Management of Osteoarthritis of the Knee (Non-Arthroplasty) (3rd Edition)
Endorsed by: AAHKS, APTA
Lateral Wedge Insoles (Recommendation Against)
Lateral wedge insoles are not recommended for patients with knee osteoarthritis.
Strong Recommendation Strong Recommendation

Although lateral heel wedges had historical support for their use in knee arthritis, contemporary studies have not shown a reliable improvement in pain relief and no contemporary studies have shown sufficient functional improvement for patients suffering from knee arthritis to recommend using lateral wedge insoles. Lateral heel wedges can be prepared as an insert for the heel alone, or included in the heel of an independent arch support (i.e. lateral heel wedge arch support of LWAS) or built into shoe itself (as was used by Hinman et al 2016). In the arthritic knee, medial compartment compression forces are commonly increased, especially in the knee with varus tibiofemoral alignment. The knee adduction moment can be calculated by gait analysis. The lateral wedge is thought to change the knee adduction moment thus relieving medial compartment pressure, hence relieving arthritic pain.

Our literature review screening culled several papers for analysis. Baker and Goggins 2007 was a high quality study finding no important differences between insole and wedged insole. 90 patients were randomized to one treatment for 6 weeks followed by a 4-week washout period and then the opposite treatment. There were no major differences in pain during either phase of the study. More musculoskeletal symptoms and more blisters occurred with neutral insoles. No patient falls were attributed to the treatment alternatives.

Felson and Parke 2019 prescreened patients to eliminate those with patellofemoral OA and biomechanical non-responders. Lateral wedge insoles reduced knee pain, but the effect of treatment was small and was considered likely of clinical significance in only a minority of patients. 21 of 83 of patients did not show sufficient biomechanical correction. Only 28% of patients in the active phase of treatment had minimally important improvement whereas 22% of patients wearing neutral insoles reached the same level of improvement. 2 patients stopped treatment while wearing lateral wedge insoles (calf pain at night and increased knee pain) and 2 stopped while wearing neutral insoles (toe blister and increased knee pain). They also looked at volume of arthritic bone marrow lesions (BML) found by MRI and saw no significant difference in BML change between study and control groups.

In Bennell 2011, 89 patients with mild to moderate knee arthritis completed follow up with lateral insoles worn daily for 12 months. 90 patients completed follow up as the control group wearing neutral insoles. Pain relief after 12 months showed no significant difference between the groups. 

In Hsieh 2016, 90 patients with Kellgren-Lawrence Grade 2 or higher radiographic changes were randomized to either a rigid insole with lateral wedge arch support (LWAS) or a soft insole with lateral wedge. Dropout rate was 20% with rigid and 15.6% with soft insoles over the 3-month long study. They concluded that patients using the soft insole LWAS had improved pain and function. However, their primary data suggests better walking time and speed going up and down stairs with rigid LWAS. Furthermore, pain was improved with soft LWAS only at the 3-month mark. Authors suggested longer term follow up for soft insoles.

Hinman 2016 evaluated an unloading shoe with stiff lateral midsole and 5 degree lateral wedge insole in comparison to a standard walking shoe. 164 patients were enrolled with 96% retention during the 6 months study. 83 patients received the unloading shoes and the control shoes. 14 of 83 stopped wearing the unloading shoes for various reasons and 8 of 81 stopped wearing the control shoes. 160 completed primary outcome measures at 6 months. There was no significant difference between groups with regard to pain or function, although both groups did show improvement.

20% of participants with the study shoes reported ankle and foot pain whereas 9% of control shoe participants did so. There was no difference in reason to discontinue treatment (unloading shoe 4% versus 2% control). Other reported adverse events were back pain, hip pain, knee pain, knee stiffness/swelling and shin/calf pain. 2 of 83 experimental group patients reported increase in knee pain with the unloading shoe and 2 or 81 control patients reported that the conventional shoe did not relieve knee pain.

Toda 2004 followed 84 knee female arthritis patients were followed for one month wearing either a hard rubber insole or urethane insole secured to the foot with a subtalar strap used for ankle sprains. 12 mm lateral wedge was manufactured for both. 17 of the 42 rubber insole patients had complications (foot pain in 8; popliteal pain in 6; low back pain in 3) versus 8 of the 42 using urethane insoles (popliteal pain in 4; foot pain in 3; low back pain in 1). All patients improved by the Lequesne Index with the urethane group achieving statistically significant improvement.

Niazi 2014 was a comparison of off-loading knee brace versus lateral wedge insole. 120 patients with both radiographic medial compartment arthritis and gene varum were randomized to either knee brace or lateral wedge insole. Pain improvement with the knee brace group was statistically significant compared to the lateral wedge insole, but clinically minor (VAS 3.97 in the study group compared to 4.53 in controls).

In Hatef 2013, 118 of 150 patients completed the 2-month long study (101 women and 17 men). Half were given LWAS and the control group wore neutral insoles. Patient compliance was much worse in the LWAS group. They noted statistically significant decline in knee pain and EKFS in women in the LWAS, but not men. Overall, there was improvement in the LWAS group. There was a much higher noncompliance rate in the LWAS than with the neutral insoles with 29 of 57 patients stopping use of the insoles by weeks 5 to 6 of an 8-week study.

We identified one potential study within our literature which addressed the question of special shoe versus a conventional shoe. Nigg 2006 evaluated a training shoe which purports to convert a flat hard surface into “natural uneven ground”, thus prompting increased muscle activity in the lower extremity. The control shoe was a standard walking shoe. 58 patients were enrolled in the study group and 67 in the control group. Both groups had one patient drop out (cumbersome shoes in the study group and increased knee pain with the control shoe). Pain with walking was improved at 12 weeks in both groups, without between groups difference. The study shoe showed increased pain relief at 3, 6 and 12 weeks. The control shoe showed increased pain relief at 3 and 12 weeks. They also reported improved balance from baseline in the study shoe at 12 weeks which was not statistically significant.

Benefits/Harms of Implementation
Although lower extremity pain might be increased with either neutral insole or LWAS, there is no significant harm to the patient trying either option.

Cost Effectiveness/Resource Utilization
A standard insole or conventional walking shoe provided equivalent improvement in pain compared to lateral wedge arch support. Although the lateral wedge modifications are more expensive, the increased cost is not prohibitive and a patient attempting self-treatment could discontinue at any time with little loss of time effort or out of pocket cost.

Acceptability
Insoles are already commercially available and have long history of orthopaedic use. They can fit or be adapted to a variety of shoes, commonly already owned by the patient. Some studies reviewed (Hsieh 2016 and Hatef 2013) also described a relatively high dropout or non-compliance rate. While this does not imply frank harm to the patient, such data do suggest distinct potential for patient dissatisfaction with the treatment.

Feasibility
Although there are no immediately obvious limits to feasibility of implementing neutral insoles or LWAS for knee pain, Felson 2019 confirmed previous studies by showing 25% of their potential subjects did not correct knee adduction when using the lateral wedge.

Future Research
We did not identify any studies of sufficient size comparing a “walking shoe” versus a random or conventional shoe. The studies reviewed by the workgroup suggest that a walking shoe or a soft neutral insole might provide some degree of pain relief. Future studies addressing this treatment option could confirm some type of specialized yet commercially available shoe as beneficial to the patient with knee arthritis.

Canes
Canes could be used to improve pain and function in patients with knee osteoarthritis.
Moderate Recommendation Moderate Recommendation

Canes have been used since antiquity for lower extremity orthopaedic disorders. With that in mind, only a small number of modern studies have formally investigated use of a cane for knee osteoarthritis. Our literature review found one high quality study (Jones 2012) showing support for use regarding moderate pain relief and another study (Van Ginckel 2019) of moderate quality showing no major improvement in pain.

Jones 2012 performed a comprehensive study of use of a cane for knee osteoarthritis. Their primary outcome was pain, but they also looked at function, general health, consumption of NSAIDs and energy expenditure. 64 patients were selected out a total of 323 patients nominated from a rheumatology clinic. The majority of possible patients (168 out of 323) refused to participate in the study. 32 patients were randomized to use of a cane for 60 days (EG, experimental group). The cane was cut to appropriate height and they received instructions on its use. The 32 control patients (CG) were instructed to maintain normal lifestyle and not to use auxiliary gait devices. At 30 and 60 days the EG patients had less pain compared to controls. The greatest improvement was in the VAS (10 cm scale): EG averaging 3.84 cm and CG 5.95 cm at 60 days). The Lequesne scale (0-24) difference was only 2.53 (CG 15.09 and EG 12.56 at 60 days). At 60 days, the study group consumed fewer NSAIDs than control.

Van Ginckel 2019 evaluated use of a cane in patients with medial compartment knee osteoarthritis and bone marrow lesions (BML) on MRI. The primary intent of their study was to identify an effect on the size of BML by using a cane. Out of 1989 potential patients (contacted by phone or online) 231 were considered eligible for radiographic screening and of those, only 79 showed arthritic changes on plain films and BML on MRI and chose to continue with the study. 40 patients were assigned to use a cane whenever walking for the next 12 weeks. 39 control patients were instructed to maintain their usual lifestyle without any gait aids. Only one patient in the control group was lost to follow up. After 3 months there was no significant improvement in BML size. Secondary information was obtained relative to clinical characteristics. There was no significant difference between the two groups with regards to knee pain (WOMAC scale) or quality of life (AQoL 6-D scale) although there was improvement in global knee pain in the group using the cane.

Benefits/Harms of Implementation
Jones 2012 addressed the increased oxygen demand for ambulation using a cane. Initially all patients had decreased ambulation distance with a cane and increased heart rate and increased oxygen consumption.bAfter 60 days patients who had been using the cane (EG) were able to ambulate a similar distance with or without the cane suggesting an ability to adapt to the cane. Also, EG patients using the cane had more normal oxygen consumption after walking again suggesting physiologic adaptation to the cane with time.

Outcome Importance
Our review shows moderate confirmation that canes can relieve the pain of knee arthritis and improve function in those patients.

Cost Effectiveness/Resource Utilization
Use of a cane is a common low-cost treatment which is readily available with many options and typically covered by third party payers.

Acceptability
Van Ginckel noted “patient vanity” as a common reason for non-compliance. 

Feasibility
Canes have been used since antiquity although comparison studies are only being produced in the last generation. There is little downside to extended use of the cane as patients appear to adapt to the increased oxygen consumption demands.

Future Research
More studies would be beneficial to the knowledge base confirming value to patients with knee osteoarthritis.

Braces
Brace treatment could be used to improve function, pain and quality of life in patients with knee osteoarthritis.
Moderate Recommendation Moderate Recommendation

The Braces recommendation has been downgraded one level because of heterogeneity.

Four high, four moderate and two low quality studies were included for review, following the application of exclusion criteria by committee (Brouwer 2006; Kirkley 1999; Callaghan 2015; Van Raaij 2010; Thoumie 2018; Hjartarson 2018; Petersen 2018; Niazi 2014; Hungerford 2013; Yu 2016). Three prospective randomized controlled trials compared bracing to control groups for treatment of symptomatic OA of the knee joint and found statistically significant and clinically meaningful improvement in patient symptoms related to symptomatic OA of the knee (Callaghan 2015; Thoumie 2018; Brouwer 2006b). Kirkley et al compared outcomes between valgus offloading brace, neoprene sleeve, and non-brace control for symptomatic OA and Varus alignment. This study reported statistically significant improvement in disease specific quality of life and function in both study groups (Kirkley 1999). The study by Brouwer et al., comparing valgus bracing to a non-braced control, and reported no significant difference in functional assessment, PRO or pain; however, clinically significant improvements were noted in walking distance (1.25km[0.15,2.35]) for the brace group. Subgroup analysis demonstrated greater positive effect of bracing in patients with varus alignment and more severe symptoms. Callaghan et al examined the effects of bracing for patella-femoral OA and found significant improvement from baseline VAS and KOOS pain scores. Finally, Hjartarson et al examined outcomes of bracing vs. placebo by removing valgus tension straps from the control group brace. In their study they reported statistically significant and clinically meaningful improvements in KOOS sub scores: symptoms, ADL, sports and recreation, and quality of life.

