Knee osteoarthritis is a leading cause of pain and disability worldwide, having a considerable impact on both the health-care system and a patient’s long-term quality of life1-4. The prevalence of osteoarthritis is rising with the aging population and the increased presence of its associated risk factors, such as obesity and low levels of physical activity1,5,6. In the United States, knee osteoarthritis affects approximately 9 million adults and has an estimated incidence of 240 per 100,000 people per year, with greater risk among women and obese individuals3,7-9. By the year 2020, osteoarthritis is expected to be 1 of the top 10 causes of disability (measured in terms of disability-adjusted life-years) in developed countries10.
Osteoarthritis typically worsens slowly over time; treatments provide symptomatic relief and may at least theoretically slow the progression of articular damage1,11. The most common complaint of patients with osteoarthritis is pain, which often affects an individual’s daily activities1,2,4. Pharmacological treatment is uniformly recommended for initial management; however, a number of options are available, and recommendations as to the efficacy of these treatments are inconsistent between knee osteoarthritis guidelines and medical providers1,4,6,12-17.
For a treatment to be approved for use in clinical practice, its efficacy and safety must be demonstrated in clinical trials. Proof of the statistical significance of trial results is established by a p value, but a more directly applicable determination for clinicians is whether or not these results are also clinically important18. The concept of the minimum clinically important difference (MCID) was introduced in 1989 by Jaeschke et al.18. This concept was developed in an effort to determine and communicate whether there was clinical relevance associated with the observed differences between treatments in a clinical trial. Jaeschke et al. defined the MCID as “the smallest difference in score in the domain of interest which patients perceive as beneficial and which would mandate, in the absence of troublesome side effects and excessive cost, a change in the patient’s management.” The MCID is typically determined through an anchor-based approach. This requires patients to respond to a questionnaire that asks an “anchor” question that assesses the extent to which they feel that their pain has improved. At the same time points, patients answer a continuous-scale patient-reported pain outcome measure, such as the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) or a visual analog scale (VAS) pain score. This allows for calculation of the amount of pain reduction on the continuous scale that is associated with a patient changing his or her opinion about clinical improvement on the anchor question19. The MCID is recognized to vary between different patient populations and the various health outcome measures used in clinical trials, but variability also may be seen among studies examining the same patient population as a result of differences in study design, study location, and treatment product administered.
Numerous clinical practice guidelines and meta-analyses have been published to guide physicians in the treatment of knee osteoarthritis; therefore, we conducted a systematic review of the literature to determine how much variation exists in the MCIDs used in the knee osteoarthritis literature.
Materials and Methods
The primary objective was to identify and describe the variability in the MCID for pain assessments used in guidelines and meta-analyses evaluating nonsurgical interventions for knee osteoarthritis. The secondary objective was to evaluate the treatment effects of nonsurgical interventions found in these guidelines and meta-analyses against a common MCID threshold from the literature.
Search and Eligibility Criteria
A systematic search (Appendix A) was conducted on September 2, 2016, in MEDLINE and Embase (via Ovid) to identify published guidelines and meta-analyses on nonsurgical treatments for knee osteoarthritis. A manual search was also performed in the National Guideline Clearinghouse web site (www.guideline.gov). The specific study eligibility criteria are included in Appendix A.
Duplicate references, conference proceedings, book chapters, and historical reviews were removed after completement of the search. The remaining titles and abstracts were screened. Full-text publications of potentially eligible abstracts were then reviewed for inclusion. An updated search was conducted on November 23, 2018 to identify any articles published between that date and the date of the original search.
For each eligible guideline and meta-analysis, the publication information, methodological details, treatment effect estimate, and reported MCID value were extracted.
When sufficient information was provided in the publication, the quality of evidence of pain outcomes was also assessed with use of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach20-22. GRADE is a tool that is used to evaluate the quality of a body of evidence through assessing the risks of bias, imprecision, inconsistency, indirectness, and potential publication bias within a body of literature23. The methodology proposed by Salanti et al., which is based on the GRADE approach, was applied to evaluate network meta-analyses24. For this assessment, interventions were categorized as nonpharmacological interventions, intra-articular injections, or oral pharmacological interventions.
Comparison of MCIDs Used for Pain Assessments
During the extraction of the MCIDs for pain outcomes, the specific pain scale, whether the comparison was a between-group or within-group difference, and whether the MCID was a single value or range of values were also recorded. The results were then compared directly with the American Academy of Orthopaedic Surgeons (AAOS) guideline, which is considered the gold standard for orthopaedic surgeons in North America7, as well as other included knee osteoarthritis guidelines. The findings were tabulated and described qualitatively.
Comparison of Treatment Effects and Their Clinical Significance
Only interventions that had a treatment effect for pain outcomes reported in the AAOS guideline and that were evaluated in at least 1 other meta-analysis or guideline were included in this specific comparison. Also, only effect estimates provided on the WOMAC pain subscale, or that listed WOMAC pain at the top of the study’s hierarchy of outcomes25, were included to limit the potential heterogeneity between the various pain scales. The WOMAC was prioritized on the basis of its priority in a previously published hierarchy that is commonly used in knee osteoarthritis meta-analyses25. All estimates for a given treatment were plotted with the range of MCIDs identified from the search to assess their clinical significance relative to common thresholds. This was also done to compare the different pharmacological therapies relative to their respective placebos, but only estimates from most recent or comprehensive meta-analyses were summarized to ensure that the most accurate and up-to-date effect estimates were presented. The standardized mean difference (SMD)26 was used to ensure that effect estimates were on a common scale. The values provided in the AAOS guideline were converted to SMD units as the estimates in that guideline were not presented on this scale (a weighted mean difference or MCID units may have been used instead). For example, if the effect estimate was presented as 0.5 MCID unit, the corresponding value on the SMD scale (x) was calculated as:
The same descriptions for clinical significance provided in the AAOS guideline were used (Appendix B). These definitions are dependent on the confidence interval (CI) surrounding the estimate; however, meta-analyses conducted via Bayesian methods calculate a credible interval (CrI)27. For the purposes of the present study, it was assumed that these 2 measures of precision were equivalent. Last, the ratios of the point estimates of effect size for nonsteroidal anti-inflammatory drug (NSAID) therapies, because of their consistently positive recommendations in the guideline literature28, were calculated relative to the effect sizes of the other pharmacological interventions.
