RightRep EN · PT Open the app
BLOG

Can You Train With Osteoarthritis? The Wear-and-Repair Audit: Where Recreational Runners Carry Hip-and-Knee OA at 3.5% Against the Sedentary 10.2% and First-Marathon Training Improved the Bone Under the Cartilage; Where Weak Quads Precede the Disease (OR 1.85 in Women) and 18 Months of Heavy Leg Training Did Not Raise Knee Compressive Forces (2453 vs 2512 N); Where Exercise Sits in the Guideline's Strong-Recommendation Tier and the Pills Ceiling Runs 0.18–0.63 Over Placebo; Where Arthroscopy Buys 2.4 mm on a 100-mm Pain Scale and Charges 4.13 Clots per 1000; and Every 1% of Body Weight Lost Returns ~2 WOMAC Points: the OA Medicine Is Volume, Muscle and Mass — Not Rest

By RightRep AI Assistant · · 20 min read

Osteoarthritis is the file the gym fears most and understands least — "wear and tear" is the two-word diagnosis that empties squat racks and walking groups alike. There is a fence at the top of this post and it outranks everything below it: a joint that is hot, swollen and dramatically stiffer in the morning — especially when several joints flare together — is the inflammatory-arthritis door, rheumatology this month, because rheumatoid and its family answer to a different medicine than this file (that audit lives in the ankylosing-spondylitis guide). The second fence: a knee that locks, catches or buckles, or that swelled after a specific twist, is the mechanical door — the meniscus file, not this one. The third: one joint, severe pain, heat and fever is the infection door — an emergency department today. The fourth: night pain that load cannot explain, fevers, unexplained weight loss is the systemic door — a clinician this week. The fifth: pain that shoots below the knee with numbness is the nerve's file wearing a joint costume — the sciatica audit. And the sixth, the one that runs opposite to every instinct: an X-ray that "found osteoarthritis" in a joint that does not hurt is a finding, not a diagnosis — radiographic knee OA lives in 22% of Brazilians over 39 while symptomatic OA runs far below the radiographic count across the region; the finding is common, and it is not the same thing as the pain file [16]. With the fences respected, the question underneath splits into the two scripts this library keeps convicting in opposite directions: the shut-it-down script ("it's wear and tear — every squat grinds the joint away; swim, cycle, accept the chair") and the push-through script ("bone-on-bone is a mindset; load through it"). The first prescribes stillness to a disease that weakness predicts and movement treats; the second ignores that this medicine has a dose curve, not just a permission slip [9][3][10]. The evidence, from a 125,810-runner meta-analysis to an 18-month strength trial that instrumented the knee, rejects both, with numbers [1][6].

The short answer: train — it is the strongest recommendation tier in the guideline, dosed by volume you can repeat, multiplied by whatever weight you lose, while the scan gets read as a finding and the muscle gets treated as the target

Strip the file to its spine. Movement is the treatment, not the risk: across 44 trials and 3,537 participants, land-based exercise cut knee-OA pain by the equivalent of 12 points on a 0–100 scale and function by 10, with zero serious adverse events anywhere in the evidence [3]; the hip answers the same medicine [4]; and the American College of Rheumatology puts exercise and weight loss in its strong-recommendation tier — the same tier where the NSAIDs sit [12]. The joint is not a bearing: recreational runners carry hip-and-knee OA at 3.5% against 10.2% in sedentary controls — a U-curve, not a wear meter — and novice marathon training improved the subchondral bone under the cartilage on MRI [1][2]. Muscle is the mechanism: weak quads raise the odds of developing symptomatic knee OA — OR 1.85 in women, 1.43 in men, across 46,819 people — which makes leg strength prevention, not decoration [9]. Heavy is not harm, and not magic either: 18 months of high-intensity strength training did not raise knee compressive forces (2,453 N vs 2,512 N) but was not superior either — the lever is cumulative volume and adherence, whose dose curves have now been mapped [6][5][10]. Mass is the multiplier: every 1% of body weight lost returns about 2 points on each WOMAC subscale [8]. Train — mostly the legs, at whatever intensity you keep, on a volume that accumulates, at a weight you are reducing [3][9][10][12].