Benefits/Harms of Implementation
Braces can provide significant pain relief and improved function to patients that have unicompartmental knee osteoarthritis. Braces can provide a subjective feeling of more normal tibiofemoral kinematics, preventing excessive strain on the affected compartment while protecting against preexistent concomitant  meniscal and chondral injuries. There is also a theoretical benefit of increased confidence in the knee during the different activities by providing a sense of security to the knee. There are almost no harms in trialing a brace besides some skin irritation, or the brace being uncomfortable.

Future Research
Future high-quality studies that are well powered are required to assess the real efficacy of unloader braces vs knee sleeves in a population with similar mechanical axis with a similar degree of OA.

Oral/Dietary Supplements
The following supplements may be helpful in reducing pain and improving function for patients with mild to moderate knee osteoarthritis; however, the evidence is inconsistent/limited and additional research clarifying the efficacy of each supplement is needed. a) Turmeric b) Ginger extract c) Glucosamine d) Chondroitin e) Vitamin D
Limited Recommendation Limited Recommendation

The Oral/ Dietary Supplements recommendation has been downgraded two levels because of inconsistency and need for additional clarity of efficacy.

The majority of 6 high quality trials (Reginster 2001, Cibere 2004, McAlindon 2004, Clegg 2006, Herrero-Beaumont 2007, Fransen 2015), the majority of 6 moderate quality trials (Noack 1994, Houpt 1999, Rindone 2000, Pavelka 2002, Giordano 2009, and Shahine 2014), and 1 low quality study that met inclusion criteria showed either improvement or no change in patient outcomes for those with osteoarthritis of the knee when taking glucosamine versus control.

The majority of 8 high quality trials (Fransen 2015, Clegg 2006, Uebelhart 2004, Reginster 2017, Morita 2018, Zegels 2013, Kahan 2009, Rondanelli 2019) and 6 moderate quality trials (Mazieres 2007, Moller 2010, Rondanelli 2019, Bourgeois 1998, Mazieres 2001, Bucsi 1998) that met inclusion criteria showed either improvement or no change in patient outcomes for those with osteoarthritis of the knee when taking chondroitin.

One high quality study (Srivastava 2016) that met inclusion criteria showed that Turmeric extract could be used over control to improve adverse events, function, and pain in patient with osteoarthritis of the knee.

One high study (Zakeri 2011) and one moderate quality study (Altman 2001) that met inclusion criteria showed that ginger extract may be used to improve pain in patients with osteoarthritis of the knee. However, there was no significant difference in function between ginger extract and control.

Three high quality studies (McAlindon 2013, Sanghi 2013, and Jin 2016) and 1 moderate quality study (Arden 2016) that met inclusion criteria showed either improvement or no significant difference in patient outcomes for those with osteoarthritis of the knee between Vitamin D and control.

Benefits/Harms of Implementation
The United States Food and Drug Administration does not hold dietary supplements to the same standards as prescription medication. As a result, variability can exist between producers of dietary supplements. The patient’s medications should be evaluated for potential drug-supplement interactions prior to initiating any dietary supplement.

Outcome Importance
Improvement in the predictability of the treatment effect for specific dietary supplements will either reduce an unnecessary out-of-pocket expense or provide patients with a relatively safe method for reducing pain and improving function in the treatment of mild to moderate osteoarthritis of the knee. Knowing which supplements have evidence-based support for effective treatment will reduce unnecessary out-of-pocket expenses and provide patients with a safe method for reducing pain and improving function in the treatment of mild to moderate osteoarthritis of the knee.

Cost Effectiveness/Resource Utilization
Dietary supplements are typically an out-of-pocket expense not covered by medical insurance. As a result, patients must individually consider the expense associated with utilization of dietary supplements. We did not have the analysis to comment on the cost-benefit ratio of dietary supplements compared to other over the counter or prescription medications.

Acceptability
The limited evidence supporting dietary supplements and out-of-pocket expense associated with their use will limit broader acceptance. For patients without a medication interaction, dietary supplements pose limited physical harm as it is typically difficult to consume a toxic quantity. Therefore, it is believed to be an acceptable means to potentially achieve benefit in the treatment of mild to moderate osteoarthritis of the knee.

Feasibility
Dietary supplements are widely available to patients. The associated out-of-pocket expense is the primary barrier to access.

Future Research
The most important future research will need to provide multiple investigations demonstrating reproducible results illustrating the effectiveness of dietary supplements. Additionally, future research is necessary to determine the appropriate dose, frequency, and duration of treatment with dietary supplements.

Topical Treatments
Topical NSAIDs should be used to improve function and quality of life for treatment of osteoarthritis of the knee, when not contraindicated.
Strong Recommendation Strong Recommendation

Twelve high quality studies (Baer 2005, Roth 2004, Conaghan 2013, Simon 2009, Kneer 2013, Rother 2013, Bookman 2004, Wadsworth 2016, Sandelin 1997, Dehghan 2019, Dehghan 2020, Rother 2007) and two moderate quality studies (Barthel 2009, Ottillinger 2001) show that topical NSAIDs could result in improved function and quality of life over placebo gel. However, inconsistent evidence suggests no significant difference in pain and adverse events between topical NSAIDs and control.

Benefits/Harms of Implementation
Topical NSAIDs are now available over the counter for patients to buy and use. Topical NSAIDs should be used with caution in certain health conditions such as stage 4-5 chronic kidney disease, coronary artery disease, and congestive heart failure. There is a risk for skin sensitivity.

Outcome Importance
Improved function and quality of life improvement are the important outcomes from Topical NSAID use.

Cost Effectiveness/Resource Utilization
Since this will now be available over the counter, prescription coverage will be less. This could pose as a barrier to those without health savings accounts or without insurance. Also, patients might not be willing to apply a topical gel multiple times a day.

Acceptability
The use of a topical gel for osteoarthritis of the knee should be an acceptable method of treatment if cost is not prohibitive and the patient does not have any skin irritation from the gel.

Feasibility
Topical NSAIDs will be available at pharmacies over the counter. The main barrier will be cost and having to apply the gel multiple times a day.

Future Research
Future research might be directed at determining if continued topical NSAID use is required to sustain
benefits or if benefits continue after usage for a defined period of time.

Supervised Exercise
Supervised exercise, unsupervised exercise, and/or aquatic exercise are recommended over no exercise to improve pain and function for treatment of knee osteoarthritis.
Strong Recommendation Strong Recommendation

Ten high quality studies were reviewed that compared a supervised exercise program to a non-exercise control (e.g., no treatment, heat only, education, usual primary care). (Christensen 2015,Holsgaard-Larsen 2018, Oliveira 2012, Willamson 2007,  de Rooij 2017, Imoto 2012, Topp 2002, Hu 2020, Kim 2013, Chen 2014) Seven of these studies found greater improvements in pain, function, or both pain and function over the non-exercise control group (Oliveira 2002, de Rooij 2017, Imoto 2012, Topp 2002, Hu 2020, Kim 2013, Chen 2014).

One high quality study and four moderate quality studies were reviewed that compared supervised exercise to a non-supervised exercise program (e.g., home program, internet-based program, exercise brochure). (McCarthy 2004, Allen 2018, Yilmaz 2019, Tunay 2010, Bennell 2014). Patients from both groups received benefit from the interventions but there were mixed results as to whether supervised exercise was superior to the non-supervised exercise programs. It appears that both supervised or non-supervised exercise programs can result in improved pain and function in people with knee osteoarthritis.

Four high quality studies and one moderate quality study were reviewed that compared aquatic exercise to either usual primary care, education, or self-management. (Kuptniratsaikul 2019, Rewald 2020, Waller 2017, Munukka 2020, Dias 2017.) Three high quality studies reported greater improvements in pain, function, or global ratings of improvement for the aquatic groups over the control groups. (Kuptniratsaikul 2019, Rewald 2020, Dias 2017) One high quality study reported increased leisure time activity for the aquatic group compared to the control. (Waller 2017) One moderate quality study compared aquatic exercise to land-based exercise. (Silva 2008) There was no difference in WOMAC pain and function scores reported between groups for this study, but the aquatic exercise group had less pain with walking compared to the land-based group. Although there may be some benefit from aquatic exercise, inconsistent results do not allow us to recommend aquatic exercise over land-based exercise at this time.

Several studies examined clinical outcomes for different modes of exercise in patients with knee osteoarthritis. Ebnezar 2012 reported some improvement in anxiety measures when comparing yoga to non-yoga exercise. (Ebnezar 2012) Other studies compared weightbearing to non-weightbearing exercise (Bennell 2020, Jan 2009), high versus low resistance training (Jan 2008), isokinetic, isometric, and isotonic exercise (Huang 2005), and leg versus hip exercise (Lun 2015) and did not find substantial differences in the mode of exercise. It appears that exercise is beneficial, but the mode of exercise may not matter as much as engaging in any exercise program.

Benefits/Harms of Implementation

Most patients can expect an improvement in pain and function with exercise. Patients may experience a temporary increase in knee pain or muscle soreness when engaging in an exercise program.

Cost Effectiveness/Resource Utilization

One study by Bove, et al, 2018 examined the cost-effectiveness of delivering physical therapy supervised exercise and manual therapy in booster (treatment sessions spread out periodically over 1 year) vs no-booster (treatment sessions delivered consecutively over 9 weeks) session approaches. It appeared that the booster delivery approach may be more cost-effective than the non-booster delivery approach.

Feasibility

Most exercise programs would be considered feasible. However, some patients may have difficulty with access to supervised exercise due to travel or co-pay concerns. Aquatic programs would not be feasible for patients who do not have access to a pool or walking tank.

Future Research

Identifying factors that could discriminate between people who would likely benefit from supervised programs vs independent exercise programs is an area of research that could improve clinical decision-making for prescribing exercise. More studies are also needed to examine delivery of exercise through telerehabilitation compared with in-person supervised programs. Studies should also examine differences in outcomes between varying modes of delivering telerehabilitation exercise programs. More research would also be beneficial in examining the role of booster session delivery of exercise programs.

Neuromuscular Training
Neuromuscular training (i.e. balance, agility, coordination) programs in combination with traditional exercise could be used to improve performance-based function and walking speed for treatment of knee osteoarthritis.
Moderate Recommendation Moderate Recommendation

The Neuromuscular training recommendation has been downgraded one level because of inconsistent evidence.

Three high quality studies (Fitzgerald 2011, Gomiero 2018, Apparao 2017) and two moderate quality studies (Bennell 2014 and Diracoglu 2005) comparing neuromuscular training combined with traditional strength and joint mobility exercise programs to strength and joint mobility exercise alone were reviewed. There were no differences in knee pain reported between groups in any of the studies. There were mixed results on function measures with two studies reporting greater improvements in self-reported function (Apparao 2017, Diracoglu 2005) and two studies reporting greater improvements in walking speed (Bennell 2014, Diracoglu 2005) for the neuromuscular training group.

