Author’s Accepted Manuscript

The role of muscle strengthening in exercise

therapy for knee osteoarthritis: A systematic review
and meta-regression analysis of randomized
trialsMuscle strengthening and knee OA

Cecilie Bartholdy, Carsten Juhl, Robin Christensen,

Hans Lund, Weiya Zhang, Marius Henriksen
www.elsevier.com/locate/semarthrit

PII: S0049-0172(16)30172-X
DOI: http://dx.doi.org/10.1016/j.semarthrit.2017.03.007
Reference: YSARH51163
To appear in: Seminars in Arthritis and Rheumatism
Cite this article as: Cecilie Bartholdy, Carsten Juhl, Robin Christensen, Hans
Lund, Weiya Zhang and Marius Henriksen, The role of muscle strengthening in
exercise therapy for knee osteoarthritis: A systematic review and meta-regression
analysis of randomized trialsMuscle strengthening and knee OA, Seminars in
Arthritis and Rheumatism, http://dx.doi.org/10.1016/j.semarthrit.2017.03.007
This is a PDF file of an unedited manuscript that has been accepted for
publication. As a service to our customers we are providing this early version of
the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting galley proof before it is published in its final citable form.
Please note that during the production process errors may be discovered which
could affect the content, and all legal disclaimers that apply to the journal pertain.
The Role of Muscle Strengthening in Exercise Therapy for Knee
Osteoarthritis: A systematic review and meta-regression
analysis of randomized trials

Cecilie Bartholdy1,6, Carsten Juhl2,3, Robin Christensen1, Hans Lund2,4, Weiya Zhang5, Marius
Henriksen1,6

Affiliations:

1: The Parker Institute, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Copenhagen, Denmark

2: SEARCH research group (Synthesis of Evidence and Research), Research Unit for Musculoskeletal Function and
Physiotherapy, Institute of Sports Science and Clinical Biomechanics, University of Southern Denmark, Odense M,
Denmark

3: Department of Rehabilitation, Copenhagen University Hospital, Herlev and Gentofte, Denmark

4: Center for Evidence-based Practice, Bergen University College, Bergen, Norway

5: Division of Rheumatology, Orthopedics and Dermatology, School of Medicine, University of Nottingham, United
Kingdom

6: Department of Physical and Occupational Therapy, Copenhagen University Hospital, Bispebjerg and Frederiksberg,
Copenhagen, Denmark

Emails:

Cecilie Bartholdy: [email protected]

Carsten Juhl: [email protected]

Robin Christensen: [email protected]

Hans Lund: [email protected]

Weiya Zhang: [email protected]

Correspondence:

Marius Henriksen, The Parker Institute, Copenhagen University Hospital, Bispebjerg and Frederiksberg, Nordre

Fasanvej 57, vej 8 indgang 19, Copenhagen, Denmark. Email: [email protected] Tel: +45 3816 4160

Running Title: Muscle strengthening and knee OA

1
ABSTRACT

Objectives: To analyze if exercise interventions for patients with knee osteoarthritis (OA) following the

American College of Sports Medicine (ACSM) definition of muscle strength training differs from other types

of exercise, and to analyze associations between changes in muscle strength, pain, and disability.

Methods: A systematic search in 5 electronic databases was performed to identify randomized controlled

trials comparing exercise interventions with no intervention in knee OA, and reporting changes in muscle

strength and in pain or disability assessed as standardized mean differences (SMD) with 95% confidence

intervals (95%CI). Interventions were categorized as ACSM interventions or not-ACSM interventions and

compared using stratified random effects meta-analysis models. Associations between knee extensor

strength gain and changes in pain/disability were assessed using meta-regression analyses.

Results: The 45 eligible trials with 4,699 participants and 56 comparisons (22 ACSM interventions) were

included in this analysis. A statistically significant difference favoring the ACSM interventions with respect

to knee extensor strength was found (SMD difference: 0.448 [95%CI 0.091 to 0.805]). No differences were

observed regarding effects on pain and disability. The meta-regressions indicated that increases in knee

extensor strength of 30 and 40% would be necessary for a likely concomitant beneficial effect on pain and

disability, respectively.

Conclusion: Exercise interventions following the ACSM criteria for strength training provide superior

outcomes in knee extensor strength but not in pain or disability. An increase of less than 30% in knee

extensor strength is not likely to be clinically beneficial in terms of changes in pain and disability.

PROSPERO: CRD42014015344

KEYWORDS: Osteoarthritis, knee, exercise, strength training, muscle strength

2
1. INTRODUCTION

Contemporary clinical guidelines for the management of knee OA emphasize non-pharmacological

approaches: weight reduction (in obese individuals), education, and exercise therapy – highlighting the

added value of exercise therapy 1. Based on several meta-analyses and systematic reviews 2-6 different

types of exercise interventions, including aerobic and strengthening exercise, are proven effective in

reducing pain and disability in patients with knee OA. Besides the beneficial effects on symptoms, strength

training is advocated as an important component of exercise interventions, as decreased muscle strength is

common in this patient group 7 8.

The underlying mechanism by which exercise is beneficial for knee OA symptoms remains unclear.

Theories have been put forward including local biomechanical and systemic effects (including anti-

inflammatory effects), yet with limited empirical evidence in support 9. In exercise programs that include or

focus on muscle, the increase in muscle strength must be considered a necessary effect, if a biological (and

plausible) link between muscle strength and clinical benefits exists. However, it is unknown if the positive

effects of exercise programs are caused by the increased muscle strength (the biomarker) or other aspects.

It therefore stands to reason to investigate if exercise interventions that aim directly at the biomarker

(muscle strength) yield superior clinical effects (i.e. decreased pain and disability) compared to

interventions with less emphasis on the biomarker (muscle strength). Earlier reviews have addressed the

role of strength training in relation to efficacy and symptoms change10 11, however it has not been explored

if strength outcome can be a possible biomarker of clinical effect, and how the association is between an

increase in muscle strength and clinical benefits (i.e. reduction of pain and disability). The American College

of Sports Medicine (ACSM) has a clear and clinically feasible definition of strength training, by which

interventions can be appraised to adhere to or not. As studies that include measures of muscle strength can

be assumed to expect muscle strength changes, a comparison between interventions designed according to

the ACSM criteria and those not seems relevant in the investigation of muscle strength as a possible

biomarker of the effect on pain and disability assessment of the possible biomarker of effect.

3
Consequently, The objectives of this study were i) to investigate whether strength training interventions

(defined as interventions designed according to the American College of Sports Medicine (ACSM) guidelines

for strength training) differs from exercise interventions not designed according to the ACSM criteria on

self-reported outcomes in the pain and disability domains as well as in muscle strength gain in patients with

knee OA, and ii) to investigate whether there is an association between muscle strength gain and changes

in the clinical outcomes pain and disability, and iii) to identify a potential critical threshold for the

‘minimum gain in muscle strength’ needed for a likely change in self-reported pain and disability to be

expected.

2. METHODS

2.1 Protocol and registration

Study selection, assessments of eligibility criteria, data extraction and statistical analyses were performed

based on a pre-specified protocol that was registered before commencing the review (PROSPERO:

CRD42014015344). The protocol was developed in accordance with the methodological guidelines from the

Cochrane Collaboration12 and the manuscript is written in accordance with the PRISMA (Preferred

Reporting Items for Systematic reviews and Meta-Analyses) statement13.

2.2. Eligibility criteria

Eligibility criteria were randomized or quasi-randomized controlled trials comparing at least one exercise

intervention with no intervention, waiting list, sham or placebo. The trial population should be diagnosed

with knee OA in one or both knees. All studies having performed an exercise intervention and reporting a

strength measurement of the lower limb, and included outcomes on self-reported pain or disability were

eligible. No restrictions regarding age, BMI, or gender were applied.

Exercise interventions were categorized as “ACSM-interventions” if they described the delivered

intervention according to the ACSM recommendation of strength training for this patient group 14: A

4
voluntary contraction against an external resistance typically performed in especially designed equipment

or with free weights. The external load should be above 40% of 1 repetition maximum (1RM) corresponding

to very light to light intensity, and the exercises performed in 2-4 sets of 8-12 repetitions; preferably to

contraction failure or muscular exhaustion. The exercise program should consist of at least 2-3 sessions per

week.

Exercise interventions that in their description were considered not to follow all of the above

definitions were categorized as “not ACSM interventions”, and include all other types of exercise

interventions. The exercise intervention categorization was done by one reviewer (CB), and the

categorization was spot checked by another reviewer (MH). Consensus on categorization was reached by

discussion.

