The Journal of Arthroplasty Vol. 13 No. 4 1998

Preoperative Physical Therapy in Primary Total Knee Arthroplasty

J e f f r e y A. R o d g e r s , M D, * K e v i n L. G a r v i n , MD,* Cra ig W. Walke r , MD, t

D e e M o r f o r d , RN, MPA,* J o s h U r b a n , MD,* a n d J o e B e d a r d , B S t

Abstract: In order to evaluate the efficacy of preoperative physical therapy for patients undergoing elective primary total knee arthroplasty, l0 patients completed 6 weeks of physical therapy before surgery (PT group). Ten patients served as controls (C group). Subjects were tested at baseline (PT only), before surgery, 6 weeks after surgery, and 3 months after surgery using the Hospital for Special Surgery knee rating scale, range of motion, thigh circumference, walking speed, Cybex II isokinetic knee flexion, and extension testing, and computed tomography scanning for cross- sectional muscle area. Hospital stay and need for physical therapy after inpatient rehabilitation were also compared. Physical therapy produced modest gains in isokinetic flexion strength in these severely arthritic knees but no difference in extension strength. The decrease in isokinetic strength after surgery was not affected by preoperative physical therapy. Muscle area did not decrease significantly for the PT group, but it did decrease for the C group after surgery. While postoperative strength differences could not be demonstrated, preoperative physical therapy preserved thigh muscle area after surgery. The clinical significance of this finding is uncertain. Consequently, this study failed to support the routine use of preoperative physical therapy in knee replacement surgery. Key words: preoperative physical therapy, range of motion, osteoarthritis.

Despite advancements and research in surgical tech- nique, prostheses, and modalities of rehabilitation, scant attention is paid to preoperative physical preparation of the patient for whom total knee arthroplasty (TKA) has been prescribed. Weakness of both the quadriceps and hamstrings, in contrast to the contralateral limb, is commonly observed in patients with osteoarthritis of the knee [1,2]. After TKA, significant further atrophy of the quadriceps has been shown histologically [3]. An age-related decline in overall muscular strength also begins in adulthood with a significant increase in the rate of

From the Departments of *Orthopaedic Surgery and y-Radiology, University of Nebraska Medical Center, Omaha, Nebraska.

Reprint requests: Jeffrey A. Rodgers, MD, Des Moines Ortho- paedic Surgeons, P.C., 6001 Westown Parkway, West Des Moines, IA 50266.

Copyright © 1998 by Churchill Livingstone ® 0883 -5403/1304-000853.00/0

strength loss beyond the age of 50 years [4]. In addition, a negative nitrogen balance develops fol- lowing major orthopaedic procedures [5]. Consider- ing this list of compounding factors, patients under- going total knee replacement begin the long journey to recovery substantially disadvantaged.

While the large body of knowledge evaluating postoperative physical therapy, including continuous passive motion (CPM) and electrical stimula- tion [6-12] continues to grow, the role of preopera- tive physical therapy in TKA has not been estab- lished, and research is extremely limited as to its efficacy [2,13]. The ability of the elderly to respond to heavy resistance strength training, however, has been clearly demonstrated [14-16]. While it may seem that patients with severe osteoarthritis may be unable to complete a successful preoperative physi- cal therapy program, well-designed studies have shown both feasibility and effectiveness of strength

414

Preoperative PTin PrimaryTKA • Rodgers et al. 415

training in patients with knee osteoarthritis [ 17,18]. The purpose of this controlled investigation was to evaluate prospectively the effects of preoperative physical therapy on short-term outcome variables following primary TKA.

Materials and Methods

Study Population

From December 1992 to August 1995, patients scheduled by the senior author of this report for unilateral primary TKA for osteoarthritis were re- cruited to participate in the study. Patients with a history of uncontrolled hypertension, cerebral aneu- rysm, unstable angina, or any other contraindica- tion to high-intensity physical exertion or testing were excluded. According to the Investigational Review Board-approved protocol, patients were assigned to the control (C) group or the physical therapy group (PT) based on their geographic avail- ability. Those from the focal metropolitan area were invited to participate in the PT group, while those living too far to attend the physical therapy sessions were invited to participate in the C group. Origi- nally, 11 patients were enrolled in the C group; one withdrew from the protocol for personal reasons, which left a group of I0 to complete the study. Group C comprised five men and five women with an average age of 65 years (range, 50-83). The original PT group enrolled 12 patients, two of these discontinued the protocol: one patient's operation was cancelled because of occult coronary artery disease and the other was unable to perform the physical testing because of fibromyalgia. The remain- ing 10 patients who completed the preoperative physical therapy protocol included four men and six women, with an average age of 70 years (range, 63-78). There was no statistical difference in the age or sex distribution of the two groups. One of the patients in the control group had undergone contra- lateral TKA 1 year before the study, otherwise all other patients had native knees.

