Osteoarthritis — Body Handbook

Musculoskeletal · §161

Osteoarthritis

Cartilage isn't a tyre. The advice you've absorbed — take it easy, save what's left, eventually get it replaced — is the wrong story, and it's costing the version of your fifties and sixties you could otherwise have. Osteoarthritis is now understood as an actively managed chronic disease, not a slow surrender, and the single best-supported treatment is the one most people are quietly told to avoid: loading the joint, on purpose, progressively, for the rest of your life. Done well, pain falls a quarter to a third inside three months, sleep stitches back together, the long walk comes back, and the trajectory of your next twenty years bends. Done by avoidance, the joint quietly takes the walk away — and walking ability in your seventies is one of the cleanest predictors of how the rest of life goes.

Respond · Daily Evidence Strong Chapter Musculoskeletal

The core plan is unglamorous and overwhelmingly evidence-backed: load the joint with strength and movement most days, drop weight if you carry extra, use topical anti-inflammatories when you need them, and treat surgery as a powerful tool you schedule on your own terms — not a last resort you stall into. Most of the rest of what you've heard about — supplements that rebuild cartilage, repeated steroid shots, platelet injections, arthroscopic clean-outs — either doesn't work or works briefly at a cost. The honest catch is that the work is daily and the payoff is months out; nothing about this is a quick win.

The old picture of osteoarthritis was a tyre wearing thin: the cushion between two bones — the cartilage — gets thinner with use, and once it's gone, it's gone. That picture is wrong in a way that matters for what you do every day. The cartilage isn't passively worn down; the whole joint becomes inflamed and starts misbehaving. The bone underneath the cartilage remodels itself, the lining of the joint runs a low-grade inflammation, the small muscles around the joint weaken, and the pain you feel doesn't come from the cartilage at all — cartilage has no nerves. It comes from the synovial lining, the bone underneath, the capsule, the ligaments, the muscles that try to brace a joint that's lost its rhythm Hunter & Bierma-Zeinstra 2019.

The practical consequence: a joint with bad-looking x-rays can be quiet, and a joint with mild-looking x-rays can hurt severely. Pain doesn't track the picture. It tracks how inflamed the lining is, how irritable the bone underneath is, how strong and coordinated the surrounding muscles are. All three are things you can actually move.

What you feel: a deep ache that gets worse with use and better with rest, a few minutes of stiffness in the morning or after sitting (under half an hour — long morning stiffness points to a different kind of arthritis), a grinding or clicking inside the joint, sometimes swelling. Knees, hips, hands, and the base of the thumb are where it shows up most. Hip osteoarthritis often broadcasts itself as groin pain or a knee that hurts even though the knee is fine — a referred pattern worth knowing.

What actually works — and how settled it is

Four major bodies have written guidelines in the last several years and they agree on the spine of the plan: structured exercise, weight management if you carry extra weight, judicious use of anti-inflammatory medication, joint replacement when conservative care stops working OARSI 2019 ACR 2019 NICE 2022 AAOS 2021. NICE went further than the others and made therapeutic exercise the explicit first thing to try, ahead of any pill.

The exercise evidence isn't a single study to argue about. It's a Cochrane review of fifty-four randomised trials in five thousand patients, all coming to the same place: pain falls by about a quarter on average, function improves by a similar amount, and the gains hold for months after the program ends as long as the person keeps moving Fransen et al. 2015. The effect size is in the same ballpark as a non-steroidal pill, without the gut, kidney, and heart cost a non-steroidal pill carries at scale.

The weight evidence is similarly hard to argue with. Each kilogram you carry adds roughly four kilograms of force across the knee when you walk. A network meta-analysis of weight-loss trials in knee osteoarthritis found that for every one percent of body weight lost, pain, stiffness, and function each improved by about two percent Panunzi et al. 2021. Diet and exercise together do more than either alone — the IDEA trial, which randomised 454 overweight adults with knee osteoarthritis for eighteen months, showed the combination cut knee compressive forces by more than two hundred newtons per step compared with exercise alone, and produced bigger gains in pain, function, mobility, and quality of life than either intervention solo Messier et al. 2013. The two levers compound; they don't substitute.