Only published data was used; no attempts to retrieve further information from the authors were

made.

2.3. Information sources and search

Studies were identified by searching following databases: MEDLINE, EMBase, CINAHL, PEDro, Cochrane

Central Register of Controlled Trials (CENTRAL). A pre-specified search strategy was developed based on a

previous systematic search in the field4 and provided in Supplementary file A. The final literature search

was performed 23 February 2015. Reference list of the included studies were scanned as well.

2.4. Study selection

Eligibility assessment was performed independently by two reviewers (CB and CJ). Firstly, by screening of

titles and abstracts with the full-text article obtained if it was judged initially eligible by at least one

reviewer. Eligibility of the full-text articles was judged independently by the same reviewers. Consensus on

inclusion based on full-text articles was reached by discussion.

5
2.5. Data collection process and data items

A data extraction sheet was developed in Microsoft Excel and data extraction was performed by one

reviewer (CB) and checked by another (CJ). The following data was extracted; authors, year of publication,

numbers of participants allocated to intervention and control, age, BMI, gender, Kellgren/Lawrence (K/L)

score, baseline and follow-up outcome measures for knee pain, disability, and muscle strength (of the

lower limb). Details of the exercise interventions in terms of; type of exercise, numbers of session per

week, number of supervised sessions, duration, resistance, sets, and repetitions were also extracted. We

extracted mean baseline and follow-up values with corresponding standard deviations (SD), or mean

changes with the corresponding SDs. If data for strength, pain, and disability was provided on more than

one scale a predefined hierarchy (as described in the protocol; PROSPERO: CRD42014015344) was used to

decide which outcome score to include. Studies with more than one exercise intervention were handled as

two or more independent comparisons and presented and analyzed as such. The control group in these

studies was then divided into the appropriate number of groups.

2.6. Risk of bias in individual studies

The Cochrane Collaboration's tool for assessing risk of bias in randomized trials was used to assess potential

bias15. A data extraction sheet was developed in Microsoft Excel with the different domains from the tool.

This was filled out independently by two reviewers (CB and CJ). Assessment was done at study level and

disagreement was resolved by consensus discussion. The following risk of bias domains was assesses as

either adequate (low risk of bias), inadequate (high risk of bias), or unclear (insufficient information):

sequence generation, concealment of allocation, blinding, incomplete outcome data, selective outcome

reporting, and other risks of bias. The scoring of the different studies was performed based on the

Cochrane handbook criteria’s for judging the different domains12.

6
2.7. Synthesis of results

The effect sizes on muscle strength and self-reported pain and disability were expressed as standardized

mean difference (SMD), by dividing the mean group difference in the change of the outcome by the pooled

SD; we applied the Hedges’s small-sample adjustment per default. If the SD was not reported it was

estimated from the reported standard error (SE), the 95% confidence interval (95% CI), inter-quartile-range

(IQR), or P-value related to the pertinent number of participants. If necessary we approximated mean score

and SD from figures in the individual reports. We evaluated the extent of inconsistency among the studies’

results, using the I2 “inconsistency index”; the I2 is interpreted as the proportion of total variation in study

estimates that is due to heterogeneity rather than sampling error 16.

Random effects models fitted to a restricted maximum likelihood (REML) model were performed for

evaluating the difference between the studies applying with the ACSM guidelines and those who do not for

all three outcomes (i.e. muscle strength, self-reported pain and disability). To assess the association

between muscle strength gain and clinical benefits (pain and disability) a meta-regression analysis model

was used with percent muscle strength gain as independent variable and SMD for pain and disability as

dependent variables (i.e., outcomes). The analyses were performed using STATA (version 13.1).

3. RESULTS

3.1. Study selection

As illustrated in Figure 1, the literature search resulted in 4,408 articles (1,120 CENTRAL, 444 CINAHL, 1527

EMBase, 1,106 MEDLINE, 211 PEDro, 3 from reference lists), with 2,251 being screened on title and abstract

after removal of duplicates. After screening of 208 records in full text, 45 articles, including 56 comparisons,

were included in the final analyses (See Figure 1).

7
3.2. Study characteristics

The 56 comparisons involved 4,699 participants allocated to either exercise or control. The participants

were on average 64 years (range 55 to 71 years), with the proportion of women ranging from 36% to 100%

between the comparisons; two comparisons did not describe gender distribution17. The mean BMI was

available in 46 (out of 56) of the comparisons with a mean of 29.6 kg/m2 and a range from 22.4 to 33.9

kg/m2 and for those that included K/L score (29 comparisons) a median of 2 (range 1 to 4) was presented.

Twenty-two interventions 18-37 followed the ACSM guidelines for strength training while 34

interventions 25 28 29 33 38-61 did not. The 22 interventions that followed the ACSM guidelines for strength

training had a median exercise period of 8 weeks (range 4 to 120) with an average of 3 sessions per week

(range 2 to 7). The median number of exercise sets was 3 for each exercise with an average of 10

repetitions and a load of minimum 40 % of 1RM. The 34 interventions not following the ACSM guidelines

for strength training had an average exercise period of 8 weeks (range 4 to 104) with a median of 3 sessions

per week (range 2 to 7). Thirty-three comparisons had no intervention as their control group, 20 compared

with a sham interventions, placebo, or attention control group, and 3 compared with a waiting list control.

All of the different comparisons reported knee extensor strength. In the ACSM group all comparisons

reported pain. WOMAC pain was used in 15 comparisons, pain during walking rated on a numeric rating

scale (NRS) was used 4 times; as for pain (NRS), brief pain inventory (NRS), and pain during motion, they

were all used once. Function was reported in 19 comparisons and of those 17 used WOMAC function, SF-36

physical component score and Lequesne Index was used once each. For the not-ACSM group all

comparisons reported pain. WOMAC pain was used in 14 comparisons, VAS (visual analog scale) pain during

walking was used 5 times, both pain during standing (NRS) and KOOS pain was used 4 times. The following

rating scales for pain were used once each: pain score, brief pain inventory(NRS), AIMS2 arthritis pain, VAS

pain, VAS pain during motion, OASI pain, and VAS present pain. A total of 30 comparisons in the not-ACSM

group reported functional outcomes. WOMAC function was reported in 16 comparisons, Lequesne Index

Score was used 4 times, KOOS function was used 3 times. The following functional scores were reported

8
once: functional incapacity score, SF-36 physical component, AIMS2 mobility level, OASI mobility, subjective

rating of daily activity, and KOOS ADL. Table 1 gives an overview of the study characteristics.

3.3. Methodological characteristics

The methodological characteristics of the comparisons showed that 36 (64%) reported using an adequate

sequence generation and 31 (55%) comparisons reported adequate allocation concealment. Blinding was

graded as adequate in only 2 (4%) of the comparisons, and in 22 (39%) of the comparison analyses

(intention to treat) were regarded adequate. The two comparisons that had adequate blinding of the

participants/personnel did so by ensuring that the personnel that did the control and exercise interventions

was blinded to who was in the opposite group, that participants was not informed of the fact that there

was two different group, and by blinding outcome assessors. Reporting of selective outcome was

considered adequate in 8 (14%) studies and the likelihood of other sources of bias were considered

adequate in 1 (2%) of the studies. A summary of the risk of bias assessments is provided in Supplementary

file B. Most comparisons delivered insufficient information to assess whether selective outcomes reporting

were an issue or if other sources of bias were present. Therefore, only a small portion of the studies gave

adequate information on this matter. Between the two subgroups (ACSM/not-ACSM) similar patterns of

risk of bias level was observed.

Data for pain outcomes was available in all comparisons while disability was not reported sufficiently for

data extraction in 6 comparisons22 35 36 54 58 . Stratified analyses effect sizes, according to the different risk of

bias domains, are presented in Supplementary file C.

3.4. Muscle strengthening vs. other exercise types

Benefits of exercise were recognized with respect to knee extensor strength, pain, and disability (Table 2)

across all studies. A statistically significant difference in favor of the ACSM exercise interventions with

respect to knee extensor strength gain was found (difference in SMD: 0.448 [95% CI 0.091 to 0.805], P=

9
0.014). No statistically significant differences between ACSM and not-ACSM interventions were found

regarding effects on pain (difference in SMD: 0.106 [95% CI -0.239 to 0.451], P= 0.547) and disability

(difference in SMD: 0.153 [95% CI -0.243 to 0.549], P= 0.449). The results from the stratified analyses are

shown in table 2.