The initial search resulted in a total of 5,936 citations. After removing duplicates, conference proceedings, book chapters, and historical reviews, 4,204 titles and abstracts were screened. Four hundred and seventy full-text articles were reviewed. After full-text screening, 33 guidelines and meta-analyses met the eligibility criteria. One additional guideline was identified from the manual search. Three eligible guidelines or meta-analyses were found from the updated search; therefore, 37 references were included in the current analysis (Fig. 1)1-3,5-7,11-14,16,17,29-53.
Twenty-two pairwise (or traditional) meta-analyses, 5 Cochrane reviews, 5 network meta-analyses, 4 guidelines (1 of which only focused on exercise interventions), and 1 health technology assessment that included a meta-analysis on intra-articular hyaluronic acid were included (Table I). Eleven of the guidelines and meta-analyses examined >1 knee osteoarthritis treatment, whereas the other 26 investigated just 1 therapy. Variations in items related to the search strategy and identification of eligible studies, inclusion and exclusion criteria, and statistical methodology were noted across these investigations. Additional publication characteristics and methodological details for each of the included guidelines and meta-analyses are summarized in Appendix B.
TABLE I -
Summary of the Included Studies (n = 37)
|Study Type and Treatment Evaluated
||No. of Studies
| Pairwise meta-analysis
| Cochrane review
| Network meta-analysis
| Health technology assessment*
| Multiple treatments
| Intra-articular hyaluronic acid
| Exercise interventions
| Dietary supplements
| Duloxetine (SNRI)
| Valgus bracing
| Weight loss
Included a meta-analysis.
SNRI = serotonin and norepinephrine reuptake inhibitor, NSAIDs = nonsteroidal anti-inflammatory drugs, PEMF = pulsed electromagnetic field, and TENS = transcutaneous electrical nerve stimulation.
The quality of evidence was evaluated for 87 effect estimates for pain across 33 publications (Appendix C); the information provided in the other 4 publications was not sufficient to perform a GRADE assessment. The quality of evidence was rated as very low for 16 effect estimates (18.4%), low for 42 (48.3%), moderate for 21 (24.1%), and high for 8 (9.2%). The effect estimates were categorized as high for 16.7% of oral pharmacological therapies, compared with 3.4% of intra-articular interventions and 10.9% of nonpharmacological interventions. The categorization of the quality of evidence by intervention type for all effect estimates is summarized in Appendix D.
Comparison of MCIDs Used for Pain Assessments
MCIDs for pain values were often presented as an SMD or mean difference (MD) on a pain-specific outcome measure (Table II). The most common scales were the WOMAC pain subscale and visual analog scale (VAS) for pain.
TABLE II -
MCIDs for Pain
||Single Values Used (No. of Meta-Analyses/Guidelines)
||Range of MCID Values Used (No. of Meta-Analyses/Guidelines)
|WOMAC, Likert (0-20)
|SF-36, bodily pain (0-100)
||Standard deviation units
||Percent improvement relative to control
|Within-group (pre-treatment to post-treatment) difference
||Standard deviation units
||Percent decrease from baseline
MCID reported in the American Academy of Orthopaedic Surgeons (AAOS) guideline.
MCID reported in the European League Against Rheumatism (EULAR) and National Institute for Health and Care Excellence (NICE) guidelines.
When the MCID was expressed as an SMD, values ranged from 0.20 to 1.23 standard deviation (SD) units. When the MCID for pain was presented on the VAS (0 to 100), values ranged from 10 to 19.9 points. MCID ranges of 8.4 to 19.9, 10 to 30, and 15 to 20 points were also reported for the VAS (0 to 100). When the MCID was on the WOMAC pain subscale (0 to 100), either 8.3 or 9.7 points were used; however, a range of 9 to 12 points was also reported for this scale. The authors of 1 meta-analysis used an MCID of 5 points for the WOMAC pain (0 to 20-point) Likert scale16. MCIDs for the Short Form-36 (SF-36) bodily pain subscale (100-point scale) and numerical rating scale (NRS, 10-point scale) were 7.8 points and 1 point, respectively. Last, the authors of a guideline used a percentage value as their MCID, stating that a ≥15% improvement relative to a control on a continuous outcome measure was clinically relevant5.
A pre-treatment to post-treatment evaluation (i.e., the improvement from baseline within an individual group) for differences in pain was conducted in 6 meta-analyses14,17,30,32,48,49. The MCIDs for these assessments also varied; a 20-point improvement on the WOMAC pain subscale (0 to 100), a 2-point difference on the VAS (0 to 10), an SMD of 0.37 SD unit, and a 10% to 20% decrease in pain from baseline were all used across these investigations. The authors of 1 meta-analysis also stated that patients who achieved a score of <32.3 on the VAS (0 to 100) after their baseline visit reached a clinically significant result48.
In the methods section of the AAOS guideline, the authors reported an MCID in SD units for 3 different pain measures (the WOMAC pain subscale, VAS for pain, and SF-36 bodily pain scale). The MCID for WOMAC pain (SMD = 0.39) was also used in 2 other publications, but the MCIDs for the VAS (SMD = 1.23) and SF-36 (SMD = 0.47) were not reported in any other publication besides the AAOS guideline (Table II). Two other knee osteoarthritis guidelines, from the European League Against Rheumatism (EULAR) and National Institute for Health and Care Excellence (NICE), were included in the present review; both used an SMD of 0.50 SD unit as the MCID.