Benefits/Harms of Implementation

Some patients can expect an improvement in function and walking speed with neuromuscular training. Some patients may experience a temporary increase in knee pain or muscle soreness when engaging in the exercise program.

Feasibility

Neuromuscular exercise programs would be considered feasible. However, some patients may have difficulty with access to supervised exercise due to travel or co-pay concerns.

Future Research

Given the mixed results between exercise programs with and without neuromuscular exercise programs, it would be beneficial for future research to identify factors that would discriminate between patients who would have better success with a neuromuscular exercise program versus those who would have better success with a traditional strength and joint mobility exercise program.

Self-Management
Self-management programs are recommended to improve pain and function for patients with knee osteoarthritis.
Strong Recommendation Strong Recommendation

Self-management programs refer to formalized training and education programs that are taught by both healthcare professionals and trained layperson instructors. They typically include several sessions over several weeks. These programs train people in several elements of self-management for osteoarthritis including medication compliance, pain management and pain coping strategies, joint protection strategies during physical activity, exercise advice, problem-solving approaches and stress management techniques.

Four high quality studies (Saffari 2018, Somers 2012, Hurley 2007, Omidi 2018) and one moderate quality study (Coleman 2012) compared self-management to usual care or no treatment. These studies reported greater improvements in pain, function, or both compared to the control groups. In addition, some of these studies reported greater improvements in quality of life, pain catastrophizing, and self-efficacy in the self-management groups (Saffari 2018, Somers 2012).

One high quality (Marconcin 2018) and three moderate quality studies examined the combined use of self-management and exercise to either groups that received self-management or exercise alone (Bennell 2016) or usual care (Yip 2007, Kao 2012). Yip et al. reported greater improvements in pain, time spent in leisure activities, and self-efficacy, compared to usual care. (Yip 2007) reported greater improvements in pain and function compared to the control groups. Bennell, et al, reported improvements in pain and function in all groups. There were no differences between groups on pain measures but the combined use of self-management (i.e., pain coping skills training) and exercise had greater improvements in function compared to those receiving only self-management or exercise (Bennell 2016).

An attempt was made to examine the literature on cognitive behavioral therapy (CBT) in the management of people with knee osteoarthritis. One high quality (Helminen 2015) and 4 moderate quality (Focht 2012, Focht 2017, Smith 2015, Lerman 2017) studies were reviewed. Control groups consisted of usual care (Helminen 2015), traditional exercise approaches for knee osteoarthritis, (Focht 2012, Focht 2017) or behavioral desensitization (Smith 2015, Lerman 2017). Inconsistency in outcome results across studies made it difficult to provide a recommendation for this intervention approach at this time.

Benefits/Harms of Implementation

Patients may expect improvements in pain and function, problem-solving abilities, and self-efficacy from participating in these programs. Engagement in exercise recommendations could result in some temporary increased knee pain or muscle soreness.

Outcome Importance

Pain, function, self-efficacy in managing osteoarthritis.

Feasibility

Self-management programs are feasible for patients provided they have appropriate access. Some patients may have limited access for participation, making the programs less feasible.

Future Research

Future research should examine delivery methods designed to increase access for patients (e.g., online delivered programs). Given that some patients with severe symptoms and disability may not be able to respond to self-management programs, future research should assess whether the outcomes of self-management programs vary with disease severity. This research should determine if there is a threshold of disease or disability severity that discriminates between responders and non-responders of this treatment approach.

 

 

Patient Education
Patient education programs are recommended to improve pain in patients with knee osteoarthritis
Strong Recommendation Strong Recommendation

Six high-quality studies (Saffari 2018, Somers 2012, Cagnin 2019, Gilbert 2018, Baker 2019, Berman 2004) thirteen moderate quality study (Brosseau 2012, Allen 2010, O’Brien 2018, Allen 2010, Bennell 2017, Marra 2017, Rezende 2017, Sandeghi 2019, Rodriguez da Silva 2017, Rini 2015, Moseng 2020, Chen 2020, Ravaud 2009) and two limited quality studies compared patient education and control. These studies reported more significant improvements in pain compared to the control groups.

Patient education programs in studies overlap with self-management programs. Patient education programs vary from patient handout, 2+ hour DVD, one-day education to multiple sessions over a month (Saffari 2018, Cagnin 2019, Brosseau 2012, O’Brien 2018, Rezende 2017, Rodriguez da Silva 2017, Rini 2015). Many studies are challenging to evaluate the effects of education because they involve exercise classes and other proven interventions (Marra 2017, Ravaud 2009). Self-management programs train people in several elements of self-management for osteoarthritis (1148), including medication compliance, pain management, and pain coping strategies, joint protection strategies (1149) during physical activity, exercise advice, problem-solving approaches, and stress management techniques. Patient education programs may not be as labor-intensive, and further work is needed to identify the amount of education needed to improve patient-related outcome measures, like pain.

Programs that focused on education are two high quality (Saffari 2018, Cagnin 2019) and four moderate quality (Brosseau 2012, O’Brien 2018, Rodriguez da Silva 2017, Rini 2015). Saffari used seven (7) group sessions over one month and provided a CD-ROM and booklet describing preventive lifestyle procedures and the importance of treatment adherence (Saffari 2018). They found improvement in SF-12 and pain scores. Cagnin used an educational session with a physical therapist who demonstrated how recommended exercises should be performed and how patients can manage their pain. They demonstrated improvement in KOOS pain scores (Cagnin 2019). Brousseau looked at education (educational pamphlet) vs. walking and education vs. walking and behavioral intervention (Brouseau 2012). There was a non-clinically significant improvement in pain in the education-only group at 12 months compared to walking and behavioral intervention. O’Brien used weight loss education, where trained telephone interviewers provided brief advice and education about the benefits of weight loss and physical activity for knee osteoarthritis immediately after randomization [O’Brien 2018]. The intervention group provided an evidence-based public health non-disease-specific telephone-based coaching service funded by the local Australian state government to support adults in making sustained lifestyle improvements, including diet, physical activity, and achieving a healthy weight and, where appropriate, access smoking cessation services. They did not find an added benefit from the coaching service over the brief telephone education in pain nor WOMAC scores. Rini compared an internet-based app (PainCoach) [http://tri.ad/projects-2/] to usual care and found a non-clinically significant reduction in VAS pain scores (Rini 2015). Rodriguez da Silva used a single day (Saturday, from 8 a.m. to 5 p.m.), which included seven lectures of 30 min by each professional team, and 60-min workshops by the physical education, physical therapy, and occupational therapy professionals, approaching the importance of their area in knee OA treatment/management. The study did not report pain scores but did note an increase in mobility with improvements in the get-up and go test. The two high-quality and four moderate-quality studies showed improved pain scores from the education given during educational sessions. Most studies (15 of the 21) incorporate education with other interventions; therefore, it is impossible to isolate the effects of education in these other 15 studies.

One high quality (Gilbert 2018) and three moderate quality (Rezende 2017, Chen 202, Ravaud 2009) used the transtheoretical model (TTM) and motivational interviewing to improve osteoarthritis treatment adherence. These studies showed improvement in WOMAC pain scores. TTM has been used successfully in other conditions that benefit from lifestyle changes [PMID: 24500864].

Future Research

Further research is needed to determine the best practice of education for reducing pain and other PROM for knee OA and the delivery method. Since many studies use different components and delivery methods and multiple interventions, it is impossible to recommend one particular educational module or particular component.

Weight Loss Intervention
Sustained weight loss is recommended to improve pain and function in overweight and obese patients with knee osteoarthritis.
Moderate Recommendation Moderate Recommendation

The Weight Loss Intervention recommendation has been downgraded one level because of inconsistent evidence.

There were 1 high (Jenkinson et.al 2009),1 moderate (Miller 2006), and 2 low strength (Focht 2005, Rejeski 2002) studies evaluating diet and exercise as weight loss interventions to treat knee osteoarthritis. Overall pain and function improved with weight loss achieved through a combination of diet and exercise.

 However, when evaluating only diet vs control, 2 high (Bliddal 2011, Christensen 2015), 2 moderate (Messier 2013, Mihalko 2018) and 2 low strength (Rejeski 2002, Fochyt 2005) there was no clear clinically significant change in patient outcomes. Specifically, Christensen et al, 2015 published a high-quality study investigating the effect of weight on symptoms of knee osteoarthritis.  They showed no significant difference in pain and function at 1 year. Bliddal et al, 2010 published another high quality study  which investigated the effect of weight loss on symptoms of knee OA in the obese patient, showing  that perceived pain (via WOMAC) was significantly lessened despite not being able to show improvement in function and quality of life at 1 year.

There were 2 moderate strength studies (Messier 2013, Mihalko 2018) which evaluated diet vs exercise, which favored exercises. To note, Messier et al  2013 published results of the IDEA trial with moderate quality study which was an attempt to determine if a 10% reduction in body weight (induced by diet, with or without exercise) would improve “clinical and mechanistic” outcomes in sedentary lifestyle patients (BMI 27 thru 41).  Interestingly, in this primary study, they were unable to show an improvement in WOMAC pain but they did show improvement in the WOMAC function subscale, and also showed improvements in the 6 minute walk test.

Given the current evidence, it is at the discretion of the surgeon as to which approach is utilized to address weight loss, however a combination of diet and exercises appears to be the preferred alternative.

Benefits/Harms of Implementation

There are no known or anticipated harms associated with implementing this recommendation.

Outcome Importance

Management of obesity and overweight through weight loss may have high impact on symptoms and overall health.

Cost Effectiveness/Resource Utilization

Weight loss presents a potentially high level of cost effectiveness compared to other surgical and nonsurgical approaches; however, the cost-effectiveness of different weight loss approaches is still to be determined.

Acceptability

Currently weight loss is a commonly utilized approach in the optimization of patient who present with obesity and knee osteoarthritis, hence there should be no issues implementing this recommendation as it does not influence a major change in clinical practice. To note, patients often are reluctant or unwilling to accept referrals for weight loss as a suggestion for a path to improvement.

Feasibility

This recommendation does not interfere with other interventions or clinical practice therefore it is deemed very feasible in a subset of overweight and obese patients.

Future Research

Future research should focus on large randomized clinical trials and should focus more on function and quality of life measures since it appears obvious that pain improved through weight loss.

Manual Therapy
Manual therapy in addition to an exercise program may be used to improve pain and function in patients with knee osteoarthritis.
Limited Recommendation Limited Recommendation

The Manual Therapy recommendation has been downgraded one level because of inconsistent evidence and to lack of internal consistency with recommendations of equal supporting evidence.

Manual therapy consists of maneuvers applied with manual force from the treating therapist to the patient’s body to improve joint mobility and/or relieve pain. The techniques may generally consist of manually applied joint mobilization techniques, manually applied joint range of motion and/or muscle stretching, and soft tissue massage. One high quality study (Fitzgerald 2016) and one moderate quality study (Deyle 2000) were reviewed that examined manual therapy combined with exercise compared to exercise alone (Fitzgerald 2016) or non-therapeutic ultrasound (placebo physical therapy) in subjects with knee osteoarthritis. (Deyle 2000) Fitzgerald, et al, reported that both groups yielded significant improvements in clinical outcomes from baseline but the manual therapy group had greater improvements in the WOMAC total score and were more likely to meet the OMERACT-OARSI Responder Criteria at the 9 week follow-up. (Fitzgerald 2016) While both groups demonstrated sustained improvements in clinical outcomes at 1 year there was no difference between groups on any measures at this timepoint. Deyle et al. reported similar findings with the manual therapy and exercise group demonstrating greater improvements at 8 weeks but no significant differences between groups at 1 year (Deyle 2000).