3.5. Association between knee extensor strength gain and pain

Of the 56 comparisons, it was not possible to calculate percent change in knee extensor strength in 3

comparisons40 48. From the meta-regression analyses of the remaining 53 comparisons a statistically

significant association between knee extensor strength change and pain was found (slope: 0.0197 SMD

[95% CI 0.006 to 0.033], P=0.005) (figure 2A), with considerable heterogeneity remaining I2= 67.4% across

studies, and an explained variance of 21.1% (Adjusted R-squared). From figure 2A it can be seen that an

increase in knee extensor strength of minimum 30% is necessary for a likely beneficial effect on pain is to

be expected (approximated intercept of the 95% prediction interval with the x-axis).

3.6. Association between knee extensor strength gain and disability

Of the 56 comparisons only extractable data on disability was available in only 47. An association between

knee extensor strength change and disability was found across the studies (slope: 0.0215 SMD [95% CI

0.007 to 0.036], P=0.006) as illustrated in Figure 2B. There were considerable heterogeneity remaining I2 =

75.2% across studies, and an explained variance of 10.5% (adjusted R-squared). From figure 2B it can be

seen that an increase in knee extensor strength of minimum 40% is necessary for a likely beneficial effect

on disability is to be expected (approximated intercept of the 95% prediction interval with the x-axis).

4. DISCUSSION

This systematic review and meta-analysis explored if changes in core clinical outcomes related to knee OA

(i.e. pain and disability) and muscle strength differed between exercise interventions specifically aiming at

10
increasing muscle strength (i.e. followed the ACSM guidelines for strengthening exercise) and other types

of exercise interventions. Furthermore it was investigated if muscle strength increase was associated with

changes in the core clinical outcomes of pain and disability.

We found that strength training as defined by ACSM criteria increased knee extensor strength significantly

more than the not-ACSM group in patients with knee OA, which corresponds well with a previous report 10

and common sense. While knee extensor strength increase was associated with change in the clinical

outcomes, no difference could be identified when comparing changes in pain and disability between the

specific strengthening exercises and other types of exercise. Thus, based on the current knowledge it is only

possible to deduce that exercise per se has beneficial clinical effects and no specific exercise protocol has

yet been proven superior to others. This falls well in line with the present recommendations on non-

pharmacological treatment of patients with knee OA, that exercise should be first treatment with no

specific type of exercise being recommended as superior 2 62-65.

The results of this meta-analysis are ambiguous; exercise interventions that follow the ACSM guidelines

for strength training results in significant increase in knee extensor strength compared to other

interventions. However, changes in clinical outcomes seem unrelated to the choice of strengthening

exercise or other types of exercise interventions. Yet, we found an association between changes in knee

extensor strength and changes in pain and disability, which suggest that strength gain is beneficial. Reduced

muscle strength especially quadriceps strength is common among patients with knee OA66 67 and therefore

strength training seems to be well indicated. Achieving an increase of 30-40% in knee extensor strength is

considerable but not impossible with this patient group; however factoring age and general condition of

this patient group (pain and disability) together with the beneficial clinical effects from other types of

exercise interventions, striving for an increase of 30-40% in knee extensor strength to add a small

additional effect in pain and disability can seem inappropriate in clinical practice.

An explanation for these ambiguous results could be that exercise in general affects the individual on

several other components than muscle strength such as; neuromuscular function, general fitness and

11
health, and psychosocial factors that are candidate mechanism to influence changes in clinical outcomes9.

This implies that at this point we are unable to clearly differentiate between the mechanisms that affect

the knee OA patient and play a key role in the changes of clinical outcomes.

As with all systematic reviews the limitations of the included studies are shared with the current

review. The overall likelihood of potential bias within each study was moderate to high and each risk of bias

element that was deemed unclear or high seemed to increase the estimated effect of the intervention. The

overall likelihood of bias across the studies was in most cases related to the difficulties of blinding

personnel and participants when performing exercise interventions, together with possible selective

outcome reporting. However, the risk of bias profiles was comparable between the ACSM studies and the

not-ACSM studies. This justifies the stratification according strength training (ACSM criteria) and it is

unlikely to be affected by potential biases because the risk of a particular bias is equally likely in each

groups. The results are therefore unlikely to change in terms of the relation between knee extensor

strength and clinical outcomes but the effect size may change based on level of bias.

It is well known that both exercise interventions and the patient group is known to be heterogeneous,

but we expected the heterogeneity to decrease when the studies were stratified according to strength

training as defined by the ACSM recommendations. Interestingly, the heterogeneity generally decreased

among the not-ACSM studies and increased among the ACSM studies (although the heterogeneity generally

was still high). This is striking as the not-ACSM studies cover a wide range of exercise types including

aquatic exercise, aerobic exercise, neuromuscular/functional exercise and individualized physiotherapy

guided exercises, which at first sight would indicate large heterogeneity. In contrast, the ACSM-studies

would be expected to be more homogenous; they were not. This observation remains unexplained.

The relatively high level of inconsistency among the exercise studies suggest that the treatment

provided varies in terms of dosage and duration to a point where it becomes difficult to compare exercise

studies and recommend general criteria for an optimal exercise intervention. Previous meta-analyses have

tried to subgroup and stratify for other factors to ascertain if a superior exercise intervention exists1 4 62 64,

12
but reach the same conclusion that the large variation in the construct of the applied exercise interventions

across studies clouds the identification of a possible association between specific treatments and clinically

relevant effects. We are far from understanding how the potential mediating factors influence pain and

disability, complicating the process of determining or designing the optimal exercise protocol with regards

to type, dosage and duration for patient with knee OA.

An important limitation in this study is the interpretation of the ACSM recommendation. We chose to

allow a 40 percent resistance for improvement of strength, because this is recommended to the elderly,

frail or beginners as starting resistance. Further, the not-ACSM interventions include a wide variety of

exercise types, which are not all obvious strengthening exercise programs. A further subgrouping of these

not-ACSM interventions could have been valuable to identify specific differences within strengthening

exercise programs. However, because of the generally poor intervention description such classification

would rely too much on subjective interpretation and thereby increase the likelihood of misclassification.

Furthermore, the not-ACSM interventions had a markedly lower heterogeneity (I2), suggesting that further

subgrouping should be of the ACSM interventions.

Another limitation is that we only included exercise studies that reported muscle strength data and it is

possible that the estimates would have been different if all exercise studies have measured and reported

muscle strength and that the different studies did not had different control groups (no intervention, waiting

list, sham or placebo). The lack of detailed description of the exercise interventions may have led to

misclassification of the studies thereby explaining some of the inconsistency found despite subgrouping the

interventions. It is reasonable to suggest that the studies that did sham or placebo are likely to have other

estimates due to less risk of performance bias. Further, an important limitation is the use of study level

variables for the association between changes in pain and disability outcomes and muscle strength; there

might be different associations within each individual study, which we cannot capture using a meta-

analytical approach. However, the observed associations in this study seem plausible.

13
 This study also has several strengths including a pre-specified protocol to dictate the process and

analyses of the review. The search and screening of studies was done systematically and according to

guidelines to minimize bias. The results from this systematic review is in line with previous meta-analysis

concluding that exercise is beneficial for the knee OA patient in terms of improved clinical outcomes and

increased strength.

5. CONCLUSION

This review supports the current recommendations on exercise as effective means for improving clinical

outcomes among patients with knee OA. We extend this knowledge by showing that the current research

evidence support that muscle strengthening interventions increases knee extensor strength significantly

more than other exercise interventions – as expected. However, exercise interventions with a focused aim

at muscle strength do seemingly not provide superior clinical outcomes when compared to other types of

exercise. Further studies could evaluate the relationship between patients with knee OA performing

exercise but not improving over 30-40% in knee extensor strength compared to those that did improve

over 30-40%.

ACKNOWLEDGEMENTS

This study was supported financially by The Danish Physiotherapists Association and The Oak Foundation.

The funders had no role in the design, analyses or decision to publish.

AUTHOR CONTRIBUTION

CB, CJ, RC and MH contributed to the conception and design of the study. Acquisition and analysis of data

was performed by CB and CJ. All authors contributed to the interpretation. CB drafted the article and all

14
authors contributed with a thorough and critically revision for important intellectual content. All authors

have approved the final version of this manuscript.

CONFLICTS OF INTEREST

None to declare.

Reference list

1. McAlindon TE, Bannuru RR, Sullivan MC, et al. OARSI guidelines for the non-surgical management of knee
osteoarthritis. Osteoarthritis and Cartilage 2014;22(3):363-88.

2. Fernandes L, Hagen KB, Bijlsma JW, et al. EULAR recommendations for the non-pharmacological core
management of hip and knee osteoarthritis. Annals of the rheumatic diseases 2013;72(7):1125-35.