Comparison of Treatment Effects and Their Clinical Significance
According to the criteria established for this comparison, only treatment effects for intra-articular hyaluronic acid (also presented as low-molecular-weight [<1,500 kDa] and higher-molecular-weight [≥1,500 kDa] variations), intra-articular corticosteroids, oral (non-selective and cyclooxygenase [COX]-2-selective) and topical NSAIDs, acetaminophen, and acupuncture were included (i.e., effect estimates for the remaining nonsurgical therapies listed in the methods were not eligible for this analysis); however, the AAOS guideline did not synthesize trials examining intra-articular corticosteroids, so the estimates reported for each individual trial were presented.
There were similarities in the effect estimates for oral NSAIDs7,30,32,34,51 (Fig. 2-A), topical NSAIDs7,30 (Fig. 2-B), and intra-articular corticosteroids7,30,32,51 (Fig. 3-A) reported across the included guidelines and meta-analyses; however, there were inconsistencies across the studies and guidelines in the estimates for intra-articular hyaluronic acid7,30-32,45,49,51 (Fig. 3-B), acetaminophen7,30,32 (Fig. 4-A), and acupuncture1,7,38,42,43 (Fig. 4-B). Figures 2, 3, and 4 also demonstrated how differences in the effect estimates and the selection of the MCID (the range of 0.20 to 0.50 SD units presented in the figures) can alter the clinical significance of these therapies.
Fig. 2-A and 2-B Comparison of treatment effect estimates (in SD units) for oral NSAIDs7,30,32,34,51 (Fig. 2-A) and topical NSAIDs7,30. (Fig. 2-B) with corresponding 95% confidence intervals. The blue line indicates the AAOS MCID for WOMAC pain (0.39), and the dotted red lines indicate the range of other reported MCIDs (0.20 to 0.50). AAOS 5 American Academy of Orthopaedic Surgeons, NSAID 5 nonsteroidal anti-inflammatory drug, SMD 5 standardized mean difference.
Figs. 3-A and 3-B Comparison of treatment effect estimates (in SD units) for intra-articular corticosteroids7,30,32,51(Fig. 3-A) and intra-articular hyaluronic acid7,30-32,45,49,51(Fig. 3-B) with corresponding 95% confidence intervals. The blue line indicates the AAOS MCID for WOMAC pain (0.39), and the dotted red lines indicate the range of other reported MCIDs (0.20 to 0.50). AAOS = American Academy of Orthopaedic Surgeons, MW = molecular weight, IA = intra-articular, SMD = standardized mean difference.
Figs. 4-A and 4-B Comparison of treatment effect estimates (in SD units) for acetaminophen7,30,32 (Fig. 4-A) and acupuncture.1,7,38,42,43 (Fig. 4-B) with corresponding 95% confidence intervals. The blue line indicates the AAOS MCID for WOMAC pain (0.39), and the dotted red lines indicate the range of other reported MCIDs (0.20 to 0.50). AAOS 5 American Academy of Orthopaedic Surgeons, NICE 5 National Institute for Health and Care Excellence, SMD 5 standardized mean difference.
The effect estimates used to compare therapies with a placebo control were 0.14 for topical NSAIDs, 0.43 for non-selective NSAIDs, 0.34 for COX-2-selective NSAIDs, 0.18 for acetaminophen, 0.32 for intra-articular corticosteroids, 0.34 for intra-articular hyaluronic acid, 0.57 for higher-molecular-weight intra-articular hyaluronic acid, and 0.23 for low-molecular-weight intra-articular hyaluronic acid (Fig. 5)30,32,45. On the basis of these estimates and the descriptions for clinical significance outlined in Appendix B only non-selective NSAIDs, COX-2-selective NSAIDs, intra-articular hyaluronic acid, and higher-molecular-weight intra-articular hyaluronic acid had precision intervals that exceeded the lowest MCID threshold (0.20 SD unit) and were, therefore, deemed “clinically significant,” whereas topical NSAIDs, acetaminophen, intra-articular corticosteroids, and low-molecular-weight intra-articular hyaluronic acid were considered “possibly clinically significant.” Compared with the AAOS MCID for WOMAC pain (0.39 SD unit), only higher-molecular-weight intra-articular hyaluronic acid was “clinically significant,” whereas non-selective NSAIDs, COX-2-selective NSAIDs, intra-articular corticosteroids, and intra-articular hyaluronic acid were “possibly clinically significant.” Compared with a more conservative MCID of 0.50 SD unit, none of the treatments represented effect estimates that were “clinically significant”; only the effect sizes of non-selective NSAIDs and higher-molecular-weight intra-articular hyaluronic acid were “possibly clinically significant.”
In Figure 6, values of <1.0 indicate that the effect estimate of the respective NSAID therapy was larger than that of the comparison treatment. Intra-articular hyaluronic acid (regardless of molecular weight), intra-articular corticosteroids, and acetaminophen all had larger effect estimates than topical NSAIDs (Fig. 6, top panel). Compared with both COX-2-selective and non-selective NSAIDs, only higher-molecular-weight intra-articular hyaluronic acid had a larger effect than the 2 forms of oral NSAID therapy (Fig. 6, middle and bottom panels).
