Knee kinematics and kinetics during locomotor activities of daily living in subjects with knee arthroplasty and in healthy control subjects.

Although many researchers examine the differences in kinematics and kinetics between healthy subjects and patients with knee arthroplasty (KA) during locomotor activities of daily living (ADLs), little attention has focused on ADL requirements as goals for postoperative knee rehabilitation.[1-10] Few data exist that support specific rehabilitative goals currently in use. Isokinetic dynamometers, for example, are frequently used to measure and enhance muscle performance of the postsurgical knee. Normative ADL data for knee angular velocities and moments are absent from the literature. Current isokinetic protocols and goals are therefore based largely on local custom.

Many publications describe knee kinematics or kinetics in level human gait.[11-15] Sagittal knee motion during gait has been studied extensively in healthy individuals, and several researchers [1,4,5,9,10] report alterations in knee range of motion (ROM) during gait in patients with KA. Gait ground reaction forces and knee joint moments can reveal differences between healthy subjects and subjects with KA.[1,5,10] Only two published studies[4,16] have examined the relationship between the velocity and range of knee motion in healthy individuals. No study has investigated knee angular velocity, joint moments, and ROM during gait or other locomotor ADLs.

Several studies of the knee joint motion and the moments that occur when a subject rises from a chair have been reported. Sagittal knee ROMs have been shown to vary with chair height.[3,17] Most investigations of the chair-rise maneuver have been devoted to information on knee joint moments. Rodosky et al[17] reported that knee joint moments produced during the sit-to-stand maneuver increased with decreasing chair height. They reported that the lowest moments occurred with a chair height greater than or equal to knee height. Similar results have been reported in patients following knee arthroplasty. These studies have not reported on the rate velocity) or range of knee motion in association with kinetic data or on gait data.

Andriacchi et al[18] emphasized the significance of stair ascent and descent in the assessment of knee joint kinematics and kinetics. They have shown that greater ROM and moment production in the sagittal plane is required at the knee during stair climbing in patients with KA than during level gait in subjects without knee disorders.[2] The functional performance at the knee required for stair climbing has had an impact on knee prosthesis design, but these studies on stair climbing have not explored the angular velocity requirements for stair ascent and descent in healthy subjects or patients with KA. To date, no study has provided an analysis of knee joint motion and moments for level gait, rising from a chair, and stair climbing from the same subjects.

The purpose of our study was to analyze knee kinematics and kinetics in patients with KA and in healthy control subjects during gait, chair, and stair locomotor ADLs. Within the control group, we hypothesized that no significant right/left asymmetries or age-related differences would be found. Between the control and KA groups, we hypothesized there would be differences in sagittal knee ROM, joint moments, and angular velocities. Finally, we provide descriptive data on these locomotor ADLs in an effort to stimulate improvements in knee rehabilitative efforts and implant designs.

Method

Subjects

Subject characteristics are listed n Table 1. The control group of 11 healthy subjects with no musculoskeletal complaints, whose ages ranged from 26 to 88 years (X = 52.5, SD = 22.1). All con could walk at least 3.2 km (2 miles) without stopping. Patients in the KA group were divided int groups: patients with unicompartmental KAs (UNI group) and patients with tricompartmental KAs (TRI group). Prior studies[1,3,8,9,19] suggest that subjects with unicompartmental KAs may perform at levels closer to those of healthy control subjects. The UNI group consisted of 5 subjects with 7 unicompartmental knees, whose ages ranged from 64 to 73 years (X = 67.1, SD = 2.8). The TRI group consisted of 10 subjects with 12 posterior cruciate ligament sparing knees, whose ages ranged from 61 to 78 years (FC = 68.6, SD = 4.9). No significant differences in height or weight were found between groups, although the KA group subjects were a mean of 11.2 kg heavier than the control group subjects.

[TABULAR DATA OMITTED]

All KA group subjects were at least 1 year postarthroplasty (range = 12-19 months) and had a mean Hospital for Special Surgery (HSS) knee assessment score[19] of 94 (range = 90-100). Each patient had an excellent clinical result (ie, each patient had an HSS knee assessment score of >90) and was considered fully rehabilitated. ]Each KA group subject had a knee ROM of [greater than or equal to] 110 degrees; a manual muscle test grade of [greater than or equal to] 4/5 (at least "Good"), and was able to walk without assistive devices, ascend and descend stairs in a reciprocal manner, and rise from a chair independently.