Physical and Functional Testing

Before surgery subjects were tested at baseline (PT only), and then at 6 weeks and 3 months after surgery. Testing and measurement included Hospi- tal for Special Surgery Knee Rating Scale (HSS Knee Score) [19], range of motion, Cybex II (Lumex Inc., Ronkonkoma, NY) isokinetic knee flexion and exten- sion testing (3 practice trials followed by 3 testing trials at 60°/s and 180°/s), walking speed (10 m normal and tandem gait), and thigh circumference.

The HSS knee score was performed by the senior investigator or his resident staff, while the remain- der of the testing was performed by a certified physical therapist. The duration of hospitalization, need for posthospitalization physical therapy, and complication rates were also compared.

Muscle Area

All patients also were examined with a "single- slice" CT examination of both thighs for muscle area using a previously described technique [16] at baseline (PT only), before surgery and at 6 weeks after surgery (Fig. 1). The cross section of interest was established using the midpoint between the center of the femoral head and the medial femoral condyle. The absolute measurement from the cen- ter of the femoral head to the midpoint of the femoral shaft was used for subsequent scans. The images were analyzed using MTRACE (University of Iowa, 1992) on a SUN image-processing station. Gray scale levels ->i corresponded to muscle and those ~<0 corresponded to fat. The bone was ex- cluded using the region of interest function. The number of pixels for each group were then counted and converted to square centimeters. Using this technique, intramuscular fat was excluded.

Intervention

The PT group completed 6 weeks of preoperative physical therapy three times per week under the direction of a certified physical therapist. Each patient's program was individualized according to their baseline physical capacity and reevaluated and advanced accordingly after 3 weeks. Exercises in- cluded stretching and warm-up, heel-slides, isomet- ric quadricep sets, straight leg raises, short-arc quad- ricep sets, standing squats, step-ups, and bicycling. Both groups received preoperative physical therapy instruction in the usual postoperative exercise pro- tocol.

All patients were reconstructed using the same posterior-stabilized cemented total knee implant (Insall-Burstein II, Zimmer, Warsaw, IN). They re- ceived the same postoperative physical therapy including ankle pumps, quadricep sets, straight leg raises, short-arc quadricep sets, heel-slides, assisted flexion, calf-stretching, hamstring-stretching, ham- string sets, hip abduction, and hip adduction exer- cise. Patients started gait-training (weight-bearing as tolerated) beginning on the first postoperative day. Depending on the patient's progress and living situation, patients were either discharged to home with instructions for a home physical therapy pro-

416 The Journal of Arthroplasty Vol. 13 No. 4 June 1998

Fig. 1. Mid-thigh "single slice" computed tomography scan analysis. (A) Representative CT scan with region of interest function activated. (B) Histogram of Gray scale distribution: -->1 represents muscle and <--0 represents fat.

gram or transferred to a geriatric rehabilitation center for supervised physical and occupational therapy. At the discretion of the senior author, outpat ient physical therapy was prescribed postop- eratively as necessary, regardless of their study group designation.

Statistical Analysis

Statistical analysis of the data included Repeated Measure Analysis of Variance to compare the trends be tween the groups and t test: Paired Two-Sample for Means for comparison of differences within each group.

Results

Physical and Functional Testing

The groups did not differ significantly with respect to extension and flexion range of motion over time, although a trend toward decreased motion was demon- strated for both groups at the 6-week evaluation (Table 1). The thigh circumference and 10-m walk times also did not differ significantly over time for either group. Hospital for Special Surgery knee rating scores improved for both groups at the 3-month follow-up with no difference in degree of improvement (Fig 2).