This systematic review confirmed the variability in the MCID for pain assessments used across guidelines and meta-analyses evaluating nonsurgical interventions for knee osteoarthritis. In some cases, the threshold for clinical significance was even inconsistent for a single scale (e.g., the WOMAC pain subscale or VAS for pain). The clinical significance associated with an effect estimate for a treatment in patients with knee osteoarthritis is dependent on which threshold is selected. If too low of an MCID is employed, an ineffective treatment may be determined to provide benefit. Alternatively, setting the MCID too high may result in rejecting treatments that are indeed efficacious. The findings also demonstrate that treatment effects of a given therapy could be influenced by either clinical factors (e.g., hyaluronic acid molecular weight) or methodological (both study and statistical) factors. The comparison of the effect estimates of various pharmacological interventions versus a common comparator (e.g., placebo) showed considerable overlap. This similarity of treatment effect size across treatments raises questions as to the validity of and rationale for the heterogeneity of the recommendations assigned to some of these treatments, with certain knee osteoarthritis guidelines recommending for, and others against, the treatments that generate similar effect sizes. Such observations might explain, at least in part, why prior research has demonstrated conflicting treatment recommendations among the various knee osteoarthritis guidelines15,28 and further reinforce the contention that the MCID should be used strictly as a supplementary tool when making treatment decisions.
One factor that could have altered these results and conclusions is that trial quality may have an impact on treatment effect estimates. As highlighted in Appendix D, the quality of evidence of the included pain outcomes was quite variable. If, for example, the authors of a given meta-analysis were to restrict their analysis to include only high-quality studies, the resulting effect size may very well be different from the effect calculated regardless of trial quality, which could then change the interpretation of clinical significance for the effect size. Another consideration that should be addressed is that the WOMAC pain scale was prioritized in this comparison. As noted in the results of the literature search, different scales have different MCIDs. If, for example, the VAS for pain had been prioritized instead of the WOMAC for this analysis, the effect sizes of the included therapies may have compared differently to the range of MCIDs reported for the VAS, which might (or might not) have changed the conclusions made about their clinical significance. An additional consideration for future research on knee osteoarthritis is the emergence of platelet-rich plasma (PRP) as a potential treatment option. This therapy has a small but growing body of evidence examining its effects in comparison with those of intra-articular saline solution as well as those of active comparators such as intra-articular hyaluronic acid. Current guidelines have not typically included PRP in their formal recommendations, perhaps because of the lack of a body of evidence to aid in decision-making, yet this body of evidence is continually growing and should be considered in future knee osteoarthritis investigations.
The authors of the included guidelines and meta-analyses often referred to the previous literature when selecting the most appropriate MCID for their research investigations. Historically, the MCID has been determined with one of a number of different methods, which can be categorized as expert-based, distribution-based, or anchor-based. An expert-based approach involves the development of a consensus panel to review and discuss the available data to determine the MCID, which may involve a voting procedure54,55. A distribution-based MCID involves the use of statistical parameters to estimate thresholds of discrimination for changes in health, as based on the early work by Cohen56 and Norman et al.57. Anchor-based MCIDs, calculated by relating changes on a continuous outcome measure to changes in a patient’s self-assessment of health status58-63, were common across the guidelines and meta-analyses included in this review; however, anchor-based MCIDs may differ for various reasons. For example, the 2 studies by Angst et al.64,65 both used a transition questionnaire to ask patients if their health was “much worse,” “slightly worse,” “equal,” “slightly better,” or “much better” relative to their baseline status, whereas Tubach et al. asked patients to assess their response to treatment on either a 5-point (“none,” “poor,” “fair,” “good,” or “excellent”) or 15-point (ranging from “a very great deal worse” to “a very great deal better”) Likert scale66. Angst et al. defined the MCID for improvement as patients who stated that their health was “slightly better” relative to baseline, whereas Tubach et al. used the scores of 75% of the patients who deemed their response to treatment as “good.”64-66 This example highlights the controversy on how much of a change in a patient’s health status should be considered a “minimal” important difference, as the definition provided by Angst et al. appears to represent a smaller change than the criteria used by Tubach et al., which was, arguably, set too high67. The MCID is recognized to be context-specific, and a patient’s expectation may change depending on factors such as the type of treatment received, the time between outcome assessments, and the patient’s age or other baseline characteristics67,68. Variation in the patient populations and anchors used to determine these clinically significant thresholds may account for the wide range of MCIDs reported in the knee osteoarthritis literature. The anchor-based MCIDs used across these studies also represented populations of patients with conditions other than osteoarthritis (e.g., ankylosing spondylitis, low back pain, fibromyalgia, peripheral neuropathy, rheumatoid arthritis, rotator cuff disease, acute pain/trauma, endometriosis-associated pelvic pain)59,61-63,69-73. A recent systematic review assessing the credibility of the MCID for patients with knee osteoarthritis suggested that the most credible estimates of the MCID were around 12 points for the WOMAC pain score, which is considerably lower than the 19.9-point VAS pain MCID referenced frequently in the assessed literature19. This range of credible MCIDs should be considered in the application of prior research findings, as clinical interpretation of the MCID may vary depending on the cutoff used. An MCID of 12 points on the VAS pain scale is a better representation of an appropriate MCID cutoff19. As the MCID for an outcome may be derived by evaluating a within-group change in health status relative to baseline, it is not clear that the same MCID is appropriate for between-group comparisons67. For example, an SMD of 0.37 has been used for both within-group and between-group differences in pain scores. Also, although the VAS for pain commonly has been used as a pain scale in the knee osteoarthritis literature, it can be assessed under different scenarios (e.g., global pain, pain at rest, pain during activity), and, in many cases, the same MCID value is used for the VAS without this consideration. Using a single threshold as a marker for clinical significance also may be problematic67. When an outcome is quantified on a continuous scale, the scores typically are averaged across all patients in the sample; therefore, if this average score (including its precision interval) does not exceed the MCID, then it would not be considered a clinically significant effect, which could potentially ignore a proportion of individual patients who did have a score greater than this value67,74. When pooled effect estimates of a given intervention are inconsistent across guidelines and meta-analyses, the inconsistency may be due to differences in search strategy, inclusion and exclusion criteria, selection of outcome measures (multiple scales are available that quantify the same outcome), or statistical analysis plan (e.g., the results of a pairwise meta-analysis versus a network meta-analysis or frequentist versus Bayesian approaches). For many nonsurgical interventions for knee osteoarthritis, recommendations for or against use vary across the knee osteoarthritis guidelines28. In the present study, the effect estimates of treatments that have received less support in the guidelines had similar, if not greater, effect sizes than those that have been more commonly recommended. For example, as shown in Figures 5 and 6, the treatment effects of intra-articular hyaluronic acid, especially higher-molecular-weight formulations, were similar to, if not greater than, those of both non-selective and COX-2-selective NSAIDs. Relevant stakeholders must come to a consensus on a standard approach to determining the clinical significance of these therapies in order to ensure that patients are not limited in their treatment options prior to invasive surgical intervention.