Benefits/Harms of Implementation

Most patients can expect an improvement in pain and function with the interventions. Patients may experience a temporary increase in knee pain or muscle soreness when engaging in a manual therapy and exercise program.

 

Cost Effectiveness/Resource Utilization

Bove, et al 2018 found manual therapy and exercise delivered with periodic booster sessions was more cost-effective than delivery of this same intervention without booster sessions. (Bove 2018) The cost effectiveness assessment was over a 2-year follow-up (Bove 2018).

Feasibility

The manual therapy interventions are feasible for patients who have access to in-person physical therapy. The exercise interventions are feasible both from an in-person and home program perspective.

Future Research

Future studies should examine ways to sustain the effects of manual therapy and exercise for extended follow-up periods. Research is needed to determine the effectiveness of self-applied manual therapy and telerehabilitation applications of manual therapy and exercise interventions for improving pain and function in patients with knee osteoarthritis.

Massage
Massage may be used in addition to usual care to improve pain and function in patients with knee osteoarthritis.
Limited Recommendation Limited Recommendation

The Massage recommendation has been downgraded one level because of inconsistent evidence and due to lack of internal consistency with recommendations of equal supporting evidence.

Two high quality study (Pehlivan 2018, Sansila 2019) and one moderate quality study (Perlman 2018) were reviewed that examined massage plus usual care to a usual care control group. Pehlivan et al. applied a leg massage focusing primarily on tissues around the knee for a total of 6 sessions over 3 weeks (Pehlivan 2018). Subjects receiving massage had greater improvements in knee pain at 4 weeks follow-up compared to the usual care control group but there was no difference between groups at 8 weeks. There were also no differences in function between groups at either time point. Perlman et al. applied a 60-minute total body massage one time per week for 8 weeks, followed by bi-weekly sessions for 52 weeks (Perlman 2018).  There was greater improvement in pain and function in the massage group at the 8-week follow-up, but these effects were not sustained for any of the longer-term follow-up timepoints.

Benefits/Harms of Implementation

Patients receiving massage may experience some short-term improvements in pain and function. Massage treatments are generally safe but common side effects may include skin redness or irritation, bruising, muscle soreness, fatigue, and nausea.

Feasibility

The interventions are considered feasible and relatively accessible to the general public. Patients who have transportation issues may have difficulty with obtaining access to treatment.

Future Research

Future research may be warranted to determine how massage could be combined with other interventions for knee osteoarthritis to sustain longer term effects on pain and function.

Laser Treatment
FDA-approved laser treatment may be used to improve pain and function in patients with knee osteoarthritis
Limited Recommendation Limited Recommendation

The Laser Treatment recommendation has been downgraded two levels because of feasibility, usage in practice and a lack of internal consistency with recommendations of equal supporting evidence.

A meta-analysis was performed using pain data from two high quality studies (Gur 2003, Nazari 2018) and one moderate quality (Marquina 2012) study examining high intensity laser treatment compared to either placebo laser treatment or no treatment groups. The results of the analysis are provided in Figure 10 in appendix. The overall findings were in favor of the laser intervention over the sham or no-treatment groups. In addition, two of these studies reported greater improvements in function. (Gur 2003, Nazari 2018) Gur 2003 also compared high dose vs. low dose laser treatment on clinical outcomes and found no significant difference between the groups. (Gur 2003)

Benefits/Harms of Implementation

Patients should expect to experience improvements in pain and function with the treatment. There have been no reports of serious side effects from laser treatment for pain control. Long-term exposure of the laser beam to the eyes can cause eye damage.

Feasibility

Access to the laser treatment may not be available in all clinics.

Future Research

Continued study of laser treatment for pain control, improving function and cost-effectiveness in people with knee osteoarthritis is encouraged.

Acupuncture
Acupuncture may improve pain and function in patients with knee osteoarthritis.
Limited Recommendation Limited Recommendation

The Acupuncture recommendation has been downgraded two levels because of inconsistent evidence and a lack of internal consistency with recommendations of equal supporting evidence.

A meta-analysis was performed using pain data from five high quality studies (Chen 2013, Suarez-Almazor 2010, Mavrommatis 2012, Berman 2004, Hinman 2014) and two moderate quality studies (Vas 2007, Berman 1999). Acupuncture treatments were either traditional (Chen 2013, Hinman 2014) or electro-acupuncture. (Suarez-Almazor 2010, Mavrommatis 2012, Vas 2007, Berman 2004, Berman 1999) Control groups consisted of either no acupuncture, (Hinman 2014) sham acupuncture, (Mavrommatis 2012, Suarez-Almazor 2010, Vas 2007, Berman 2004) sham TENS (Chen 2013), or usual care. (Berman 1999) The meta-analysis also accounted for the degree of blinding effectiveness of the studies. The results of the meta-analysis can be seen in Figure 11 in the appendix. The overall findings were in favor of acupuncture for reducing pain in subjects with knee osteoarthritis. There appeared to be no effect in two studies where blinding was effective. In studies where there was no blinding or the effects of blinding were unclear, there were greater effects favoring acupuncture. This prompted our decision to apply a limited strength of recommendation in favor of acupuncture for pain control.

A similar meta-analysis was performed using the same studies for measures of function. The results of this meta-analysis can be seen in Figure 12 in appendix. The overall findings were in favor of acupuncture for improving measures of function in subjects with knee osteoarthritis. However, the effects of blinding effectiveness on the results were similar to that described above for pain. Again, this prompted our decision to apply a limited strength of recommendation in favor of acupuncture for improving function.

Some investigators examined variations in delivery of acupuncture treatment. Ju et al. examined high intensity vs low intensity electro-acupuncture and found no difference between these approaches for pain but possibly better improvements in function favoring the high intensity group. (Ju 2015) Others found no meaningful differences between using 2-point, 4-point, or 6-point acupuncture approaches. (Qi 2016, Taechaarpornkul 2009).

Benefits/Harms of Implementation

Many patients receive benefit such as reduced pain and improved function but not all patients respond favorably to treatment. The treatment should be administered by a certified acupuncture practitioner. Common side effects can include soreness and minor bleeding or bruising where needles are inserted. Risk of infection is low if proper procedures are followed. Patients who have a bleeding disorder, pacemaker, or could be pregnant may not be safe candidates for acupuncture and should consult with their physician before having the treatment.

Acceptability

The interventions are considered acceptable, but some individuals may not be enthusiastic about having needles inserted into their skin.

Feasibility

The intervention is feasible provided there is access to a trained practitioner. Not all clinics can provide such access.

Future Research

Continued research is encouraged, with more studies that improve blinding effectiveness in the methodology and studies that may identify patient characteristics that could discriminate between responders and non-responders of this treatment approach.

 

 

Transcutaneous Electrical Nerve Stimulation
Modalities that may be used to improve pain and/or function in patients with knee osteoarthritis include: a) Transcutaneous Electrical Nerve Stimulation (pain)
Limited Recommendation Limited Recommendation

The Transcutaneous Electrical Nerve Stimulation recommendation has been downgraded two levels because of inconsistent evidence and a lack of internal consistency with recommendations of equal supporting evidence.

A meta-analysis was performed using pain data from two high quality studies (Palmer 2014, Inal 2016) and one moderate quality study (Atamaz 2012) in which Transcutaneous Electrical Nerve Stimulation (TENS) was compared to sham TENS in subjects with knee osteoarthritis. Blinding effectiveness was considered fair in all three studies. The results of the meta-analysis can be seen in Figure 15 in the appendix. The overall findings were in favor of receiving TENS for reducing pain in subjects with knee osteoarthritis. A similar meta-analysis was performed using the same studies for measures of function. The results of this meta-analysis can be seen in Figure 16 in the appendix. The overall findings did not favor the use of TENS to improve measures of function in subjects with knee osteoarthritis.

Benefits/Harms of Implementation

TENS units are small, portable, non-invasive devices that could provide pain relief and may allow some patients to reduce pain medication use. Some patients may experience skin irritations or allergic reactions to the adhesive pads used to deliver the stimulation. Use of TENS is not recommended in people with pacemakers and women who are pregnant should not apply TENS in the abdominal or pelvic regions.

Acceptability

The interventions appear to be acceptable to most patients however some patients may not like the sensation of electrical or electro-magnetic energy being applied to their bodies.

Future Research

Continued research with larger randomized trials that examine long-term effectiveness of the interventions is warranted. Studies that identify factors discriminating between responders and non-responders to the interventions would also be important.

Percutaneous Electrical Nerve Stimulation
Modalities that may be used to improve pain and/or function in patients with knee osteoarthritis include: a) Percutaneous Electrical Nerve Stimulation (pain and function) b) Pulsed Electromagnetic Field Therapy (pain)
Limited Recommendation Limited Recommendation

The Percutaneous Electrical Nerve Stimulation/Pulsed Electromagnetic Field Therapy recommendation has been downgraded one level because of feasibility issues.

One high quality study was reviewed that examined the use of Percutaneous Electrical Nerve Stimulation (PENS) combined with a Cox-2 inhibitor to sham PENS combined with a Cox-2 inhibitor in subjects with knee osteoarthritis. (He 2019) The results indicated greater improvements in pain and function measures in subjects receiving PENS compared to sham PENS.

One high quality study was reviewed that examined the use of a wearable Pulsed Electomagnetic Field (PEMF) device for pain management in subjects with knee osteoarthritis. (Bagnato 2016) Subjects were randomized to either the PEMF group or a sham PEMF group. PEMF was applied 12 hours per day for a period of 4 weeks. The results indicated greater improvement in WOMAC pain and VAS pain scores for subjects receiving PEMF over sham PEMF. There was no difference between groups on WOMAC function scores.

Benefits/Harms of Implementation

PENS is a minimally invasive procedure. Patients may experience reduction in pain and improvement in function after receiving the intervention. Because the intervention involves the insertion of very thin needles, side effects are similar to acupuncture and include bleeding, bruising, or skin irritation at the insertion site. Infection or nerve damage are possible but very rare side effects.

PEMF devices are generally safe and can reduce pain and inflammation. The treatment often results in an increase in blood flow which, in some circumstances, could temporarily trigger an increase in pain and discomfort and oxidative stress. Some patients may report fatigue and loss of energy, sleep disturbances, dizziness, and increased urination. PEMF treatment may induce a decrease in blood pressure and heart rate so caution must be taken in patients with cardiovascular deficiencies. A fall in blood sugar levels can also result in some people with PEMF treatment so caution should be taken in patients who have difficulty regulating blood sugar. PEMF may affect blood coagulation and is not recommended in people who are taking anti-coagulant therapies.

Acceptability

The interventions appear to be acceptable to most patients however some patients may not like the sensation of electrical or electro-magnetic energy being applied to their bodies.

Feasibility

PENS is feasible but requires a practitioner trained in the technique which could limit access for some patients. PEMF is not widely used in clinics treating patients for knee osteoarthritis and thus could limit access for some patients.

Future Research

Continued research with larger randomized trials that examine long-term effectiveness of the interventions is warranted. Studies that identify factors discriminating between responders and non-responders to the interventions would also be important.