Table 1. Anthropometric and Physical Testing Data

Range of M o t i o n

E x t e n s i o n F l ex ion

M e a n Range M e a n Range

Th igh Circumference

I n v o l v e d U n i n v o l v e d (cm) (cm)

W a l k Time

10 m 10 mT (s) (s)

HSS Score

Score Range

Control Preop 8 0-25 i13 7 7 - i 4 4 6 weeks 9 0 -28 104 8 4 - i 2 6 3 m o n t h s 6 0-20 113 80-117

PT Baseline 4 0 -10 112 99-130 Preop 7 0-25 112 85-128 6 weeks 5 2-15 i01 85-125 3 m o n t h s 4 0-10 109 9 5 - i 2 0

52 52 l0 35 54 40-67 52 51 13 34 5 i 50 9 32 85 68-97

52 53 9 35 60 44-79 53 53 10 33 51 51 12 36 52 51 10 26 87 79-95

HSS, Hospital for Special Surgery; Preop, preoperat ive; PT, physical test.

ioo 80

60

40

20

• Control M PT

0 base 3 mo

F i g . 2. H o s p i t a l for S p e c i a l S u r g e r y k n e e r a t i n g scores .

Isokinetic Testing Data

Table 2 summar izes the isokinetic peak torque data for bo th groups. At preopera t ive (C) and baseline (PT) evaluat ion, the involved knee was significantly weake r for bo th groups in flexion and extension at 60°/s (Figs 3, 4) and for flexion at 180°/s for the C group only. Significant s t rength i m p r o v e m e n t wi th training was demons t ra ted for the PT group wi th a 5 ft lb increase in flexion strength at 60°Is (17%, P = .01) f rom baseline to preoperat ive. Extension s t rength improved a m e a n of 2 ft lb but was not significant (Figs 5, 6).

Both groups demons t ra ted decreased peak- to rque at 60°/s in the involved knee at the 6 -week follow- up. The C group decreased 28% (P = .06) in flexion and 30% (P = .05) in extension, while the PT group

Preoperative PTin PrimaryTKA • Rodgers et al. 417

60 deg./sec. ft.lb~

100

80

60

40

20

0 Flexion Extension

F i g . 3. P r e o p e r a t i v e p e a k f l e x i o n a n d e x t e n s i o n t o r q u e :

c o n t r o l g r o u p (60° /s ) .

decreased 40% (P = .007) in flexion and 30% (P = .02) in extension. The 3 -mon th testing re~ vealed recovery to baseline s trength for both groups at 60°Is and 180°/s. Values improved for the PT group f rom 6 weeks to 3 mon ths 9 ft lb (36%) in flexion (P = .02) and 13 ft. lb (33%) in extension (P = .002). The C group failed to demons t ra te a statistically significant i m p r o v e m e n t during this in- terval. Overall, repeated measures analysis of vari- ance revealed no significant difference be tween the groups over time.

Muscle Area

The cross-sectional muscle area of the thigh failed to change significantly for the PT group f rom base-

Table 2. C y b e x I s o k i n e t i c Testing

Peak Torque (ft-lb)

F l ex ion Extension

60 d e g / s 180 deg / s 60 deg / s 180 deg / s

U n i n v I n v o l v e d Un inv I n v o l v e d Un inv I n v o l v e d U n i n v I n v o l v e d

Control Pre-op 6 w e e k 3 m o n t h

PT Baseline Pre-op 6 w e e k 3 m o n t h

P vaIues paired N e s t

42a 32b 43 25e 42 33f

36g 30h 43 35k 38 251 39 34m

a vs. b 0.03 c vs. d 0.02 b vs. e 0.06 e vs. f ns g vs. h 0.05 i vs. j ns h vs. k 0.01 k vs. 1 0.007 I vs. in 0.02

32c 26d 8 i n 57o 46p 42q 34 22 79 44r 48 33 29 28 85 56s 50 38

26i 21j 70t 51u 45v 35x 29 25 73 53y 46 37 30 13 71 40z 46 28 30 24 67 53aa 43 37

n vs. o 0.02 p vs. q ns o vs. r 0.05 r VS. S ns t vs. u 0.03 v vs. x ns u vs. y ns y vs. z 0.02 z vs. aa 0.002

418

ft.ibs.

80

60

40

20

0

The Journal of Arthroplasty Vol. 13 No. 4 June 1998

60 deg./se¢.

Flexion Extension Fig. 4. Baseline peak flexion and extension torque: physi- cal therapy group (60°/s).

60 deg./sec. ft-lbs

60

50

45

40

3 5 ~ "base pre-op 6wk* 3mo

Fig. 6. Extension peak torque over time.