Each subject was tested with the same protocol and completed the protocol without difficulty. One female control subject did not complete the gait portion of the protocol because of equipment malfunction.

Instrumentation

The data-acquisition area was a 9.1 x 12.2-M (30 x 40-ft) carpeted room, on the ground floor. It was equipped with a four-camera bilateral kinematic data-acquisition system based on [Selspot.sup.TM] II cameras.(*) Ground reaction forces were measured simultaneously with two Kistler piezoelectlic force platforms[dagger] mounted on a granite foundation. Two computers, a DEC PDP 11/60[double dagger] and a DEC MicroVAX [II.sup.TM] workstation,[double dagger] were used for data acquisition. processing, and analysis.

The [Selspot.sup.TM] system detects active infrared light-emitting diodes (LEDs). Sixty-four LEDs embedded in rigid plastic arrays were anchored to the 11 body segments with polypropylene molds and straps (Fig. 1). Arrays were mounted on both feet, shanks, and thighs and on the pelvis, trunk, arms, and head. The viewing volume was about 1.8 m per side. The data sampling rate was 153 frames per second.

A modified version of the [TRACK.sup.C] [NEWTON.sup.C] software packages[subsections] was used for analysis of bilateral kinematic and kinetic data. The [TRACK.sup.C] computer program permits automatic calculation of the six degrees of freedom individual body segments (three translational and three rotational), producing three-dimensional trajectory data. This technique, which uses automatic error detection routines within the software, computes joint angles using the (3-1-2) Cardan angles method.[20] The modified [NEWTON.sup.C] software combines the [TRACK.sup.C], results with force-plate data and estimates of body segment mass and inertial properties to determine net joint torques.[21]

Test Protocol

Stair-climbing, gait, and chair-rising data were collected on the same day for all subjects. Two trials of all activities, separated by several minutes, rest, were collected for each lower extremity. All activities were paced in an attempt to make certain comparisons within and across subjects and were not confounded by major differences in movement velocity. Paced cadences were chosen to match normal rates.[1-7,11-14,21-24] Thus, only during gait could differing stride lengths affect locomotor velocity; velocity differences can confound kinematic and kinetic comparisons.

For the stair ascent and descent trials, right and left modular staircases with four steps per staircase were utilized. The second and third steps (18 x 28 cm, height x depth) of each staircase were positioned over the force platform. The staircases were separated by 0.5 cm, allowing the right and left-side forces to be measured independently by the two force plates. Kinematic and kinetic data were acquired in the mid-stair areas, when the subjects were on steps 2 and 3. The subject was positioned in the staircase mid-line to avoid crossover of the feet to the opposite force platform. Subjects ascended and descended the stairs at a pace of 80 steps per minute, moving to the beat of a metronome. After the subjects were able to accomplish the task smoothly, data were collected over four separate trials. Two trials began with the right foot moving first, and two trials began with the left foot moving first.

For the gait trials, each subject began to walk 3 to 4 m away from the force platform. The investigator aligned the subject with the mid-line of the platform, and the subject was instructed to walk toward a marker on a wall across the room. The subject practiced walking at a pace of 120 steps per minute. and two trials with right-limb force-plate strikes and two trials with left-limb force-plate strikes were then performed. A successful trial occurred when the complete gait cycle was seen in the viewing volume and the subject accurately struck the force platform. Subjects were not aware of the force platform location.

The sit-to-stand trials were performed with armless and backless chairs of adjustable height.[21] For this study, chair height was set at the level of each subject's knee height, as measured from the floor to the medial tibiofemoral joint line with the subject in a standing position. The chair was positioned off the force plates, so that each foot was on a separate force plate, to allow independent force measurements. The subject was seated prior to data collection, and the feet were, positioned so that each ankle was in 18 degrees of dorsiflexion. There positions were then marked for subsequent trials. The subject was asked to cross his or her arms and hold them against the chest to prevent upper-limb participation, and then to arise from the chair within 1.2 seconds.[21-23] After several practice trials, two data-acquisition trials were completed.