 

 

Extracorporeal Shockwave Therapy
Extracorporeal shockwave therapy may be used to improve pain and function for treatment of osteoarthritis of the knee.
Limited Recommendation Limited Recommendation

The Extracorporeal shockwave Therapy recommendation has been downgraded two levels because of inconsistent evidence and a lack of internal consistency with recommendations of equal supporting evidence.

A meta-analysis was performed using pain data from three high quality studies (Zhong 2019, Ediz 2018, and Uysal 2020) in which Extracorpeal Shockwave Therapy (ESWT) was compared to sham ESWT in subjects with knee osteoarthritis. The results of the meta-analysis can be seen in Figure 19 in the appendix. The overall findings were in favor of receiving ESWT for reducing pain in subjects with knee osteoarthritis. In addition, four high quality studies reported greater improvements in function scores in subjects receiving ESWT compared to the sham group at 4 to 12 weeks but not at 1-year follow-up. (Zhao 2013, Ediz 2018, Zhong 2019, Uysal 2020).

Benefits/Harms of Implementation

Patients receiving this treatment may experience improvements in pain and function. Side effects may include redness or mild bruising, swelling, pain, numbness or tingling in the treated area, migraine headaches, and syncope.

Acceptability

Generally tolerated well but some patients may experience side-effects listed above.

Feasibility

The intervention is considered feasible but may not be widely available in clinics managing patients with knee osteoarthritis and therefore may not be accessible for all patients.

Future Research

Future studies should examine methods to sustain longer term effects of the intervention (e.g., optimal dosage, use of booster interventions), investigate the potential of ESWT as an osteoarthritis disease-modifying agent, and identify characteristics of patients that may discriminate between responders and non-responders of the intervention. The mechanisms of action of these interventions in human subjects is not well understood. Future studies investigating the mechanisms of action of these interventions in human subjects would be helpful.

Oral NSAIDs
Oral NSAIDs are recommended to improve pain and function in the treatment of knee osteoarthritis when not contraindicated.
Strong Recommendation Strong Recommendation

Among the 34 high-quality, 23 moderate-quality, and 1 low-quality studies that met the inclusion criteria, non-selective and selective clyclooxygenase-2 (COX-2) oral nonsteroidal anti-inflammatory drugs (NSAID) consistently improved pain and function compared to controls in the treatment of osteoarthritis of the knee (Reginster 2017; Lee 2017; Gordo 2017; Strand 2017; Essex 2014; Kongtharvonskul 2016; Altman 2015; Gibofsky 2014; Ishijima 2014; Conaghan 2013; Essex 2012; Singh 2012; Elsaman 2016; Schnitzer 2011; Kivitz 2004; Fleischmann 1997; Lee 1986; Davies 1999; Sandelin 1997; Puopolo 2007; Gibofsky  2003; Bensen 1999; Kivits 2002; Clegg 2006; Sangdee 2002; Sheldon 2005; Tannenbaum 2004; Lehmann 2005; Rother 2007; Simon 2009; Svensson 2006; Schnitzer 2010; Doherty 2011; McKenna  2001 (a); Paul 2009; Bolten 2015; Essex 2016; Ekman 2014; Ohtori 2013; Selvan 2012; Pavelka 2007; Ehrich 1999; Lee 1985; Dwicandra 2018; Asmus 2014; Smugar 2006; Bingham 2007; Altman 1998; Schnitzer  1999; Birbara 2006; Williams 2001; Miceli 2004; Mckenna 2001 (b); Pincus 2004; Lohmander 2005; Schnitzer 2005b; Williams 2000; Fleischmann 2006). Although meta-analysis of non-selective oral NSAIDs compared to controls demonstrated a meaningful reduction in pain, the results need to be interpreted with caution due to the relatively high degree of heterogeneity. The meta-analysis of non-selective oral NSAIDs compared to controls demonstrated a meaningful improvement in function with an acceptable degree of heterogeneity. In terms of selective COX-2 oral NSAIDs, the meta-analysis of celecoxib, the only available selective COX-2 oral NSAID on the United States market, demonstrated a meaningful reduction in pain and improved function with an acceptable degree of heterogeneity. The comparison of non-selective and selective COX-2 oral NSAIDs shows no significant difference in the effectiveness between the types of oral NSAIDs (Gordo 2017; Essex 2014; Essex 2016; Essex 2012; Bensen 1999; Kivits 2002; Kivitz 2004; Puopolo 2007; Hochberg 2011; Mckenna 2001 (b); Schnitzer 2005b; Malik 2017). Although NSAIDs effectively reduce pain and improve function in the treatment of osteoarthritis of the knee, providers should consider patient comorbidities, the type of NSAID administered, dose, and duration of administration. In fact, the United States Food and Drug Administration (FDA) has a black-box warning for NSAIDs citing an increased risk of serious cardiovascular thrombotic events and serious gastrointestinal events. Therefore, we recommend the lowest effective dose for the shortest duration possible for the patient. Although selective COX-2 oral NSAIDs were developed to reduce gastrointestinal adverse events compared to non-selective oral NSAIDs, meta-analysis did not reveal a significant reduction in gastrointestinal adverse events.

Benefits/Harms of Implementation

Although oral NSAIDs are widely used to treat osteoarthritis of the knee, providers must recognize the specific risks associated with each medication. Specific patient contraindications need to be assessed on an individual basis (Example: patient with cardiac conditions have an increased risk of myocardial infarction). 

Outcome Importance

The most important consideration will be removal of oral narcotics from the medications prescribed in the treatment of osteoarthritis of the knee. This becomes particularly significant due to the rise of the opioid epidemic in the United States.

Cost Effectiveness/Resource Utilization

Evidence based decision making in selecting the optimal systemic treatment for the treatment of symptomatic knee osteoarthritis should result in improved pain and function. For a given systematic treatments as effectiveness increases without raising the risk of adverse events so will its cost-effectiveness. To date, the most cost-effective systemic treatment is still to be determined.

Acceptability

Currently oral NSAIDs are commonly utilized approach in treating symptomatic knee osteoarthritis, hence there should be no issues implementing this recommendation as it does not influence a major change in clinical practice, and it provides further evidence to support and guide these practices. 

Feasibility

This recommendation may be implemented immediately having a potential positive impact in clinical practice.

Future Research

Most important future research will provide high quality investigation through either prospective randomized trials or prospective cohort studies to establish efficacy within specific subgroups and populations to tailor systemic medications increasing efficacy and decreasing risk of adverse effects

Oral Acetaminophen
Oral Acetaminophen is recommended to improve pain and function in the treatment of knee osteoarthritis when not contraindicated.
Strong Recommendation Strong Recommendation

Among the 4 high-quality and 3 moderate quality studies that met the inclusion criteria, oral acetaminophen consistently improved pain and function compared to controls in the treatment of osteoarthritis of the knee (Herrero-Beaumont 2007; Doherty 2011; Reed 2018; Prior 2014; Micelli 2004; Pincus 2004; Altman 2007). The meta-analysis of oral acetaminophen compared to controls demonstrated a meaningful reduction in pain and improved function with no evidence of confounding heterogeneity. Overall, acetaminophen is considered a safe medication with no evidence of significantly increased adverse events among the included studies. However, the United States FDA has a black-box warning for acetaminophen secondary to concern of overdose leading to hepatoxicity or death. When oral acetaminophen was compared to NSAIDs, the use of oral NSAIDs provided a significant reduction in pain and improved function. As a result, providers may consider using oral NSAIDs instead of acetaminophen when a contraindication to oral NSAIDs does not exist in the patient.

Benefits/Harms of Implementation

Although oral acetaminophen is widely used to treat osteoarthritis of the knee, providers must recognize the specific risks associated with each medication. Specific patient contraindications to need to be assessed in an individual basis (Example: patient with renal failure).

Outcome Importance

The most important consideration will be removal of oral narcotics from the medications prescribed in the treatment of osteoarthritis of the knee. This becomes particularly significant due to the rise of the opioid epidemic in the United States.

Cost Effectiveness/Resource Utilization

Evidence based decision making in selecting the optimal systemic treatment for the treatment of symptomatic knee osteoarthritis should result in improved pain and function. For a given systematic treatments as effectiveness increases without raising the risk of adverse events so will its cost effectiveness. To date, the most cost-effective systemic treatment is still to be determined.

Acceptability

Currently acetaminophen is commonly utilized approach in treating symptomatic knee osteoarthritis, hence there should be no issues implementing this recommendation as it does not influence a major change in clinical practice, and it provides further evidence to support and guide these practices. 

Feasibility

This recommendation may be implemented immediately having a potential positive impact in clinical practice.

Future Research

Most important future research will provide high quality investigation through either prospective randomized trials or prospective cohort studies to establish efficacy within specific subgroups and populations to tailor systemic medications increasing efficacy and decreasing risk of adverse effects.

Oral Narcotics (Recommendation Against)
Oral narcotics, including tramadol, result in a significant increase of adverse events and are not effective at improving pain or function for treatment of osteoarthritis of the knee.
Strong Recommendation Strong Recommendation

Among the 5 high-quality and 2 moderate quality studies that met the inclusion criteria, oral narcotic medications are not an effective treatment to reduce pain and improve function in osteoarthritis of the knee (Serrie 2017; Afilalo 2010; Mayorga 2016; Fishman 2007; Fleischmann 2001; Burch 2007; Babul 2004). In fact, the use of narcotics to treat osteoarthritis of the knee is consistently associated with a significantly high risk of adverse events. Due to the lack of efficacy and increase of adverse event, we would recommend against the use of narcotics for the treatment of osteoarthritis of the knee. Given the effective and relatively safe alternatives of oral NSAIDs and acetaminophen, oral narcotics should be avoided when the provider is considering the recommendation of an oral medication.

Benefits/Harms of Implementation

The removal of oral narcotics from the medications prescribed in the treatment of osteoarthritis of the knee will have further beneficial effects if patients fail non-surgical management and eventually progress to replacement of the knee, as they have been associated with adverse events after surgery as well.

Outcome Importance

The most important consideration will be removal of oral narcotics from the medications prescribed in the treatment of osteoarthritis of the knee. This becomes particularly significant due to the rise of the opioid epidemic in the United States.

Cost Effectiveness/Resource Utilization

Evidence based decision-making in selecting the optimal systemic treatment for the treatment of symptomatic knee osteoarthritis should result in improved pain and function. For a given systematic treatments as effectiveness increases without raising the risk of adverse events so will its cost effectiveness. To date, the most cost-effective systemic treatment is still to be determined.

Acceptability

Due to the lack of efficacy and increase of adverse event, a recommendation against the use of narcotics for the treatment of osteoarthritis of the knee was made and this could present some resistance from patients who have failed oral NSAIDs and acetaminophen. Nonetheless, patients should be counseled on the risks associated with narcotics and their lack of efficacy for the purpose of treating knee osteoarthritis.

Feasibility

This recommendation may be implemented immediately having a potential positive impact in clinical practice.

Future Research

Most important future research will provide high quality investigation through either prospective randomized trials or prospective cohort studies to establish efficacy within specific subgroups and populations to tailor systemic medications increasing efficacy and decreasing risk of adverse effects.

Hyaluronic Acid (Recommendation Against)
Hyaluronic acid intra-articular injection(s) is not recommended for routine use in the treatment of symptomatic osteoarthritis of the knee.
Moderate Recommendation Moderate Recommendation

The Hyaluronic Acid recommendation was downgraded one level due to a lack of generalized results.