Control - l l - PT .,¢..

line to preoperat ive (Table 3) for either the involved or uninvolved extremity. Involved thigh muscle area decreased f rom a mean of 105.3 cm 2 to 94.0 cm 2 (not significant) for the PT group, while the area decreased from 1 I2.5 cm 2 to 90.1 cm 2 (P : .04) for the C group (Fig. 7). The uninvolved extremity did not change significantly 6 weeks after surgery for ei ther group. Again, repeated measures analysis of variance failed to demonstra te a difference be- tween groups.

Hospitalization and Physical Therapy Utilization

Acute hospital stays averaged 5 days (range, 3-9 days) for the C group and 6 days (range, 3-12 days) for the PT group. Rehabilitation unit stays were required for four patients in the control group (mean stay, 6 days) and for six patients in the PT group (mean stay, 4 days). Overall hospitalization (acute and rehabilitation) did not differ be tween groups and averaged 8 days for the PT group and 7 days for the C group.

The need for addi t ional ou tpa t i en t physical therapy was also not different be tween groups (6 of

ft-lbs

36

34

32

30

28

26

60 deg./sec.

24 base

Control 41" PT

pre-op* 6wk* 3mo

Fig. 5. Flexion peak torque over time.

I0 in the C group and 7 of 10 in the PT group). No patient in either group developed clinically evident deep venous thrombosis nor did they require fob low-up knee manipulat ion for poor range of mo- tion. In a follow-up "exit interview," 9 of l0 of the patients in the PT group said they felt the preopera- tive physical therapy helped them prepare for sur- gery, and they would do it again if they were to have the opposite knee reconstructed.

Discussion

Functionally, preoperat ive physical therapy does not appear to have a significant effect on range of mot ion or maximal walking speed. Likewise, there was no difference in the degree of improvement for the HSS knee scores. This ins t rument relies on estimations of pain and subjective measures of function, strength, and stability [19]. Subtle differ- ences would be difficult to detect using this mea- sure, especially considering the t r emendous impact of surgery alone.

Isokinetic strength testing is a reproducible and convenient means of assessment of strength in many pathologic conditions {20]. Peak torque is the variable that is traditionally assessed in isokinetic studies and has p roven most reliable in research applications [2i]. The observed weakness of the involved knees at baseline in both flexion and extension is consistent with current literature [I,20]. While the PT group produced gains in flexion strength with training, these gains did not translate to the immediate postoperative period. The PT group's strength gains f rom 6 weeks to 3 months after surgery were significant, but the C group's gains were not significant. One would not expect a delayed effect of preoperative physical therapy. These differences most likely represent small sample size error and the inability of this study to demon- strate a statistically significant difference for the C

Preoperative PTin PrimaryTKA • Rodgers et al. 419

Table 3. Computed Tomograpby Muscle Area

I n v o l v e d U n i n v o l v e d

Total T h i g h M u s c l e c m 2 c m 2

I n t r a m u s c u l a r Total T h i g h M u s c l e I n t r a m u s c u l a r Fat cm 2 % M u s c l e c m 2 c m 2 Fat c m 2 % M u s c l e

Cont ro l Preop 6 weeks

PT Basel ine Preop 6 w e e k s

P va lues pa i red t- test

261 .56 112 .52a 244 .32 90 .12b

285 .86 I 0 8 . 7 1 c 284 .95 105 .31d 266 .67 94 .00e

a v s . b c v s . d d v s . e

25.62 43 .02 262 .04 113.85 23.85 43.45 27 .66 36.89 251 .97 111.35 21.85 44 .19

23 .49 38.03 298 .73 117.04 23 .8 I 39.18 24 .84 36.96 295 .50 114.43 22 .88 38.72 25.15 35.25 271 .80 110.44 22. l 1 40 .63

0 .04 *no signif icant differences n.s. n.S.

group. Repeated measures analysis showed no differ- ence between the groups, supporting this conten- tion.

Quadricep strength is essential to immediate post- operative rehabilitation and progress in weight- bearing [22]. Later in rehabilitation, the emergence of symmetrical and uniform gait also depends on increased quadricep strength [ 1 ]. Extension strength, however, failed to improve with training. Patello- femoral pain during testing may have limited perfor- mance, although in a study of isokinetic perfor- mance of patients with osteoarthritis of the knees, Lankhorst et al. [20] felt the influence of pain on torque was minimal [20].