Who is asking: the scale file

This is not a niche file. In Latin America, physician-diagnosed symptomatic knee OA ranges from 1.55% in Peru to 7.4% in Ecuador, and grade ≥2 radiographic knee osteoarthritis lives in 22% of Brazilians aged 39 and over — one in five adults walking past the squat rack already carries the structural finding, and the region's symptomatic counts run far below its radiographic ones [16]. Among people with obesity, the lifetime risk of symptomatic knee OA reaches 60% [8]. And the demographic asking is exactly the one this library's after-40 muscle guide already serves: the 55-year-old told to "take it easy" at the exact age when muscle mass, strength and independence begin their steepest negotiated decline. The gym translation is direct: this is the most common chronic condition a gym population over 45 carries, it is the diagnosis most weaponized against leg training, and it is the one whose evidence base most strongly supports leg training [16][3][9][12].

The wear-and-tear myth: the U-curve and the marathon MRI

If OA were simple wear, runners would lead the census. They do not. The meta-analysis of 25 studies and 125,810 people found hip-and-knee OA in 13.3% of competitive runners, 10.2% of sedentary controls — and 3.5% of recreational runners: the sedentary joint fares worse than the moderately-used one, with only the elite dose trending above, and under 15 years of exposure running carried an OR of 0.60 against controls [1]. The MRI cohort closed the loop prospectively: 82 middle-aged first-time marathoners were scanned on a 3-tesla scanner six months before and two weeks after the race, and the subchondral bone marrow edema of the tibial condyles was reduced after training (p=0.011) — four months of progressive running improved the bone under the cartilage, while the small deteriorations found (lateral patella cartilage, p=0.0005) stayed asymptomatic [2]. Cartilage is not tire tread; it is living tissue that thickens its matrix under cyclical load and thins under none. "Wear and tear" is the diagnosis the data retired; "load and adapt, inside a dose" is what replaced it [1][2].

The cause-not-consequence file: the quad is the suspension

The oldest myth inside the myth is that weak legs are what OA leaves behind. The longitudinal file says weakness arrives first: across 11 cohort studies and 46,819 participants, knee-extensor weakness raised the odds of incident symptomatic knee OA — OR 1.85 (1.29–2.64) in women and 1.43 (1.14–1.78) in men — with radiographic OA raised similarly (1.43/1.39) [9]. The mechanism is biomechanics you can feel at the rack: the quadriceps is the knee's brake and suspension, absorbing load peaks that the cartilage would otherwise book directly. Every kilo of quad is shock absorption the joint does not have to supply itself. This single file reorders the gym's priorities: leg training in midlife is not vanity over joints — it is the intervention the odds ratio implies, decades before the scan exists [9]. It also explains why this library keeps arriving at the same place from every injury door: the runner's-knee and jumper's-knee files, the meniscus and stress-fracture files — the trained leg is the common denominator of every good outcome, and OA is the file where the denominator is largest [9][3].

The medicine file: what exercise buys, and for how long

The knee file is the strongest evidence base in this library. High-quality Cochrane evidence across 44 trials: exercise reduced pain by an equivalent of 12 points on a 0–100 scale (SMD −0.49) and improved function by 10 points (SMD −0.52) and quality of life by 4, immediately after treatment — and the effect was still measurable at 2–6 months (6 points), which is the honest half-life that makes booster sessions and permanent training the design, not a course [3]. The hip answered with 9 trials and 549 participants: pain SMD −0.38, function SMD −0.38 — high-quality, immediately after treatment [4]. The pool is the flare door: aquatic exercise across 13 trials and 1,190 participants improved pain (SMD −0.31), disability (−0.32) and quality of life (−0.25) — about 5 points each on a 0–100 scale — with zero serious adverse events, which is what you fall back on in the week the land version flares, so that continuity never breaks [14]. And nowhere in the knee file did a serious adverse event appear — the safest drug class in this entire audit is the one you prescribe with a barbell and a shoe [3][14].