Twenty-eight studies (17 high-strength (Chevalier 2010, Petterson 2018, Maheu 2019, Neustadt 2005, Baltzer 2009, Lundsgaard 2008, Altman 2004, Huang 2011, van der Weergen 2015, Altman 2009, Day 2004, Jorgensen 2010, Henrotin 2017, Henderson 1994, Hangody 2018, Saccomanno 2016, Altman 1998)  and 11 moderate-strength (Jubb 2003, Navarro-Sarabia 2011, Farr 2019, Kahan 2003, Kahan 2003, Karlsson 2002, Hermans 2019, Huskisson 1999, Heybeli 2008, Petrella 2006, Takamura 2018, Wobig 1998)) assessed intraarticular hyaluronic acid (HA) injections when compared to controls. A comparison of patients from these studies and from studies validating the MCIDs were used to judge clinical significance. Results revealed that patients were demographically comparable for WOMAC and VAS pain as well as WOMAC function based on age, baseline pain scores, BMI, weight, and gender. Meta-analysis in meaningfully important difference (MID) units showed that the effect was less than 0.5 MID units, indicating a low likelihood that an appreciable number of patients achieved clinically important benefits after intraarticular HA injection (Guyatt et al.). When we differentiated high- versus low- molecular weight viscosupplementation (three high, two moderate and two low quality studies), our analyses demonstrated no significant differences among different viscosupplementation formulations. Crosslinking features of the viscosupplemtation product was assessed in two high quality studies. In patients with OA, there was no difference between cross-linked and non-cross-linked HA.

Some studies demonstrated a statistical benefit with the use of HA but could not reach the significance for a minimally clinical meaningful difference, leading to the conclusion that viscosupplementation can represent a viable option for some patients that failed other treatments when appropriately indicated.  The number needed to treat to see a tangible benefit from HA was 17 patients. Furthermore, this difference was most evident at 6 weeks and 3 months. Most of the studies that exist in the literature evaluate low to moderate arthritic knees (Kellgren Lawrence of I-III) with worse results in patients with severely affected knees (KL IV).

The 2013 edition of this guideline strongly recommended against the use of viscosupplementation. In contrast to this updated version, the 2021 version found that statistically significant improvements were associated with high-molecular cross-linked hyaluronic acid but when compared to mid-range molecular weight, statistical significance was not maintained. This newer analysis did not demonstrate clinically relevant differences when compared to controls. However, as previous research reported benefits in their use, the group felt that a specific subset of patients might benefit from its use.

Benefits/Harms of Implementation

There are no major known or anticipated harms associated with implementing this recommendation.

Cost Effectiveness/Resource Utilization

The cost-effectiveness of different intra-articular therapies is still to be determined, in comparison to other treatment strategies and among different intra-articular alternatives.

Acceptability

Currently intra-articular treatments are commonly utilized approach in treating symptomatic knee osteoarthritis, hence there should be no issues implementing this recommendation as it does not influence a major change in clinical practice, and it provides further evidence to support and guide these practices.

Feasibility

These recommendations do not interfere with other interventions or clinical practice therefore it is deemed very feasible in patients with symptomatic knee osteoarthritis.

Future Research

Future research in this area should embrace detailed osteoarthritis characterization including sub-group analyses and osteoarthrosis severity stratification. Furthermore, using clinically relevant outcomes and controls for bias are warranted along with cost-effectiveness analysis.

Intra-articular Corticosteroids
Intra-articular (IA) corticosteroids could provide short-term relief for patients with symptomatic osteoarthritis of the knee.
Moderate Recommendation Moderate Recommendation

The Intra-Articular Corticosteroids recommendation has been downgraded one level because of potential risk in accelerating osteoarthritis from injections.

Our search found 18 high (Campos 2017, Cai 2019, Erturk 2016, de Campos 2013, Shrestha 2018, Mendes 2019, Yilmaz 2019, Chao 2010, Raynauld 2003, McAlindon 2017, Henrikson 2015, Neilsen 2018, Riis 2017, Arden 2014, Delgado-Enciso 2019, Smith 2003, Soriano-Maldonado 2016) and 6 moderate quality studies (Conaghan 2018, Langworthy 2019, Gaffney 1995, Yavuz 2012, Yilmaz 2019, Jones 1996) comparing intra-articular corticosteroids to control to treat knee osteoarthritis. Overall pain and function improved with intra-articular corticosteroids; however, it is important to note that such effect lasted only up to 3 months. When we differentiated intra-articular corticosteroids extended versus immediate release (one high, two moderate quality studies) (Bodick 2015, Conaghan 2018 and Langworthy 2019), our analyses demonstrated that, extended release IA steroids can be used over immediate release to improve patient outcomes (Moderate strength recommendation).

Benefits/Harms of Implementation

The AAOS Patient Safety Committee recommends avoiding musculoskeletal corticosteroid injections for two weeks before and one week after COVID vaccine administration.

Outcome Importance

Pain and function improvement through intra-articular therapies for the treatment of knee osteoarthritis may have high impact on symptoms and overall health.

Cost Effectiveness/Resource Utilization

The cost-effectiveness of different intra-articular therapies is still to be determined, in comparison to other treatment strategies and among different intra-articular alternatives.

Acceptability

Currently intra-articular treatments are commonly utilized approach in treating symptomatic knee osteoarthritis, hence there should be no issues implementing this recommendation as it does not influence a major change in clinical practice, and it provides further evidence to support and guide these practices.

Feasibility

These recommendations do not interfere with other interventions or clinical practice therefore it is deemed very feasible in patients with symptomatic knee osteoarthritis.

Future Research

Future research in this area should embrace detailed osteoarthritis characterization including sub-group analyses and osteoarthrosis severity stratification. Furthermore, using clinically relevant outcomes and controls for bias are warranted along with cost-effectiveness analysis

Platelet-rich Plasma
Platelet-rich plasma (PRP) may reduce pain and improve function in patients with symptomatic osteoarthritis of the knee.
Limited Recommendation Limited Recommendation

The Platelet-Rich Plasma recommendation has been downgraded two levels because of inconsistent evidence.

There were two high (Rayegani; 2014, Gormeli; 2017) and one moderate (Akan; 2018) study with 30 people per group comparing PRP vs. control. There were mixed results in the studies for pain and function. A meta-analysis was not performed due to heterogeneity. Two studies (Akan; 2018, Gormeli; 2017) looked at PRP in severe OA with mixed results. Two studies (Rayegani; 2014, Gormeli; 2017) looked at change in all stages of OA at a six-month timeframe. The studies had mixed results. One study (Gormeli; 2017) looked at Kellgren-Lawrence 1-3 stage OA with improvement in IKDC and EQ-VAS. Therefore, due to the heterogeneity of results and the difference in early and late stage OA results, we downgraded the recommendation to Limited from Strong. We feel these recommendations may change with future research on the use of PRP in different levels of severity of OA.

The number of PRP injections had mixed results with the studies with three PRP injections (Akan; 2018, Gormeli; 2017) having positive results outcomes for pain and function. Studies with one and two PRP injections had mixed results, with the positive being less likely clinically significant changes in pain and function. Further research should be done to determine the number of PRP injections for treatment of KOA. Currently, three IA-PRP injections appear to have more favorable results.

Adverse events from PRP injections has been investigated in one high-quality study (Huang; 2018) reported adverse events for PRP vs. control. They reported hypertension and proteinuria were treatment-related side-effects. These met Common Toxicity Criteria grade ≥3. This raises questions on the safety of PRP, which needs further evaluation. Therefore, the strength of recommendation was downgraded to Limited.

When evaluating the effectiveness of PRP vs HA, there were eight high-quality studies (Sanchez; 2012, Vaquerizo; 2013, Filardo; 2015, Gormeli; 2017, Cole; 2017, Buendia-Lopez; 2018, Di Martino; 2019, Yaradilmis; 2020) and six (6) medium quality studies (Spakova; 2012, Raeissadat; 2015, Lana; 2016, Duymus; 2017, Raeissadat; 2017, Ahmad; 2018) and one low-quality studies (Sanchez; 2008) that investigated IA-PRP vs. IA-HA. Four studies were included in a meta-analysis of Total WOMAC results at the 9- OR 12-months mark. This analysis showed a clinically significant difference for IA-PRP over IA-HA. The results between IA-PRP vs IA-HA diverged after 6 months. Most studies showed similar results between IA-PRP and IA-HA at six months, except one (Yaradilimis 2020) where the LR-PRP total WOMAC was better at all time points than the IA HA. Both the patients in the IA HA and IA PRP improved in total WOMAC at six months. Patients in the IA-PRP-arms maintained improvement after 6 months at the 9- OR 12-months mark for total WOMAC vs. IA-HA which started to have a worsening score. The standard is to inject IA-HA every six months in patients with painful KOA. The preparation of the PRP (LR-PRP vs LP-PRP) was noted to be different with the LR-PRP had higher MID values than LP-PRP vs. IA-HA. The research highlights the prolonged effect of IA-PRP over IA-HA, though both appear to be equivalent at 6 months.

Patient-related outcome measures (OARSI-OMERACT responders, percentage of subjects meeting a percentage reduction in VAS Pain OR WOMAC Pain scores) (Sanchez; 2008, Sanchez; 2012, Vaquerizo; 2013, Buendia-Lopez; 2018) more often favorited IA-PRP at both the six month and 12-month time frame. Further, research is needed using standardized PROM’s to investigate the effectiveness of IA-PRP to determine if more patients will benefit from IA-PRP at six months over IA-HA.

Adverse events were higher in the PRP group than IA HA, both local soreness and injection pain (two studies (Spakova; 2012, Yaradilmis; 2020)) and one study (Huang; 2018)) systemic events (proteinuria and hypertension). One study (Vaquerizo; 2013) did not find a difference comparing any adverse event, and one study (Raeissadat; 2017) did not see a difference for minor injection-site adverse events. Therefore, there appears to be more studies finding IA-PRP to have more adverse events vs. IA-HA, more research is need to determine if the adverse events outweigh the benefit of IA-PRP over IA-HA at 9 and 12 months. This is another reason for the downgrade in evidence from Strong to Limited.

Comparisons between IA-PRP and IA-CS, there were three high (Joshi Juber; 2017, Khan; 2018, Nabi; 2018) and one moderate quality study (Huang; 2019). One study (Joshi Juber; 2017) was KL IV end-stage OA and did not find a difference. One study (Khan; 2018) was repeat injections every other month (0, 2, 4 months) with follow up at six months in KL II OAK with no difference. One study (Nabi; 2018) used patients with KL II-III given three injections one month apart showed improvement at three months (one month after the last injection) and six months (4 months after the final injection). One study (Huang; 2019) did three PRP injections every three weeks on KL I-II OAK showed improvement in pain and function at six months (4 months after last injection) and 12 months (10 months after last injection). Therefore, the IA-PRP given in three injections evaluated at 4 months post last injection is more likely to show benefit in KL II and III stages of KOA. More research is needed to evaluate long-term benefits of IA-PRP vs IA-CS over a two- or five-year period to determine if IA-PRP is cartilage sparing vs IA-CS concern for possible cartilage damage over time.