3. Zhang W, Moskowitz RW, Nuki G, et al. OARSI recommendations for the management of hip and knee
osteoarthritis, Part II: OARSI evidence-based, expert consensus guidelines. Osteoarthritis and Cartilage
2008;16(2):137-62.

4. Juhl C, Christensen R, Roos EM, et al. Impact of exercise type and dose on pain and disability in knee
osteoarthritis: A systematic review and meta-regression analysis of randomized controlled trials. Arthritis
and Rheumatology 2014;66(3):622-36.

5. Fransen M, McConnell S. Exercise for osteoarthritis of the knee. The Cochrane database of systematic
reviews 2008(4):CD004376.

6. Roddy E, Zhang W, Doherty M, et al. Evidence-based recommendations for the role of exercise in the
management of osteoarthritis of the hip or knee--the MOVE consensus. Rheumatology (Oxford)
2005;44(1):67-73.

7. Slemenda C, Brandt KD, Heilman DK, et al. Quadriceps weakness and osteoarthritis of the knee. Annals of
internal medicine 1997;127(2):97-104.

8. van der Esch M, Holla JF, van der Leeden M, et al. Decrease of muscle strength is associated with
increase of activity limitations in early knee osteoarthritis: 3-year results from the cohort hip and cohort
knee study. Archives of physical medicine and rehabilitation 2014;95(10):1962-8.

9. Beckwee D, Vaes P, Cnudde M, et al. Osteoarthritis of the knee: why does exercise work? A qualitative
study of the literature. Ageing Res Rev 2013;12(1):226-36.

15
10. Zacharias A, Green RA, Semciw AI, et al. Efficacy of rehabilitation programs for improving muscle
strength in people with hip or knee osteoarthritis: A systematic review withmeta-analysis. Osteoarthritis
and Cartilage 2014;22(11):1752-73.

11. Lange AK, Vanwanseele B, Fiatarone Singh MA. Strength training for treatment of osteoarthritis of the
knee: A systematic review. Arthritis Care and Research 2008; 15(10).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/594/CN-00898594/frame.html.

12. Higgins JP, Green S. Cochrane handbook for systematic reviews of interventions In: Higgins JP, Green S,
eds.: The Cochrane Collaboration, 2011.

13. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA statement for reporting systematic reviews and
meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med
2009;6(7):e1000100.

14. Garber CE, Blissmer B, Deschenes MR, et al. American College of Sports Medicine position stand.
Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and
neuromotor fitness in apparently healthy adults: guidance for prescribing exercise. Medicine and science in
sports and exercise 2011;43(7):1334-59.

15. Higgins JP, Altman DG, Gotzsche PC, et al. The Cochrane Collaboration's tool for assessing risk of bias in
randomised trials. BMJ 2011;343:d5928.

16. Higgins JP, Thompson SG, Deeks JJ, et al. Measuring inconsistency in meta-analyses. BMJ
2003;327(7414):557-60.

17. Gur H, Cakin N, Akova B, et al. Concentric versus combined concentric-eccentric isokinetic training:
Effects on functional capacity and symptoms in patients with osteoarthrosis of the knee. Archives of
physical medicine and rehabilitation 2002;83(3):308-16.

18. Anwer S, Alghadir A. Effect of isometric quadriceps exercise on muscle strength, pain, and function in
patients with knee osteoarthritis: a randomized controlled study. Journal of Physical Therapy Science 2014
May;26(5):745-748 2014.

19. Baker KR, Nelson ME, Felson DT, et al. The efficacy of home based progressive strength training in older
adults with knee osteoarthritis: a randomized controlled trial. Journal of Rheumatology 2001;28(7):1655-
65.

20. Bennell KL, Hinman RS, Metcalf BR, et al. Efficacy of physiotherapy management of knee joint
osteoarthritis: a randomised, double blind, placebo controlled trial. Annals of the rheumatic diseases 2005;
(6). http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/846/CN-00511846/frame.html

http://ard.bmj.com/content/64/6/906.full.pdf.

21. Bennell KL, Hunt MA, Wrigley TV, et al. Hip strengthening reduces symptoms but not knee load in
people with medial knee osteoarthritis and varus malalignment: a randomised controlled trial.

16
Osteoarthritis and cartilage / OARS, Osteoarthritis Research Society 2010; (5).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/002/CN-00753002/frame.html.

22. Borjesson M, Robertson E, Weidenhielm L, et al. Physiotherapy in knee osteoarthrosis: effect on pain
and walking. Physiotherapy research international : the journal for researchers and clinicians in physical
therapy 1996;1(2):89-97.

23. Bruce-Brand RA, Walls RJ, Ong JC, et al. Effects of home-based resistance training and neuromuscular
electrical stimulation in knee osteoarthritis: A randomized controlled trial. BMC musculoskeletal disorders
2012;13(118).

24. Foroughi N, Smith RM, Lange AK, et al. Lower limb muscle strengthening does not change frontal plane
moments in women with knee osteoarthritis: A randomized controlled trial. Clinical biomechanics (Bristol,
Avon) 2011; (2). http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/096/CN-
00789096/frame.html.

25. Jan M, Lin J, Liau J, et al. Investigation of clinical effects of high- and low-resistance training for patients
with knee osteoarthritis: a randomized controlled trial. Physical therapy 2008;88(4):427-36.

26. Jorge RTB, de Souza MC, Chiari A, et al. Progressive resistance exercise in women with osteoarthritis of
the knee: a randomized controlled trial [with consumer summary]. Clinical Rehabilitation 2014 Jul 3:Epub
ahead of print 2014.

27. Lim B, Hinman RS, Wrigley TV, et al. Does knee malalignment mediate the effects of quadriceps
strengthening on knee adduction moment, pain, and function in medial knee osteoarthritis? A randomized
controlled trial. Arthritis & Rheumatism: Arthritis Care & Research 2008;59(7):943-51.

28. Lim JY, Tchai E, Jang SN. Effectiveness of Aquatic Exercise for Obese Patients with Knee Osteoarthritis: A
Randomized Controlled Trial. PM and R 2010;2(8):723-31.

29. Lin DH, Lin CHJ, Lin YF, et al. Efficacy of 2 non-weight-bearing interventions, proprioception training
versus strength training, for patients with knee osteoarthritis: A randomized clinical trial. Journal of
Orthopaedic and Sports Physical Therapy 2009;39(6):450-57.

30. McKay C, Prapavessis H, Doherty T. The Effect of a Prehabilitation Exercise Program on Quadriceps
Strength for Patients Undergoing Total Knee Arthroplasty: A Randomized Controlled Pilot Study. PM and R
2012; (9). http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/518/CN-00899518/frame.html.

31. Mikesky AE, Mazzuca SA, Brandt KD, et al. Effects of strength training on the incidence and progression
of knee osteoarthritis. Arthritis and Rheumatism 2006; (5).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/295/CN-00568295/frame.html.

32. Sayers SP, Gibson K, Cook CR. Effect of high-speed power training on muscle performance, function, and
pain in older adults with knee osteoarthritis: a pilot investigation. Arthritis care & research 2012; (1).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/153/CN-00805153/frame.html.

17
33. Salli A, Sahin N, Baskent A, et al. The effect of two exercise programs on various functional outcome
measures in patients with osteoarthritis of the knee: A randomized controlled clinical trial. Isokinetics and
Exercise Science 2010;18(4):201-09.

34. Rooks DS, Huang J, Bierbaum BE, et al. Effect of preoperative exercise on measures of functional status
in men and women undergoing total hip and knee arthroplasty. Arthritis and Rheumatism 2006; (5).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/296/CN-00568296/frame.html.

35. Swank AM, Kachelman JB, Bibeau W, et al. Prehabilitation before total knee arthroplasty increases
strength and function in older adults with severe osteoarthritis. Journal of Strength and Conditioning
Research 2011; (2). http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/215/CN-
00780215/frame.html.

36. Weidenhielm L, Mattsson E, Broström LA, et al. Effect of preoperative physiotherapy in

unicompartmental prosthetic knee replacement. Scandinavian journal of rehabilitation medicine 1993; (1).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/189/CN-00092189/frame.html.

37. Weng MC, Lee CL, Chen CH, et al. Effects of different stretching techniques on the outcomes of
isokinetic exercise in patients with knee osteoarthritis. The Kaohsiung journal of medical sciences 2009; (6).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/101/CN-00723101/frame.html.

38. An B, Dai K, Zhu Z, et al. Baduanjin alleviates the symptoms of knee osteoarthritis. Journal of alternative
and complementary medicine (New York, NY) 2008; (2).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/126/CN-00638126/frame.html

http://online.liebertpub.com/doi/pdfplus/10.1089/acm.2007.0600.