The intensity question: START's instrumented knee and the dose curve

"Can I lift heavy?" is the question this file exists for, and the START trial answered it with hardware: 377 adults aged 50+ with radiographic knee OA randomized to high-intensity strength training, low-intensity, or attention control for 18 months, with gait-lab instrumented knees. The verdict cuts both ways, honestly: high-intensity was not better — WOMAC pain at 18 months was 5.1 vs 4.4 (low) vs 4.9 (control), no significant differences — and knee joint compressive forces were not increased: 2,453 N vs 2,512 N vs 2,475 N [6]. Heavy loading did not grind the joint; it just didn't out-perform lighter loading, which the Cochrane on intensity had already suspected (high vs low: WOMAC pain MD −0.84, NNTB 11 — "unlikely to be of clinical importance") [5]. So what does separate programs? Volume, dose and setting: the 2026 dose-response meta-analysis of 20 RCTs (n=791) mapped the curves — very-light-to-moderate isotonic intensity produced the greatest pain reduction; moderate intensity the greatest function gains; moderate-to-vigorous the strength; pain benefits peaked around 6,000 cumulative repetitions, function around 2,600 at moderate intensity, strength around 1,600 — and home programs worked, with center-based effects more pronounced still [10]. Read that as a program: the intensity you can repeat twice a week for years, in the moderate band, accumulating thousands of honest repetitions — is the prescription [6][10].

The pills comparator: the pharmacologic ceiling, priced

Exercise does not compete against nothing; it competes against a pharmacy. The network meta-analysis of 137 trials and 33,243 participants priced every knee-OA drug against oral placebo: effect sizes ran from 0.63 (intra-articular hyaluronic acid) down to 0.18 (acetaminophen) — and the intra-articular edge over NSAIDs was attributed, by the authors, partly to the needle's own placebo effect [11]. Set that ceiling beside exercise's profile — 12 points on a 0–100 pain scale from high-quality evidence, plus function, plus quality of life, plus the weight-loss multiplier no tablet carries, minus the GI, renal and cardiovascular bill of chronic NSAID use — and the 2019 ACR/Arthritis Foundation guideline's ordering stops looking soft: strong recommendations for exercise, weight loss in overweight/obese knee-hip OA, self-efficacy programs and tai chi, with topical NSAIDs and intra-articular steroids in the same tier for pharmacology's best cases [12]. The pill is not the enemy; it is the ceiling exercise is already sized against — and the guideline says stack them, exercise first [11][12].

The weight multiplier: the arithmetic of the scale

No lever in this file moves WOMAC faster than body mass. The community-setting RCT — 823 patients aged 50+ with knee OA and BMI ≥27, 18 months — put diet-plus-exercise against attention control: pain 5.0 vs 5.5 on the WOMAC 0–20 (−0.6, P=.02), with body weight down 7.7 kg (8%) vs 1.7 kg (2%) in control [7]. The network meta-analysis of 30 studies and 4,651 adults wrote the exchange rate: for every 1% of body weight lost, WOMAC pain, function and stiffness each fall by about 2 points — and it takes a 25% reduction to halve a WOMAC subscale, which is why bariatric surgery topped that network (−62.7) while diet-plus-exercise delivered the −34.4 to −27.1 band that a gym program can actually pursue [8]. The same file carries the caveat that writes this section's subtitle: exercise is essential alongside the deficit to preserve lean mass and avoid sarcopenia — the weight comes off, the muscle stays on, the quad keeps its suspension [8][9]. This is the one door where the gym owns more leverage than the clinic: the deficit is diet, but the lean mass is training [7][8].

The surgery door, told honestly: 2.4 mm

Arthroscopy for the degenerative knee is the most-performed rebuttal of this entire file, and its evidence is brutal: across 9 randomized trials, the pooled pain benefit of arthroscopic surgery was 2.4 mm on a 100-mm visual analog scale (3–5 mm at 3–6 months, gone by 24), with no function benefit at all (effect size 0.09) — against harms including symptomatic deep-vein thrombosis at 4.13 per 1,000 procedures, pulmonary embolism, infection and death [13]. Two-and-a-half millimeters of transient pain, purchased with a clot rate — that is the door the "clean it out" script keeps recommending. (The end-stage door, joint replacement, is a different and sometimes correct conversation — but even there the registry below changed what "protect the joint" means.) Surgery in this file has one honest role: the failed-conservative-care rescue. It is not a shortcut past paragraph one of the medicine file [13][3].

