Secondary Logo

Exercise Improves Early Functional Recovery After Total Hip Arthroplasty

Gilbey, Helen, J.*; Ackland, Timothy, R.*; Wang, Allan, W.**; Morton, Alan, R.*; Trouchet, Therese; Tapper, Jeff

Clinical Orthopaedics and Related Research: March 2003 - Volume 408 - Issue - p 193-200

The purpose of this prospective, randomized study was to apply an 8-week customized exercise program to patients (Group E) scheduled for total hip arthroplasty, followed by a postsurgery exercise program, and show the effect on functional recovery compared with control subjects (Group C) who received no additional exercise apart from routine in-hospital physical therapy. Strength, range of motion, and physical function tests were completed by 57 patients at Week 8 and Week 1 before surgery and at Weeks 3, 12, and 24 postoperatively. No differences between the exercise and control groups were observed at baseline. By 1 week before surgery, patients in Group E had shown significant improvements for Western Ontario and McMaster Universities Osteoarthritis Index (total score, stiffness, and physical function components), and combined hip strength. Patients in Group E had improved hip flexion range of motion in the diseased hip compared with patients in Group C. Significant differences in outcome measures between Group E and Group C were observed throughout the postoperative phase from Weeks 3 to 24. The current study showed that customized perioperative exercise programs are well tolerated by patients with end-stage hip arthritis, and are effective in improving early recovery of physical function after total hip arthroplasty.

From the *Department of Human Movement and Exercise Science and the **Division of Orthopaedic Surgery, The University of Western Australia, Crawley, Western Australia; the School of Occupational Therapy, Curtin University of Technology, Perth, Western Australia; and the Department of Physiotherapy, Sir Charles Gairdner Hospital, Nedlands, Western Australia.

Reprint requests to Timothy R. Ackland, PhD, Department of Human Movement and Exercise Science, The University of Western Australia, 35 Stirling Highway, Crawley, Western Australia 6009. Phone: 61-8-9380-2668; Fax: 61-8-9380-1039; E-mail:

Received: May 1, 2001.

Revised: July 19, 2001; January 24, 2002; June 12, 2002.

Accepted: August 9, 2002.

DOI: 10.1097/01.blo.0000053168.71678.2b

Individuals with advanced hip arthritis are more likely to have a lower level of muscle strength, 15,16,20,22 exercise tolerance, and reduced aerobic capacity 8,13,14,18 in comparison with healthy individuals. Increased pain and deterioration of the hip also lead to decreased mobility and reduced independence in some activities of daily living. 4,9

Clinical evidence suggests that fit, strong patients generally rehabilitate more quickly after surgery compared with patients who are less fit. 6,19,21,24 However, little scientific evidence exists to support the efficacy of presurgery and postsurgery exercise programs in patients having total hip arthroplasty. 11,17,25 Wijgman et al 25 reported that preoperative physical therapy and instruction were not useful for patients before total hip arthroplasty. In their sample of 31 patients, few differences were observed between the patients and control subjects on measures such as the visual analog pain scale and the Harris hip score. In a published critique, however, Oosterveld 17 argued that there were too few objective measures in the study and that patients in the experimental group tended to experience less pain than the control subjects at almost every testing occasion. The Bramlett OrthoPACE™ management system (Walk Inc, Birmingham, AL) recorded data on presurgery exercise in 92 patients having primary total hip arthroplasty. 2 More than 90% of patients who were treated with their program of specific education and a prescribed rehabilitation exercise regimen had an average length of hospital stay of 4.2 days compared with an historic cohort of 7.2 days, suggesting that presurgery exercise can improve early rehabilitation after total hip arthroplasty.

The purpose of the current prospective, randomized study was to determine if structured presurgery and postsurgery exercise programs can improve muscular strength and early functional recovery of patients after total hip arthroplasty compared with control subjects who had the normal course of care.

Back to Top | Article Outline


Seventy-six patients with end-stage hip arthritis were recruited during 24 months beginning in January 1997. All patients met the following criteria: (1) diagnosis of osteoarthritis, posttraumatic arthritis, inflammatory arthritis, osteonecrosis, or Paget’s disease of the hip; (2) chronic pain and disability, unresponsive to conservative treatment; and (3) stable health and fit for anesthesia. Patients who had a history of infection in the hip, significant neuromuscular disease, malignancy in the area of the hip, poor general health, or required revision hip surgery or bilateral hip replacements were excluded. Of the 141 patients contacted, 14 did not meet the inclusion criteria, 23 could not commit to the time requirements, and surgery was scheduled too soon for the remaining 28.