39. Beaupre LA, Lier D, Davies DM, et al. The effect of a preoperative exercise and education program on
functional recovery, health related quality of life, and health service utilization following primary total knee
arthroplasty. Journal of Rheumatology 2004;31(6):1166-73.

40. Fransen M, Crosbie J, Edmonds J. Physical therapy is effective for patients with osteoarthritis of the
knee: a randomized controlled clinical trial. Journal of Rheumatology 2001; (1).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/314/CN-00329314/frame.html.

41. Gür H, Cakin N, Akova B, et al. Concentric versus combined concentric-eccentric isokinetic training:
effects on functional capacity and symptoms in patients with osteoarthritis of the knee. Archives of Physical
Medicine & Rehabilitation 2002;83(3):308-16.

42. Huang M, Lin Y, Yang R, et al. A comparison of various therapeutic exercises on the functional status of
patients with knee osteoarthritis. Seminars in Arthritis & Rheumatism 2003;32(6):398-406.

43. Huang MH, Yang RC, Lee CL, et al. Preliminary results of integrated therapy for patients with knee
osteoarthritis. Arthritis and Rheumatism 2005; (6).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/412/CN-00532412/frame.html.

18
44. Huang MH, Lin YS, Lee CL, et al. Use of ultrasound to increase effectiveness of isokinetic exercise for
knee osteoarthritis. Archives of physical medicine and rehabilitation 2005; (8).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/703/CN-00523703/frame.html.

45. Hurley MV, Walsh NE, Mitchell HL, et al. Clinical effectiveness of a rehabilitation program integrating
exercise, self-management, and active coping strategies for chronic knee pain: A cluster randomized trial.
Arthritis Care and Research 2007;57(7):1211-19.

46. Koli J, Multanen J, Kujala UM, et al. Effect of Exercise on Patellar Cartilage in Women with Mild Knee
Osteoarthritis. Medicine and science in sports and exercise 2015.

47. Kuptniratsaikul V, Tosayanonda O, Nilganuwong S, et al. The Efficacy of a Muscle Exercise Program to
Improve Functional Performance of the Knee in Patients with Osteoarthritis. Chotmaihet thangphaet
[Journal of the Medical Association of Thailand] 2002; (1).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/799/CN-00506799/frame.html.

48. Lund H, Weile U, Christensen R, et al. A randomized controlled trial of aquatic and land-based exercise
in patients with knee osteoarthritis. Journal of rehabilitation medicine 2008; (2).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/957/CN-00639957/frame.html.

49. Maurer BT, Stern AG, Kinossian B, et al. Osteoarthritis of the knee: isokinetic quadriceps exercise versus
an educational intervention. Archives of Physical Medicine & Rehabilitation 1999;80(10):1293-99.

50. Multanen J, Nieminen MT, Hakkinen A, et al. Effects of high-impact training on bone and articular
cartilage: 12-month randomized controlled quantitative MRI study. Journal of Bone and Mineral Research
2014; (1). http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/009/CN-00980009/frame.html.

51. O'Reilly SC, Muir KR, Doherty M. Effectiveness of home exercise on pain and disability from
osteoarthritis of the knee: a randomised controlled trial. Annals of the rheumatic diseases 1999; (1).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/586/CN-00163586/frame.html.

52. Peloquin L, Bravo G, Gauthier P, et al. Effects of a cross-training exercise program in persons with
osteoarthritis of the knee. A randomized controlled trial. Journal of Clinical Rheumatology 1999; (3).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/957/CN-00354957/frame.html.

53. Quilty B, Tucker M, Campbell R, et al. Physiotherapy, including quadriceps exercises and patellar taping,
for knee osteoarthritis with predominant patello-femoral joint involvement: randomized controlled trial.
Journal of Rheumatology 2003; (6). http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/020/CN-
00438020/frame.html.

54. Røgind H, Bibow-Nielsen B, Jensen B, et al. The effects of a physical training program on patients with
osteoarthritis of the knees. Archives of physical medicine and rehabilitation 1998; (11).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/012/CN-00157012/frame.html.

55. Salacinski AJ, Krohn K, Lewis SF, et al. The effects of group cycling on gait and pain-related disability in
individuals with mild-to-moderate knee osteoarthritis: a randomized controlled trial. Journal of

19
Orthopaedic and Sports Physical Therapy 2012; (12).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/719/CN-00906719/frame.html.

56. Schilke JM, Johnson GO, Housh TJ, et al. Effects of muscle-strength training on the functional status of
patients with osteoarthritis of the knee joint. Nursing research 1996;45(2):68-72.

57. Sekir U, Gur H. A multi-station proprioceptive exercise program in patients with bilateral knee
osteoarthrosis: functional capacity, pain and sensoriomotor function. A randomized controlled trial. Journal
of sports science & medicine 2005;4(4):590-603.

58. Talbot LA, Gaines JM, Huynh TN, et al. A home-based pedometer-driven walking program to increase
physical activity in older adults with osteoarthritis of the knee: A preliminary study. Journal of the American
Geriatrics Society 2003;51(3):387-92.

59. Thomas KS, Muir KR, Doherty M, et al. Home based exercise programme for knee pain and knee
osteoarthritis: Randomised controlled trial. British Medical Journal 2002;325(7367):752-55.

60. Trans T, Aaboe J, Henriksen M, et al. Effect of whole body vibration exercise on muscle strength and
proprioception in females with knee osteoarthritis. The Knee 2009; (4).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/171/CN-00722171/frame.html.

61. Villadsen A, Overgaard S, Holsgaard-Larsen A, et al. Immediate efficacy of neuromuscular exercise in

patients with severe osteoarthritis of the hip or knee: A secondary analysis from a randomized controlled
trial. Journal of Rheumatology 2014; (7).
http://onlinelibrary.wiley.com/o/cochrane/clcentral/articles/124/CN-00995124/frame.html.

62. Jansen MJ, Viechtbauer W, Lenssen AF, et al. Strength training alone, exercise therapy alone, and
exercise therapy with passive manual mobilisation each reduce pain and disability in people with knee
osteoarthritis: a systematic review. Journal of Physiotherapy (Australian Physiotherapy Association)
2011;57(1):11-20.

63. McAlindon TE, Biggee BA. Nutritional factors and osteoarthritis: Recent developments. Current opinion
in rheumatology 2005;17(5):647-52.

64. Fransen M, McConnell S, Harmer AR, et al. Exercise for osteoarthritis of the knee. The Cochrane
database of systematic reviews 2015;1:CD004376.

65. McAlindon TE, Bannuru RR, Sullivan MC, et al. OARSI guidelines for the non-surgical management of
knee osteoarthritis. Osteoarthritis and cartilage / OARS, Osteoarthritis Research Society 2014;22(3):363-88.

66. Felson DT. An update on the pathogenesis and epidemiology of osteoarthritis. Radiol Clin North Am
2004;42(1):1-9, v.

67. van Dijk GM, Dekker J, Veenhof C, et al. Course of functional status and pain in osteoarthritis of the hip
or knee: a systematic review of the literature. Arthritis Rheum 2006;55(5):779-85.

20
FIGURE LEGENDS

Figure 1. Flow of randomized controlled trials included in the systematic review and meta-analysis.

Figure 2. Meta-regression analyses (REML): standardized mean differences (SMD) of the individual studies

according to pain (A) and disability (B) at different percentages of changes in knee extensor strength. From

approximation (visual) of the intercept of the 95% prediction interval with the x-axes, the lower limit of

percent change needed in knee extensor strength to achieve significant change in pain and disability is

~30% and 40%, respectively.