Joint replacement, for end-stage disease, is one of the highest-value operations in modern medicine. Pooled registry data on knee replacements: about 93 percent still in place at fifteen years, 90 percent at twenty, 82 percent at twenty-five Evans et al. 2019. Hip replacements with modern bearings extrapolate to roughly the same range out past twenty-five years Evans et al. 2019. The pain and function gains for someone with severe end-stage disease are dramatic — these aren't marginal procedures.

What the next ten years look like if you wait it out

The under-managed version of osteoarthritis is rarely a single bad event. It's a slow tightening of the radius of your life. The walk that used to be an hour becomes the loop around the block. The stairs become the elevator. The trip you were going to take stops being scheduled because the airport's a lot of walking. Your knee or hip wakes you up at 2am most nights because lying still hurts in the second half of sleep Hunter & Bierma-Zeinstra 2019. The afternoon you used to give to the garden becomes the afternoon you give to the chair. You don't notice the decade go this way; it happens in quarter-percent increments.

People around you start to register it before you do. Your partner offers you the seat in the cafe. Your kids start checking with each other before suggesting the walk after dinner. Your friends stop including you in the things that involve being on your feet for three hours. You become someone with a bad knee, and the bad knee becomes the size of your week.

The metabolic spiral is real. Pain takes the activity, which takes the fitness, which takes the muscle, which takes the metabolic flexibility, which takes the weight back from any progress you made. Joint pain is one of the leading reasons adults over forty-five lose sleep in the second half of the night, and broken sleep is its own engine for mood, attention, and weight. The cognitive bandwidth that chronic pain steals is large and largely unmeasured by the person whose head it's in.

The mortality signal is the line you don't get back. In a population study of more than seventeen thousand adults with hip or knee osteoarthritis, the difficulty-walking score predicted death over follow-up, and people who'd started using a walking aid were about half again as likely to die over the follow-up window as those who hadn't Hawker et al. 2014. The arrow points the other way too: maintained physical function tracks with cardiovascular fitness, with falls prevention, with the muscle that lets you stay in your own home into your eighties. This is the part of the entry that earns its loss-aversion framing. A joint isn't just a joint — it's the gate to the activity that runs the body.

The plan, in order

There's a specific sequence that does the work, and it's worth stating plainly because most people are told a vaguer version that lets them slide.

Load the joint progressively, most days. Aim for thirty to sixty minutes, three to five times a week. Strength work for the muscles around the affected joint is the most important piece — quadriceps and hip strength for the knee, hip abductors and trunk for the hip, grip and intrinsic hand work for the thumb. Add aerobic work (walking, cycling, swimming, elliptical — pick what your joint tolerates) and balance work. Twelve weeks is the minimum dose to feel a real shift. The pattern that works at scale — twelve supervised sessions over eight weeks plus a small amount of education up front — is delivered as a structured program called GLA:D in countries that have built it out; in the US ask a physiotherapist for a knee or hip OA program with the same structure.
If you're carrying extra weight, lose some. Five to ten percent of body weight is the threshold where the evidence converts cleanly to symptom relief. For knee OA the math is roughly four kilograms of force across the knee per kilogram of body weight; the inflammation lever runs in parallel.
Use topical anti-inflammatories as needed. Topical diclofenac gel or solution gives knee and hand OA pain relief comparable to swallowed non-steroidals, with a fraction of the gut, heart, and kidney exposure Derry et al. 2016. The 2019 American College of Rheumatology guideline strongly recommends them, ahead of oral pills, for that reason Kolasinski et al. 2020.
Add oral non-steroidals for flares, not continuously. Effective; carry real risks above age sixty or with kidney, gut, or cardiovascular issues. Lowest dose, shortest duration that does the job. A talk with your doctor about COX-2 selectivity and gastric protection is worth having.
Use a corticosteroid injection sparingly — for a specific reason. A wedding next month. A holiday. A patch of life where you can't load the joint and need to. One injection bridges. Don't accept a schedule of one every three months indefinitely — that pattern was tested in a two-year randomised trial and showed greater cartilage loss than saline, with no better pain control at two years McAlindon et al. 2017.
Schedule joint replacement when it makes sense, not when you're forced to. Severe pain that interferes with sleep, walking distance dropping, function declining despite real engagement with steps 1–4: that's the point where surgery converts to a high-value option. Going in fit — strong quadriceps, decent cardiovascular shape — produces dramatically better recoveries than going in deconditioned after years of avoidance.