Institutional ethics approval was attained, and all individuals attending the familiarization session consented to participate in the study. Patients became familiar with test procedures before random allocation was made to the exercise or control group. Baseline musculoskeletal testing and disease-specific outcome measures were done approximately 8 weeks before surgery, and then were repeated the week before surgery and postoperatively at Weeks 3, 12, and 24. Patient demographics, comorbidities, and prior surgical procedures were compiled from the patient’s hospital files. Most patients (87%) presented at baseline with at least one comorbidity, and although no attempt was made to grade their severity, the most common medical problems were hypertension (50%), asthma (24%), cancer (19%), and ischemic heart disease (19%). Patients allocated to the exercise and control groups were well matched.

Back to Top | Article Outline

Musculoskeletal Testing Procedures

Strength of bilateral thigh flexor and extensor musculature was measured using the Kinetech hip and knee machine (Keylink Physical Care, Blair Athol, South Australia) as peak torque (Nm) with the mean speed of the lever arm at 120°/second (± 5°/second). Isometric thigh abduction strength (kg) was measured using a calibrated cable tensiometer.

The quantification of strength may be complicated by numerous factors including pain, reduced range of motion (ROM) of the hip, and in particular, apprehension in assuming the required test position. 22 Therefore, the Kinetech hip and knee machine was fitted with a specifically designed trunk stabilization unit 5 that allowed patients to walk into the test position and complete the assessment in a weightbearing posture. Body weight was supported partially by the stabilization unit, thereby reducing the stress on the support leg during lower limb strength testing, which allowed a more accurate assessment of the noninvolved side.

A Velcro strap attached to the backrest was placed around the torso of the patient to isolate movements of the hip. Patients received no verbal feedback on performance until the end of the session when all the tests had been completed.

The patient stood with the hips in 0° flexion and 0° abduction and with the knees fully extended for measurement of isometric thigh abduction strength. Two Velcro straps connected by a short length of wire cable were used for this test. One strap was placed around the distal thigh of the patient with the lower edge located 5 cm above the superior aspect of the patella. The other strap was attached to the upright support of the hip machine on the contralateral side and a cable tensiometer was fitted onto the wire. From this position the patient was asked to contract the abductor musculature maximally for a count of three.

All strength tests were repeated two to five times until the scores reached a plateau or decreased, or the patient had increased pain in the joint. A minimum of 30-seconds rest was given between trials, and the highest score was recorded.

Back to Top | Article Outline

Analysis of Strength Data

The performance of most activities of daily living requires the combined effects of lower extremity muscle strength. Therefore, to reduce the number of strength variables, a combined strength score that represented bilateral functionally of related muscle groups was computed. Because of differences in the relative force generating capacity of different muscle groups, it was appropriate to standardize the strength scores of each muscle group to a common scale before summation. 7 Therefore, each of the six thigh strength measures (thigh flexion, extension, and abduction for both limbs) was converted to a z-score, and the mean of these was calculated to produce a combined strength score.

Back to Top | Article Outline

Measurement of Hip Flexion Range of Motion

The patient stood in the Kinetech hip machine with the upper body supported, while a Leighton Flexometer (Leighton Inc, Spokane, WA) was strapped onto the lateral surface of the thigh approximately 5 cm above the joint margin of the knee. The tester held the support leg in extension and the patient was instructed to flex the contralateral hip by raising the knee as high as possible. Two trials were completed for the involved and noninvolved hips and the higher score for each hip was recorded.

Back to Top | Article Outline

Disease-Specific Outcomes Assessment

The effect of training on a patient’s symptoms from the affected hip was evaluated using the Western Ontario and McMaster Universities Osteoarthritis Index, 3 a validated disease-specific questionnaire. The index consists of 24 questions, divided into three domains: pain, stiffness, and difficulty with physical function. Individual questions were assigned a score between 0 points (no pain, stiffness, or difficulty with physical functions) and 4 points (extreme pain, stiffness, or difficulty with physical functions). Results were normalized to produce a total score between zero (best) and 100 (worst). Since 1988 the scoring of this instrument has had some changes, 10,12 however, the original scoring system developed by Bellamy et al 3 was used in the current study.