The strengths of the current study were that a systematic and reproducible search was conducted within the most comprehensive electronic databases (PubMed/MEDLINE and Embase) and that a manual search for references that were not indexed in these databases was also performed. The most patient-important and most commonly assessed outcome among patients with knee osteoarthritis (i.e., pain) was examined. Meta-analyses and guidelines were only included if treatments were evaluated against a common comparator (placebo or control) in order to ensure comparability between studies when assessing the MCIDs and treatment effect sizes. One of the limitations of the present review was the inclusion of only studies that were published in English. It is unclear how the results might compare with the knee osteoarthritis literature in other countries, and it is possible that other reported MCIDs for this patient population were missed. Only pain outcomes were examined; however, treatment decisions also may be based on other variables, such as costs and safety. Also, as only evaluations of nonsurgical therapies were included, it cannot be assumed that the results and conclusions of the literature search are also representative of a population of patients with knee osteoarthritis that requires surgical intervention. The comparison to published MCID values aimed to provide clinical insights into the treatment effects seen; however, the present study was limited by the inability to determine which of the published MCIDs is most appropriate to inform clinical significance within this patient population. For this reason, >1 MCID cutoff point was included in the comparison.
The review of the literature identified considerable variability in the MCIDs for pain assessments used across guidelines and meta-analyses evaluating nonsurgical interventions for knee osteoarthritis. Such a wide range of MCIDs may lead to conflicting treatment recommendations, potentially rejecting treatments that are efficacious or accepting treatments that are not effective for individuals in this patient population. Additional research is required to determine why certain pharmacological therapies are more consistently recommended in the knee osteoarthritis guidelines than others as the findings suggest substantial similarities in their effect estimates for pain outcomes. Relevant stakeholders must come to a consensus on a standard approach for determining the clinical significance of these therapies to ensure that patients are provided with all of the effective treatment options that may abrogate the need for invasive surgical intervention.
Supporting material provided by the authors is posted with the online version of this article as a data supplement at jbjs.org (http://links.lww.com/JBJSREV/A480)
2. Altman RD, Bedi A, Karlsson J, Sancheti P, Schemitsch E. Product differences in intra-articular hyaluronic acids for osteoarthritis of the knee. Am J Sports Med. 2016 Aug;44(8):2158-65. Epub 2015 Nov 17.
3. Jevsevar D, Donnelly P, Brown GA, Cummins DS. Viscosupplementation for osteoarthritis of the knee: a systematic review of the evidence. J Bone Joint Surg Am. 2015 Dec 16;97(24):2047-60.
4. Trojian TH, Concoff AL, Joy SM, Hatzenbuehler JR, Saulsberry WJ, Coleman CI. AMSSM scientific statement concerning viscosupplementation injections for knee osteoarthritis: importance for individual patient outcomes. Br J Sports Med. 2016 Jan;50(2):84-92.
5. Loew L, Brosseau L, Wells GA, Tugwell P, Kenny GP, Reid R, Maetzel A, Huijbregts M, McCullough C, De Angelis G, Coyle D; Ottawa Panel. Ottawa Panel evidence-based clinical practice guidelines for aerobic walking programs in the management of osteoarthritis. Arch Phys Med Rehabil. 2012 Jul;93(7):1269-85. Epub 2012 Mar 12.
6. Wang SY, Olson-Kellogg B, Shamliyan TA, Choi JY, Ramakrishnan R, Kane RL. Physical therapy interventions for knee pain secondary to osteoarthritis: a systematic review. Ann Intern Med. 2012 Nov 6;157(9):632-44.
7. Jevsevar DS, Brown GA, Jones DL, et al. The American Academy of Orthopaedic Surgeons evidence-based guideline on: treatment of osteoarthritis of the knee, 2nd edition. J Bone Joint Surg Am. 2013;95(20):1885-1886.
8. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H, Deyo RA, Gabriel S, Hirsch R, Hochberg MC, Hunder GG, Jordan JM, Katz JN, Kremers HM, Wolfe F; National Arthritis Data Workgroup. Estimates of the prevalence of arthritis and other rheumatic conditions in the United States. Part II. Arthritis Rheum. 2008 Jan;58(1):26-35.
9. Losina E, Weinstein AM, Reichmann WM, Burbine SA, Solomon DH, Daigle ME, Rome BN, Chen SP, Hunter DJ, Suter LG, Jordan JM, Katz JN. Lifetime risk and age at diagnosis of symptomatic knee osteoarthritis in the US. Arthritis Care Res (Hoboken). 2013 May;65(5):703-11.
10. Puig-Junoy J, Ruiz Zamora A. Socio-economic costs of osteoarthritis: a systematic review of cost-of-illness studies. Semin Arthritis Rheum. 2015 Apr;44(5):531-41. Epub 2014 Oct 31.
11. McCarthy CJ, Callaghan MJ, Oldham JA. Pulsed electromagnetic energy treatment offers no clinical benefit in reducing the pain of knee osteoarthritis: a systematic review. BMC Musculoskelet Disord. 2006 Jun 15;7:51.
12. Health Quality O, Medical Advisory Secretariat. Intra-articular viscosupplementation with hylan G-F 20 to treat osteoarthritis of the knee: an evidence-based analysis. Ont Health Technol Assess Ser. 2005;5(10):1-66. Epub 2005 Jun 1.
