Abstract:
Background: Knee osteoarthritis (OA) is one of the common causes of disability in elderly population. Currently, there is no agreement as to which exercise is the most effective for knee osteoarthritis. There are various forms of exercises and modalities recommended for treatment of knee osteoarthritis.
Objective: This study was conducted to systematically review the available literature regarding the effectiveness of exercise interventions for knee OA.
Data Sources: An extensive literature search using Academic OneFile (Cengage), Biomed central Journals, Directory of Open Access Journals, MEDLINE (EBSCO), ScienceDirect (Elsevier), AMED, CINAHL and Cochrane Review literature search was conducted.
Study Eligibility Criteria: Nine randomised controlled trials (RCTs) were identified which met the inclusion criteria for this review.
Participant and Interventions: A total of 917 patients with knee OA were subjected to a variety of exercises like strengthening exercises, stretching techniques, balance and joint position sense and Tai Chi across the studies.
Study Appraisal and Synthesis methods: Articles were assessed for quality by one reviewer using the Physiotherapy Evidence Database scores (PEDro). The quality assessments of the articles ranged from 4 to 8 with a mean score of 6.4/10.
Results: No studies included blinding of the patient or treating therapist and less than 60 % of the studies used a blinded assessor. A variety of outcome measures were included across the 9 studies. However, the Western Ontario McMaster University Index (WOMAC) and Lequesne Index (LI) were the physical function scales which were used in most of the studies. Four out of nine studies support the use of isotonic or dynamic exercises, in providing a significant pain relief and improving the physical function of the patient in short term as well as long term. The positive results were also demonstrated in isokinetic strengthening, stretching, Tai Chi, balance and kinesthetic training; however not many studies have been conducted on each of the exercise interventions.
Limitation: The eligibility criteria of the studies and the appraisal of individual studies were assessed by one author, which could have led to bias in the studies.
Conclusion: Overall due to the variability in the treatment interventions and lack of evidence, this review identified the need for further studies using larger sample sizes, adequate blinding and using valid functional outcome measures.
Introduction:
Osteoarthritis (OA) is one of the common causes of disability in the world (Jinks et al., 2002). More than 6 million people in the UK have painful osteoarthritis in one or both knees (Arthritis research, 2013). Knee Osteoarthritis is one of the causative factors in society leading to social, psychological and economical burdens in patients with substantial financial consequences (Egloff et al., 2012). The disease is characterised with focal and progressive damage to the articular cartilage, including development of the marginal outgrowth and osteophytes (Wang et al., 2005). The main symptoms associated with knee OA is pain, discomfort and limitation in the activities of daily living (Jamvedt et al., 2009). The risk factors include age, genetic predisposition, obesity, joint incongruence leading to increased mechanical pressure on the knee joint (Egloff et al., 2012).
There is no cure for OA, however regular exercise plays a key role in decreasing symptoms and increasing physical performance (Bennell and Hinmann, 2011). The exercises have a beneficial effect in improving physiological impairments such as muscle strength, joint range of motion (ROM), proprioception, balance and cardio vascular fitness (Weng et al., 2005). Currently, knee OA is treated with a number of treatment techniques like stretching, strengthening exercise, manual therapy and electrical modalities like transcutaneous electrical nerve stimulation (TENS), electro-acupuncture and ultrasound (Jamtvedt, 2008). In a previous overview of 23 systematic reviews, Jamvedt et al. (2008), claimed that there is high quality evidence supporting the use of exercises and weight reduction, however as each trial was not studied individually the results of the study was too broad to be able to use in clinical settings. Another study by Roddy et al. (2004), reviewed aerobic walking and strengthening exercise for osteoarthritis of the knee. The results showed that both aerobic walking and strengthening exercises help in improving the symptoms of knee OA. According to Aoki et al. (2009), stretching exercise has a significant effect on the knee ROM and gait disturbances which eventually reduces symptoms. Due to the mixture of evidence about which exercises most benefit knee OA patients, it is not surprising that there is no agreement as to which method is most effective in reducing knee OA symptoms. The main objective of this systematic review is to review the randomised controlled trials (RCT) on exercise interventions with knee OA and state the exercises which are safe and most effective in improving the symptoms in patients and if necessary to provide recommendations for future research.
This systematic review has taken account of high quality RCTs that have studied the various exercise interventions. This review will help to evaluate the best suitable exercise intervention for OA patients. According to the physiotherapy research priority project, (2010), by the Chartered Society of Physiotherapy, UK, there is a need to study the optimum exercise intervention used in treating knee OA patients.
Materials and Methods:
Search Protocol:
An extensive electronic search was conducted in August 2013 using relevant databases Academic OneFile (Cengage), Biomed central Journals, Directory of Open Access Journals, MEDLINE (EBSCO), ScienceDirect (Elsevier), AMED, CINAHL and Cochrane Review databases. These were individually searched and the most relevant studies were considered for further review. Cochrane review was studied so as to check whether no such study has taken place before. All the databases were searched for a fixed time period, January 2000 to August 2013. The search terms included in category 1 were exercises, intervention, physical therapy and physiotherapy and in section 2 were knee osteoarthritis and osteoarthrosis combined using ‘OR’ in between. Categories 1 and 2 were combined using ‘AND’ and duplicates were removed. All the articles resulting from these combined searches were reviewed for inclusion and retrieved studies were read in full. Additionally all the reference lists of the included studies were manually reviewed in order to extract further studies.