PRP is defined in LR-PRP and LP-PRP. There may be a difference in the effectiveness in knee osteoarthritis between these two preparations. Currently, there is limited data from one direct comparison (Yaradilimis; 2020) and our meta-analysis (Figure 45) of IA-PRP and IA-HA that would demonstrate that intra-articular LR-PRP vs. LP-PRP for KOA is more likely to demonstrate a benefit at 9 and 12 months. The number of studies is limited, therefore determining the better choice between LR-PRP vs. LP-PRP is still inconclusive, but at this time appears to favor LR-PRP.

Benefits/Harms of Implementation

Some patients may experience temporary local soreness or injection pain.

Outcome Importance

Pain and function improvement through intra-articular therapies for the treatment of knee osteoarthritis may have high impact on symptoms and overall health.

Cost Effectiveness/Resource Utilization

The cost-effectiveness of different intra-articular therapies is still to be determined, in comparison to other treatment strategies and among different intra-articular alternatives.

Acceptability

Currently intra-articular treatments are commonly utilized approach in treating symptomatic knee osteoarthritis, hence there should be no issues implementing this recommendation as it does not influence a major change in clinical practice, and it provides further evidence to support and guide these practices.

Feasibility

These recommendations do not interfere with other interventions or clinical practice therefore it is deemed very feasible in patients with symptomatic knee osteoarthritis.

Future Research

Future research in this area should embrace detailed osteoarthritis characterization including sub-group analyses and osteoarthrosis severity stratification. Furthermore, using clinically relevant outcomes and controls for bias are warranted along with cost-effectiveness analysis. Specifically, to platelet rich plasma it will be of outmost importance to include comprehensive platelet rich plasma characterization and description of platelet rich plasma preparation protocol.

Denervation Therapy
Denervation therapy may reduce pain and improve function in patients with symptomatic osteoarthritis of the knee.
Limited Recommendation Limited Recommendation

The Denervation Therapy recommendation has been downgraded two levels because of inconsistent evidence and bias.

For the denervation therapies, there are 2 high quality studies (Radnovic et al 2017 and Mendes et al 2019) and 2 moderate quality studies (McAlindon et al 2017 and El-Hakeim et al 2018) comparing denervation technique with placebo.

One high quality study (Radnovic et al 2017) specifically evaluated the efficacy of cryoneurolysis in comparison to placebo control in patients with knee OA. It was found that the group receiving cryoneurolysis had improved total WOMAC, WOMAC stiffness, WOMAC pain, WOMAC physical function and in VAS pain compared to placebo control group.

Another high-quality study (Mendes et al 2019) evaluated the efficacy of chemical ablation in comparison to placebo control in patients with knee OA. It was found that the group receiving chemical denervation had improved in WOMAC pain compared to placebo control group. Another moderate quality study (McAlindon et al 2017) comparing the efficacy of chemical ablation in comparison to placebo control in patients with knee OA found no major difference between the two groups.

One moderate quality study (El-Hakeim et al 2018) specifically evaluated the efficacy of thermal ablation in comparison to placebo control in patients with knee OA. It was found that the group receiving thermal ablation had improved WOMAC total, WOMAC function and VAS pain compared to placebo control group.

One high quality study (Davis et al) and one moderate evidence study (Davis et al 2018) compared IA HA to thermal ablation in patients with knee OA.  The first study (Davis et al 2018) showed worse Oxford Knee Score, Global Perceived Index and Numeric Rating Scale in the HA group compared to the thermal ablation group, while the second study (Davis et al 2018) showed worse Oxford Knee Score, Change in Medication Use (mg) from Baseline, Knee pain-Numeric Rating scale and Mean Reduction in average NRS score in the HA group compared to the thermal ablation.

One high quality study (Gulec et al 2017) compared unipolar to bipolar radiofrequency ablation of the knee in patients with knee OA. In patients with OA, Bipolar intra-articular pulsed radiofrequency thermocoagulation may be used over Unipolar intra-articular pulsed radiofrequency thermocoagulation to improve patient pain.

One moderate quality study (Sari et al 2018) compared IA steroids to thermal ablation of the knee in patients with knee OA. The study showed worse WOMAC total, WOMAC function, WOMAC stiffness and worse VAS pain in the IA steroids group compared to thermal ablation group

In summary, our analysis demonstrates that denervation therapy may reduce pain and improve function in patients with symptomatic osteoarthritis of the knee.

Benefits/Harms of Implementation

There are no major known or anticipated harms associated with implementing this recommendation anticipated.

Outcome Importance

Pain and function improvement through denervation therapies for the treatment of knee osteoarthritis may have high impact on symptoms and overall health.

Cost Effectiveness/Resource Utilization

The cost-effectiveness of different denervation therapies is still to be determined, in comparison to other treatment strategies and among different denervation alternatives.

Acceptability

Currently denervation treatments are commonly utilized approaches in treating symptomatic knee osteoarthritis, hence there should be no issues implementing this recommendation as it does not influence a major change in clinical practice, and it provides further evidence to support and guide these practices.

Feasibility

These recommendations do not interfere with other interventions or clinical practice therefore it is deemed very feasible in patients with symptomatic knee osteoarthritis.

Future Research

Future research in this area should embrace detailed osteoarthritis characterization including sub-group analyses, osteoarthrosis severity stratification, and end stage disease in patients unable to have total knee arthroplasty (e.g. due to age or comorbities). Furthermore, using clinically relevant outcomes and controls for bias are warranted along with cost-effectiveness analysis.

Lavage/Debridement (Recommendation Against)
Arthroscopy with lavage and/or debridement in patients with a primary diagnosis of knee osteoarthritis is not recommended.
Moderate Recommendation Moderate Recommendation

There were four studies that met the inclusion criteria for this recommendation.  There was one high strength (Moseley et al 2002), two moderate strength (Kirkley et. Al 2008, Kalunian et. Al 2000), and one low quality (Saeed et.al 2015).

Kirkely et al 2008 compared arthroscopic surgery which, included lavage and debridement combined with physical therapy and medical treatment versus physical therapy and medical treatment alone. The outcome measures utilized were the total Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score, Short Form-36 (SF-36) Physical Component Summary score, McMaster–Toronto Arthritis Patient Preference Disability Questionnaire (MACTAR), and the Arthritis Self-Efficacy Scale (ASES) and standard-gamble utility scores. Six patients assigned to surgery elected not to have the procedure; data from these patients were analyzed, according to the intention to- treat principle, with data from the surgery group. Out of all potential outcomes, only two were statistically significant in favor of surgery. In summary, this randomized controlled trial demonstrated no benefit of arthroscopic lavage and debridement compared to physical therapy and medical treatment for osteoarthritis of the knee.

Kalunian et al. 2000 compared arthroscopic lavage (3000ml) with placebo (250ml). The study was performed at 4 different institutes and included a large number of enrolled patients from one institution with intra-articular crystals in their knee. The arthroscopes used were less than usual caliber in size ranging from 1.7mm to 2.7mm. Outcome measures were WOMAC scores at 12 months. There were not any statistically significant differences in aggregate WOMAC scores between the two treatment groups. The study concludes that irrigation may be helpful in a small subset of patients, especially those with crystals.

Mosley et al 2002 study is an RCT comparing arthroscopic debridement, arthroscopic lavage, versus placebo / sham surgery. The study provides strong evidence that knee arthroscopy with or without debridement is not better and appears to be equivalent to a placebo procedure in improving knee pain and self-reported function. However, the study raised questions regarding its limited sampling (mostly male veterans) as well as the number of potential study participants who declined randomization into a treatment group. They also used a non-validated Knee Specific Pain score. Also, patients with substantial malalignment (varus or valgus deformity) and those with advanced disease, who might have a poorer response to surgical intervention were included in the trial.

Saeed et al 2015 compared HA injections versus arthroscopic debridement in patients with OA in an RCT where only the pain component of the knee society score was utilized. In short term follow up of 6 months, arthroscopy failed to show better pain outcome than injections.

Most of the studies excluded patients with meniscal tear, loose body, or other mechanical derangement, with concomitant diagnosis of osteoarthritis of the knee. The present recommendation does not apply to such patients.

Benefits/Harms of Implementation

Owing to lack of strong evidence in support of clinical benefits from lavage surgery coupled with increased risks from surgery, the workgroup decided to recommend against arthroscopic debridement and/or lavage in patients with a primary diagnosis of osteoarthritis of the knee.

Future Research

Most studies included in this review had compared lavage to variety of nonoperative modalities. However, considering emergence of new modalities and technology it would be worthwhile to do high quality study with arthroscopic lavage done with a standardized size instrument comparing with nonoperative measures which could include PRP injections, nerve ablations, etc. Recent introduction of nanoscopes also warrants further studies of use of such devices in an office setting and comparing it to cost effectiveness of doing procedures in the operating room.

Partial Meniscectomy
Arthroscopic partial meniscectomy can be used for the treatment of meniscal tears in patients with concomitant mild to moderate OA who have failed physical therapy or other nonsurgical treatments.
Moderate Recommendation Moderate Recommendation

The three studies discussed below compare outcomes following arthroscopic partial meniscectomy with physical therapy and demonstrate that knee arthroscopy with partial meniscectomy is as effective as physical therapy.  In PICO 5, this work group recommended supervised or unsupervised exercise as opposed to no exercise to improve pain and function in patients with knee osteoarthritis.  Currently, there are no studies that compare outcomes (knee pain and function) following arthroscopic partial meniscectomy versus physical therapy alone in patients who have failed to improve with an initial course of physical therapy.  It is important to clearly define the appropriate indications for arthroscopic partial meniscectomy in patients with knee OA.  This procedure should be considered in patients with mild-to-moderate knee OA and an MRI confirmed meniscal tear who have previously failed appropriate conservative treatment such as physical therapy, corticosteroid injections, and a course of non-steroidal anti-inflammatory medications.

Katz et al (2013) conducted a multicenter, randomized, controlled trial of symptomatic patients over the age 45 or older with a meniscal tear and evidence of mild-to-moderate knee osteoarthritis to determine efficacy of arthroscopic partial meniscectomy compared to standardized physical therapy in this patient population.   Three hundred fifty-one patients were randomly assigned to surgery and postoperative physical therapy or to a standardized physical therapy regimen (with the option to cross over to surgery at the discretion of the patient and surgeon).  The patients were evaluated at 6 and 12 months and the primary outcome was the difference between the groups with respect to the change in the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) physical function score.  The mean improvement in WOMAC score at 6 months was similar between the groups.  At 6 months, 51 patients who were randomized to physical therapy alone (30%) had undergone surgery.  The authors concluded that in their intention-to-treat analysis, there were no significant differences in functional improvement 6 months after randomization; however, 30% of patients in the physical therapy alone group underwent surgery. These patients were analyzed in their original group, based on the intention-to-treat analysis.  

Van de Graaf et al. (2018) performed a multicenter randomized clinical trial in the Netherlands to determine whether physical therapy is inferior to arthroscopic partial meniscectomy (APM) for improving patient-reported knee function in patients with meniscal tears.  Three hundred twenty-one patients were randomly assigned to APM or a predefined physical therapy protocol. Patients were excluded if they had locking of the knee, prior knee surgery, instability caused by anterior or posterior cruciate ligament rupture, severe osteoarthritis (Kellgren Lawrence score of 4), and a BMI > 35 kg/m2.  Change in patient-reported knee function on the International Knee Documentation Committee Subscale Knee form (IKDC) over a 24-month period was used as the primary outcome.  In the PT group, 47 patients (29%) had APM during the 24-month follow-up period.  The authors noted a similar level of improvement in knee function between the APM and PT groups.  They concluded that PT was noninferior to APM for improving patient-reported knee function over a 24-month follow-up period in patients with nonobstructive meniscal tears. 