The single hardest thing about this list is that the work is daily and the payoff is months out. Most people are looking for the version that works in a week. There isn't one. The closest thing to a quick win is a topical gel; the rest is patient, accumulating.

What most advice gets wrong

"Don't exercise — you'll wear it out faster." The most damaging single thing patients are still told. Loading the joint progressively is the best-studied non-drug treatment we have; the bigger risk by far is the avoidance cascade — weight gain, weakness, sleep loss, deconditioning Fransen et al. 2015. Some discomfort during loading is expected and does not mean damage. A sharp catching pain is different and worth checking, but the dull ache that fades after movement is the signal you're on the right side of the curve.

"There's a supplement that rebuilds cartilage." Glucosamine and chondroitin, alone or together, were tested head-to-head against placebo and a real anti-inflammatory in the NIH's GAIT trial (n=1,583): they didn't beat placebo on the primary outcome Clegg et al. 2006. The 2019 ACR guideline strongly recommends against them Kolasinski et al. 2020. Collagen peptides are in the same general territory: marketing pressure, soft evidence. Money better spent on a physiotherapist.

"My x-ray is bad, so I need surgery now / mild, so I'm fine." The picture and the pain don't track each other reliably. Severe radiographic disease can be silent; mild radiographic disease can be debilitating. Decisions about surgery should be made on function and quality of life, not on the image — the image is one piece of context, not the verdict.

"Joint replacement is a last resort to put off as long as possible." Modern arthroplasty is one of the most reliable operations going — about ninety percent of knee and hip replacements are still in place at fifteen to twenty years Evans et al. 2019 Evans et al. 2019. Stalling into a deconditioned body with severe quadriceps loss produces worse recoveries than going in fit. "When it makes sense" is a better decision rule than "as late as humanly possible."

"Osteoarthritis is just for the elderly." Half of people who tear an ACL or have substantial meniscus damage in their twenties or thirties have radiographic knee OA inside ten to twenty years, regardless of whether they had surgery Lohmander et al. 2007. If you've had a meaningful joint injury, you have a stake in this protocol now, not in twenty years.

Where this goes wrong in practice

The injection treadmill. A steroid shot every three months feels like managing the disease. It's not — it's managing today's pain with a tool that wears off in weeks and may make the underlying structure worse over years when used repeatedly McAlindon et al. 2017. The repeated injection slot is where the exercise and the weight work should have been. If you find yourself on the every-three-months schedule, the question to ask is what would have to change for that to stop being the only intervention you're getting.

Loading shy of the dose that does anything. Most patients who "tried physical therapy and it didn't work" did six sessions of gentle range-of-motion and then went home. The dose that's been tested is twelve sessions of supervised progressive strength and neuromuscular work over eight weeks, with maintenance after Skou & Roos 2017. If your program looks lighter than that, you're not running the experiment the evidence describes.

The expensive biologic detour. Platelet-rich plasma injections cost between five hundred and a couple thousand dollars per cycle, are usually not covered by insurance, and in the highest-quality randomised trial — RESTORE, n=288, three weekly PRP versus saline injections, twelve-month follow-up — did not significantly reduce pain or slow cartilage loss compared with saltwater Bennell et al. 2021. Stem-cell injections are in worse evidence territory and cost more. The marketing is loud; the trials are not.