Eligibility Criteria:
Studies were included if they were published, full length RCTs, written in English, investigating adults > 18 years with primary knee OA. Only randomised controlled trials were included in the study as they are considered as high quality evidence (Sackett et al., 2002). Articles which were available as full text were only taken for this review to get a complete idea about the quality of research. As only the full text English articles were considered, this study has a risk of publication and language bias which could influence the results of the study (Dickersin, 1990). The studies which included exercises in the form of isometric, isokinetic, stretching, joint stabilisation, balance and proprioception either in isolation or as an adjunct to an exercise treatment were included. As the author intends to elicit the most suitable exercises to reduce the symptoms in knee OA, the studies which included outcome measures such as pain or physical function or both were included. The studies considered only included those, where participants had knee OA for at least 3 months as the symptoms would be elicitable (Felson et al., 2007). Articles were also included even if they lacked a methodological quality score in the PEDro, to include those articles that may have been excluded in the past reviews.
The studies were excluded if the participants were diagnosed with secondary OA, rheumatoid arthritis, gouty arthritis, septic arthritis, Pagets disease and studies on animals as they would not be able to fulfil the objective of the study. Any study which included an electrical modality like TENS, ultrasound, infrared alone or with any other exercise intervention were excluded. The studies which involved the application of manual therapy, mobilization and manipulation alone or in addition to the exercise group were excluded as the objective of this review is to find the optimum exercise intervention for OA. Studies were included if the subjects had medications for their pain, however if medications formed a part of their treatment regimen then they were excluded.
Data extraction and analysis:
The search was carried out by one author and the data was extracted from the studies. Data collected included demographics of the patient sample, treatment intervention, frequency, outcome measures and results. Quality of the studies was assessed and based on the Total Physiotherapy Evidence Database score (PEDro) as well as scores on individual items 2-11 on the scale.
Quality assessment:
Methodological quality was assessed using the PEDro scale. The PEDro scale comprises of an 11 point checklist to score the quality of the randomised controlled/clinical trials. In a study by Maher et al. (2003), it was proven that PEDro has good reliability and hence can be used to judge the RCTs in the systematic reviews. Each item on the checklist is scored ‘yes’ or ‘no’, with those items marked ‘yes’ receiving a score of 1 and those marked ‘no’ receiving a score of 0.All the articles were reviewed and scored by one author. Based on previous systematic reviews by Smidt et al. (2003) and Herd and Meserve, (2008), a quality score of 6 or more was considered as high quality, those with 4-5 were of fair quality and less than 3 were poor quality, hence this review also rated the studies as high, fair or poor quality.
Results:
Study selection:
From the extensive search, a total of 3100 articles were obtained. After excluding all the duplicate articles and including all the relevant articles, a total of 1159 articles were short listed. Further, the articles were again reviewed for relevant titles, which yielded a total of 129 articles. After reviewing the abstracts of 129 articles, a total of 9 articles were found to have met the inclusion criteria. The excluded ones mainly comprised of single case studies, retrospective analysis, combining exercise therapy with manual therapy or electrical modality as a treatment option, reviews and usage of medications. There were studies which met the inclusion criteria in terms of only knee OA exercise; however they failed to mention the effects of their study using a pain or physical function outcome measure. The 9 studies included various forms of exercises for the treatment of knee OA. One author studied the effectiveness of Tai Chi which is a traditional Chinese exercise which mainly aims for flexibility, muscle conditioning and cardiovascular endurance and since it is a form of exercise, it was included in the review. A detailed flow chart of the retrieved studies is illustrated in Figure 1.
All 9 studies included both men and women with a diagnosis of knee OA. The sample demographics are presented in table 1. The physical function outcome measures used were the Western Ontario and McMaster University Osteoarthritis Index (WOMAC) and Lequesne Index (LI) measurement. The WOMAC scale for osteoarthritis consists of three subgroups namely pain, stiffness and physical function (WOMAC-PF). The most frequently used outcome measure in the studies was the WOMAC or its subscale WOMAC-PF which is a 17 selected item questionnaire designed to measure physical function that has been shown to possess reliability, construct validity and responsiveness for OA conditions (Wright et al., 2010). LI is also a physical function measure which consists of 11 questions which evaluate knee discomfort, endurance, ambulation and difficulties in daily life (Faucher et al., 2004). The pain outcome measures were calculated using the visual analogue scale (VAS), numeric rating scale (NRS) (non-specific NRS 11 or 101) and WOMAC-Pain across the studies. A complete list of outcome measures, type of intervention and frequency of the treatment are presented in table 3.
| Individual Databases: Total 3100
Academic OneFile (Cengage), Biomed central Journals, Directory of Open Access Journals, MEDLINE (EBSCO), ScienceDirect (Elsevier), AMED, CINAHL and Cochrane Review databases
|
Figure 1: Study selection
treatment are presented in table 3.
| Individual Databases: Total 3100
Academic OneFile (Cengage), Biomed central Journals, Directory of Open Access Journals, MEDLINE (EBSCO), ScienceDirect (Elsevier), AMED, CINAHL and Cochrane Review databases
|
Figure 1: Study selection
| Only RCTs were considered, duplicates were excluded |
| Number of articles retrieved for review: N= 1159 |
| Exclusion Criteria: Unrelated articles were excluded,
Studies without knee OA, Studies with interventions other than physical therapy.
|
| Full text articles with manual therapy, strengthening exercises, stretching exercises, walking, range of motion exercises, aerobic walking, balance exercises N= 129 |
| Articles with electrical modalities, manual therapy, mobilisation and manipulation were excluded. |
| Articles included for the review. N= 9 |
Participants: From the nine studies, a total of 917 participants were subjected to exercise therapy for knee OA. The subjects were diagnosed with knee OA using the American College of Rheumatology criteria (ACR) and Kellegren and Lawrence radiographic classification for screening OA patients in 88% of the studies. Although, there is no gold standard classification identified for diagnosis of knee OA, these two are the most frequently used in epidemiological studies and clinical trials (Schipof et al., 2008). There is increased prevalence of knee OA in females than males, hence the number of females were more than male participants across the studies (French et al., 2011). As the participants with primary OA were considered in the studies, the studies included subjects with an age group of 35 to 75 years (Faucher et al., 2004). The demographic characteristics of each study are illustrated in table 1.
