Article Text

Extended report
Postoperative effects of neuromuscular exercise prior to hip or knee arthroplasty: a randomised controlled trial
  1. Allan Villadsen1,
  2. Søren Overgaard2,
  3. Anders Holsgaard-Larsen2,
  4. Robin Christensen1,3,
  5. Ewa M Roos1
  1. 1Research Unit for Musculoskeletal Function and Physiotherapy, Institute of Sports Science and Clinical Biomechanics, University of Southern Denmark, Odense, Denmark
  2. 2Orthopaedic Research Unit, Department of Orthopaedics and Traumatology, Odense University Hospital, Institute of Clinical Research, University of Southern Denmark, Odense, Denmark
  3. 3Musculoskeletal Statistics Unit, The Parker Institute, Copenhagen University Hospital at Frederiksberg, Copenhagen, Denmark
  1. Correspondence to Allan Villadsen, Research Unit for Musculoskeletal Function and Physiotherapy, Institute of Sports Science and Clinical Biomechanics, University of Southern Denmark, Campusvej 52, Odense DK-5230, Denmark; avilladsen{at}health.sdu.dk

Abstract

Objective To investigate the postoperative efficacy of a supervised programme of neuromuscular exercise prior to hip or knee arthroplasty.

Methods In this assessor-blinded randomised controlled trial, we included 165 patients scheduled for hip or knee arthroplasty due to severe osteoarthritis (OA). An 8-week preoperative neuromuscular supervised exercise programme was delivered twice a week for 1 h as adjunct treatment to the standard arthroplasty procedure and compared with the standard arthroplasty procedure alone. The primary outcome was self-reported physical function measured on the activities of daily living (ADL) subscale in the Hip disability and Osteoarthritis Outcome Score (HOOS) and the Knee injury and Osteoarthritis Outcome Score (KOOS) questionnaires for patients with hip and knee OA, respectively. Primary endpoint was 3 months after surgery.

Results 165 patients randomised to the two groups were on average 67±8 years, 84 (51%) had hip OA and 92 (56%) were women. 153 patients (93%) underwent planned surgery and were evaluated postoperatively. There was no statistically significant difference in effects between hip or knee patients (p=0.7370). Three months postoperatively, no difference was found between groups for ADL (4.4, 95% CI −0.8 to 9.5) or pain (4.5, 95% CI −0.8 to 9.9). However, there was a statistically significant difference indicating an effect of exercise over the entire period (baseline to 3-months postoperatively) (p=0.0029).

Conclusions Eight weeks of supervised neuromuscular exercise prior to total joint arthroplasty (TJA) of the hip or knee did not confer additional benefits 3 months postoperatively compared with TJA alone. However, the intervention group experienced a statistically significant short-term benefit in ADL and pain, suggesting an earlier onset of postoperative recovery.

Trial registration ClinicalTrials NCT01003756.

  • Osteoarthritis
  • Rehabilitation
  • Treatment
  • Orthopedic Surgery

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Total joint arthroplasty (TJA) of the hip or knee is considered to be a successful orthopaedic procedure for pain relief and improvement in physical function in patients with severe osteoarthritis (OA).1 Treatment efficacy is important as worldwide more than 1.5 million procedures are currently performed annually; these numbers are increasing and these procedures are costly.2 Taking into account that patients’ expectations are not always met3—9% and 20% report no improvement or worse pain after hip and knee arthroplasty, respectively,4 and full physical function is not regained5—additional investigations into ways to improve treatment outcomes are warranted. Traditionally, exercise therapy has been delivered postoperatively for the purpose of rehabilitation. However, exercise could be delivered prior to surgery and serve as adjunct treatment in perioperative care. A recent systematic review and meta-analysis found moderate quality evidence that preoperative exercise and education programmes improve self-reported physical function 3 weeks after total hip arthroplasty (THA).6 For total knee arthroplasty (TKA), however, there is no evidence to support postoperative effects of preoperative exercise.6 Sufficiently powered high-quality trials are warranted.7

The exercise interventions applied in OA research are heterogeneous, making comparison across studies challenging.8 Furthermore, the content of exercise programmes is often insufficiently described, suggesting poor therapeutic validity, which may result in negative study findings.7 Studies applying previously tested protocols are warranted.

The aim of this paper was to evaluate the 3-month postoperative effects of TJA surgery in combination with a preoperative 8-week individualised, feasible neuromuscular exercise programme-total joint replacement (NEMEX-TJR)9 compared with TJA surgery alone, evaluated on the basis of self-reported physical function and joint-related pain.