Data Analysis

The [TRACK.sup.C]/[NEWTON.sup.C] software Produced data in a format available for graphical interpretation. Joint moments and ground reaction forces were normalized to percentage of body weight in newtons (N-m/%BW) to permit body-mass-normalized comparisons among subjects.[7,9] The graphics software used in the data analysis allowed for manual or automatic retrieval of minimum and maximum values for an entire data-acquisition cycle. All data presented in this study were manually selected and verified. The data were then analyzed with Statistical Analysis System (version 6.03) statistical software.[parallel] A multivariate analysis of variance (MANOVA) for repeated measures was used for all inferential statistics presented in the "Results" section to protect the overall alpha level at .05; when multivariate significance was found, univariate contrasts were performed.

Results

The repeated-measures MANOVA was first performed to determine intertrial reliability for the control group. Analysis of all sagittal-plane parameters between trials 1 and 2 showed no statistically significant differences (F<3.98, P>.05) for all activities (Tab. 2).

[TABULAR DATA OMITTED]

The control group was subdivided by age to assess possible differences before comparison with the KA group. Analysis of all sagittal-plane parameters was done by age, with the young group having ages less than 40 years and the older group having ages greater than 60 years. No statistically significant differences (F<4.13, P>.05) were found for sagittal ROM, moments, angular velocities, and vertical forces for all activities between the young and elderly groups (Tab. 3).

[TABULAR DATA OMITTED]

Similarly, the data were analyzed to assess the validity of combining the control group subjects' bilateral knee performance data in this study. There were no statistically significant differences (F< 2.9, P>.05) in any variable between the right and left extremities in the control subjects (Tab. 4).

[TABULAR DATA OMITTED]

In addition, the MANOVA revealed no statistical significance (F<2.6, P>.05) for sagittal ROM, moments, angular velocities, and vertical forces between the UNI group subjects and the TRI group subjects. Subsequently, these groups were combined and classified simply as the KA group.

Sagittal knee ROM for the KA group was significantly different than for the control group for all activities (Hotelling-Lawley F=5.83, P<.05). The largest mean sagittal knee excursions ([+ or -] SD) occurred during stair ascent (91.8 [degrees] [+ or -] 10.4 [degrees] and 82.5 [degrees] [+ or -] 11.4 [degrees] for the control and KA groups, respectively). The lowest mean knee ROMs were recorded during gait for both groups, with control group subjects having a mean ROM of 63.3 [+ or -] 8.1 degrees and KA group subjects hiving a mean ROM of 56.1 [+ or -] 5.1 degrees (Fig. 2).

Normalized peak knee sagittal moment differences between the control and KA groups for all activities were not statistically significant (F<3.62, P>.05). Both groups used similar patterns of moment production. Stair descent moments were the largest, with control group subjects producing 11.9 [+ or -] 2.9 N [multiplied by] m/%BW and KA group subjects producing 11.3 [+ or -] 3.0 N [multiplied by] m/ %BW. The lowest peak moments occurred during gait (7.4+/-2.9 N-m/ %BW and 6.2 +/- 1.9 N-m/,%BW for the control and KA groups, respectively) (Fig. 3).

Knee angular velocity measurements Were analyzed over three phases: loaded (stance), extension (swing), and flexion (swing). The maximum swing phase velocities for both extension and flexion were statistically different between the control and KA groups only for gait (F=7.86 and 6.74, respectively; P<.05). The largest swing phase extension velocities (374.6 [+ or -] 46.3 [degrees]/s and 320.6 [+ or -] 31.7 [degrees]/s for control and KA groups, respectively) occurred during gait trials (Fig. 4). The largest swing phase flexion velocities (374..0 [+ or -] 54.5 [degrees]/s and 359.6 [+ or -] 48.7 [degrees]/s in the control and KA groups, respectively (Fig. 5) were recorded in stair ascent.