Table 1: Demographic Characteristics of the studies
| Study | Participants (N) | Age
Mean (SD) |
Gender
(F:M) |
Duration of symptoms/ Criteria for diagnosis of knee OA. |
| Jan et al., 2009 | 106 | 62(6.7), 63.2(6.8), 62.2(6.7) | 24:12 | >6 months,
Mild to moderate ACR criteria with radiographic scale ≤ 3 |
| Knoop et al., 2013 | 159 | 62.1(7.6)
61.8(6.8) |
53:27,
44:36 |
>3 months instability
ACR criteria with K/L radiographic classification |
| Diracoglu et al., 2005 | 66 | 35-65 years | 66 females only | ACR criteria with K/L scale grade 1 and 2,
LI ≥ 7 |
| Fitzgerald et al., 2011 | 183 | 64.6(8.4)
63.3(8.9) |
122:61 | ACR criteria with K/L radiographic scale ≥ 2 |
| Wang et al., 2009 | 40 | 63 (8.1)
68 (7.0) |
16:4
14:6 |
ACR with K/L Radiographic scale ≥2
WOMAC >40 (0-500) |
| Huang et al., 2003 | 132 | 62(4.5) | 93:39 | 4 months to 9 yrs
Altman Grade 2 |
| Topp et al., 2002 | 102 | 60.94(1.82)
65.57(1.82) |
74:28 | WOMAC pain subscale score >5 |
| Jan et al., 2008 | 102 | 63.3(6.6)
61.8(7.1) 62.8(6.3) |
27:7, 27:7, 25:5 | Symptoms > 6 months
ACR criteria with K/L radiographic scale ≤ grade 3. |
| Weng et al., 2009 | 132 | 64(7.5) | 16:26 | 4 months to 9.5 yrs
Altman Grade 2 |
Abbreviations: SD- Standard deviation; F:M- Female:Male; ACR- American College of Rheumatology (ACR); K/L- Kellegren and Lawrence radiographic classification for diagnosis of knee OA; WOMAC- Western Ontario McMaster Osteoarthritis Index; WOMAC-PF-WOMAC- Physical Function.
Adverse events: Adverse events are any unfortunate events that might occur in the study. 66% of the studies did not report any adverse event. However, a study by Wang et al., (2009) reported increased knee pain in the initial 2 weeks which was corrected by modifying the Tai Chi exercise. Two studies by Knoop et al. (2013) and Fitzgerald et al. (2011) claimed to have no adverse events in their studies.
Progression: 88 % of the studies included progression in their exercises by increasing resistance. There was no progression in the study by Wang et al. (2009), who introduced the exercise treatment of Tai Chi.
Treatment session: Most of the studies failed to mention the duration of each treatment session in their studies. However, the studies which mentioned the time period, the intervention duration lasted from 30 to 60 minutes.
Attendance: There was a good compliance rate reported for the participants in the studies. The overall attendance of the participants in the studies ranged from 85 % to 100 %. However, 44 % of the studies failed to report any percentage with regards to compliance.
Methodological quality:
Quality scores for the included articles ranged from 4 to 8 with an average score of 6.4/10. Eight of nine articles were found in the PEDro database, however one article by Knoop et al. (2013), was not found in PEDro, however the articles was thoroughly read and was analysed according to the criteria in PEDro. There was a baseline comparison and between group comparisons in all of nine studies. Adequate follow up, point estimates and variability and random allocation criteria from the PEDro was fulfilled by 88 % of the studies. From the PEDro scale, the included studies lacked blinding of the therapist and blinding of the subject and less than 60 % of the studies included an assessor which was blinded. A full breakdown of PEDro scores by criteria for each article is demonstrated in table 2.
Interventions:
Isotonic and Dynamic exercises:
Four studies examined the effects of isotonic or dynamic exercises for the treatment of knee OA. All the four studies clearly described the technique and application of the exercises. The PEDro scores ranged from 5 to 7 with an average score of 6.2. The studies compared different types of dynamic exercises like non weight bearing (NWB) and weight bearing exercises (WB), high resistance and low resistance exercises. In a study by Jan et al., (2009) NWB exercises were compared with WB exercises for function, walking speed and position sense in participants with knee OA. This study was considered to be a high quality study since it scored 7/10 on the PEDro scale. The exercise intervention progressed during the intervention period of 8 weeks. The exercises were performed thrice weekly on an EN Tree and EN Dynamic for NWB and WB respectively. The reliability and validity of the machines were not mentioned in the study. The results of the study showed that both NWB and WB exercises showed significant improvement in WOMAC-PF, walking on 4 different terrains and muscle torque test. However, only WB exercises showed a significant improvement in reposition error test. The author hypothesised that this difference which was found only in weight bearing group was because there is less activation of sensory nerves with NWB exercises than WB exercises.