The arthroscopy that didn't help. Arthroscopic clean-outs for osteoarthritis and arthroscopic trims for the degenerative meniscus tears that often accompany OA were both tested against sham surgery in randomised trials — a kind of test that's hard to do and impossible to fake. Neither beat the sham Moseley et al. 2002 Sihvonen et al. 2013. NICE, the AAOS, and the ACR all recommend against them NICE 2022 AAOS 2021 Kolasinski et al. 2020. The procedure persists in some places anyway.

Stalling surgery into a deconditioned body. The "I'm not ready" patient who avoids surgery for five extra years by sitting more and walking less arrives at the operating table with poor quadriceps, lost cardiovascular fitness, and added weight. The operation still works, but the recovery is harder and the outcomes are worse. Earlier-than-you-think can be better than later-than-you-thought.

Joint-specific notes

Most of this entry is the knee story because that's where the burden is largest and the trials are deepest. The other joints aren't identical.

Hip. Same protocol applies — progressive loading is the spine — but the weight-loss effect is smaller than for the knee, and the pain often broadcasts as groin pain or as a knee that hurts when the knee itself is fine. Hip replacement is the most reliable joint replacement we do, with thirty-year survivorship now extrapolating into the ninety-percent range with modern implants Evans et al. 2019.

Hand and base of the thumb. Affects roughly one in five people over fifty-five, with a clear female predominance and a marked rise through life. The plan shifts: splinting of the affected joint plus targeted hand-therapy exercise is the mainstay, with topical anti-inflammatories layered on for flares; the weight argument doesn't apply because hands don't carry your body weight, but the inflammation argument still partly does Kolasinski et al. 2020. Surgery for the thumb base is a different operation from a knee or hip replacement and is reserved for severe cases.

Post-injury (any joint, any age). If you've had a meaningful ACL or meniscus injury, you're in this entry's audience now — the risk of developing radiographic osteoarthritis in that joint within ten to twenty years sits near fifty percent regardless of how the injury was treated Lohmander et al. 2007. The strength work and the weight management aren't preventive in a vague sense; they're the specific lever against your specific risk.

When the picture isn't osteoarthritis — red flags

For the standard medications: topical anti-inflammatories are usually fine for almost everyone. Oral non-steroidals carry the usual cautions — kidney disease, peptic ulcer history, blood thinners, uncontrolled hypertension, established cardiovascular disease — and the lowest dose for the shortest time is the right setting. Worth a real conversation with your doctor, not a hand-wave.

What it costs, where to find it

The cheap parts of the plan are the parts that work the most. Walking, body-weight strength work at home, and a tube of topical diclofenac will run you under ten dollars a month. A structured physiotherapy program is a few hundred to a few thousand dollars depending on country and coverage; in countries with built-out delivery (Denmark, Australia, parts of Canada and the UK) you can ask specifically for a hip or knee osteoarthritis exercise program built off the GLA:D protocol Skou & Roos 2017. In the US the equivalent is usually six to twelve sessions of physical therapy ordered by your doctor; ask for progressive strength and neuromuscular work specifically, not range-of-motion alone.

Injections add up: a steroid shot is typically a hundred to a few hundred dollars and usually covered; hyaluronic acid series run hundreds to over a thousand; PRP is five hundred to a few thousand, often not covered. Joint replacement in the US lists at tens of thousands of dollars and is typically covered against your deductible; in most public health systems it's covered with a queue. The pharmacological-weight-loss option (a GLP-1 like semaglutide) costs a few hundred dollars a month if paying cash and is currently off-label for osteoarthritis itself, though the trial evidence is recent and strong Bliddal et al. 2024.

The bottleneck for most people is not money — it's access to a structured exercise program with someone watching the progression. Find that person before you book any of the injections.

What changes when you actually do this

First two weeks. Nothing dramatic on the pain — the system needs reps to adapt. What you notice is that twenty minutes of progressive loading doesn't make the joint worse the next day, which contradicts the story you'd been quietly running. The fear of moving starts to dissolve.