Table 1. Characteristics of randomized controlled trial included in the meta-analysis

ACSM Outcome
Participants Participan
Author, Intervention Control / Outcome s
(interventio ts
year (Exercise) Intervention not- Strength pain &
n) (control)
ACSM function
An, B. et n = 14 n = 14 Baduanjin no Not- Isokinetic,  WOMA
al, Age: 65.4 Age: 64.6 (traditional intervention ACSM knee C
2008 % female: % female: Chinese aerobic extension functio
100 100 exercise) n
BMI: 25.7 BMI: 25.4 8 weeks  WOMA
K/L grade: K/L grade: 5 times a week C pain
NA NA NA sets, NA
reps.
Load: NA
Anwer, S. N= 21 N= 21 Isometric Ultrasound ACSM Isometric,  WOMA
& Age: 54.9 Age: 56 exercise therapy as knee C
Alghadir, % female: % female: 5 weeks per the extension functio
A., 69 69 5 times a week patient’s n
2014 BMI: 26.5 BMI: 27.1 1-3 sets, 10 reps. request, 7  Pain on
K/L grade: K/L grade: Load: max min. NRS
NA NA isometric
contraction
Baker, K. N= 23 N= 23 Home based attention ACSM Isokinetic  WOMA
et al, Age: 69 Age: 68 progressive control, 7 knee C
2001 % female: % female: strength training home visits extension functio
73.9 82.6 program and a n
BMI: 31 BMI: 32 4 month nutritional  WOMA
K/L grade: 3 K/L grade: 3 times a week booklet C pain
3 2 sets, 12 reps.
Load: light to
hard intensity
(modified Borg
Scale, 10 point,
3-8)

21
Beaupre, N= 65 N= 66 Education, no Not- Isometric  WOMA
L. A. et al, Age: 67 Age: 67 mobility and intervention ACSM knee C
2004 % female: % female: progressive extension functio
60 50 resistance n
BMI: 32 BMI: 31 training program  WOMA
K/L grade: K/L grade: 4 weeks C pain
NA NA 3 times per
week
3 sets, 20 reps.
Load: NA
Bennell, N= 73 N= 67 multifaceted sham ACSM Isometric  WOMA
K. L. et al, Age: 67.4 Age: 69.8 physiotherapy ultrasound knee C
2005 % female: % female: program and light extension functio
68 66 12 weeks application n
BMI: 29.3 BMI: 28.9 7 times per of non-  WOMA
K/L grade: 3 K/L grade: week therapeutic C pain
3 3 sets, 5-10 reps. gel
Load: max
isometric
contraction
Bennell, N= 45 N= 44 Hip no ACSM Isometric  WOMA
K. L. et al, Age: 64.5 Age: 64.6 strengthening intervention knee C
2010 % female: % female: 12 weeks extension functio
51.1 45.5 5 times per n
BMI: 27.5 BMI: 28.4 week  WOMA
K/L grade: 3 K/L grade: 3 sets, 10 reps. C pain
3 Load:
10RM/hard
Börjesson N= 34 N= 34 Strength and no ACSM Isokinetic  Pain
, M. & Age: 64 Age: 64 ROM training intervention knee during
Robertso % female: % female: 5 weeks extension walking
n, E., 50 50 5 times per (NRS)
1996 BMI: 28.4 BMI: 27.7 week
K/L grade: K/L grade: 2 sets, 10 reps.
NA NA Load: max
isometric
contraction
Bruce- N= 10 N= 6 Resistance standard ACSM Isometric  WOMA
Brand, R. Age: 63.4 Age: 65.2 training care knee C
A. et al, % female: % female: 6 weeks (education, extension functio
2012 40 50 3 times per weight loss, n
BMI: 33.9 BMI: 31.7 week pharmacolo  WOMA
K/L grade: K/L grade: 3 sets, 10 reps. gical C pain
NA NA Load: Hard (Borg therapy,
14) physio
therapy)
Foroughi, N= 26 N= 28 High intensity sham- ACSM Isokinetic  WOMA
N. et al, Age: 66 Age: 65 resistance exercise knee C
2011 % female: % female: training control extension functio
100 100 6 month (same n
BMI: 31.4 BMI: 32.7 3 times per exercises,  WOMA
K/L grade: 2 K/L grade: week minimal C pain
2 3 sets, 8 reps. resistance)
Load: 80% of
1RM

22
Fransen, N= 83 N= 43 Individual and Waiting list Not- Isometric  WOMA
M. et al, Age: 67.3 Age: 66.1 group based control ACSM knee C
2001 % female: % female: physical therapy extension functio
76 67 (ROM, strength, n
BMI: 30 BMI: 28.3 aerobic)  WOMA
K/L grade: K/L grade: 8 weeks C pain
NA NA 5 times per
week ( 3 at
home)
NA sets, 20-40
reps
Load: some
effort
Gür, H. et N= 8 N= 6 Concentric- no Not- Isokinetic  Pain
al, Age: 55 Age: 57 eccentric intervention ACSM knee during
2002 % female: % female: isokinetic extension standin
NA NA resistance g on
BMI: 31.6 BMI: 32.3 training 8 weeks NRS
K/L grade: K/L grade: 3 times per
2/3 2/3 week
6 sets, 12 reps.
Load: hard
N= 9 N= 6 Concentric no Not- Isokinetic  Pain
Age: 56 Age: 57 isokinetic intervention ACSM knee during
% female: % female: resistance extension standin
NA NA training g on
BMI: 32.1 BMI: 32.3 8 weeks NRS
K/L grade: K/L grade: 3 times per
2/3 2/3 week
6 sets, 12 reps.
Load: hard
Huang, N= 33 N= 33 Isokinetic no Not- Isokinetic  Leques
M. et al, Age: 62 Age: 62 muscle- intervention ACSM knee ne
2003 % female: % female: strengthening extension functio
70 70 exercise nal
BMI: NA BMI: NA 8 weeks Index
Altman Altman 3 times per  VAS
Grade II Grade II week Pain
1-6 sets, 5 reps. during
Load: 60% of walking
1RM
N= 33 N= 33 Isotonic muscle- no Not- Isokinetic  Leques
Age: 62 Age: 62 strengthening intervention ACSM knee ne
% female: % female: exercise extension functio
70 70 8 weeks nal
BMI: NA BMI: NA 3 times per Index
Altman Altman week  VAS
Grade II Grade II 1-6 sets, 5 reps. Pain
Load: 60% of during
1RM walking
N= 33 N= 33 Isometric no Not- Isokinetic  Leques
Age: 62 Age: 62 muscle- intervention ACSM knee ne
% female: % female: strengthening extension functio
70 70 exercise nal

23
 BMI: NA BMI: NA 8 weeks Index
Altman Altman 3 times per  VAS
Grade II Grade II week Pain
1-6 sets, 5 reps. during
Load: 60% of walking
1RM
Huang, N= 30 N= 30 Isokinetic no Not- Isokinetic  Leques
M. et al, Age: 62 Age: 62 muscular intervention ACSM knee ne
2005 aug % female: % female: strengthening extension functio
77 77 exercises nal
BMI: NA BMI: NA 8 weeks Index
Altman Altman 3 times per  VAS
Grade II Grade II week Pain
1-6 sets, 5 reps. during
Load: 60% of walking
1RM
Huang, N= 35 N= 35 Isokinetic no Not- Isokinetic  Leques
M. et al, Age: 65 Age: 65 muscular intervention ACSM knee ne
2005 dec % female: % female: strengthening extension functio
81 81 exercises nal
BMI: NA BMI: NA 8 weeks Index
Altman Altman 3 times per  VAS
Grade II Grade II week Pain
1-6 sets, 5 reps. during
Load: 60% of walking
1RM
Hurley, N= 278 N= 140 Education and usual care Not- Quadriceps  WOMA
M. V. et Age: 66.9 Age: 67 individualized ACSM maximal C
al, 2007 % female: % female: exercise voluntary functio
71 69 programs contraction n
BMI: 30 BMI: 30.3 6 weeks  WOMA
K/L grade: K/L grade: 2 times per C pain
NA NA week
NA sets, NA reps
Load: individual
arrangements
Jan, M. et N= 34 N= 30 High-resistance no ACSM Isokinetic  WOMA
al, Age: 63.3 Age: 62.8 exercise intervention knee C
2008 % female: % female: 8 weeks extension functio
79 83 3 times per n
BMI: 24.1 BMI: 24 week  WOMA
K/L grade: 2 K/L grade: 3 sets, 8 reps. C pain
2 Load: 60% of
1RM
N= 34 N= 30 Low-resistance no Not- Isokinetic  WOMA
Age: 61.8 Age: 62.8 exercise intervention ACSM knee C
% female: % female: 8 weeks extension functio
79 83 3 times per n
BMI: 24 BMI: 24 week  WOMA
K/L grade: 2 K/L grade: 10 sets, 15 reps. C pain
2 Load: 10% of
1RM
Jorge, R. N= 29 N= 31 Progressive waiting list ACSM 1 RM tests  WOMA
T. B. et al, Age: 61.7 Age: 59.9 resistance control C
2014 % female: % female: exercise