Loaded angular velocities were lower for the KA group than for the control group for all activities (F>5.74, P [less than or equal to] .0l) except stair descent (F=4.35, P>.05). The largest loaded knee angular velocities were recorded for chair rising (ie, 140.1 [+ or -] 34.1 [degrees]/s and 117.6 [+ or -] 27.0 [degrees]/s for control and KA groups, respectively). The stair descent activity showed significantly different and lower peak loaded angular velocities of 27.2+/-15.3'/s and 40.9 [+ or -] 11.6 [degrees]/s for the control and KA groups, respectively (Fig. 6).

Vertical forces did not differ between the control and KA groups for any activity (F<3.23, P>.05). Mean vertical forces were greatest during stair descent, being 139% [+ or -] 28% of body weight (%BW) for the control group and 128% [+ or -] 24%BW for the KA group. The chair-rise activity, which involves no unilateral lower-limb forces, produced the lowest mean peak vertical forces of all activities, being 65% [+ or -] 9%BW for the control group and 60% [+ or -] 5%BW for the KA group (Fig. 7).

Discussion

There are no generally accepted criteria defining when a patient with KA reaches maximum benefit from physical therapy. Although we do not know the details of each KA group subject's physical therapy program, we know that all KA group subjects were discharged from physical therapy after being considered fully rehabilitated by their orthopedist and physical therapist. All KA group subjects had excellent HSS scores. One important finding of our study, therefore, is that a gait laboratory can detect and quantify functional performance decrements not revealed by clinical examination. For example, knee torque and angular velocity during locomotion - variables that were significantly decreased from "excellent" or normal values in our KA group subjects - cannot be measured without laboratory instrumentation.

Reliability

Kadaba et al[12] reported high intertrial reliability ("coefficients of multiple determination") in gait assessment when the same subject was tested on separate days. Our study also demonstrated high intertrial reliability for same-day gait, chair, and stair assessments (Tab. 2). Because each subject was evaluated during a single visit, limb marker placements were not changed during testing. Trials were also accepted only when two investigators observed the activity, reviewed the LED displays, and agreed that the subject had completed the maneuver successfully. The TRACK[C] software minimized the introduction of system errors into both data acquisition and processing.[23]

Age Comparisons

The effects of age differences on gait assessment have been discussed previously in the literature. Murray et al[15] found differences in gait characteristics between healthy men aged 20 to 25 years and men aged >65 years, but they did not find differences between the young men and healthy men aged 50 to 65 years. Our study also demonstrated no differences in kinematic and kinetic data for the healthy control group subjects when compared by age. This finding might be related to the testing protocol: The pace requirement for the activities tested minimized differences in velocities, both linear and angular. Range of motion and moment production might have then been less similar between the young and old groups than if the subjects had been permitted to perform the activities at each individual's desired pace, because the older group (including the KA group subjects) would be expected to walk more slowly. For the testing protocol used in this study, age did not appear to influence the results. Power analysis of the values in Table 3 suggest that to find a significant age effect with an alpha level of .05 and a beta level of .2, 34 and 81 subjects per age group would be needed even to detect the maximum mean difference (swing extension velocity and sagittal ascent ROM, respectively) for the standard deviations observed. Thus, combining young and older subjects into a single "healthy" group appears to the reasonable. Future studies should include more subjects to increase the age comparisons' statistical power.

Left-Right Symmetry 

Comparisons

Symmetry of knee motion had been difficult to study in the past, primarily because of the use of unilateral data-acquisition systems, which preclude simultaneous right and left knee comparisons during a given activity. Hannah and Morrison,[24] however, found high sagittal-plane knee motion symmetry in the time and frequency domains in a group of 12 healthy subjects. Our system permits simultaneous bilateral kinematic and kinetic data collection for each activity tested. This capability allowed ready comparison of right and left knee motion differences. The healthy individuals tested displayed a high degree of sagittal-plane symmetry for all measured variables in the gait, stair, and chair ADLs (Tab. 4). Indeed, there was nc) clear trend of left or right knees providing maximum differences, and statistical power analysis of the values in Table 4 suggest that even if a sufficient sample size were available to detect statistical significance, the differences would not be clinically meaningful. Thus, previous assumptions of knee kinematic symmetry from unilateral data used for comparisons with pathologic motion are further supported.