Six to twelve weeks. The pain is meaningfully lower — about a quarter to a third on average across thousands of patients in structured programs Fransen et al. 2015 Skou & Roos 2017. Morning stiffness goes from ten minutes to two. Stairs stop being something you brace for. If you also pulled the weight lever and dropped a few percent, the WOMAC pain and function scores moved by roughly twice that Panunzi et al. 2021.

Six months. The texture of your week has changed. The walk after dinner has come back. You sleep through more nights than you don't — the joint stops waking you in the second half of sleep. The afternoon you used to give up to the chair you give back to whatever you were doing before. People around you stop offering you the seat in the cafe.

Two to five years. The trajectory has bent. The patient who would have escalated through injection cycles into a deconditioned arthroplasty is on a different curve: functional muscle, an active week, and the option to delay or to schedule surgery on their own terms, not the disease's. If surgery does become the right call, going in fit produces a recovery that looks more like four months back to life than twelve.

The decade. The arc that runs through the next decade of well-managed osteoarthritis runs through the activity you keep doing — and the activity carries the rest of the body's metabolic, cognitive, and social health with it. The walking ability you protect in your sixties is the same walking ability that predicts how the rest of life goes Hawker et al. 2014. The joint isn't just a joint.

The interventions that do most of the work here have entries of their own: strength training, structured walking, weight loss, sleep, and — for end-stage disease — joint replacement. Worth their own reads. So is the post-injury thread: if you've had an ACL or meniscus injury, the trajectory toward osteoarthritis isn't fixed, and the protocol above is the same protocol that bends the curve.

Substance and claimed effects

Osteoarthritis (OA) is the most prevalent joint disorder worldwide, affecting approximately 595 million people globally in 2020 (7.6% of the population) and now the 10th leading cause of years lived with disability, up from 14th in 1990 GBD 2021. It is a whole-joint disease — historically framed as "wear-and-tear" cartilage degeneration but now understood as coordinated pathology of articular cartilage, subchondral bone, synovium, meniscus, ligaments, and periarticular muscle, with low-grade inflammation driving symptoms and progression Hunter & Bierma-Zeinstra 2019. Knee OA is dominant (~365 million prevalence cases globally in 2020), followed by hand and hip GBD 2021. The entry covers the consequences readers actually experience or decide about: joint pain (often deep, mechanical, worse with loading and use); morning and post-rest stiffness (typically <30 minutes, distinguishing OA from inflammatory arthritis); loss of mobility, range, and walking function; the cascade into deconditioning, sleep disruption, low mood, and weight gain; the role of weight management; the central evidence on progressive loading exercise (the single best-supported non-surgical intervention); the realistic effects of intra-articular injections (corticosteroid, hyaluronic acid, PRP); and what to expect from joint replacement when conservative measures stop working.

Evidence by addressing question

Mechanism

OA is no longer modeled as passive cartilage wear. Subchondral bone remodelling — marrow lesions visible on MRI, increased turnover, sensory nerve ingrowth — often precedes cartilage loss and is a major driver of pain, given that articular cartilage itself is aneural Hunter & Bierma-Zeinstra 2019. Synovitis (low-grade inflammation of the joint lining) is present in early disease and predicts progression. Chondrocytes shift to a catabolic phenotype, upregulating matrix-degrading enzymes (MMP-13, ADAMTS-5) and proinflammatory mediators (IL-1β, IL-6, TNF-α, prostaglandins), while cellular senescence and mechanical overload amplify the cycle. The pain mechanism is composite: nociceptive input from the synovium, subchondral bone, joint capsule, ligaments, and periarticular muscle; central sensitisation in a meaningful subset of patients; and the contribution of effusion, muscle weakness, and altered gait. The "structural-symptom mismatch" is therefore not a paradox — substantial radiographic disease can be asymptomatic, while a knee with mild radiographic changes can produce severe pain when synovitis, bone-marrow lesions, and muscle weakness coincide.