Methods

Study design

A two parallel group assessor-blinded randomised controlled trial (RCT) (stratified according to the affected joint) was conducted. Conditions and methods were approved by The Committee for Biomedical Research Ethics for the Region of Southern Denmark, identifier: S-20090099. The trial is registered with ClinicalTrials.gov, identifier: NCT01003756 with the primary endpoint being 3 months after surgery.

Patients

Eligible patients were those who understood Danish, were at least 18 years of age and scheduled for primary unilateral THA or TKA due to severe symptomatic OA at Odense University Hospital, Denmark. Exclusion criteria were current or previous fractures in or adjacent to the joint, inflammatory arthritis and comorbidity, for example, severe heart disease and neurological deficits, contraindicating exercise and testing. Patients were ineligible if scheduled for bilateral TJA in the same procedure or for geographical reasons, for example, where logistics made attendance at exercise sessions unrealistic. Patient interest and eligibility were screened by the principal investigator via telephone and patient records. Written information was given to the participant in the clinic by the scheduling surgeon who was not otherwise involved in the study. Informed written consent was obtained on the day of baseline testing.

Randomisation

Allocation was concealed and conducted by the principal investigator after baseline assessment using sequentially numbered, opaque and sealed envelopes. The allocation sequence was stratified by gender, municipality (n=6), hip or knee and blocked in groups of four to allow for similar recruitment rates into both groups. The sequence and envelopes were produced manually (no number generator was used) by a person not otherwise affiliated with the trial. The allocation was performed either with the patient present or over the telephone. There is a treatment guarantee in the Danish Health Care System that one will be operated on within 1 month of being scheduled for TJR. Entering this study meant that all patients accepted an additional waiting time of up to 5 weeks in comparison with this guarantee. After randomisation, the additional waiting time applied only to patients randomised to the exercise intervention.

Surgical procedures

Patients with hip OA received uncemented implant (Corail/Pinnacle, DePuy International Ltd, Leeds, UK) using a posterolateral approach. For patients with knee OA, various implants were used: (PFC-Sigma Fixed Bearing, DePuy International Ltd, Leeds, UK), (Triathlon System, Stryker Orthopaedics, Mahwah, New Jersey, USA), (Signature, Biomet Orthopedics, Warsaw, Indiana, USA) (Oxford, Biomet Orthopedics, Warsaw, Indiana, USA), all cemented and inserted via an incision medial to the patella.

Intervention

Patients in the intervention group received the standard preoperative educational package in addition to attending a NEMEX programme for 8 weeks prior to surgery (EX+TJA). The control group received only the standard preoperative educational package (TJA).

The NEMEX-TJR exercise programme was delivered in a group setting for 1 h twice a week. It consisted of a 10-minute aerobic warm-up on a stationary exercise bike followed by a circuit programme with four main focus areas: core stability/postural control, postural orientation, lower extremity muscle strength and functional exercises. During all exercises, the supervising physiotherapist focused on the quality of the movement (sensorimotor control), including knee over foot alignment, which was evaluated as good quality movement. Progression in training level was made when the patient exhibited good sensorimotor control and good quality movement (as visually inspected by the physiotherapist) and with minimal exertion and control of the movement (as determined by the patient). A pain rating after exercise indicating unacceptable pain (more than 5 on a 0–10 visual analogue scale (VAS)) indicated the need for a reduction in training level. The exercise programme was identical to the previously published programme by Ageberg et al.9 The supervising physiotherapist in our trial attended a training session of the Ageberg cohort. A record was kept on the number of sessions attended. We stated a priori that attendance at 12 exercise sessions was considered good compliance. Patients randomised to exercise started immediately after randomisation and joined an existing group. Group size varied between 6 and 12 patients depending on the inclusion rate and patients leaving for surgery.

A standard preoperative educational package with written information on the operating procedure, expected postoperative progression and a leaflet on various exercises was handed out at the clinic when the patient was scheduled for THA or TKA. Furthermore, 1 week prior to surgery, patients participated in a 3-hour in-clinic information session led by health professionals. No limitations were given with regard to change of exercise habits or seeking other treatment.

Postoperative rehabilitation is offered to all patients within their municipality. The duration and intensity vary with municipality, therefore, we stratified for municipality.10

Assessment

Data collection points for this report were baseline, 6-weeks and 3-months postoperatively (primary endpoint). Four assessors conducted the physical testing and all underwent the same laboratory training. The assessors were blinded to group allocation, and the patients were instructed not to mention the allocation. The assessors had no access to previously obtained data.