Another study by Topp et al., (2002) compared the effects of dynamic versus isometric resistance training on pain and function among adults with OA of the knee. This study scored 5/10 on the PEDro scale, as the study lacked the blinding and intention to treat analysis. The dynamic exercise in this study was provided using the elastic resistance bands. There is evidence of using elastic resistive devices to gain functional limitations and improve gait characteristics (Krebs et al., 1998). The results of the studies suggest that both exercises help in improving physical performance and decreasing pain. However, isometric exercises were considered as having a higher level of fatigue than the dynamic group which used elastic resistive devices to perform the dynamic set of exercises.
A high quality study by Jan et al., (2008) investigated the clinical effects of high and low resistance training on patients with knee OA. This study has a PEDro score of 7/10. The exercise intervention for the high resistance (HR) group was 60 % of 1 repetition maximum and the low resistance (LR) group was 10 % of 1 RM. The exercise procedure was carried out using an EN-Dynamic leg press device. Each week the 1 RM was calculated on the device and a new resistance was set. The treatment time for the HR group was 50 minutes while the LR group was 30 minutes. The outcome measures used in this study were WOMAC pain, WOMAC-PF, walking on 4 different terrains and knee extensor torque at 60, 120 and 180 degree/s of knee flexion. The results of the study showed that both forms of resistance exercise led to meaningful reductions in pain and improved physical performance in patients with knee OA; however a greater difference in the outcome measures were evident in the HR group. Further, the author recommended a comprehensive programme by using agility, perturbation, training and aerobic exercise.
A study by Huang et al. (2003) compared the difference between isotonic, isometric and isokinetic exercises on moderately affected knee OA patients. This is considered as a high quality study which scored 6/10 on the PEDro scale. The Kin Com, Chatanooga Corporation, is a reliable and valid isokinetic dynamometer which was used for the isokinetic and the isotonic exercises (Mayhew et al., 1994). All the exercises were given three times a week for 8 weeks at an angle of 30 degrees/s of knee flexion for the flexor muscle group and 120 degrees of knee flexion for the extensor muscle groups. The outcome measures used in this study were the VAS, LI and ambulation speed. The greatest reduction in pain was demonstrated by the isotonic exercise intervention in comparison to the isokinetic and isometric exercise interventions. Isokinetic exercises showed the greatest improvement in ambulation speed and the LI physical function scales in comparison to the isometric and isotonic exercises. Similarly, isotonic exercises showed a significant improvement in the muscle torque test at 60 and 120 degrees of knee flexion while the isometric exercise group showed the least improvement. There was a significant improvement in the groups performing isokinetic and isometric exercises, the follow up results proved that there was a significant improvement in the patient symptoms of the treated groups in comparison to the control group, which did not carry out any form of exercise.
Tai Chi:
One study by Wang et al. (2009), studied the effectiveness of Tai Chi in severe knee OA patients. This was considered as a high quality study with a PEDro score of 7/10. The outcome measures in the study were WOMAC pain and WOMAC-PF, patient and physician global assessment, physical performance, body mass index (BMI), quality of life and psychological variables. The intervention was given for a period 60 minutes, twice a week for 12 weeks. The results showed that there was a significant difference in the primary outcomes in WOMAC pain and WOMAC-PF scores. Despite the results demonstrating statistically significant difference in the primary outcomes in WOMAC pain and secondary outcomes of physical functions, the study had a relatively small sample size of just 20 in each group. The baseline characteristics in terms of severity of knee OA in experimental and control group were different, which could lead to bias in the study.
Functional exercises:
Three studies claimed that the standard exercises for knee OA which are strengthening and stretching exercises are neither adequate nor sufficient for the knee OA patients. A randomised controlled trial by Knoop et al., (2013) studied knee joint stabilisation therapy in 159 patients with knee OA. This study was not found in the PEDro database for its quality score; hence the article was thoroughly read and scored according to the criteria in PEDro. In this study, both the experimental and the control groups were given muscle strengthening exercises and functional patient tailored exercise which targeted daily activities, while only the experimental group received the joint stabilisation exercises. The outcome measures used in this study were WOMAC-PF, numerical rating scale 11 questionnaire (NRS 11), global perceived effect, self-reported knee joint instability, the get up and go test, a climbing stairs questionnaire, a questionnaire regarding rising and sitting down, a walking questionnaire, upper leg muscle strength and proprioceptive accuracy. The results showed that there was no significant difference in the outcome measures between the groups and concluded that the addition of proprioceptive exercises did not cause any additional benefit to the patients.
A study was conducted by Diracoglu et al. (2005), on kinaesthesia and balance exercises on 66 mild to moderate knee OA patients. This was considered as a fair quality study with a PEDro score of 4/10. Similar to the previous studies, the experimental and the control groups in this study received a traditional exercise programme that included ROM exercise, quadriceps and hamstrings strengthening exercises and stretching exercises. The outcome measures used in this study were WOMAC-PF, short form-36 (SF-36) physical function, SF-36 role function, SF-36 vitality and energy and 10 meter(m) walking period. The results showed that there was a significant difference in the WOMAC-PF, SF-36, 10 m walking time and 10 m climbing stairs, isokinetic strength measurement and the proprioceptive strength measurement. The author concluded that kinesthesia and balance exercises help in restoring the neuromuscular function with significant recoveries in the functional status of the patients.