Functional Effects of

Knee Arthroplasty

Knee ROM in the sagittal plane was significantly less in the KA group than in the control group for all activities. Although all KA group subjects were pain-free, freedom from pain apparently entity does not ensure the attainment of normal knee angular excursions. post patients with KA have some limitation of ROM because of their preoperative osteoarthritis, and the surgical procedure can create scar tissue in the healing process, further limiting ROM. A combination of these factors most likely accounts for the decreased motion seen postoperatively. This combination of factors has clinical implications for patients with both knee and hip arthritis. Patients with KA need to increase their hip motion to compensate for their decrease in knee motion, despite the fact that our KA group subjects had sufficient passive range of motion to accomplish all ADLs. Consequently, it may be prudent to address severe arthritic disease of the hip prior to KA.

Although knee excursions were different between the KA and control subjects, both groups demonstrated a similar pattern of knee joint motion in all activities. This finding suggests that although a mechanism exists to limit knee flexion and extension, the basic motor control mechanisms governing knee motion were intact.

Somewhat surprisingly, greater sagittal ROM was required for stair ascent than for stair descent (Fig. 2). Apparently, the loaded limb's stance phase motion during ascent is always less than during descent, but in order for the unloaded swinging limb to ascend, it requires more knee flexion to clear the step than in descent. McFayden and Winter[25] examined aspects of this motion only during the stance phase of gait. This greater flexion in descent than in ascent was consistently found in all subjects and was related to the movement of their center of gravity. When climbing the stairs, subjects were required to raise their center of gravity to the next step and then actively carry it forward to safely clear the next step.[25] In ascent, the hip is more extended as the foot comes off the floor, requiring the knee to flex more to clear the step. While descending stairs, however, subjects must actively propel their center of gravity forward and then resist gravity while lowering to the next step. Because the center of gravity is more forward in descent, the swinging limb has the more flexed hip and less knee flexion is required to clear the step. This phenomenon should be studied further.

The sagittal moments described in this report reflect the maximum net flexion moments during the midstance phase (Fig. 3). These moments are normally resisted primarily with the knee extensors. The mean knee moments produced by the KA group subjects were lower for each activity as compared with the control group, but not statistically different. This trend of lower knee moments could be secondary to postoperative quadriceps femoris muscle motor unit alterations that led to inability to attain desired extension forces,[26] but it could also be due to differences in triceps surae muscle activation or upper-body compensatory motions.[3]

Consistent with a report by Andriacchi et al,[18] the largest flexion moments were recorded for stair descent. Similarly, the vertical ground reaction forces were significantly higher for stair descent than for gait trials for both the control and KA groups, presumably related to the higher potential and kinetic energy of descent. That KA group subjects used smaller knee moments than did control group subjects to perform identical locomotor tasks may be explained by compensatory upper-body motions.[3,7,9] Such compensatory motions accommodate the decreased knee ROM, velocity, and muscle performance by substituting other methods of advancing and lifting the lower limbs. Altered upper-body motions can also bring the body's anteroposterior (AP) center of mass position closer to the AP knee joint center, thus decreasing the net knee flexion moment that must be resisted by the extensors (Fig. 8). These data support the importance of examining knee kinematic and kinetic data during stair climbing. in addition to more typically examined gait activities.[1]

Implications for Rehabilitation

Exercises

Sapega,[27] along with many other workers, has decried the lack of quantified functional goals for knee exercise Programs, suggesting that studies similar to ours be done to guide clinical decision making. At present, many therapists use the contralateral knee's peak torque value or regression equations[28] as the standard for determining "normal" knee extensor torque values during knee exercise.[29] Neither approach addresses the locomotor ADLs or other functional knee kinematic and kinetic requirements. The results of our study, if replicated on large representative samples of healthy subjects and with other subjects with knee impairments, could help guide postoperative knee rehabilitation toward more functional knee performance goals.

Angular velocity. Unloaded (swing phase) knee angular velocities, both extension and flexion, measured for the control and KA groups were significantly different in this study only for the gait trials. Gait produced the largest swing phase extension velocities. Brinkman and Perry,[4] who reported angular velocities for gait similar to those of this study, also found that although angular velocities improved after KA, these measurements remained abnormally low 1 year after surgery. In our study, swing phase velocity abnormalities may even have been minimized as a result of the pacing requirement designed to obtain, comparability among subjects in both control and KA groups.