Outcomes

The intervention was evaluated by patient self-assessment, a range of functional performance measures and muscle function in terms of maximal power of hip and knee muscles. The primary outcome was the activities of daily living (ADL) subscale of the Hip disability and Osteoarthritis Outcome Score (HOOS) and the Knee injury and Osteoarthritis Outcome Score (KOOS) for patients with hip and knee OA, respectively.11–13 Secondary outcomes were the HOOS/KOOS Pain, Symptoms, Sport and Recreation and joint-related Quality of Life subscales and the EuroQol 5 Dimension Health Questionnaire (EQ5D).14 ,15 The HOOS and KOOS are scored on a 0–100 worst to best scale. EQ5D consists of two parts: EQ5D index and VAS, which are scored 0–1 and 0–100, respectively. Questionnaires were sent out via mail, and patients were instructed to fill them out at home and return them on the day of clinical assessment or via mail.

Three exploratory performance measures (20 m walk, five timed repeated chair stands and maximal knee bends per 30 s) and four unilateral single-joint muscle power variables (single-joint knee extension, hip extension and hip abduction tests with a linear encoder (MuscleLab Power, Ergotest Technology, Langesund, Norway)) plus multi-joint leg extension press (Nottingham Power Rig, Nottingham University, Nottingham, UK) were applied.16

Statistical analysis

To allow for separate analysis of patients with knee and hip OA, 74 knee patients and 74 hip patients were required to detect a change of 10 points on the HOOS/KOOS ADL subscale (SD 15, power=0.80 and α=0.05). We anticipated around 10% loss to follow-up. A statistical analysis plan was drafted prior to breaking the code of the allocation sequence and data analysis. The project statistician (RC) who performed all the analyses was masked to group allocation and affected joint. The data for primary and secondary outcomes were analysed according to the intention-to-treat principle with the baseline observation carried forward in cases where data were missing. Data were checked for completeness and normality, and descriptive statistics were calculated for patient characteristics. To evaluate the longitudinal effects of the intervention, all outcome change values were analysed using a multilevel repeated measures random effects model with subject as the random effect factor and based on a restricted maximum likelihood estimate. The change value was the response variable, and the baseline values of group, joint and time were covariates. Assessment of these baseline values was of interest; hence, all interaction terms of group, joint and time were included. Data from all assessed time points were included. This one statistical model holds all between-group comparisons at all assessment points and allows for evaluation of the average effect over the time period from baseline to 3months postoperatively. Data for functional performance and maximal muscle power (exploratory outcomes) were analysed as observed (baseline observation not carried forward) by analysis of covariance. The SAS statistical package (V.9.2; SAS institute Inc, Cary, North Carolina, USA) was used for all statistical models. A two-sided p value of 0.05 was used to provide evidence against the null hypothesis.

Results

In total, 628 patients were operated on from 4 January 2010 to 21 March 2011. Of these, 499 patients were eligible, with 334 declining participation due to unwillingness to wait longer for surgery or logistical constraints such as travelling distance or lack of transportation. Of the 499 patients who were eligible, 165 patients were randomised to the two groups. The inclusion of knee patients was somewhat slower and took 14 months compared with 12 months for the hip patients. These patients were on average 67±8 years, 84 (51%) had hip OA and 92 (56%) were women (table 1). The 334 patients unwilling to participate were on average 4 years older (95% CI 2.3 to 5.6), 58% had hip OA and 60% were women.

Table 1

Baseline characteristics of study participants*

All patients, in both groups, received the educational material. Of the 84 patients in the intervention group, 62 (74%) attended the pre-specified goal of 12 or more exercise sessions. One hundred and fifty-three patients attended the preoperative 3-hour information session and underwent surgery.

The median time from baseline to surgery was 9.6 weeks (IQR from 8.9 to 10.7 weeks) in the intervention group and 7.1 weeks (IQR from 6.1 to 8.1 weeks) in the control group (data not shown). Figure 1 depicts the flow of patients through this trial.