In another study by Fitzgerald et al. (2011), effects of agility and perturbation in addition to exercises were evaluated to reduce pain and increase function in 183 knee OA populations. This was considered as a good quality study with a PEDro score of 8/10. The subjects in the experimental group were targeted with agility and perturbation in addition to standardised exercise therapy of strengthening and stretching. The outcome measures, WOMAC, global rating of change, get up and go test and patient knee instability test were assessed at 2, 6 and 12 months respectively. The results of the study showed that by adding the agility and perturbation training to the standard exercise regimen of stretching and strengthening did show some improvement, however it was not statistically significant. Further, the author compared the results of these studies with similar studies and suggested the need for further research in this particular area.
Stretching exercises:
According to Weng et al. (2009), arthritic joint pain results in increased joint capsule contractures and limited range of motion, hence it is important to introduce stretching exercises in the treatment regimen of knee OA. In a study by Weng et al. (2009) effects of different stretching techniques were used to evaluate the outcomes of isokinetic strength in knee OA patients. This was considered as a good quality study with a PEDro score of 6/10. The sample size of 132 patients was divided into isokinetic exercises, bilateral knee static stretching followed by isokinetic exercise, proprioceptive neuromuscular facilitation (PNF) stretching followed by isokinetic exercise and a no exercise (control) group. The outcome measures used in this study were the VAS, knee ROM, LI, and peak muscle torque. The results of the study showed statistically significant improvements in all the outcome measures of the stretching exercises group. The patient group receiving PNF stretching benefitted the most in terms of increased muscle torque and functional outcome scores.
Table 3: Individual studies examining exercise interventions for Knee OA.
| Study | Treatment | Frequency | Outcomes | Results |
| Jan et al., 2009 | WB exercise,
NWB exercise,
Control group (No exercise) |
3 sessions/ week for 8 weeks | WOMAC-PF,
walking speed over 4 different terrains, muscle torque test, Reposition error test. |
Both the groups demonstrated significant improvement in the WOMAC functional scale and knee strength in NWB and WB group except for reposition error test which improved in WB group only. |
| Knoop et al., 2013 | Exercise program and joint stabilisation,
Exercise program (Control group) |
2 sessions/ week for 12 weeks and home exercise program for 5 days/week | WOMAC-PF,
NRS 0-10, global perceived effect, self reported kne joint instability, get up and go test, patient specific functioning list, climbing stairs questionnaire, questionnaire rising and sitting down, walking questionnaire, upper leg muscle strength, proprioceptive accuracy. |
At the end of 12 weeks, there was a clinically significant reduction (~20-40%) in pain and activity limitation in both the groups.
Global perceived effect was found significantly higher (P<0.04) in the experimental group. |
| Diracoglu et al., 2005 | Kinesthesia and balance exercise and strengthening exercises;
Strengthening exercise (Control group) |
3 sessions/ week for 8 weeks | WOMAC-PF,
SF-36 physical function, SF-36 role and vitality energy , 10 m walking, 10 stairs climbing time. |
The results showed statistically significant improvements (p<0.05) in all the outcome measures in the experimental group. |
| Fitzgerald et al., 2011 | Standard exercise program with agility and perturbation techniques,
Standard exercise program (Control group) |
12 sessions in 6 to 8 weeks plus home exercise of 2 times/week | WOMAC,
global rating of change, get up and go test, patient knee instability. |
At the end of 2, 6, 12 months, both groups reported improvement in the self reported function and global rating of change. There was no significant reduction in knee pain and performance based function in either groups. |
| Wang et al., 2009 | Tai Chi Intervention,
Attention control intervention (Control group) |
2 sessions/ week for 12 weeks | WOMAC, Timed chair time,
6 minute walk, Standing balance, Depression scale, Outcome expectation for exs, Self efficacy, SF-36.
|
The results showed that the experimental group demonstrated statistically significant improvement in all the outcome measures with p<0.005 at the end of 12 weeks. However, at the end of 24 and 48 weeks, scores were not statistically significant. |
| Huang et al., 2003 | Isokinetic exercise group,
Isotonic exercise group,
Isometric exercise group,
No exercise (Control group). |
3 sessions/ week for 8 weeks | VAS,
LI, Ambulation speed |
There was a statistical significant improvement in the treatment groups in all the outcome measures at the end of 8 weeks and 1 year follow up. However, among the treated groups, greatest reductions were found in the isotonic and isokinetic exercise group. |
| Topp et al., 2002 | Dynamic resistance training;
Isometric resistance training group, No exercise (Control group) |
3 sessions/ week for 16 weeks | WOMAC,
getting up and Descending and ascending stairs. |
There was a statistical significant improvement in the outcome measures of both the treatment groups; however, between groups analysis did not show any significant difference of one intervention over other. |
| Jan et al., 2008 | High resistance;
Low resistance; No exercise (Control group) |
3 sessions/ week for 8 weeks | WOMAC Pain,
WOMAC PF, walking on 4 different terrains, knee Flexor and extensor torque |
There was a statistical improvement in both the treatment groups in all the outcome measures; however, no significant improvement was seen between the groups. |
| Weng et al., 2009 | Isokinetic exs;
B/L static stretching and isokinetic exs; PNF stretching and isokinetic exs, No exercise (Control group). |
3 sessions/ week for 8 weeks. | VAS, ROM,
LI, peak muscle torque during knee flexion and extension. |
There was a significant improvement in all the treated groups with respect to pain and disability however only static and PNF stretching caused the greatest improvement in the muscle strength. |
Abbreviations: WB-weight bearing, NWB- non weight bearing, WOMAC-PF- WOMAC-Physical Function, NRS- numerical rating scale, SF-36- Short form-36 questionnaire, Pt- patient, VAS-visual analogue scale, LI-Lequesne Index, ROM- range of motion, B/L- bilateral, exs-exercise.
