givens-Heiss, PT, OCS, is Senior Clinician, Harmarville Outpatient Rehabilitation Center, 351 W Beau St. Washington, PA 15301-4613, and Instructor, Department of Physical Therapy, Wheeling Jesuit College, Wheeling, WV 26003. She was a master's degree candidate at the time of this study at the MGH Institute of Health Professions.

WA Hodge, MD, is Assistant in Orthopaedics, Massachusetts General Hospital, 5 Longfellow P1, Boston, MA 02114.

RW Mann, ScD, is Whitaker Professor of Biomedical Engineering, Department of Mechanical Engineering, Massachusetts Institute of Technology; Director, MIT Newman Laboratory for Biomechanics and Human Rehabilitation; and Director, Harvard-MIT Rehabilitation Engineering Center. Cambridge, MA 02139.

This study and the surgical procedure were both approved by the human studies committees of Massachusetts General Hospital and the Massachusetts Institute of Technology.

This research was supported in part by US Department of Education Grant #1 HI33P90005, Veterans Administration Medical Center Grant #REUVDHT9898, and the Arthritis Foundation.

This article was submitted September 27, 1991, and was accepted May 18, 1992.