24
 100 100 program functio
BMI: 30.6 BMI: 31.4 12 weeks n
K/L grade: 2 K/L grade: 2 times per  WOMA
1 week C pain
2 sets, 8 reps.
Load: 50%-70%
of 1RM
Koli, J. et N= 38 N= 40 aerobic/step- group Not- Isometric  KOOS
al, Age: 58 Age: 59 aerobic meeting on ACSM knee functio
2015 % female: % female: 12 month healthy extension n
100 100 3 times per lifestyle and  KOOS
BMI: 27.1 BMI: 26.7 week stretching ( pain
K/L grade: 2 K/L grade: NA sets, NA 1 every 3
2 reps. month)
Load: NA
Krasilshc N= 8 N= 8 Progressive no Not- Isokinetic,  WOMA
hikov, O. Age: 58.38 Age: combined intervention ACSM knee C
et al, % female: 58.25 resistance and extension functio
2011 100 % female: aerobic exercise n
BMI: 28.6 100 8 weeks  WOMA
K/L grade: BMI: 26.5 3 times per C pain
NA K/L grade: week
NA 2 sets, NA reps.
Load: NA
Kuptnirat N= 199 N= 193 Quadriceps no Not- Isometric  Functio
saikul, V. Age: 67,9 Age: 67,6 muscle intervention ACSM knee nal
et al, % female: % female: strengthening extension incapac
2002 79.4 76.7 and education ity
BMI: NA BMI: NA 8 weeks score
K/L grade: K/L grade: 2 times per  Pain
2.5 2.5 week score
NA sets, NA
reps.
Load: NA
Lim, B. N= 26 N= 26 Quadriceps no ACSM Isometric  WOMA
W. et al, Age: 67.2 Age: 66.6 strengthening intervention knee C
2008 % female: % female: 12 weeks extension functio
50 46 5 times per n
BMI: 28.2 BMI: 30.3 week  WOMA
K/L grade: 4 K/L grade: 2-3 sets, 10 reps. C pain
4 Load: Light
N= 27 N= 28 Quadriceps no ACSM Isometric  WOMA
Age: 64.1 Age: 60.8 strengthening intervention knee C
% female: % female: 12 weeks extension functio
63 61 5 times per n
BMI: 29 BMI: 28.4 week  WOMA
K/L grade: 1 K/L grade: 2-3 sets, 10 reps. C pain
1 Load: Light
Lim, J. et N= 26 N= 24 Aquatic exercise attention Not- Isokinetic  SF-36
al, Age: 65.7 Age: 63.3 8 weeks control ACSM knee Physical
2010 % female: % female: 3 times per (education) extension compo
88 88 week nent
BMI: 27.9 BMI: 27.7 NA sets, NA  Brief
K/L grade: K/L grade: reps. pain
≥2 ≥2 Load: 65% of

25
 max HR invento
ry
(NRS)
N= 25 N= 24 Land-based attention ACSM Isokinetic  SF-36
Age: 67.7 Age: 63.3 exercise control knee Physical
% female: % female: 8 weeks (education) extension compo
84 88 3 times per nent
BMI: 27.6 BMI: 27.7 week  Brief
K/L grade: K/L NA sets, NA pain
≥2 grade:≥2 reps. invento
Load: 40-60% of ry
1RM (NRS)
Lin, D. et N= 36 N= 36 Proprioceptive no Not- Isokinetic  WOMA
al, Age: 63.7 Age: 62.2 training intervention ACSM knee C
2009 % female: % female: 8 weeks extension functio
69 72 3 times per n
BMI: 23.9 BMI: 24.7 week  WOMA
K/L grade: 3 K/L grade: NA sets, NA C pain
3 reps.
Load: NA
N= 36 N= 36 Strength training no ACSM Isokinetic  WOMA
Age: 61.6 Age: 62.2 8 weeks intervention knee C
% female: % female: 3 times per extension functio
67 72 week n
BMI: 23.7 BMI: 24.7 4 sets, 6 reps.  WOMA
K/L grade: 3 K/L grade: Load: min 50% C pain
3 of 1RM
Lund, H. N= 27 N= 27 Aquatic exercise no Not- Isokinetic  KOOS
et al, Age: 65 Age: 70 (strength, intervention ACSM knee functio
2008 % female: % female: aerobic, balance, extension n
83 66 stretching)  KOOS
BMI: 27.4 BMI: 26.1 8 weeks pain
K/L grade: K/L grade: 2 times per
NA NA week
NA sets, NA
reps.
Load: 40% of
1RM(in one
exercise)
N= 25 N= 27 Land-based no Not- Isokinetic  KOOS
Age: 68 Age: 70 exercise intervention ACSM knee functio
% female: % female: (strength, extension n
88 66 aerobic, balance,  KOOS
BMI: 23.7 BMI: 26.1 stretching) pain
K/L grade: K/L grade: 8 weeks
NA NA 2 times per
week
NA sets, NA
reps.
Load: 40% of
1RM(in one
exercise)
Maurer, N= 57 N= 56 Strength training education Not- Isokinetic  WOMA
B. T. et al, Age: 66.3 Age: 64.5 8 weeks (disease, ACSM knee C

26
1999 % female: % female: 3 times per joint extension functio
47 36 week protection, n
BMI: NA BMI: NA 3 sets, 3 reps. nutrition,  WOMA
K/L grade: K/L grade: Load: NA pain coping) C pain
NA NA
McKay, C. N= 10 N= 12 Lower-body placebo ACSM Isometric  WOMA
et al, Age: 63.5 Age: strength training control knee C
2012 % female: 60.58 6 weeks (upper body extension functio
50 % female: 3 times per strength n
BMI: 35.03 66.7 week training)  WOMA
K/L grade: BMI: 2 sets, 8 reps. C pain
NA 33.78 Load: 60% of
K/L grade: 1RM
NA
Mikesky, N= 113 N= 108 Strength training range of ACSM Isokinetic  WOMA
A. E. et Age: 69.4 Age: 68.6 30 month motion knee C
al, 2006 % female: % female: 3 times per (same extension functio
57 59 week schedule as n
BMI: 29.6 BMI: 29 3 sets, 8-10 reps. ST)  WOMA
K/L grade: 2 K/L grade: Load: 8-12RM C pain
2
Multanen N= 38 N= 40 High impacts, attention Not- Isometric  WOMA
, J. et al, Age: 58 Age: 59 multidirectional control ACSM knee C
2014 % female: % female: aerobic and (group extension functio
100 100 step-aerobic meeting n
BMI: 27.1 BMI: 26.7 jumping exercise every 3  WOMA
K/L grade: 2 K/L grade: program month) C pain
2 12 month
3 times per
week
NA sets, NA
reps.
Load: NA
O’Reilly, N= 108 N= 72 Graded no Not- Isometric  WOMA
S. C. et al, Age: 61.94 Age: strengthening intervention ACSM knee C
1999 % female: 62.15 exercise extension functio
64.8 % female: program n
BMI: NA 68.1 6 month  WOMA
K/L grade: BMI: NA 7 times per C pain
NA K/L grade: week
NA NA sets, 20 reps.
Load: NA
Péloquin, N= 59 N= 65 combination of attention Not- Isometric  AIMS2
L. et al, Age: 65.64 Age: aerobic, control ACSM knee Mobilit
1999 % female: 66.43 strength, stretch (education 2 extension y level
71.2 % female: 12 weeks times  AIMS2
BMI: 29.79 69.2 3 times per monthly, arthritis
K/L grade: 2 BMI: week total 6 pain
29.77 NA sets, NA session)
K/L grade: reps.
2 Load: NA
Quilty, B. N= 43 N= 44 Physiotherapy No Not- Isometric  WOMA
et al, Age: 66.8 Age: 66.7 (taping, posture intervention ACSM knee C
2003 % female: % female: correction, extension functio
100 100 strength

27
 BMI: 30.2 BMI: 30 training, weight n
K/L grade: K/L grade: loss and  VAS
1/2 1/2 footwear advice) pain
10 weeks
7 times per
week
5 sets, 10 reps.
Load: max
contraction (only
for one exercise)
Rooks, D. N= 22 N= 23 combination of education ACSM Isokinetic  WOMA
S. et al, Age: 65 Age: 69 aquatic and control knee C
2006 % female: % female: land-based (handout extension functio
50 57 exercise and 1 n
BMI: 35.7 BMI: 33.9 (strength, telephone  WOMA
K/L grade: K/L grade: flexibility) call) C pain
NA NA 6 weeks
3 times per
week
2 sets, 8-12 reps.
Load: light
Røgind, N= 11 N= 12 Physiotherapy no Not- Isometric  VAS
H. et al, Age: 69.3 Age: 73 training intervention ACSM knee weight-
1998 % female: % female: (strength, extension bearing
90.9 91.7 flexibility,
BMI: 27.4 BMI: 26.8 balance/coordin
K/L grade: 3 K/L grade: ation)
3 3 month
6 times per
week
3 sets, Min. of
10 reps.
Load: 70% of
their individual
ability
Salacinski N= 19 N= 18 Cycling (aerobic) waiting list Not- Isokinetic  WOMA
, A. J. et Age: 55.1 Age: 60.6 12 weeks control ACSM knee C
al, 2012 % female: % female: 2 times per extension functio
79 67 week n
BMI: 22.4 BMI: 25.7 NA sets, NA  WOMA
K/L grade: 2 K/L grade: reps. C pain
2 Load: 70-75% of
max HR
Salli, A. et N= 23 N= 24 Isokinetic no ACSM Isokinetic  WOMA
al, Age: 55.73 Age: 58.3 8 weeks intervention knee C
2010 % female: % female: 3 times per extension functio
83 79 week n
BMI: 31.5 BMI: 5 sets, 10 reps.  VAS
K/L grade: 2 32.82 Load: 70% of pain
K/L grade: 1RM during
2 motion