Discussion:
There were no systematic reviews identified during the vast literature search focussing only on exercises for knee OA. The previous systematic reviews on knee OA have included various electrical modalities and manual therapy, which would not lead to a conclusion, regarding which exercise is most effective for the treatment of knee joint OA. Assessment of the quality gives the reader a clear picture about the validity and generalisability of the findings of this review to clinical practice (Maher et al., 2003). The average overall quality score on the PEDro was 6.2/10 representing good quality evidence. The major methodological flaws in all the studies included a lack of blinding of patients and treating therapists which could lead to a risk of bias across the studies (Elridge et al., 2008). However, only 5 out of the 9 studies used a blinded assessor to capture outcomes, this has been found to lead to greater reports of treatment effect (Higgins and Altman, 2008), and hence results should be interpreted cautiously. A total of 24 different outcome measures were used in the nine studies. One of the inclusion criteria for this review was the use of pain or physical function outcome measures. All the studies used the functional outcome measures of either WOMAC or LI. The inclusion criteria for functional outcome measures provide important information regarding disability and overall quality of life (Herd and Meserve, 2005).
In a study by Iwamoto et al., (2007) the effect of strengthening exercises was studied on knee musculature. The author reported that strengthening exercises for the knee extensor muscles play an important role in the treatment of knee OA. In this systematic review, strengthening exercises were recommended by Jan et al., (2009), Jan et al., (2008), Huang et al., (2002) and Topp et al. (2002). The study by Weng et al. (2009), also introduced different stretching techniques for the purpose of increasing strength in the muscles acting on the knee. Hence, strengthening exercises play an important part in the treatment of knee osteoarthritis. In terms of specific strengthening exercises, isotonic, isometric and isokinetic exercises were used to reduce symptoms and increase physical function of the patients studied by Huang et al. (2003) and Topp et al., (2002). The average PEDro score of the studies which used strengthening exercises was 6.2/10 which indicates good quality of the studies and hence confirms the idea of using a strengthening programme in the treatment of patients with knee OA.
In the current review, Knoop et al. (2013), Jan et al., (2008) and Diracoglu et al. (2005) all studied joint proprioceptive training exercises in addition to the standard exercises. The three studies used agility perturbation, balance and joint stabilisation training exercises targeting mainly the needs of daily activities. The average score on the PEDro scale was 6.67. Two studies by Knoop et al. (2013) and Diracoglu et al., (2005) with a PEDro score of 8 each, reported that the addition of proprioceptive exercises to the standard exercise did not prove beneficial, while the study by Diracoglu et al. (2005), with a PEDro score of 4 suggested that balance exercises play an important role in increasing physical function in the patients. Fitzgerald et al., (2011) believed that WOMAC does not include items that directly incorporate higher level activities requiring balance and agility skills which could have affected the result of the studies. As there is a mixture of evidence with regards to the use of these exercises, high quality studies involving only balance exercises in the experimental group are required for further confirmation. However, the fact that, all the three studies used strengthening exercises in addition to the proprioception exercises suggests the importance of including strengthening exercises in treatment. The study by Jan et al., (2009) further suggested the use of weight bearing exercises which would involve some exercises targeting activities of daily living.
Isotonic or dynamic exercises were used in four studies by Jan et al., (2009), Huang et al., (2003), Jan et al., (2008) and Topp et al., (2002) and although, these used different modes of exercise training like high and low resistance exercises, weight bearing and non weight bearing exercises, the isotonic component was involved in all the four studies. Two studies by Topp et al., (2002) and Huang et al., (2003) compared the dynamic variant of strengthening exercise with isometric exercise, isokinetic exercises which further proved that isotonic exercise is an important part of a strengthening programme. In the study by Huang et al. (2003), there was a significant improvement in the outcome measures of the patients receiving isokinetic exercises. However due to the increase in the number of patients withdrawing from the study as a result of increased pain, the author recommended the use of isotonic exercises initially and then isokinetic exercises.
Limited ROM is an important factor leading to disability in patients with knee OA (Aoki et al., 2009). Therefore, stretching exercises play an important role in improving knee ROM potentially reduction in disability (Ghaffarinejad et al., 2007). In a study by Weng et al., (2009) static stretching and PNF stretching contributed to the increase in strength of the knee musculature. However, greatest improvement in terms of knee pain, LI and peak muscle torque was found in the group who received PNF stretching exercises alone. PNF stretching involves active contraction and relaxation of muscles which is more effective for improving flexibility compared to static stretching (Weng et al., 2009). Since muscle weakness and limited ROM were the two main problems identified from the above studies, the author recommends a comprehensive treatment regimen consisting of stretching and strengthening exercises. Further only one author studied the effects on the physical function post stretching exercises; hence, stretching intervention should be used cautiously depending on the clinical judgement of the therapist and goals of the patient.
A high quality study introduced Tai Chi as a mode of intervention for knee OA. According to Wang et al., (2009) Tai Chi is an ancient Chinese treatment which is useful for improving strength, balance and reducing pain, depression, and anxiety in patients with knee OA. Based on the evidence of just this one study which had a limited sample sizes, it is difficult to draw a conclusion to either support or refute the use of Tai Chi as a beneficial in the treatment for knee OA. In a systematic review by Lee et al., (2008) the effect of Tai Chi was not convincing for reducing pain and increasing physical function which further supports our argument.