At the assessment 1 week before surgery each patient completed a Patient Satisfaction Questionnaire, which required patients to rate their current level of overall pain and level of general health in comparison with that experienced at baseline. Patients used a Likert scale that incorporated five statements ranging from much better to much worse.

Back to Top | Article Outline

Exercise Intervention Program

Patients were required to do two supervised clinic-based and two home-based exercise sessions each week for 8 weeks before surgery. Patients returned to the clinic after the 3-week test after surgery, once their wound had healed, and continued clinic-based exercise sessions until 12 weeks after surgery. Participation in the exercise program beyond this time was optional, although patients were encouraged to continue with the home-based program. The clinic exercise session was approximately 1 hour with a 30-minute aerobic and strength program followed by a 30-minute program of mobility and gait training in the hydrotherapy pool.

After a 5-minute warm-up on a cycle, arm, or rowing ergometer, the exercise program sought to improve lower limb strength and range of hip flexion motion (resistance exercises included heel raise, leg flexion and extension, thigh flexion and extension, and isometric thigh abduction). In addition, exercises to improve trunk, shoulder, and arm strength were included to help the patient with bed to chair transfers and walking with a walker or crutches (resistance exercises included trunk flexion and trunk rotation, forearm curls, seated body raise, and hip hikes). Isotonic exercise programs began with one set of 10 repetitions, increasing to three sets of 10 repetitions as patients improved, whereas the isotonic protocol required patients to do two sets of 10 repetitions with the resistance developed and held for 5 seconds.

The hydrotherapy sessions included walking at a depth approximately midabdominal (5 minutes), stretching exercises (5 minutes), mobility and strength exercises (10 minutes), and water cycling and running (10 minutes). Physiologic principles of progressive overload and specificity consistent with the American College of Sports Medicine 1 guidelines for strength training were incorporated. After initial training, patients were able to exercise independently, however, no more than three patients were supervised by the therapist at any time.

Patients were instructed in the home-based training program during the first clinic session. This program was tailored individually for the patient depending on the level of mobility, degree of pain, age, and the amount of help available at home. Each patient was given an instruction booklet and a home exercise log book, which they were asked to complete each week. Adherence to the home-based program was monitored weekly and any problems were resolved.

The perioperative exercise programs were done in the Physical Therapy department of a University teaching hospital. Program cost, including a full time exercise physiologist and administrative support was approximately $400 (United States) per patient in Group E.

Back to Top | Article Outline

Research Design

At baseline a series of one-way analyses of variance (ANOVA) were done to determine if initial differences existed between the two subject groups for all the variables assessed. The effects of the intervention strategy on research parameters were analyzed using a series of three-factor ANOVA (Group × Age × Time) with repeated measures on the Time main effect. Statistical power tests were computed for each variable to ensure sufficient power existed in the event that the null hypothesis was to be accepted. A Student’s t test (two-tailed) was applied to determine differences between the experimental and control groups for the number of presurgery comorbidities and postsurgery complications.

Back to Top | Article Outline


Eight of the 76 patients (11%) withdrew before surgery. Six surgeries were canceled because of medical reasons (stroke, two; infection, two; other illness, two), and two patients in Group E chose to postpone surgery after completion of the 8-week presurgery exercise program because of markedly reduced pain and improved function. Demographic data are presented in Table 1 for the 68 patients who had surgery as part of this research.



Degenerative osteoarthritis was the diagnosis for total hip arthroplasty in 59 patients; other conditions included osteonecrosis of the femoral head, posttraumatic osteoarthritis, inflammatory arthritis, and Paget’s disease. Previous total hip arthroplasty on the contralateral side had been done in 21% of patients and all were functioning well at the time of testing. The contralateral hip was symptomatic in 10% of patients, whereas 24% had symptomatic knee involvement, and symptomatic back pain was apparent in 32% of the sample.

Surgery was done by one of four orthopaedic surgeons. The posterior approach was used in most surgeries (n = 55) with the remainder by the lateral approach. No differences between the groups were observed for the operative procedures of implant fixation and approach, nor in the number of postoperative complications.