13. Jordan KM, Arden NK, Doherty M, Bannwarth B, Bijlsma JW, Dieppe P, Gunther K, Hauselmann H, Herrero-Beaumont G, Kaklamanis P, Lohmander S, Leeb B, Lequesne M, Mazieres B, Martin-Mola E, Pavelka K, Pendleton A, Punzi L, Serni U, Swoboda B, Verbruggen G, Zimmerman-Gorska I, Dougados M; Standing Committee for International Clinical Studies Including Therapeutic Trials ESCISIT. EULAR Recommendations 2003: an evidence based approach to the management of knee osteoarthritis: Report of a Task Force of the Standing Committee for International Clinical Studies Including Therapeutic Trials (ESCISIT). Ann Rheum Dis. 2003 Dec;62(12):1145-55.
14. Wang ZY, Shi SY, Li SJ, Chen F, Chen H, Lin HZ, Lin JM. Efficacy and safety of duloxetine on osteoarthritis knee pain: a meta-analysis of randomized controlled trials. Pain Med. 2015 Jul;16(7):1373-85. Epub 2015 Jun 5.
15. Altman RD, Schemitsch E, Bedi A. Assessment of clinical practice guideline methodology for the treatment of knee osteoarthritis with intra-articular hyaluronic acid. Semin Arthritis Rheum. 2015 Oct;45(2):132-9. Epub 2015 May 7.
16. Moyer RF, Birmingham TB, Bryant DM, Giffin JR, Marriott KA, Leitch KM. Valgus bracing for knee osteoarthritis: a meta-analysis of randomized trials. Arthritis Care Res (Hoboken). 2015 Apr;67(4):493-501.
17. Zeng C, Wei J, Li H, Wang YL, Xie DX, Yang T, Gao SG, Li YS, Luo W, Lei GH. Effectiveness and safety of glucosamine, chondroitin, the two in combination, or celecoxib in the treatment of osteoarthritis of the knee. Sci Rep. 2015 Nov 18;5:16827.
18. Jaeschke R, Singer J, Guyatt GH. Measurement of health status. Ascertaining the minimal clinically important difference. Control Clin Trials. 1989 Dec;10(4):407-15.
19. Devji T, Guyatt GH, Lytvyn L, Brignardello-Petersen R, Foroutan F, Sadeghirad B, Buchbinder R, Poolman RW, Harris IA, Carrasco-Labra A, Siemieniuk RAC, Vandvik PO. Application of minimal important differences in degenerative knee disease outcomes: a systematic review and case study to inform BMJ
Rapid Recommendations. BMJ Open. 2017 May 11;7(5):e015587.
20. Neumann I, Santesso N, Akl EA, Rind DM, Vandvik PO, Alonso-Coello P, Agoritsas T, Mustafa RA, Alexander PE, Schünemann H, Guyatt GH. A guide for health professionals to interpret and use recommendations in guidelines developed with the GRADE approach. J Clin Epidemiol. 2016 Apr;72:45-55. Epub 2016 Jan 6.
21. Guyatt G, Gutterman D, Baumann MH, Addrizzo-Harris D, Hylek EM, Phillips B, Raskob G, Lewis SZ, Schünemann H. Grading strength of recommendations and quality of evidence in clinical guidelines: report from an American College of Chest Physicians task force. Chest. 2006 Jan;129(1):174-81.
22. Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, Brozek J, Norris S, Falck-Ytter Y, Glasziou P, DeBeer H, Jaeschke R, Rind D, Meerpohl J, Dahm P, Schünemann HJ. GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol. 2011 Apr;64(4):383-94. Epub 2010 Dec 31.
23. Alhazzani W, Guyatt G. An overview of the GRADE approach and a peek at the future. Med J Aust. 2018 Sep 1;209(7):291-2.
24. Salanti G, Del Giovane C, Chaimani A, Caldwell DM, Higgins JP. Evaluating the quality of evidence from a network meta-analysis. PLoS One. 2014 Jul 3;9(7):e99682.
25. Juhl C, Lund H, Roos EM, Zhang W, Christensen R. A hierarchy of patient-reported outcomes for meta-analysis of knee osteoarthritis trials: empirical evidence from a survey of high impact journals. Arthritis. 2012;2012:136245. Epub 2012 Jun 26.
26. Higgins JPTGS. Cochrane handbook for systematic reviews of interventions version 5.1.0. The Cochrane Collaboration; Cambridge, 2011.
27. Stark PB, Tenorio L. A primer of frequentist and Bayesian inference in inverse problems. In: Large-scale inverse problems and quantification of uncertainty. John Wiley & Sons; New Jersey, 2010. p 9-32.
28. Nelson AE, Allen KD, Golightly YM, Goode AP, Jordan JM. A systematic review of recommendations and guidelines for the management of osteoarthritis: the Chronic Osteoarthritis Management Initiative of the U.S. Bone and Joint Initiative. Semin Arthritis Rheum. 2014 Jun;43(6):701-12. Epub 2013 Dec 4.
29. Arrich J, Piribauer F, Mad P, Schmid D, Klaushofer K, Müllner M. Intra-articular hyaluronic acid for the treatment of osteoarthritis of the knee: systematic review and meta-analysis. CMAJ. 2005 Apr 12;172(8):1039-43.
30. Bannuru RR, McAlindon TE, Sullivan MC, Wong JB, Kent DM, Schmid CH. Effectiveness and implications of alternative placebo treatments: a systematic review and network meta-analysis of osteoarthritis trials. Ann Intern Med. 2015 Sep 1;163(5):365-72.
31. Bannuru RR, Natov NS, Dasi UR, Schmid CH, McAlindon TE. Therapeutic trajectory following intra-articular hyaluronic acid injection in knee osteoarthritis—meta-analysis. Osteoarthritis Cartilage. 2011 Jun;19(6):611-9. Epub 2011 Apr 9.