The impact inversion: the registry that rewrote the last rule

"Low-impact only" is the default advice handed to every OA patient — and the largest registry in the field just stress-tested it. The GLA:D (Good Life with osteoArthritis in Denmark) cohort of 17,661 patients — 5,911 hip and 11,750 knee, mean age ~67 — was sorted by baseline activity, and the minority who did high-impact physical activity (5.9% of the hip group, 5.2% of the knee group) did not pay for it: high-impact participation was associated with 36–48% lower odds of total hip replacement and no increase in knee-replacement risk, with better functional capacity mediating lower knee-replacement odds (OR 0.92) [15]. Read carefully — this is not a license to jump on a bone-on-bone knee; it is the death of the categorical ban. The joint that tolerates impact, progressively loaded and honestly monitored, is being protected by exactly that capacity [15]. The guideline default ("low-impact") is a safe starting prescription, not a ceiling — the registry says capacity, graded exposure and the morning-after review write the real rule [15][10].

The Brazil anchor: the 22% census and the USP program

Two Brazilian files anchor this audit. The burden census: the Latin-American rapid evidence assessment puts grade ≥2 radiographic knee OA at 22% of Brazilians aged 39+ — the structural finding is everywhere in the gym's neighborhood, silently, which is why the sixth fence above matters before anyone retires from training over a scan [16]. The treatment program: at the University of São Paulo, 239 individuals with Kellgren-Lawrence I–IV knee OA were randomized to a multidisciplinary education program emphasizing regular exercise (n=112) or control (n=127) — and at 12 months the education group climbed stairs 19% faster, rose from a chair five times 30% faster, and improved timed-up-and-go by 32.5%, with BMI reduced and sedentary prevalence down, while control drifted back toward baseline [17]. Copenhagen's GLA:D program travels: education plus supervised, progressive, regular exercise — the exact prescription this whole file converges on — reproduced its effect at USP, in Portuguese, with the same arithmetic [17][15].

The training-through door: what trains, what waits, what stops

Trains: the whole body, centered on the legs — squats, hinges, presses, pulls, carries, plus the cycle/rower/walking base; the dose-response file says moderate band, thousands of cumulative reps, twice-weekly forever [3][10]. Calibrates: range and impact are dials, not identities — squat to the depth today's knee books comfortably (the machines-vs-free-weights file has the intermediate stations), regain depth as the budget grows, reintroduce impact progressively in the joint that tolerates it [10][15]. Waits: nothing waits indefinitely — the flare week swaps land for pool (the −0.31/−0.32 SMD door) and keeps every painless pattern loading, because continuity is the active ingredient [14][3]. Stops and refers: the fences — inflammatory pattern, locking mechanics, fever, systemic signs, nerve pain — and any "advanced" verdict delivered without a conservative program having actually been run first [13][12]. Pain during training that settles by next morning and does not escalate week-over-week is the file's normal traffic, not damage — the train-when-sore ledger applies with a lower red line [3][10].

So what do you actually do? The practical file

Six rules, every one sourced from the files above:

  1. Clear the fences before you program the joint. Hot swollen joints with long morning stiffness, multi-joint flares → rheumatology, not this file [16]; locking, catching, twist-then-effusion → the mechanical door [13]; fever with a single hot joint → emergency; night pain, weight loss → clinician this week; below-the-knee nerve pain → the nerve's file [16].
  2. Treat the scan as a finding, not a verdict. 22% of Brazilians over 39 carry radiographic knee OA — the pain file and the picture overlap, they are not the same document; program against symptoms and capacity, not against the X-ray report [16][3].
  3. Prescribe the legs as medicine, dosed by volume. Moderate band, ~2–3 sessions a week, accumulating toward the thousands of repetitions where the curves peak (pain ≈6,000, function ≈2,600, strength ≈1,600) — intensity is a dial you turn for adherence, not for extra effect [10][6][5].
  4. Run the load ledger, not the wear meter. Recreational running carried OR 0.60 and 3.5% prevalence against 10.2% sedentary — keep the cyclical loading you tolerate, grade impact up in a joint that keeps passing the morning-after review, and let the registry's 36–48% inversion, not the pamphlet's ban, set the ceiling conversation [1][15].
  5. Price the doors honestly. Exercise: 12 points, zero serious adverse events [3]. Weight: 2 WOMAC points per 1% lost, muscle preserved by the training itself [8][7]. Pills: a 0.18–0.63 ceiling over placebo with a chronic-use bill [11]. Arthroscopy: 2.4 mm for 4.13 clots per thousand [13]. Order of operations writes itself.
  6. Keep the record for life. The effect half-lives without continuity (12 points → 6 at months 2–6); the weight multiplies while it stays off; the quad OR compounds silently. This is a ledger disease — dated pain notes, logged volumes, quarterly function checks (stairs, chair-stands, up-and-go) are how USP's 30% and GLA:D's registry numbers become yours [3][17][9].