After surgery, 11 patients (five in the exercise group and six control subjects) were not assessed because of social (vacation) or clinical (superficial wound infection, thrombosis) reasons. The prospective results of this study therefore are based on the 57 patients (32 in the exercise group and 25 control subjects) with complete preoperative and postoperative data sets. Compliance with the exercise protocols was excellent. Patients attended an average 97% of the scheduled exercise sessions, and documented that they completed 95% of home-based sessions. There were no training related injuries.

No differences were found between the two groups for any of the baseline parameters for patient demographics, number of comorbidities, combined hip strength, hip flexion ROM, and the self-administered questionnaires. The data initially were treated using a three-factor ANOVA, but as no significant main effect for age (p > 0.05) was observed, the data were collapsed and subsequent analyses were done using a two-factor ANOVA (Group × Time).

Back to Top | Article Outline

Presurgery Changes in Subject Outcome Parameters

Results from the assessment done 1 week before surgery revealed that the exercise group had improved scores from baseline for each of the variables assessed. Significant differences between the two groups were observed for Western Ontario and McMaster Universities Osteoarthritis Index total score (p = 0.02) (Fig 1), and for the domains of stiffness (p = 0.014) and physical function (p = 0.035).

Fig 1.

Fig 1.

Analysis of data for the domain of pain indicated that although the patients in Group E had lower mean scores than control subjects 1 week before surgery, the difference between groups was not significant (p = 0.096). However, responses to the Patient Satisfaction Questionnaire revealed that participation in the presurgery exercise program had a positive effect on reducing overall pain levels perceived by the exercise group before total hip arthroplasty. More than 63% of the patients in Group E, compared with only 13% of Group C subjects rated their general level of pain as somewhat better or much better compared with that experienced at baseline. In addition, 67% of the patients in the exercise group rated their general health as much better or somewhat better in comparison with baseline, whereas only 26% of control subjects gave a similar response.

Analysis of the presurgery combined strength scores (Fig 2) revealed differences between patients in Group E compared with Group C, 1 week before surgery. Patients in the exercise group had increased hip strength scores from baseline between 13% and 37%, and improvements between 12% and 28% for the noninvolved side. Mean strength levels for patients in the control group remained the same as at baseline, or were decreased slightly. The level of strength in both hips of patients in Group E was significantly higher (p = 0.04) than the control subjects. Mean combined strength scores attained at baseline for patients in both groups were recorded in the negative because the calculation of this variable was based on the mean of all hip strength scores throughout the study, which followed the two groups from 8 weeks before surgery (baseline) to 24 weeks after surgery. However, it is the relative change in strength scores that are important rather than the numerical value.

Fig 2.

Fig 2.

No group differences were observed for hip ROM at baseline 8 weeks before surgery (Fig 3); however, greater ROM scores were measured for patients in the exercise group 1 week before surgery, compared with the control group subjects. The presurgery intervention effected an improvement in the mean ROM of the diseased and healthy joints by 21% and 11% respectively, whereas ROM in the control group remained mostly unchanged.

Fig 3.

Fig 3.

Back to Top | Article Outline

Postsurgery Changes in Subject Outcome Parameters

Results from the Western Ontario and McMaster Universities Osteoarthritis Index revealed that the three domains of pain, stiffness, and physical function showed a significant (p ≤ 0.001) improvement over time with patients in the exercise and control groups having improved scores at Week 3 after surgery in comparison with baseline. Significant differences (p < 0.01) between the groups were observed for total score (Fig 1) and the domain of physical function at all postsurgery times with the exercise group having improved scores compared with the control group. No group differences were observed for the domain of stiffness 3 weeks after surgery, however, the group mean scores separated once patients began their rehabilitation programs. The surgical procedure had an overwhelming effect on pain scores that were only marginally influenced by exercise intervention.

Not all patients were able to participate in the bilateral thigh strength tests 3 weeks after surgery so data are presented only for two postsurgery times. The results (Fig 2) show significant (p < 0.01) improvements made by all patients after total hip arthroplasty, with the presurgery group differences in strength maintained through 24 weeks after surgery (p < 0.01).

By Week 3 after surgery mean hip flexion ROM of the surgically treated hip remained above baseline levels for patients in Group E (Fig 3), and significant differences (p < 0.05) between these patients and control subjects were observed. With rehabilitation, patients in the exercise group increased the rate of improvement in hip ROM (p < 0.01) compared with those in the control group.