Evidence — does it actually matter clinically

OA carries quantifiable mortality signal beyond the joint. In a population-based cohort of ~17,500 patients with symptomatic hip/knee OA, walking disability (per-unit increase in Health Assessment Questionnaire walking score, HR 1.30, 95% CI 1.22–1.39) and walking-aid use (HR 1.51, 95% CI 1.34–1.70) were independently associated with all-cause mortality Hawker et al. 2014. Radiographic knee OA has been associated with elevated cardiovascular mortality (HR ~1.43). The mechanism is plausible — reduced physical activity from joint pain compounds cardiometabolic risk — and is reversible in principle: maintained physical function tracks with reduced mortality risk.

Protocol — exercise

Progressive loading exercise is the single best-supported non-pharmacological intervention. The Cochrane review of 54 RCTs (5,362 participants, knee OA) found land-based therapeutic exercise reduced pain by 12 points/100 (95% CI 10–15) and improved physical function by 10 points/100 (95% CI 8–13) immediately post-treatment, with effects sustained at 2–6 months (pain reduction 6 points/100) Fransen et al. 2015. Effect sizes for pain are comparable to those reported for NSAIDs, without the gastrointestinal, cardiovascular, or renal risk profile. The OARSI 2019 guidelines designate structured land-based exercise (strengthening, cardio, or neuromuscular) as a "core treatment" alongside patient education and weight loss Bannuru et al. 2019. NICE NG226 (2022) made therapeutic exercise the explicit first-line intervention, ahead of pharmacological options, for all OA patients NICE 2022. The 2019 ACR/Arthritis Foundation guideline strongly recommends exercise for knee, hip, and hand OA Kolasinski et al. 2020. The real-world delivery model with the strongest pragmatic evidence is GLA:D (Good Life with osteoArthritis in Denmark): a structured program of 2–3 education sessions plus 12 supervised neuromuscular exercise sessions over 8 weeks; in pooled longitudinal data from ~28,000 patients across Denmark, Canada, and Australia, pain intensity improved by 26–33%, walking speed by 8–12%, chair-stand ability by 18–30%, and joint-related quality of life by 12–26%, with reduced painkiller use and fewer days of sick leave at 12 months Skou & Roos 2017. The mechanism is multimodal: increased quadriceps and hip strength reduces dynamic joint loading; improved neuromuscular control corrects malalignment-driven loading patterns; aerobic exercise reduces systemic inflammation; load-induced cartilage matrix metabolism shifts toward anabolism. The dose-response: most trials use 30–60 minutes, three times weekly, for at least 8–12 weeks; benefits attenuate but persist with maintained activity.

Protocol — weight management

Each kilogram of body weight translates to approximately 4 kg of force across the knee in walking; obesity is one of the largest modifiable risk factors for knee OA (RR ~2.8 in men, ~4.4 in women). Network meta-analysis of weight-loss interventions in knee OA found that for every 1% weight loss, WOMAC pain, function, and stiffness scores decreased by approximately 2% Panunzi et al. 2021. The IDEA trial (Messier et al., JAMA 2013, n=454 overweight/obese adults with knee OA, 18 months) demonstrated that intensive diet plus exercise reduced peak knee compressive loads by >200 N per step and produced significantly greater pain reduction, function, mobility, and quality of life than either intervention alone — a clean factorial demonstration that the two effects compound Messier et al. 2013. The STEP-9 trial (n=407, semaglutide 2.4 mg vs placebo, 68 weeks) demonstrated that pharmacologically-induced weight loss in obese patients with knee OA produces both substantial weight loss and clinically meaningful WOMAC pain reduction (41.7% greater than placebo), with corresponding physical-function gains Bliddal et al. 2024. The signal is mechanistic and dose-dependent: weight loss reduces both biomechanical loading and adipose-driven systemic inflammation. Hip OA shows a weaker but still real weight-loss effect; hand OA, less so (the load argument doesn't apply, but inflammation may).