32. Bannuru RR, Schmid CH, Kent DM, Vaysbrot EE, Wong JB, McAlindon TE. Comparative effectiveness of pharmacologic interventions for knee osteoarthritis: a systematic review and network meta-analysis. Ann Intern Med. 2015 Jan 6;162(1):46-54.
33. Bjordal JM, Johnson MI, Lopes-Martins RA, Bogen B, Chow R, Ljunggren AE. Short-term efficacy of physical interventions in osteoarthritic knee pain. A systematic review and meta-analysis of randomised placebo-controlled trials. BMC Musculoskelet Disord. 2007 Jun 22;8:51.
34. Bjordal JM, Ljunggren AE, Klovning A, Slørdal L. Non-steroidal anti-inflammatory drugs, including cyclo-oxygenase-2 inhibitors, in osteoarthritic knee pain: meta-analysis of randomised placebo controlled trials. BMJ. 2004 Dec 4;329(7478):1317. Epub 2004 Nov 23.
35. Brien S, Prescott P, Lewith G. Meta-analysis of the related nutritional supplements dimethyl sulfoxide and methylsulfonylmethane in the treatment of osteoarthritis of the knee. Evid Based Complement Alternat Med. 2011;2011:528403. Epub 2011 Feb 17.
36. Christensen R, Bartels EM, Astrup A, Bliddal H. Effect of weight reduction in obese patients diagnosed with knee osteoarthritis: a systematic review and meta-analysis. Ann Rheum Dis. 2007 Apr;66(4):433-9. Epub 2007 Jan 4.
37. Colen S, van den Bekerom MP, Mulier M, Haverkamp D. Hyaluronic acid in the treatment of knee osteoarthritis: a systematic review and meta-analysis with emphasis on the efficacy of different products. BioDrugs. 2012 Aug 1;26(4):257-68.
38. Corbett MS, Rice SJ, Madurasinghe V, Slack R, Fayter DA, Harden M, Sutton AJ, Macpherson H, Woolacott NF. Acupuncture and other physical treatments for the relief of pain due to osteoarthritis of the knee: network meta-analysis. Osteoarthritis Cartilage. 2013 Sep;21(9):1290-8.
39. Fransen M, McConnell S, Harmer AR, Van der Esch M, Simic M, Bennell KL. Exercise for osteoarthritis of the knee. Cochrane Database Syst Rev. 2015 Jan 9;1:CD004376.
40. Juhl C, Christensen R, Roos EM, Zhang W, Lund H. 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 Rheumatol. 2014 Mar;66(3):622-36.
41. Jüni P, Hari R, Rutjes AW, Fischer R, Silletta MG, Reichenbach S, da Costa BR. Intra-articular corticosteroid for knee osteoarthritis. Cochrane Database Syst Rev. 2015 Oct 22;10:CD005328.
42. Manheimer E, Cheng K, Linde K, Lao L, Yoo J, Wieland S, van der Windt DA, Berman BM, Bouter LM. Acupuncture for peripheral joint osteoarthritis. Cochrane Database Syst Rev. 2010 Jan 20;1:CD001977.
43. Manheimer E, Linde K, Lao L, Bouter LM, Berman BM. Meta-analysis: acupuncture for osteoarthritis of the knee. Ann Intern Med. 2007 Jun 19;146(12):868-77.
44. Richette P, Chevalier X, Ea HK, Eymard F, Henrotin Y, Ornetti P, Sellam J, Cucherat M, Marty M. Hyaluronan for knee osteoarthritis: an updated meta-analysis of trials with low risk of bias. RMD Open. 2015 May 14;1(1):e000071.
45. Rutjes AW, Jüni P, da Costa BR, Trelle S, Nüesch E, Reichenbach S. Viscosupplementation for osteoarthritis of the knee: a systematic review and meta-analysis. Ann Intern Med. 2012 Aug 7;157(3):180-91.
46. Rutjes AW, Nüesch E, Sterchi R, Jüni P. Therapeutic ultrasound for osteoarthritis of the knee or hip. Cochrane Database Syst Rev. 2010 Jan 20;1:CD003132.
47. Rutjes AW, Nüesch E, Sterchi R, Kalichman L, Hendriks E, Osiri M, Brosseau L, Reichenbach S, Jüni P. Transcutaneous electrostimulation for osteoarthritis of the knee. Cochrane Database Syst Rev. 2009 Oct 7;4:CD002823.
48. Schneider H, Maheu E, Cucherat M. Symptom-modifying effect of chondroitin sulfate in knee osteoarthritis: a meta-analysis of randomized placebo-controlled trials performed with Structum(®). Open Rheumatol J. 2012;6:183-9. Epub 2012 Jul 25.
49. Strand V, McIntyre LF, Beach WR, Miller LE, Block JE. Safety and efficacy of US-approved viscosupplements for knee osteoarthritis: a systematic review and meta-analysis of randomized, saline-controlled trials. J Pain Res. 2015 May 7;8:217-28.
50. Warden SJ, Hinman RS, Watson MA Jr, Avin KG, Bialocerkowski AE, Crossley KM. Patellar taping and bracing for the treatment of chronic knee pain: a systematic review and meta-analysis. Arthritis Rheum. 2008 Jan 15;59(1):73-83.
51. Jevsevar DS, Shores PB, Mullen K, Schulte DM, Brown GA, Cummins DS. Mixed treatment comparisons for nonsurgical treatment of knee osteoarthritis: a network meta-analysis. J Am Acad Orthop Surg. 2018 May 1;26(9):325-36.
52. Xu Z, Luo J, Huang X, Wang B, Zhang J, Zhou A. Efficacy of platelet-rich plasma in pain and self-report function in knee osteoarthritis: a best-evidence synthesis. Am J Phys Med Rehabil. 2017 Nov;96(11):793-800.