Back to Top | Article Outline


Patients with end-stage arthritis of the hip have lower levels of hip muscle strength compared with age and gender-matched norms. Research 20 also has shown strength deficits of 51% (thigh extensors) and 68% (thigh flexors) between healthy and diseased hips before surgery.

This physical deconditioning may have an impact on patients’ early rehabilitation and functional recovery after total hip arthroplasty. In a prospective study of 65 patients having primary total hip arthroplasty, Wang et al 23 reported that composite hip strength, measured before surgery, was a significant independent predictor of the length of hospital stay. The Bramlett OrthoPACE™ patient management system 2 used preoperative exercise programs and achieved a reduced length of hospital stay compared with an historic control.

The current randomized prospective study examined the impact of presurgery and postsurgery customized exercise programs on patients having primary total hip arthroplasty. These exercise programs were well tolerated by all patients, preoperatively and in the early stages after surgery. Compliance with the exercise sessions was high in this elderly, disabled group of patients.

Significant improvements in hip strength and ROM of the involved hip were achieved for the exercise group compared with the control group before surgery. These benefits were associated with significantly better physical function scores for patients in the exercise group compared with the control group, and better scores on the patient satisfaction questionnaire. Improved physical function before surgery provided a sound base from which to rehabilitate the patient after surgery.

At Week 3 after surgery, the mean Western Ontario and McMaster Universities Osteoarthritis Index total and physical function scores attained by patients in the exercise group were significantly improved (p < 0.01) compared with subjects in the control group. This suggests that the benefits of the presurgery exercise program had a sustained effect during the early postoperative phase. Improved functional capability has the potential to reduce the patients’ need for caregivers or other social support services after surgery.

The postoperative testing and exercise program commenced at Week 3 and continued to Week 24 after surgery. Not only did patients in the exercise group continue to significantly outperform the control group during this period, but the differences between groups also tended to increase. The diverging graphs for hip strength in particular (Fig 2) seemed to indicate that the exercise program led to a greater rate of improvement and had accelerated the rate of rehabilitation after surgery.

A limitation of the current study was that the preoperative and postoperative assessments were made without the assessor being blinded as to the treatment group. However, potential measurement bias was minimized by the assessor’s use of standard instructions for each of the test procedures.

Another possible limitation of the current study was the extensive exclusion criteria. Participation in this study required all patients to be well motivated, capable of exercising, and prepared to commit the time and effort to attend clinic sessions. Poorly motivated patients would be unlikely to benefit from the perioperative program. However, these programs may have been of greater benefit to the more elderly and frail who may have had extensive physical deconditioning before surgery. The patient who is older and more chronically disabled may achieve proportionally greater gains through participation in such an exercise program than was experienced by the group of patients studied.

This study examined only early functional recovery after total hip arthroplasty. An additional limitation of the study was that no evaluation was done at a longer followup (12 months or 2 years after surgery). Shih et al 20 reported that thigh flexor strength in the operative limb remained between 79% and 89% weaker than the healthy contralateral limb 12 months after surgery. Although the relative benefits of a perioperative exercise program no longer may be apparent 12 months after surgery, the physical gains in the first 6 months would have significantly benefited the patient in terms of independent living, ambulatory ability, and return to recreational activities and employment.

The mean Western Ontario and McMaster Universities Osteoarthritis Index score for control subjects 24 weeks after surgery still did not match that achieved by the patients in the exercise group at Week 12. These data indicate that perioperative exercise accelerated functional recovery after total hip arthroplasty by at least 3 months. In young and active patients who need to return to employment, or elderly and more frail patients who might place greater demands on rehabilitation services, such perioperative programs are likely to be cost effective.

Results from the current study indicate that an 8-week presurgery exercise program improves levels of pain, stiffness, physical function, hip flexion ROM, and muscle strength in patients with end-stage hip disease. Postoperative exercise rehabilitation not only maintained the functional advantage for the 6-month duration of the study, but also increased the rate of improvement of patients who participated in the program compared with patients who received normal care.