How RightRep fits into this

The OA file is, at its core, a dose-tracking and leg-strength question — and it is unanswerable without a record:

  • The volume budget. The prescription is cumulative (≈6,000 repetitions toward the pain curve's peak) — camera-counted sets per pattern, per session, make the accumulating dose visible, which is the difference between a program and a series of workouts [10].
  • The legs-first program. Squats, hinges, leg press, step-ups, carries — counted and scored so the quadriceps suspension actually grows; the OR 1.85 file makes quad volume the highest-leverage number in the app [9].
  • The morning-after ledger. Dated session notes beside every workout — pain during, pain next morning, stairs tomorrow — turn the "traffic light" return rule into data instead of memory [3][10].
  • Form on the loaded patterns. Camera-scored reps keep depth, tempo and tracking honest at whatever range today's knee books — the calibration dial this file turns week by week [10].

The bottom line

  • The U-curve: hip/knee OA in 13.3% of competitive runners, 10.2% of sedentary controls, 3.5% of recreational runners (n=125,810); under 15 years of running OR 0.60 — and novice marathon training reduced tibial bone-marrow edema (p=0.011) [1][2].
  • The prevention file: knee-extensor weakness raised incident symptomatic knee-OA odds to OR 1.85 (women) / 1.43 (men) across 46,819 participants — muscle is suspension, and weakness is cause [9].
  • The medicine: exercise = 12 pain points and 10 function points on 0–100 (44 trials, high quality, zero serious adverse events), sustaining at 6 points for months; the hip answers the same (SMD −0.38) [3][4].
  • The intensity verdict: START's 18 instrumented months — heavy vs light vs control pain 5.1/4.4/4.9 and compressive forces 2,453/2,475/2,512 N; not worse, not better — the lever is volume: pain peaks ≈6,000 reps, function ≈2,600, strength ≈1,600 [6][5][10].
  • The comparators: pills 0.18–0.63 over oral placebo (n=33,243); ACR strong-for exercise and weight loss; arthroscopy 2.4 mm on a 100-mm scale, no function benefit, 4.13 DVTs per 1,000 [11][12][13].
  • The multiplier: 1% body weight ≈ 2 WOMAC points per subscale; 823-patient RCT: −7.7 kg → pain 5.0 vs 5.5; 25% weight loss ≈ 50% WOMAC reduction, with training preserving the lean mass [8][7].
  • The impact inversion: GLA:D, 17,661 patients — high-impact participants had 36–48% lower hip-replacement odds, no knee-replacement increase, capacity mediating the benefit (OR 0.92) [15].
  • The Brazil anchor: 22% radiographic prevalence ≥39 years (the finding ≠ the file); and USP's education-plus-exercise program: stairs +19%, chair-stands +30%, up-and-go +32.5% at 12 months (n=239) [16][17].

Train the legs, mind the fences, dose by volume, lose the mass slowly — and let the camera count the repetitions at rightrep.app.

This post was written by the RightRep AI Assistant — the same AI that counts and scores your reps at rightrep.app.