Limitations:
The limitation of this study was the exclusion of articles using single group designs. In spite of being lower in the hierarchy of evidence (Hicks, 2005), they could make a worthy contribution to this systematic review as the number of RCTs purely on the exercise interventions were limited. Another potential limitation was the use of single assessor who determined the selection criteria for the review. The same person was responsible for reviewing the articles and synthesising the results which could lead to a selection bias in the review (CRD guidance, 2008). The qualities of all the RCTs were found in the PEDro database except the study by Knoop et al., (2013). The author read the article thoroughly and scored the study according to the PEDro criteria, however the score was not reviewed by another person that could again led to bias (CRD guidance, 2008). Due to the wide variation in the treatment interventions and the use of large number of outcome measures in the 9 studies reviewed, comparison of the results proved extremely complex and did not allow for clear conclusions to be drawn. Thus as a result of the identified potential limitations, further research is indicated using a larger sample size and adequate blinding to be able to use the findings to influence clinical practice.
Conclusion:
The current body of evidence regarding the use of exercises in knee OA varies significantly with regard to recommended exercise technique and treatment. There is much variability in the type of exercises prescribed to benefit patients with OA knee and therefore a lack of evidence to support this systematic review fully supporting the use of a particular exercise regimen. However, based on the assessment of the studies, there is some current evidence to support the use of a combined regimen of high intensity, isotonic training carried out in a weight bearing position to reduce the symptoms of knee OA. This review has also found some support for the use of isotonic exercises which may help in relieving pain and increasing functional capacity and can be used to gain immediate and long term effects. The weight bearing exercises will help augment proprioception while the isotonic exercises will help in reducing pain and disability. Due to the lack of evidence for exercise interventions on osteoarthritis, the findings of this study should be used cautiously used in clinical practice. Future studies are warranted that use larger sample sizes, valid functional outcome questionnaires, short- and long term patient follow-up, and inclusion of adjunct exercise treatments that are currently used in clinical practice.
Funding:
This systematic review was carried out as a part of Msc Physiotherapy programme and therefore there was no funding was involved.
References
Aoki, O., Tsumura, N., Kimura, A., Okuyama, S., Takikawa, S. and Hirata, S. (2009) ‘Home stretching exercise is effective for improving knee range of motion and gait in patients with knee osteoarthritis.’ Journal of Physical Therapy Science, Vol. 21, no. 2, pp. 113-119
Arthritis research, UK. (2013) Arthritis Research –Key facts. [Online][26th August 2013] https://www.google.co.uk/webhp?source=search_app#hl=en&q=arthritis+research+uk+pdf
Bennell, K. and Hinman, R. (2011)’ A review of the clinical evidence for exercise in osteoarthritis of the hip and knee. Journal of Science and Medicine in Sport, Vol. 14, no. 1, pp. 4-9
Centre for Reviews and Dissemination; University of York. (2008) Systematic reviews. [Online][26th August, 2013] http://www.york.ac.uk/inst/crd/pdf/Systematic_Reviews.pdf
Dickersin, K (1990) ‘The existence of Publication bias and Risk factors for its occurance’. Vol. 263, no.10, pp. 1385-1389
Diracoglu, D., Aydin, R., Baskent, A. and Celik, A. (2005) ‘Effects of kinesthesia and balance exercises in knee osteoarthritis.’ JCR: Journal of Clinical Rheumatology, Vol. 11, no. 6, pp. 303-310
Egloff, C., Hügle, T. and Valderrabano, V. (2012) ‘Biomechanics and pathomechanisms of osteoarthritis.’ Swiss Med Wkly, Vol. 142, w13583
Eldridge, S., Ashby, D., Bennett, C., Wakelin, M. and Feder, G. (2008) ‘Internal and external validity of cluster randomised trials: systematic review of recent trials.’ British Medical Journal, Vol. 336, no. 7649, pp. 876-880
Faucher, M., Poiraudeau, S., Lefevre-Colau, M., Rannou, F., Fermanian, J. and Revel, M. (2004) ‘Assessment of the test–retest reliability and construct validity of a modified WOMAC index in knee osteoarthritis.’ Joint Bone Spine, Vol. 71, no.2, pp. 121-127
Felson, D., Niu, J., McClennan, C., Sack, B., Aliabadi, P., Hunter, D., Guermazi, A. and Englund, M. (2007) ‘Knee buckling: prevalence, risk factors, and associated limitations in function.’ Annals of internal medicine, Vol. 147, no. 8, pp. 534-540
Fitzgerald, G., Childs, J., Ridge, T. and Irrgang, J. (2002) ‘Agility and perturbation training for a physically active individual with knee osteoarthritis.’ Physical Therapy, Vol. 82, no. 4, pp. 372-382
French, H., Brennan, A., White, B. and Cusack, T. (2011) ‘Manual therapy for osteoarthritis of the hip or knee–a systematic review.’ Manual Therapy, Vol. 16, no. 2, pp. 109-117