In 2007, Herrlin et al performed a prospective randomized study to compare knee function and physical activity following arthroscopic partial meniscectomy followed by supervised exercise or supervised exercise alone in patients with non-traumatic medial meniscal tear.  Ninety patients were evaluated using the Knee Injury and Osteoarthritis Outcomes Score (KOOS), the Lysholm Knee Scoring Scale, and Tegner Activity Scale and a Visual Analog Scale (VAS) for pain prior to the intervention and after 8 weeks of exercise and 6 months following intervention. The authors found that after the intervention, both groups reported decreased knee pain, improved knee function, and a high satisfaction (p<0.0001).  They, therefore, concluded that arthroscopic partial meniscectomy was not superior to supervised exercise alone in terms of reduced knee pain, improved knee function, and improved quality of life. 

Benefits/Harms of Implementation

Given the risks associated with surgical intervention, only appropriately indicated patients should be considered for partial meniscectomy in the setting of mild to moderate knee osteoarthritis. 

Future Research

We did not identify any studies that compare outcomes (i.e. knee pain and function) following arthroscopic partial meniscectomy versus physical therapy alone in patients who have failed to improve with an initial course of physical therapy, non-steroidal anti-inflammatory medications (NSAIDs), and a possible intra-articular steroid injection.  The three studies reviewed by the work group demonstrate that knee arthroscopy with partial meniscectomy is as effective as physical therapy.  Future studies should seek to compare outcomes in patients with mild to moderate knee osteoarthritis and an MRI confirmed meniscal tear who have undergone partial meniscectomy after failing to improve with a course of conservative treatment (NSAIDs, steroid injection, and physical therapy) versus those who have undergone partial meniscectomy without a dedicated course of conservative treatment.

Tibial Osteotomy
High tibial osteotomy may be considered to improve pain and function in properly indicated patients with unicompartmental knee osteoarthritis.
Limited Recommendation Limited Recommendation

The Tibial Osteotomy recommendation has been downgraded one level because of inconsistent evidence.

High tibial osteotomy (HTO) has been used for pain relief of medial compartment knee osteoarthritis. Realigning the varus knee provides mechanical decompression of the medial compartment. An osteotomy line is created in the proximal tibial and either a wedge defect is created by opening the medial cortex and held open with a wedge or plate and screw hardware, or a lateral wedge is removed and secured commonly with staples or wires. In the Nerhus 2017 study, patients continued to show improvement 6 and 12 months post-operatively. Historical studies have reported pain reduction with survival rates approximately 70% at 10 years (“survival” usually interpreted with endpoint conversion to replacement) (van Outeren cites Brouwer 2014 and Niinimaki 2012).

Many studies available for review by the workgroup compared various techniques of osteotomy in randomized studies. Ogawa 2019 found osteotomy distal to the tibial tubercle to be superior to proximal osteotomy for opening wedge procedure with regard to patellofemoral pain. Arthroscopic evaluation of the knee joint at time of osteotomy and second look at time of hardware removal showed less patellar and trochlear degeneration with the distal osteotomy group.

Nerhus 2017 saw no significant functional difference between surgical patients randomized to either opening or closing wedge, with all showing improvement.

Duivenvoorden 2014 reported improved HSS scores from 71 to 81 at 6 years post-op. VAS scores improved from 6.1 baseline to a statistically significant difference at follow up of 4.0 in the opening wedge patients and 3.2 with the closing wedge (albeit no statistical difference between the groups). It should be noted that patients lost to follow up started with a VAS score 6.6, thus tempering analysis of late results.

Brouwer 2006 performed a prospective randomized trial comparing closing wedge and opening wedge techniques. Closing wedge was secured with two surgical staples and opening wedge with a Puddu plate. After one-year VAS score had improved from 6.1 to 3.6.

Van Outeren 2017 is perhaps the closest attempt to a large randomized control trial between surgery and non-operative management. However, this is still not a highest quality randomized control trial. The researchers gathered two different groups of patients at two different hospitals. The first group underwent randomization to valgus bracing versus usual care. The second group was randomized to HTO with either opening wedge or closing wedge osteotomy.  The groups were matched for baseline characteristics. They found HTO more effective in pain reduction compared to non-operative methods. VAS changed from baseline 6.2 in the surgery group to 3.8 post-op. Control group improved from 6.4 to 5.0. Function was improved only in comparison of surgical patients to usual care treatment. 

The Wu 2017 study evaluated people with bilateral OA with pain around medial part of the knee. The more degenerative knee got proximal tibial osteotomy, and the other knee got usual non-operative care. The study authors included a table of individual patient data, which allowed a model that controlled for differences in baseline knee society function scores between the knees to be run. With this model, the odds ratio of achieving satisfactory knee society function scores (defined as score >= 80) with osteotomy vs. non-operative treatment was 7.51(CI 1.094, 51.6).

Benefits/Harms of Implementation

As with any osteotomy surgery, bone healing and hardware complications can occur. Incomplete osteotomy can lead to unexpected fracture at the point of correction (Getgood 2011). Neurologic injury is feared but uncommon. 2 of 35 patients in both closing wedge and opening wedge groups under Nerhus et al. 2017. Duivenvoorden had only one patient with peroneal nerve injury (closing wedge group) and Brouwer only 1 of 92 (also closing wedge group). Hardware removal appeared more common with medial opening wedge techniques. Limb length discrepancy is also increased with the opening wedge technique (Kim et al. 2016). Van Outeren had non-union of the osteotomy site only in 2 opening wedge patients. Duivenvoorden found opening wedge HTO to have more complications. However, non-union of the osteotomy was more common with their closing wedge technique (Duivenvoorden 2014). Brouwer on their other hand concluded their close wedge technique to have achieved more accurate correction with less morbidity.

The increase valgus angle across the knee will change patellar tracking, as noted above by Ogawa et al. 2019. In Song et al. 2012 no patient at baseline reported anterior knee pain. Although no significant difference was seen between their closing wedge and opening wedge groups, 30% of their patients had various levels of anterior knee pain, including 9% with moderate and 6% with severe anterior knee pain (minimum follow up 3 years).

Nerhus 2017 at two years follow up reported only 1 (closing wedge group) of their total 70 patients as revision to arthroplasty. Duivenvoorden 2014 had six year follow up and reported conversion to TKA in 10 of 45 closing wedge HTO patients and 3 of 45 in the opening wedge group.

Outcome Importance

Despite the lack of a true RCT comparing HTO to non-operative management, the studies reviewed by the workgroup all agree with the premise that pain is reduced by HTO. There appear to be relatively equal outcomes whether HTO is achieved with lateral or opening wedge surgical technique. Opening wedge osteotomy distal to the tibial tubercle appears to be preferred, according to the single study examining this aspect of surgical technique (Ogawa 2019).

Cost Effectiveness/Resource Utilization

Hardware removal was relatively common, and the second surgery will add expense to the total endeavor. For reference, Brouwer reported hardware removal in 11 of 47 closing wedge cases and 27 of 45 with opening wedge HTO. Nerhus reported metal removal in 4 of 35 closing wedge cases and 8 of 35 in the opening wedge group.

Conversion to knee arthroplasty is the most expensive late effect of the procedure. As years increase post-op, conversion rates to TKA increase. No studies were identified which adequately addressed the cost-benefit of knee replacement options versus osteotomy.

Acceptability

In select patients with isolated medial compartment arthritis, there appear to be adequate short-term results with regard to pain and function in those patients undergoing surgery.

The quality of HTO results will be dependent upon the skill and experience of the operating surgeon and the cooperation of a patient to understand he or she is entering into a post-op state with altered anatomy and a prognosis imperfectly known in comparison to other options.

Feasibility

Van Outeren rightly questions the benefits of surgical treatment versus brace treatment. Their study had only one-year follow-up and 25 of 60 patients initially in the brace group dropped. They were not optimistic for long term utility and/or compliance of brace treatment.

Future Research

Previous studies have addressed outcomes when converting UKA to TKA and likewise for converting HTO to TKA. Studies looking at functional improvement and pain relief of HTO in direct comparison to knee replacement surgery (total or partial) would be of great value to our collective knowledge base, especially if such a study could track patients from baseline into HTO versus UKA and then long term to TKA. It is unlikely such a study could be performed prospectively over such a long time, yet data from a registry might prove fruitful.

Dry Needling
In the absence of reliable evidence, it is the opinion of the workgroup that the utility/efficacy of dry needling is unclear and requires additional evidence.
Consensus Recommendation Consensus Recommendation

Two high quality studies examining the effectiveness of dry needling in combination with either exercise versus sham dry needling and exercise, (Sanchez, 2019) or dry needling combined with manual therapy and exercise versus manual therapy and exercise alone (Dunning, 2018) were reviewed. Sanchez et al. 2019 found no difference in clinical outcomes of pain or function between treatment groups. In contrast, Dunning et al. found greater improvements in measures of pain and function in the group receiving dry needling. The inconsistency in the results of these studies has prompted the workgroup not to make a recommendation for or against dry needling at this time. Additional evidence will be required before a recommendation can be made.

Future Research

Continued research with larger studies to examine the effectiveness of dry needling for reducing pain and improving function in patients with knee osteoarthritis is warranted.

Free Floating Interpositional Devices
In the absence of reliable or new evidence, it is the opinion of the work group not to use free-floating (un-fixed) interpositional devices in patients with symptomatic medial compartment osteoarthritis of the knee.
Consensus Recommendation Consensus Recommendation

One study met inclusion criteria, and no additional studies were available for review since the prior edition OAK CPG was published.  The single study was a case series and retrospective review of outcomes in patients receiving the surgical intervention for isolated medial compartment OA. The study indicated high reoperation rates in the patients who were followed, with 32% of patients being revised to total knee arthroplasty during the study period.  Regarding pain and functional improvement, the study reported no statistical difference in preoperative and postoperative Knee Society Scores. Given the lack of evidence to support use, the AAOS workgroup modified the grade of this recommendation to consensus, because of the high revision rates in this study, and the potential harm associated with surgical intervention (anesthesia risks, VTE, infection, and reoperation).

Cost Effectiveness/Resource Utilization

No economic analyses or resource utilization studies have been reported on this treatment option.

Future Research

Future research should be aimed at producing level one randomized control trials to define clinical efficacy and risk of complication.


ACKNOWLEDGEMENTS

Guidelines Work Group:

  • Yale Fillingham, MD, Co-Chair
  • Robert H. Brophy, MD, FAAOS, Co-Chair
  • Yogesh Kolwadkar, MD
  • William Macaulay, MD, FAAOS
  • Nicolas S. Piuzzi, MD
  • Lynn Fisher, MD
  • G. Kelley Fitzgerald, PT, PhD, FAPTA
  • Eric Steifel, MD, FAAOS
  • Jorge Chahla, MD, PhD
  • Thomas Trojian, MD
  • Mary K. Mulcahey, MD, FAAOS
  • Andrew W. Ryan, MD, FAAOS
  • Reda Tolba, MD


Non-Voting Members:

  • Gregory Brown, MD, PhD, FAAOS
  • Creighton Tubb, MD, FAAOS


AAOS Staff:

  • Jayson Murray, MA
  • Kaitlyn Sevarino, MBA, CAE
  • Nicole Nelson, MPH
  • Patrick Donnelly, MPH
  • Jennifer Rodriguez
  • Tyler Verity