53. Dai WL, Zhou AG, Zhang H, Zhang J. Efficacy of platelet-rich plasma in the treatment of knee osteoarthritis: a meta-analysis of randomized controlled trials. Arthroscopy. 2017 Mar;33(3):659-670.e1. Epub 2016 Dec 22.
54. Bellamy N, Kirwan J, Boers M, Brooks P, Strand V, Tugwell P, Altman R, Brandt K, Dougados M, Lequesne M. Recommendations for a core set of outcome measures for future Phase III clinical trials in knee, hip, and hand osteoarthritis. Consensus development at OMERACT III. J Rheumatol. 1997 Apr;24(4):799-802.
55. Pham T, van der Heijde D, Altman RD, Anderson JJ, Bellamy N, Hochberg M, Simon L, Strand V, Woodworth T, Dougados M. OMERACT-OARSI initiative: Osteoarthritis Research Society International set of responder criteria for osteoarthritis clinical trials revisited. Osteoarthritis Cartilage. 2004 May;12(5):389-99.
56. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale: Lawrence Erlbaum Associates; 1988.
57. Norman GR, Sloan JA, Wyrwich KW. Interpretation of changes in health-related quality of life: the remarkable universality of half a standard deviation. Med Care. 2003 May;41(5):582-92.
58. Crossley KM, Bennell KL, Cowan SM, Green S. Analysis of outcome measures for persons with patellofemoral pain: which are reliable and valid? Arch Phys Med Rehabil. 2004 May;85(5):815-22.
59. Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, Haythornthwaite JA, Jensen MP, Kerns RD, Ader DN, Brandenburg N, Burke LB, Cella D, Chandler J, Cowan P, Dimitrova R, Dionne R, Hertz S, Jadad AR, Katz NP, Kehlet H, Kramer LD, Manning DC, McCormick C, McDermott MP, McQuay HJ, Patel S, Porter L, Quessy S, Rappaport BA, Rauschkolb C, Revicki DA, Rothman M, Schmader KE, Stacey BR, Stauffer JW, von Stein T, White RE, Witter J, Zavisic S. Interpreting the clinical importance of treatment outcomes in chronic pain clinical trials: IMMPACT recommendations. J Pain. 2008 Feb;9(2):105-21. Epub 2007 Dec 11.
60. Ehrich EW, Davies GM, Watson DJ, Bolognese JA, Seidenberg BC, Bellamy N. Minimal perceptible clinical improvement with the Western Ontario and McMaster Universities Osteoarthritis Index questionnaire and global assessments in patients with osteoarthritis. J Rheumatol. 2000 Nov;27(11):2635-41.
61. Salaffi F, Stancati A, Silvestri CA, Ciapetti A, Grassi W. Minimal clinically important changes in chronic musculoskeletal pain intensity measured on a numerical rating scale. Eur J Pain. 2004 Aug;8(4):283-91.
62. Tashjian RZ, Deloach J, Porucznik CA, Powell AP. Minimal clinically important differences (MCID) and patient acceptable symptomatic state (PASS) for visual analog scales (VAS) measuring pain in patients treated for rotator cuff disease. J Shoulder Elbow Surg. 2009 Nov-Dec;18(6):927-32. Epub 2009 Jun 16.
63. Todd KH, Funk KG, Funk JP, Bonacci R. Clinical significance of reported changes in pain severity. Ann Emerg Med. 1996 Apr;27(4):485-9.
64. Angst F, Aeschlimann A, Michel BA, Stucki G. Minimal clinically important rehabilitation effects in patients with osteoarthritis of the lower extremities. J Rheumatol. 2002 Jan;29(1):131-8.
65. Angst F, Aeschlimann A, Stucki G. Smallest detectable and minimal clinically important differences of rehabilitation intervention with their implications for required sample sizes using WOMAC and SF-36 quality of life measurement instruments in patients with osteoarthritis of the lower extremities. Arthritis Rheum. 2001 Aug;45(4):384-91.
66. Tubach F, Ravaud P, Baron G, Falissard B, Logeart I, Bellamy N, Bombardier C, Felson D, Hochberg M, van der Heijde D, Dougados M. Evaluation of clinically relevant changes in patient reported outcomes in knee and hip osteoarthritis: the minimal clinically important improvement. Ann Rheum Dis. 2005 Jan;64(1):29-33. Epub 2004 Jun 18.
67. Bannuru RR, Vaysbrot EE, McIntyre LF. Did the American Academy of Orthopaedic Surgeons osteoarthritis guidelines miss the mark? Arthroscopy. 2014 Jan;30(1):86-9.
68. King MT. A point of minimal important difference (MID): a critique of terminology and methods. Expert Rev Pharmacoecon Outcomes Res. 2011 Apr;11(2):171-84.
69. Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM. Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain. 2001 Nov;94(2):149-58.
70. Gerlinger C, Schumacher U, Faustmann T, Colligs A, Schmitz H, Seitz C. Defining a minimal clinically important difference for endometriosis-associated pelvic pain measured on a visual analog scale: analyses of two placebo-controlled, randomized trials. Health Qual Life Outcomes. 2010 Nov 24;8:138.
71. Lee JS, Hobden E, Stiell IG, Wells GA. Clinically important change in the visual analog scale after adequate pain control. Acad Emerg Med. 2003 Oct;10(10):1128-30.
72. Rowbotham MC. What is a “clinically meaningful” reduction in pain? Pain. 2001 Nov;94(2):131-2.
73. Wells GA, Tugwell P, Kraag GR, Baker PR, Groh J, Redelmeier DA. Minimum important difference between patients with rheumatoid arthritis: the patient’s perspective. J Rheumatol. 1993 Mar;20(3):557-60.
74. Brozek JL, Guyatt GH, Schünemann HJ. How a well-grounded minimal important difference can enhance transparency of labelling claims and improve interpretation of a patient reported outcome measure. Health Qual Life Outcomes. 2006 Sep 27;4:69.