Back to Top | Article Outline


1. American College of Sports Medicine: Guideline for Exercise Testing and Prescription. Philadelphia, Lea & Febiger 1995.
2. Beadling L: Patient management system lowers TJR costs. Orthop Today 2:24–25, 1996.
3. Bellamy N, Buchanan W, Goldsmith C, Campbell J, Sitt L: Validation study of WOMAC: A health status instrument for measuring clinically important patient-relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 15:1833–1840, 1988.
4. Bunning R, Materson R: A rational program of exercise for patients with OA. Semin Arthritis Rheum 21(Suppl 2):33–43, 1991.
5. Cahalan DT, Johnson ME, Liu S, Chao EYS: Quantitative measurements of hip strength in different age groups. Clin Orthop 246:136–145, 1989.
6. Carswell S, Holman B, Thompson J, Walker W: Acceptable level of aerobic power for patients undergoing elective surgery. J Physiol 285:13, 1978.
7. Chandler J, Duncan P, Kochersberger G, Studenski S: Is lower extremity strength gain associated with improvement in physical performance and disability in frail, community-dwelling elders? Arch Phys Med Rehabil 79:24–30, 1998.
8. Danneskiold-Samsoe B, Lyngeberg K, Risum T, Telling M: Effect of water exercise therapy given to patients with rheumatoid arthritis. Scand J Rehab Med 19:31–35, 1987.
9. Harris W: Traumatic arthritis of the hip after dislocation and acetabular fractures: Treatment by mold arthroplasty. J Bone Joint Surg 51A:737–755, 1969.
10. Hip and Knee Registry: Massachusetts: Center for Outcomes Research. Worcester, University of Massachusetts Medical Center 1997.
11. King L: Case study: Physical therapy management of hip osteoarthritis prior to total hip arthroplasty. J Orthop Sports Phys Ther 26:35–38, 1997.
12. Laupacis A, Bourne R, Rorabeck C, et al: The effect of elective total hip replacement on health-related quality of life. J Bone Joint Surg 75A:1619–1626, 1993.
13. Lyngberg K, Danneskiold-Samsoe B, Halskov O: The effect of physical training on patients with rheumatoid arthritis: Changes in disease activity, muscle strength, and aerobic capacity: A clinically controlled minimized cross-over study. Clin Exp Rheumatol 6:253–260, 1988.
14. Minor M, Hewett J, Webel R, Dreisinger T, Kay D: Exercise tolerance and disease related measure in patients with rheumatoid arthritis and osteoarthritis. J Rheumatol 15:905–911, 1988.
15. Murray M, Brewer B, Zuege R: Kinesiologic measurement of functional performance before and after McKee-Farrar total hip replacement. J Bone Joint Surg 54A:237–256, 1972.
16. Neumann D: An electromyographic study of the hip abductor muscles as subjects with a hip prosthesis walked with different methods of using a cane and carrying a load. Phys Ther 79:1163–1173, 1999.
17. Oosterveld F: No positive effect of preoperative exercise therapy and instruction in patients about to undergo hip arthroplasty. Netherlands Tijdschrift Voor Geneeskunde 138:2520, 1994.
18. Ries M, Groff G, Philbin E: Cardiovascular fitness and health in patients with end-stage osteoarthritis. Arthritis Rheum 38:799–805, 1995.
19. Shelbourne D, Klootwyk T, DeCarlo M: Update on accelerated rehabilitation after anterior cruciate ligament reconstruction. J Orthop Sports Phys Ther 15:303–308, 1992.
20. Shih C, Du Y, Lin Y, Wu C: Muscular recovery around the hip joint after total hip arthroplasty. Clin Orthop 302:115–120, 1994.
21. Shilling JA, Molen MT: Physical fitness and its relationship to preoperative recovery in abdominal hysterectomy patients. Heart Lung 13:639–644, 1984.
22. Vaz M, Kramer J, Rorabeck C, Bourne R: Isometric hip abductor strength following total hip replacement and its relationship to functional assessments. J Orthop Sports Phys Ther 18:526–531, 1993.
23. Wang A, Ackland T, Hall S, Gilbey H, Parsons R: Functional recovery and timing of hospital discharge after primary total hip arthroplasty. Aust NZ J Surg 68:580–584, 1998.
24. Wheatley W, Krome J, Martin D: Rehabilitation programmes following arthroscopic menisectomy in athletes. Sports Med 21:447–456, 1996.
25. Wijgman A, Dekkers G, Waltje E, Krekels T, Arens H: No positive effect of preoperative exercise therapy and teaching inpatients to be subjected to hip arthroplasty (Dutch). Nederlands Tijdschrift Voor Geneeskunde 138:942–952, 1994.
© 2003 Lippincott Williams & Wilkins, Inc.