One can also not ignore the effect of specificity of training on strength measurement. The physical therapy program utilized predominantly closed- chain, isotonic exercise, while the testing consisted of open-chain isokinetic measurement. This factor has not been investigated for diseased individuals, but overwhehning evidence supports exercise-type specificity regarding isokinetic versus isotonic exer- cise for normal individuals [21]. Consequently,

cm2 120 ~ . _ 115 - I10 ~

105 ~,

100 i "I 911

i

i [ I

pre-op 6wk ~

Uninvolved Control

4 - FT

Involved Control

-O- PT

Fig. 7. Computed tomography muscle area over time.

more significant differences in strength may have been present, but not measured. However, we are not aware of an isotonic strength testing apparatus that is as reliable, safe, and convenient as the Cybex II isokinetic testing device [23].

Overall, the recovery of Both groups to baseline strength By the 3-month evaluation was quite remarkable. While Berman measured isokinetic strength after TKA, his initial postoperative measure- ment was performed from 3 to 6 months after surgery. This is the first study to evaluate isokinetic strength of all patients 3 months after surgery and document recovery to baseline. It is important to note that this "recovery" may be more a function of pain relief (allowing a better isokinetic test) than true strength gain.

The computed tomography (CT) muscle area data provide more convincing evidence of the positive effect of preoperative physical therapy. While preop- erative physical training failed to produce muscle area increase, these data suggest that, after surgery, muscle thigh area may be preserved by preoperative physical training. In addition, this study is the first to definitively measure the significant muscle area changes that occur following TKA. It is important to note that the relatively large changes in muscle area demonstrated By the CT analysis did not correlate with the thigh circumference measurement.

In this climate of increasing pressure to limit costs and decrease utilization of medical resources, all new treatment interventions will need to be scruti- nized in this light. In this limited study, no savings in terms of decreased hospital stay or need for post- hospitalization physical therapy could be demon- strated for the physical therapy. Many intangible variables beyond the control of this study influence these measures including confounding medical prob- lems, family expectations, living arrangements, and even day of the week of surgery.

420 The Journal of Arthroplasty Vol. 13 No. 4 June-1998

Few investigators have studied the effect of pre- operat ive physical therapy in TKA. Weidenhie lm et al. [2] invest igated preopera t ive isometric exer- cise in patients scheduled for un i compar tmen ta l knee rep lacement [2]. These authors demons t ra ted decreased self-selected walking speed preopera- tively wi th i m p r o v e m e n t in pa in and perceived stability, bu t no difference in strength, range of mot ion, or oxygen cost of walking. Three m on ths after surgery, no differences could be demons t ra ted be tween the C and PT group. These authors con- cluded their s tudy did not show any major benefi t f rom the the rapy tested. D 'Lima et al. [13] recently invest igated two preopera t ive physical the rapy pro- grams before TKA. They also failed to demons t ra te the value of physical therapy (either s t rength train- ing or aerobic conditioning) using the HSS knee score, Quality of Well-Being survey, and the Arthri- tis Impac t M e a s u r e m e n t Scale. These ins t ruments are not designed to measure the effects of physical therapy and are probably not sensitive enough to demons t ra te significant change.

This s tudy is the first to evaluate the effect of preopera t ive physical the rapy on TKA using objec- tive m e a s u r e m e n t techniques in the immedia te postopera t ive period, but it is not wi thou t its limita- tions. Randomiza t ion was not possible under the constraints of available funding for the physical therapy sessions in a single location. The concept of "prerandomiza t ion" has been validated previously, h o w e v e r [24]. Our exper imenta l design takes this concept of group ass ignment before rec ru i tment one step fur ther by basing it on geographic avail- ability.

While this s tudy failed to produce convincing evidence of the benefi t of preopera t ive physical therapy in TKA using strength m e a s u r e m e n t and funct ional parameters , it did demons t ra te accu- rately the decrease in muscle area following the procedure. In addition, physical the rapy m a y help limit this atrophy. The clinical significance of this finding is uncertain.

While 9 of 10 patients in the PT group felt that preopera t ive the rapy was beneficial, the objective data do not suppor t rout ine use. The inconvenience and expense of a preopera t ive the rapy p rog ram cannot be justified based on this study.

Acknowledgments

We thank Liz Ruby, of the UNMC Depar tmen t of Preventat ive and Societal Medicine for her statisti-

cal analysis, and the UNMC Depar tments of Physical Therapy and Radiology for their assistance in this study.