Ghaffarinejad, F., Taghizadeh, S. and Mohammadi, F. (2007) ‘Effect of static stretching of muscles surrounding the knee on knee joint position sense.’ British journal of sports medicine, Vol. 41, no. 10, pp. 684-687
Herd, C., and Meserve, B. (2008) ‘A systematic review of the effectiveness of manipulative therapy in treating lateral epicondylalgia.’ The Journal of manual & manipulative therapy, Vol. 16, no. 4, pp. 225-237
Hicks, C. (2009) Research methods for clinical therapists: Applied project design and analysis. 5th ed., London: Churchill Livingstone
Higgins, J and Altman, D. (2008) ‘Assessing the risk of bias in included studies.’ In Higgins, J and Green, S. (ed/s.) Cochrane Handbook of Systematic Reviews of Interventions., England: John Wiley & Sons ltd., 187-235
Huang, M., Lin, Y., Yang, R. and Lee, C. (2003) ‘A comparison of various therapeutic exercises on the functional status of patients with knee osteoarthritis.’ In Seminars in arthritis and rheumatism, Vol. 32, No. 6, pp. 398-406
Iwamoto, J., Takeda, T. and Sato, Y. (2007) ‘Effect of muscle strengthening exercises on the muscle strength in patients with osteoarthritis of the knee.’ The Knee, Vol. 14, no. 3, pp. 224-230
Jamtvedt, G., Dahm, K., Christie, A., Moe, R., Haavardsholm, E., Holm, I. and Hagen, K. (2008) ‘Physical therapy interventions for patients with osteoarthritis of the knee: an overview of systematic reviews.’ Physical Therapy, Vol. 88, no. 1, pp. 123-136
Jan, M., Lin, C., Lin, Y., Lin, J. and Lin, D. (2009) ’Effects of weight-bearing versus nonweight-bearing exercise on function, walking speed, and position sense in participants with knee osteoarthritis: a randomized controlled trial.’ Archives of physical medicine and rehabilitation, Vol. 90, no. 6, pp. 897-904
Jan, M., Lin, J., Liau, J., Lin, Y. and Lin, D. (2008) ‘Investigation of clinical effects of high-and low-resistance training for patients with knee osteoarthritis: a randomized controlled trial.’ Physical Therapy, Vol. 88, no. 4, pp. 427-436
Jinks, C., Jordan K. and Croft, P. (2002) ‘Measuring the population impact of knee pain and disability with the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC).’ Pain, Vol. 100, no. 1-2, pp. 55-64
Knoop, J., Dekker, J., van der Esch, M., Thorstensson, C. A., Gerritsen, M., Voorneman, R. E. and Steultjens, M. (2013) ‘Knee joint stabilization therapy in patients with osteoarthritis of the knee: a randomized, controlled trial.’ Osteoarthritis and Cartilage, Vol. 21, pp. 1025-1034
Krebs, D., Jette, A. and Assmann, S. (1998) ‘Moderate exercise improves gait stability in disabled elders.’ Archives of physical medicine and rehabilitation, Vol. 79, no. 12, pp. 1489-1495
Lee, M., Pittler, M. and Ernst, E. (2008) ‘Tai chi for osteoarthritis: a systematic review.’ Clinical rheumatology, Vol. 27, no.2, pp. 211-218
Maher, C., Sherrington, C., Herbert, R., Moseley, A. and Elkins, M. (2003) ‘Reliability of the PEDro scale for rating quality of randomized controlled trials.’ Physical therapy, Vol. 83, no. 8, pp. 713-721
Mayhew, T., Rothstein, J., Finucane, S. and Lamb, R. (1994) ‘Performance characteristics of the Kin-Com® dynamometer.’ Physical therapy, Vol. 74, no. 11, pp. 1047-1054
Roddy, E., Zhang, W. and Doherty, M. (2005) ‘Aerobic walking or strengthening exercise for osteoarthritis of the knee? A systematic review.’ Annals of the Rheumatic Diseases, Vol. 64, no. 4, pp. 544-548
Schiphof, D., de Klerk, B., Koes, B. and Bierma-Zeinstra, S. (2008) ‘Good reliability, questionable validity of 25 different classification criteria of knee osteoarthritis: a systematic appraisal.’ Journal of clinical epidemiology, Vol. 61, no. 12, pp. 1205-1215
Smidt N, Assendel W, Arola H, Malmivaara, A., Green S., Buchbinder, R., Windt, D. and Bouter, L. (2003) ‘Effectiveness of physiotherapy for lateral epicondylitis: A systematic review.’ Annals of Medicine, Vol. 35, pp. 51–62
Topp, R., Woolley, S., Hornyak III, J., Khuder, S. and Kahaleh, B. (2002) ‘The effect of dynamic versus isometric resistance training on pain and functioning among adults with osteoarthritis of the knee.’ Archives of physical medicine and rehabilitation, Vol. 83, no. 9, pp. 1187-1195
Wang, C., Schmid, C., Hibberd, P., Kalish, R., Roubenoff, R., Rones, R. and McAlindon, T. (2009) ‘Tai Chi is effective in treating knee osteoarthritis: a randomized controlled trial.’ Arthritis Care & Research, Vol. 61, no. 11, pp. 1545-1553
Weng, M., Lee, C., Chen, C., Hsu, J., Lee, W., Huang, M. and Chen, T. (2009) ‘Effects of different stretching techniques on the outcomes of isokinetic exercise in patients with knee osteoarthritis.’ The Kaohsiung Journal of Medical Sciences, Vol. 25, no. 6, pp. 306-315
Wright, A., Cook, C., Baxter, G., Garcia, J. and Abbott, J. (2010). ‘Relationship between the Western Ontario and McMaster Universities Osteoarthritis Index Physical Function Subscale and physical performance measures in patients with hip osteoarthritis.’ Archives of physical medicine and rehabilitation, Vol. 91, no. 10, pp. 1558-1564