Figure 1

Flow diagram of patients participating in this study. *Comorbidities (n=97); previous fracture in or adjacent to the joint (n=13) (1 knee); inflammatory arthritis (n=11) (5 knee); revision arthroplasty (n=7) (4 knee); previously enrolled with another joint (n=9) (7 knee); unicompartmental replacement (knee) (n=27); bilateral procedure in same session or within 3 months (n=16); necrosis of the femoral head (hip) (n=6); neurological disorders (n=6), hemiparesis (n=2), Parkinson's disorder (n=2), dementia (n=2); dysplasia of the femoral head (n=1); possible cancer metastasis in proximal femur (n=1). **The Danish Healthcare System has a 1-month treatment guarantee. Entering this study meant all patients accepting an additional wait of up to 5 weeks in comparison to the treatment guarantee. After randomisation, this additional wait applied only for patients randomised to the 8-week exercise intervention. The control group was operated on when originally scheduled.

Hip versus knee OA

We found no effect over time depending on the joint affected (hip or knee), as demonstrated by a non-significant value of the interaction term time×joint×group (p value=0.4639). That is, the additional effect of exercise seen over the whole time period, baseline to 3months postoperatively, did not differ between patients with hip or knee OA (figure 2A). However, the main effect of the explanatory variable ‘group’ was highly significant (p value=0.0029), indicating an effect of preoperative exercise over the time period. With no interaction of joint×group (p=0.7370) seen, the approach of ignoring the affected joint is acceptable (figure 2B). The results are thus given for all patients together, regardless of their having hip or knee OA.

Figure 2

Mean change (SE) in activities of daily living for the intervention and control group stratified for the affected joint (A) and for the two groups when ignoring the affected joint (B) at baseline, prior to surgery, 6 weeks and 3 months after surgery.

Efficacy analysis

Six weeks after surgery

Both groups improved immediately after surgery (6-weeks postoperatively) (figure 2). The intervention group demonstrated statistically significant greater improvement in ADL (p=0.0488) and Pain (p=0.0472) than did the control group (5.6, 95% CI 0.03 to 10.3 and 5.4, 95% CI 0.1 to 10.8, respectively) (table 2 and figure 2B). Patients in the intervention group reported significantly greater improvement in self-reported general health measured with the EQ5D VAS (7.6, 95% CI 2.1 to 13.0).

Table 2

Mean difference within groups and difference between groups (95% CI)

Three months after surgery—primary endpoint

For ADL and pain, we found no statistically significant difference (4.4, 95% CI −0.8 to 9.5 and 4.5, 95% CI −0.8 to 9.9, respectively) at 3 months after surgery (table 2 and figure 2B). Additionally, there was no difference in self-reported general health measured with the EQ5D (table 2).

Of the exploratory outcomes, single-joint hip extension and single-joint hip abduction displayed statistically significant differences in favour of the intervention group (∼15% and ∼35% improvement, respectively; see online supplementary table S1).

Adverse events

One patient with hip OA discontinued the exercise intervention after experiencing an increase in pain (figure 1). Two patients from the control group, one with hip and one with knee OA, developed a deep periprosthetic infection and had the prosthetic components removed prior to the 6-week postoperative assessment (figure 1).

Discussion

This randomised, assessor-blinded, controlled trial evaluated the efficacy of a supervised 8-week preoperative NEMEX programme as adjunct treatment for TJA compared with TJA alone in patients with severe hip or knee OA. At 3 months postoperatively (primary endpoint), no additional benefits were seen from the preoperative exercise. Seen over the entire time period from baseline to 3  months after surgery, this previously validated and feasible exercise programme resulted in an earlier onset of postoperative recovery in self-reported ADL and pain compared with the standard TJA procedure. Although no statistically significant differences were observed at 3 months postoperatively, the statistically significant differences in ADL, pain and QoL at 6 weeks postoperatively support the greater overall improvement seen from baseline to 3 months after surgery for the intervention group. This high quality sufficiently powered RCT will add to the existing few, small trials studying the effects of preoperative exercise on the postoperative outcomes following TJR.6

Explanation of the results

We aimed to improve ADL and reduce pain postoperatively by adding a clinically feasible NEMEX programme as adjunct treatment prior to THA and TKA surgery. A statistical analysis plan including all assessment points was outlined, allowing evaluation at 3months postoperatively and improvement from baseline to 3 months after surgery in the same analysis. Our results demonstrated no difference in the effect of preoperative exercise at 3 months after surgery. However, when considering improvement from baseline to 3 months postoperatively, neuromuscular exercise therapy was associated with greater overall improvement and earlier onset of postoperative recovery. This earlier improvement may be valued by some patients, their surgeons and other decision makers.

One-third of the participants in this study were aged less than 65 years, and hence, potentially still working. It is unknown if earlier onset of recovery was beneficial in terms of faster return to work. Other likely benefits include the reduced need for inpatient and outpatient postoperative rehabilitation services.