References

1. Berman AT, Bosacco S J, Israelite C: Evaluation of total knee arthroplasty using isokinetic testing. Clin Orthop 271:106, I991

2. Weidenhielm L, Mattsson E, Brostrom L, Wersall- Robertsson E: Effect of preoperative physiotherapy in unicompartmental prosthetic knee replacement. Scand J Rehab Med 25:33, I993

3. Martin TP, Gundersen LA, Blevins FT, Coutts RD: The influence of functional electrical stimulation on the properties of vastus lateralis fibers following total knee arthroplasty. Scand J Rehab Med 23:207, i991

4. Larsson L, Grimby G, Karlsson J: Muscle strength and speed of movement in relation to age and muscle morphology. J Appl Physio146:451, 1978

5. Michelsen CB, Askanazi J, Grump FE, Elsyn D, Kinney JM, Stinchfield FE: Changes in metabolism and muscle composition associated with total hip replacement. J Trauma 19:29, 1979

6. Gotlin RS, Hershkowitz S, Juris PM, Gonzalez EG, Scott WN, Insall JN: Electrical stimulation effect on extensor lag and length of hospital stay after total knee arthroplasty. Arch Phys Med Rehab 75:957, 1994

7. Haug J, Wood LT: Efficacy of neuromuscular stimula- tion of quadriceps femoris during continuous passive motion following total knee arthroplasty. Arch Phys Med Rehab 69:423, 1988

8. Nadler SF, Malanga FA, Zimmerman JR: Continuous passive motion in the rehabilitation setting. Am J Phys Med Rehab 72:162, 1995

9. Nielsen PT, Rechnagel K, Nielsen S: No effect of continuous passive motion after arthroplasy of the knee. Acta Orthop Scand 59:580, 1988

10. Ritter MA, Gandolf VS, Holston K: Continuous pas- sive motion versus physical therapy in total knee arthroplasty. Clin Orthop 244:239, I989

11. Ververeli PA, Sutton DC, Hearn SL, Booth RE, EIozack W J, Rothman RR: Continuous passive motion after total knee arthroplasty: analysis of cost and benefits. Clin Orthop 321:208, 1995

12. Wasilewski SA, Woods LC, Torgerson WR, Healy WL: Value of continuous passive motion in total knee arthroplasty. Orthopedics I3:291, 1990

13. D'Lima DD, Colwell DW, Morris BA, Hardwick ME, Kozin F: The effect of preopeative exercise on total knee replacement outcomes. Clin Orthop 326:174, 1996

14. Fiatarone MA, Marks EC, Ryan ND, Meredith CN, Lipsitz LA, Evans WJ: High-Intensity strength train- ing in nonagenarians. JAMA 263:3029, 1990

Preoperative PTin PrimaryTKA • Rodgers et al. 421

15. Fisher NM, Pendergrast DR, Calkins E: Muscle reha- bilitation in impaired elderly nursing home residents. Arch Phys Med Rehab 72:18I, 1991

16. Frontera WR, Meredith CN, O'Reilly KP, Knuttgen HG, Evans W J: Strength conditioning in older men: skeletal muscle hyper t rophy and improved function. J Appl Physiol 64:1038, 1988

17. Fisher NM, Pendergrast DR, Gresham GE, Calkins E: Muscle rehabilitation: its effect on muscular and funct ional per formance of pat ients wi th knee osteoarthritis. Arch Phys Med Rehab 72:367, 1991

18. Minor MA, Hewett JE, Webel RR, Anderson SK, Kay DR: Efficacy of physical conditioning exercise in patients with rheumatoid arthrotis and osteoarthritis. Arthritis Rheum 32:1396, 1989

i9. Insall JN, Ranawat CS, Aglietti P, Shine J: Comparison

of four models of total knee-replacement prostheses. J Bone Joint Surg [Am] 58:754, i976

20. Lankhorst GJ, VandeStadt RJ, VanderKorst JK: The relationships of functional capacity, pain, and isomet- ric and isokinetic torque in osteoarthrosis of the knee. Scand J Rehab Med 17:167, 1985

21. Morrissey MC, Harman EA, Johnson M J: Resistance training modes: specificity and effectiveness. Med Sci Sports Exerc 27:648, 1995

22. Krackow KA: The technique of total knee arthro- plasty, p. 388. CV Mosby, St. Louis, MO, 1990

23. Almekinders LC, Oman J: Isokinetic muscle testing: is it clinically useful]? J Am Acad Orthop Surg 2:221, 1994

24. Chang RW, Falconer J, Stulberg SD, Arnold W J, Dyer AR: Prerandomization: an alternative to classic random- ization. J Bone Joint Surg [Am] 72:1451, 1990