The loaded (stance phase) knee angular velocities provide more direct information concerning knee exercise regimens. For example, although stair descent required the greatest moment production and generated the greatest vertical ground reaction force in both groups, It demonstrated the lowest loaded angular velocities. This result probably occurred because during stair descent, eccentric contraction (extensor torques developed during loaded knee flexion) was required to decelerate the knee to prevent buckling, whereas the other activities required concentric quadriceps femoris muscle contraction to accelerate the body against gravity. This could explain why stair descent comparisons between the control and KA groups differed from other activities (Figs. 2, 4-6), as the knees of the KA group subjects were less able to perform effective deceleration. Although swing ph" angular velocities were as high as 433 [degrees]/s, no loaded angular velocity was greater than 180*/s in any subject. Thus, rehabilitation regimens that use exercise at > 180 [degrees]/s, whether concentrically or eccentrically, cannot be justified on grounds of locomotor ADL demands.

Knee moments. These data suggest that peak knee extensor muscle exercise goals, to accomplish ADL function, might include torques of 10 to 15 N [multiplied by] m/%BW. Indeed, the KA group's data could similarly be used to establish goals and limits for the patient following KA, as all patients tested had excellent clinical results. Rehabilitation goals, however, should probably acknowledge a "safety factor" that could be well in excess of the data presented here, as we obtained these values only during well-protected, controlled indoor activities. Preliminarily, we suggest a safety factor of 2 for knee torque goals (eg, 20-30 N [multiplied by] m/%BW at up to 180 [degrees]/s). This twofold safety factor is commonly used in engineering design and is well within (by [greater than or equal to] 30%) previous reports of healthy subjects' (including the elderly) peak knee extension isokinetic capacity.[28,30,31] We did not, however, measure these subjects' isokinetic muscle performance, and our results could be idiosyncratic to the subjects tested. Future studies should examine similar healthy control subjects and subjects with knee joint impairments with motion analysis, as we have presented, and with isokinetic measures. Further studies are needed to determine the appropriate "margin of safety" for locomotor knee torques and velocities before firm conclusions are offered on function-based goals for isokinetics or other exercises.

Conclusions

We have presented kinematic and kinetic knee data for selected ADLs in both a healthy control group and a post-KA group. No significant left-right or age-related differences were found in the control group, and high trial-to-trial reliability was found in both groups for all data.

Stair climbing appeared to be the most demanding activity, requiting the greatest knee ROM and largest knee moment production. We agree with previous researchers who have suggested that stair climbing be considered in the assessment of postoperative knee function. These in vivo knee performance data suggest that current exercise regimens could be better tailored to knee angular velocity and torque used during gait and ADLs, if future empirical studies are performed to address these questions directly. Our results do not provide the conclusive evidence needed to dictate changes in present rehabilitation exercises, but they do suggest a strong need for further investigations of current practices.