N= 24 N= 24 Isometric no Not- Isokinetic  WOMA

Age: 57.1 Age: 58.3 8 weeks intervention ACSM knee C

28
 % female: % female: 3 times per extension functio
83 79 week n
BMI: 32.65 BMI: NA sets, NA  VAS
K/L grade: 2 32.82 reps. pain
K/L grade: Load: NA during
2 motion

Sayers, S. N= 12 N= 11 High-speed attention ACSM 1RM leg  WOMA

P. et al, Age: 66.9 Age: 68.4 power training control press, knee C
2012 % female: % female: 12 weeks (stretching 3 extension functio
75 73 3 times per times a n
BMI: 28.4 BMI: 30.8 week week)  WOMA
K/L grade: 2 K/L grade: 3 sets, 12-14 C pain
2 reps.
Load: 40% of
1RM
N= 10 N= 11 slow-speed attention ACSM 1RM leg  WOMA
Age: 65.9 Age: 68.4 strength training control press, knee C
% female: % female: 12 weeks (stretching 3 extension functio
80 73 3 times per times a n
BMI: 33.1 BMI: 30.8 week week)  WOMA
K/L grade: 2 K/L grade: 3 sets, 8-10 reps. C pain
2 Load: 80% of
1RM
Schilke, J. N= 10 N= 10 Isokinetic no Not- Isokinetic  OASI
M. et al, Age: 64.5 Age: 68.4 strength training intervention ACSM knee mobilit
1996 % female: % female: 8 weeks extension y
85 85 3 times per  OASI
BMI: NA BMI: NA week pain
K/L grade: K/L grade: 6 sets, 5 reps.
NA NA Load: max
contraction
Sekir, U. N= 12 N= 10 Proprioceptive no Not- Isometric  Subjecti
& Gür, H., Age: 59 Age: 62 training intervention ACSM knee ve
2005 % female: % female: 6 weeks extension rating
75 70 2 times per of daily
BMI: 32.7 BMI: 29.9 week activity
K/L grade: K/L grade: 1-3 sets, NA  VAS
2/3 2/3 reps. pain in
Load: NA standin
g
Swank, A. N= 36 N= 35 Rehabilitation usual care ACSM Isokinetic  VAS
M et al, Age: 63.1 Age: 62.6 resistance knee pain
2011 % female: % female: program and extension during
67 63 usual care walking
BMI: 35.9 BMI: 32.9 4-8 weeks
K/L grade: K/L grade: 3 times per
NA NA week
1-2 sets, 10 reps.
Load: moderate
Talbot, L. N= 17 N= 17 Walking education Not- Isometric  VAS
A. et al, Age: 69.59 Age: program in ACSM knee present
2003 % female: 70.76 addition to extension pain
76.5 % female: education

29
 BMI: 31.01 76.5 12 weeks
K/L grade: 3 BMI: 7 times per
32.63 week
K/L grade: NA sets, NA
2 reps.
Load: NA
Thomas, N= 470 N= 316 Home exercise no Not- Isometric  WOMA
K. S. et al, Age: 62 Age: 61.7 (strength, ROM, intervention ACSM knee C
2002 % female: % female: locomotor extension functio
62 65 function) n
BMI: 27.95 BMI: 24 month  WOMA
K/L grade: 28.12 7 times per C pain
NA K/L grade: week
NA NA sets, NA
reps.
Load: NA
Trans, T. N= 18 N= 17 Whole body no Not- Isometric  WOMA
et al, Age: 58.7 Age: 61.1 vibration intervention ACSM knee C
2009 % female: % female: exercise on a extension functio
100 100 balance board n
BMI: 29.1 BMI: 30.2 (VibF)  WOMA
K/L grade: K/L grade: 8 weeks C pain
NA NA 2 times per
week
1 set, 6-9 reps.
Load: NA
N= 17 N= 17 Whole body no Not- Isometric  WOMA
Age: 61.5 Age: 61.1 vibration intervention ACSM knee C
% female: % female: exercise on a extension functio
100 100 stable platform n
BMI: 29.2 BMI: 30.2 (VibM)  WOMA
K/L grade: K/L grade: 8 weeks C pain
NA NA 2 times per
week
1 set, 6-9 reps.
Load: NA
Villadsen, N= 41 N= 40 Neuromuscular basic Not- Knee  KOOS
A. et al, Age: 67.1 Age: 65.1 exercise and education ACSM extension ADL
2014 % female: % female: basic education  KOOS
61 60 8 weeks Pain
BMI: 30.8 BMI: 33.4 2 times per
K/L grade: K/L grade: week
NA NA 2-3 sets, 10-15
reps
Load: NA
Weidenhi N= 20 N= 20 Physiotherapy no ACSM Isokinetic  VAS
elm, L. et Age: 64 Age: 63 training intervention knee Pain
al, % female: % female: 5 weeks extension during
1993 55 45 3 times per walking
BMI: 30.1 BMI: 29.1 week
K/L grade: K/L grade: 2 sets, 10 reps.
NA NA Load: Light
Weng, M. N= 33 N= 33 isokinetic muscle no ACSM Isokinetic  Leques
et al, Age: 64 Age: 64 strengthening intervention knee ne
2009 % female: % female: 8 weeks extension

30
 80 80 3 times per index
BMI: NA BMI: NA week  VAS
Altman Altman 1-6 sets, 10 reps. Pain
Grade II Grade II Load: 60% of during
1RM walking
ADL: Activities of Daily Living; AIMS2: Arthritis Impact Measurements Scale 2; HR: hear rate; KOOS: Knee injury and
Osteoarthritis Outcome Score; NRS: Numeric Rating Scale; OASI: Osteoarthritis Screening Index; RM: Repetition
maximum; SF-36: Short Form Health Survey; VAS: Visual Analog Scale; WOMAC: Western Ontario and McMaster
Universities Osteoarthritis

31
Table 2. Stratified meta-analyses according to type of exercise intervention administered.

Analyses Number of SMD Tau² I²

Random effects analyses comparisons (95% Confidence
Interval)
Knee extensor strength
Overall effect 56 0.59 (0.39 to 0.75) 0.25 80.9 %
Sensitivity (fixed effects 56 0.41 (0.34 to 0.47) 0.16 -
analysis)
ACSM guideline
Yes 22 0.83 (0.49 to 1.17) 0.55 86.9 %
No 34 0.38 (0.27 to 0.50) 0.04 42.2 %
Pain
Overall effect 56 0.57 (0.42 to 0.73) 0.24 80.9 %
Sensitivity (fixed effects 56 0.39 (0.33 to 0.45) 0.14 -
analysis)
Yes 22 0.62 (0.32 to 0.93) 0.42 84.2 %
No 34 0.52 (0.35 to 0.68) 0.14 73.2 %
Function
Overall effect 50 0.56 (0.39 to 0.73) 0.27 83.2 %
Sensitivity (fixed effects 50 0.34 (0.28 to 0.40) 0.17 -
analysis)
ACSM guideline
Yes 19 0.64 (0.28 to 1.00) 0.53 86.9%
No 31 0.49 (0.33 to 0.65) 0.42 71.8 %

32
Figure 1. Flow of randomized controlled trials included in the systematic review and meta-analyses.

Figure 2. Weighted random effects meta-regression analyses: standardized mean differences (SMD) of the

individual studies according to pain (A) and disability (B) at different percentages of changes in knee

extensor strength. From approximation (visual) of the intercept of the 95% prediction interval with the x-

axes, the lower limit of percent change needed in knee extensor strength to achieve significant change in

pain and disability is ~30% and 40%, respectively.

33
34