It has been suggested that the minimally important clinical change may depend on the disease, patient group and intervention in question.17–19 Differences of a minimum of 10 points out of 100 are generally accepted as clinically important in self-reported questionnaires including the KOOS.11 ,12 ,20 We observed differences between groups of 5–8 points for ADL, pain and QoL at 6 weeks postoperatively. These differences may prove clinically important when exercise is seen in the light of the combination with TJA, a much more sizeable and costly intervention than exercise therapy. Unfortunately, no measures of patient's global perceived effect were included that may have helped interpretation of these results.21

Although recommended as primary outcome measures in clinical trials, patient-reported outcomes do not provide the full picture of the patient's physical limitations.22 ,23 We included a test battery of functional performance measures and specific muscle tests to complement patient reports,16 and found statistically significant differences between groups in hip abduction and hip extension power. These joint movements are of functional importance for normal walking,24–26 but no concurrent improvement was found in the functional performance measures. It is of particular note, however, that hip patients from the intervention group performed markedly better on the operated side. For quadriceps muscle function, known to be reduced in THA and TKA patients,27 ,28 no effects of neuromuscular exercise were found. This is not surprising since this intervention is specifically targeting dynamic alignment and functional stability and not muscle power per se. This may partly explain the results for the tests performed in the standing position (hip abduction and extension), since these require more core and lower limb stability.

Strengths and limitations

We conducted an RCT and reported it according to the CONSORT statement29 ,30 with a rigorous study design, adequate sample size to allow for separate analysis for hip and knee OA patients and a clinically feasible and therapeutically valid intervention. Inclusion criteria were kept wide to reflect daily clinical practice. However, our study has limitations. No measures were taken to comply with the possible risk of attention bias. However, we believe the potential risk of attention bias introduced during the preoperative intervention to be minimal following surgery. Despite succeeding in including a higher proportion of eligible patients than in comparable studies, we still included only 30% of those eligible, which diminishes the external validity of the study.

The blinding of assessors was attempted through patient discretion and restriction of access to previously obtained data. We did not measure the success of assessor blinding; however, the primary and secondary outcomes were self-reported and thus not subjected to possible assessor bias. Delivering a placebo treatment would have been optimal but was not deemed realistic due to the nature of the exercise and the difficulty in designing a credible placebo intervention.

Conclusion and clinical implications

At the primary endpoint, 3 months after surgery, there was no additional effect of an 8-week preoperative NEMEX programme in combination with TJA compared with TJA alone. However, from baseline to 3 months after surgery, the overall longitudinal improvements seen in ADL and pain were statistically significantly greater and occurred earlier after surgery in the intervention group receiving preoperative exercise. Preoperative neuromuscular exercise constitutes a viable adjunct therapy to hip or knee arthroplasty of interest to individual patients willing to engage in preoperative exercise to achieve earlier onset of postoperative recovery.

Acknowledgments

We would like to thank the patients for their willingness to participate. We are grateful to the Department of Rehabilitation at Svendborg Hospital for providing us with research venues and equipment. Thanks to Professor Stefan Lohmander for his intellectual input.

References

View Abstract

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:

Footnotes

  • Handling editor Tore K Kvien

  • Contributors AV, SO, AHL and ER were responsible for concept and design; AV was responsible in collection and assembly of data and drafting of the article; AV and RC were responsible for the analysis of data; AV, SO, AHL, RC and ER were responsible for interpretation of data; SO, AHL, RC and ER were responsible for critical revision of the article for important intellectual content; AV, SO, AHL, RC and ER were responsible for final approval of the article; RC was responsible for statistical expertise; EMR takes responsibility for the integrity of the work as a whole from inception to finished article. EMW, Research unit for Musculoskeletal Function and Physiotherapy, Institute of Sports Science and Clinical Biomechanics, University of Southern Denmark, Odense, Denmark. eroos@health.sdu.dk

  • Funding This study was financed by the Region of Southern Denmark (095351 and 0912609), The Danish Rheumatism Association (R71-A1039-B382) and TrygFonden (7-10-0094). Protesekompagniet, a private Danish corporation supplying orthopaedic utilities, funded the exercise machines used for evaluation of muscle power. RC is an employee of The Parker Institute that is supported by grants from the Oak Foundation. The funders had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript.

  • Competing interests None.

  • Ethics approval The Regional Scientific Ethical Committee for Southern Denmark.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement The author group is willing to share data on request to the corresponding author. Each request will be treated individually.