References

[1] Andriacchi TP, Galante JO, Fermier RW. The influence of total knee-replacement design on walking and stair climbing. J Bone Joint Surg [Am]. 1982;64:1328-1335 [2] Andriacchi TP, Stanwyck S. Galante JO, Knee biomechanics and total knee replacement. J Arthroplasty. 1986;1:211-219. [3] Berger RA, Elbaum LH, Schenkman ML. et al, Functional differences following unicompartmental total knee replacements. Read at the 36th Annual Meeting of the Orthopaedic Research Society; February 1990; New Orleans. La2. [4] Brinkman JR, Perry J. Rate and range of knee motion during ambulation in healthy and arthritic subjects. Phys Ther. 1985;65:1055-1060. [5] Chao EYS, Laughman RK, Stauffer RN. Biomechanical gait evaluation of pre and postoperative total knee replacement patients, Arch Orthop Traumat Surg. 1980;97:309-317. [6] Kettelkamp DB, Nasca R. Biomechanics and knee replacement arthroplasty. Clin Orthop. 1973;94:8-14. [7] Laughman RK, Stauffer RN, Ilstrup DM, Chao EYS. functional evaluation of total knee replacement. J Orthop Res. 1984;2:307-313. [8] Mikosz RP, Andriacchi TP, Mahoney JJ, Rosenberg AG. Functional differences in patients with total and unicompartmental knee replacements for osteoarthritis of the knee. Read at the 36th Annual Meeting of the Orthopaedic Research Society; February 1990; New Orleans, La. [9] Weinstein JN, Andriacchi TP, Galante JO. Factors influencing walking and stair climbing following unicompartmental knee arthroplasty. J Arthroplasty. 1986;1:109-115. [10] Whittle MW, Jefferson R. Functional biomechanical assessment of the Oxford meniscal knee. J Arthroplasty. 1989;4:231-243. [11] Chao EYS, Laughman RK, Schneider E, Stauffer RN. Normative data of knee joint motion and ground reaction forces in adult level walking. J Biomech. 1983;16:219-233. [12] Kadaba MP, Ramakrishnan HK, Wootten ME, et al. Repeatability of kinematic, kinetic, and electromyographic data in normal adult gait. J Orthrop Res. 1989;7:849-860. [13] Kettelkamp DB, johnson RJ, Smidt GL, et al. An electrogoniometric study of knee motion in normal gait. J Bone Joint Surg [Am]. 1970;52:775-790. [14] Laubenthal KN, Smidt GL, Kettelkamp DB. A quantitative analysis of knee motion during activities of daily living. Phys Ther. 1972;52: 34-42. [15] Murray MP, Kory RC, Clarkson BH. Walking patterns in healthy old men. J Gerontol. 1969; 24:169-178. [16] Finley FR, Karpovich PV. Electrogoniometric, analysis of normal and pathological gaits. Res Q (Suppl) 1964;5:379-384. [17] Rodosky MW, Andriacchi TP, Andersson GBJ. The influence of chair height on lower limb mechanics during rising. J Orthop Res. 1989:7:266-271. [18] Andriacchi TP, Andersson GBJ, Fermier RW, et al. A study of lower-limb mechanics during stair climbing. J Bone Joint Surg [Am]. 1980:62:749-757. [19] Insall JN, Ranawat CS, Aglietti P, Shine J. A comparison of four models of total knee-replacement prostheses. J Bone Joint Surg [Am]. 1976;58:754-765. [20] Tupling SJ, Pierrynowski MR. Use of Cardan angles to locate rigid bodies in three-dimensional space. Med Biol Eng Comput. 1987:25:527-532. [21] Ikeda ER, Schenkman ML, Riley PO, Hodge WA. Influence of age on dynamics of rising from a chair. Phys Ther. 1991;71:473-481. [22] Schenkman ML, Berger RA, Riley PO, et al. Whole-body movements during rising to standing from sitting. Phys Ther. 1990:70:638-648. [23] Riley PO, Schenkman ML, Mann RW, Hodge WA. Mechanics of a constrained chair rise. J Biomech. 1991;24:77-85. [24] Hannah RE, Morrison JB. Kinematic symmetry of the lower limbs. Arch Phys Med Rehabil. 1984;65:155-158 [25] McFayden BE, Winter DA. An integrated biomechanical analysis of normal stair ascent and descent. J Biomech. 1988;21:733-744. [26] Krebs DE, Staples WH, Cuttita D, Zickel RE. Knee joint angle: its relationship to quadriceps femoris activity in normal and postarthrotomy limbs. Arch Phys Med Rehabil. 1983;64:441-447. [27] Sapega AA. Muscle performance evaluation in orthopaedic practice. J Bone Joint Surg [Aml 1990;72:1562-1574. [28] Gross MT, Credle JK, Hopkins LA, Kollins TM. Validity of knee flexion and extension peak torque prediction models. Phys Ther. 1990;70:3-10. [29] Krebs DE. Isokinetic, electrophysiologic and clinical function relationships following tourniquet-aided arthrotomy. Phys Ther. 1989; 69:803-815 [30] Knapik JJ, Ramos MU. Isokinetic and isometric torque relationships in the human body. Arch Phys Med Rehabil. 1980;61:64-67. [31] Lankhorst GJ, Van de Stadt RJ. Van der Korst, JK. The relationships of functional capacity, pain and isometric and isokinetic torque in osteoarthrosis of the knee. Scand J Rehabil Med. 1985;17:167-172.