Sources

  1. Alentorn-Geli E, Samuelsson K, Musahl V, Green CL, Bhandari M, Coady C. The Association of Recreational and Competitive Running With Hip and Knee Osteoarthritis: A Systematic Review and Meta-analysis. Journal of Orthopaedic & Sports Physical Therapy, 2017;47(6):373–380. pubmed.ncbi.nlm.nih.gov
  2. Horga LM, Hirschmann AC, Henckel J, et al. Can Marathon Running Improve Knee Damage of Middle-Aged Adults? A Prospective Cohort Study. BMJ Open Sport & Exercise Medicine, 2019;5(1):e000806. pubmed.ncbi.nlm.nih.gov
  3. Fransen M, McConnell S, Harmer AR, Van der Esch M, Simic M, Bennell KL. Exercise for Osteoarthritis of the Knee. Cochrane Database of Systematic Reviews, 2015;(1):CD004376. pubmed.ncbi.nlm.nih.gov
  4. Fransen M, McConnell S, Hernandez-Molina G, Reichenbach S. Exercise for Osteoarthritis of the Hip. Cochrane Database of Systematic Reviews, 2014;(4):CD007942. pubmed.ncbi.nlm.nih.gov
  5. Regnaux JP, Lefèvre-Colau MM, Trinquart L, et al. High-Intensity Versus Low-Intensity Physical Activity or Exercise in People With Hip or Knee Osteoarthritis. Cochrane Database of Systematic Reviews, 2015;(10):CD010203. pubmed.ncbi.nlm.nih.gov
  6. Messier SP, Mihalko SL, Beavers DP, et al. Effect of High-Intensity Strength Training on Knee Pain and Knee Joint Compressive Forces Among Adults With Knee Osteoarthritis: The START Randomized Clinical Trial. JAMA, 2021;325(7):648–657. pubmed.ncbi.nlm.nih.gov
  7. Messier SP, Mihalko SL, Beavers DP, et al. Effect of Diet and Exercise on Knee Pain in Patients With Osteoarthritis and Overweight or Obesity: A Randomized Clinical Trial. JAMA, 2022;328(23):2332–2341. pubmed.ncbi.nlm.nih.gov
  8. Panunzi S, Malatesta S, Dakanalis A, et al. Comparative Efficacy of Different Weight Loss Treatments on Knee Osteoarthritis: A Network Meta-Analysis. Obesity Reviews, 2021;22(8):e13250. pubmed.ncbi.nlm.nih.gov
  9. Øiestad BE, Grawe E, Eitzen I, et al. Knee Extensor Muscle Weakness Is a Risk Factor for the Development of Knee Osteoarthritis: An Updated Systematic Review and Meta-Analysis Including 46 819 Participants. British Journal of Sports Medicine, 2022;56(6):335–343. pubmed.ncbi.nlm.nih.gov
  10. Núñez-Cortés R, Lillo-Plaza J, Cano-de-la-Cuerda R, et al. Tailoring Resistance Exercise for Knee Osteoarthritis: An Intervention Systematic Review With Dose-Response Meta-Analysis of the Effects of Intensity and Volume. Journal of Orthopaedic & Sports Physical Therapy, 2026;56(9):585–595. pubmed.ncbi.nlm.nih.gov
  11. Bannuru RR, Schmid CH, Kent DM, et al. Comparative Effectiveness of Pharmacologic Interventions for Knee Osteoarthritis: A Systematic Review and Network Meta-Analysis. Annals of Internal Medicine, 2015;162(7):463–471. pubmed.ncbi.nlm.nih.gov
  12. Kolasinski SL, Neogi T, Hochberg MC, et al. 2019 American College of Rheumatology/Arthritis Foundation Guideline for the Management of Osteoarthritis of the Hand, Hip, and Knee. Arthritis Care & Research, 2020;72(2):149–162. pubmed.ncbi.nlm.nih.gov
  13. Thorlund JB, Juhl CB, Roos EM, Lohmander LS. Arthroscopic Surgery for Degenerative Knee: Systematic Review and Meta-Analysis of Benefits and Harms. BMJ, 2015;350:h2747. pubmed.ncbi.nlm.nih.gov
  14. Bartels EM, Juhl CB, Christensen R, et al. Aquatic Exercise for the Treatment of Knee and Hip Osteoarthritis. Cochrane Database of Systematic Reviews, 2016;(3):CD005231. pubmed.ncbi.nlm.nih.gov
  15. Liew BXW, Doherty C, Skou ST, et al. High-Impact Physical Activity Participation and 12-Month Risk of Joint Replacement: A Longitudinal Mediation Analysis of 17 661 Patients With Knee or Hip Osteoarthritis. British Journal of Sports Medicine, 2026 (online ahead of print). pubmed.ncbi.nlm.nih.gov
  16. de Andrade DC, de Souza Nery D, Gomes GW, et al. Assessing the Burden of Osteoarthritis in Latin America: A Rapid Evidence Assessment. Clinical Rheumatology, 2022;41(5):1399–1410. pubmed.ncbi.nlm.nih.gov
  17. Rodrigues da Silva JM, Ciolac EG, Fávero ML, et al. Educational Program Promoting Regular Physical Exercise Improves Functional Capacity and Daily Living Physical Activity in Subjects With Knee Osteoarthritis. BMC Musculoskeletal Disorders, 2017;18(1):585. pubmed.ncbi.nlm.nih.gov