Evidence — corticosteroid injections

Intra-articular corticosteroids (typically triamcinolone or methylprednisolone) provide short-term pain relief, generally over 1–4 weeks, with diminishing effect by 12 weeks. The McAlindon JAMA 2017 trial (n=140, triamcinolone 40 mg vs saline every 3 months for 2 years) tested whether repeated injection altered structural progression: it found greater cartilage volume loss in the steroid arm (−0.21 mm vs −0.10 mm) and no significant pain benefit vs saline at 2 years McAlindon et al. 2017. The clinical takeaway is that corticosteroid injections are reasonable as occasional, short-term symptomatic rescue (e.g., to enable a wedding, a holiday, or a course of physical therapy), but repeated injection at fixed intervals is no longer evidence-supported and may accelerate cartilage loss. ACR 2019 conditionally recommends them; OARSI 2019 conditionally recommends for short-term knee OA pain; AAOS notes the limited durability Kolasinski et al. 2020 Bannuru et al. 2019 AAOS 2021.

Evidence — hyaluronic acid injections

Hyaluronic acid (viscosupplementation) has been controversial: trials are mixed, effect sizes are small, and committee positions have diverged. AAOS 2021 (3rd edition non-arthroplasty knee OA guideline) recommends against routine use, citing the inability to identify a patient subgroup that consistently benefits AAOS 2021. ACR 2019 conditionally recommends against in knee OA Kolasinski et al. 2020. OARSI 2019 gives a conditional recommendation only for knee OA without comorbidities Bannuru et al. 2019. NICE 2022 recommends against NICE 2022. Some meta-analyses find a modest short-term effect, particularly with high molecular weight formulations, but effect sizes are below most accepted minimum clinically important differences for knee pain. Practical reality: still widely used in the US, particularly when insurance prefers HA over surgery; reimbursement, not efficacy, drives much of the utilization.

Evidence — PRP and biologics

Platelet-rich plasma has been the subject of substantial commercial interest. The pre-RESTORE literature was mixed-quality and showed small positive signals. The RESTORE trial (Bennell et al., JAMA 2021, n=288, three weekly PRP vs saline injections, 12-month follow-up) was the highest-quality test to date: mean pain change −2.1 (PRP) vs −1.8 (saline) on an 11-point scale; medial tibial cartilage volume change −1.4% vs −1.2%; neither difference statistically significant Bennell et al. 2021. Major society guidelines (ACR 2019, AAOS 2021) recommend against routine PRP for knee OA Kolasinski et al. 2020 AAOS 2021. Stem-cell injections (bone marrow concentrate, adipose-derived, expanded MSCs) lack high-quality placebo-controlled trials and remain experimental; no major guideline endorses them. The risk for the reader is paying $1,000–$5,000+ per injection cycle for an intervention that, in the best controlled tests, performs no better than saline.

Evidence — joint replacement

For end-stage knee or hip OA with persistent disabling pain despite optimised non-surgical care, joint replacement (arthroplasty) is one of the highest-value interventions in modern medicine. Pooled registry data: knee replacements show ~93.0% construct survivorship at 15 years, 90.1% at 20 years, 82.3% at 25 years Evans et al. 2019 (knee). Hip replacement registry data: ~89% at 15 years, ~58% at 25 years using older bearing surfaces; modern bearings extrapolate to ~92% at 25 years and ~92% at 30 years Evans et al. 2019 (hip). Effect on pain and function is large — total knee and hip arthroplasty produce some of the biggest pre-post QALY gains in surgery — though ~15–20% of TKA patients report persistent pain or dissatisfaction. Risks: surgical complications (DVT, infection ~1%), revision over time, recovery period of 3–12 months.

Evidence — pharmacological adjuncts

Topical NSAIDs (diclofenac gel or solution) provide pain relief comparable to oral NSAIDs in knee and hand OA with substantially lower systemic exposure and GI risk Derry et al. 2016. The 2019 ACR guideline gives topical NSAIDs a strong recommendation for knee and hand OA; oral NSAIDs received an upgrade to strong recommendation in 2019 (from conditional in 2012) Kolasinski et al. 2020. Acetaminophen is conditionally recommended but effect sizes are small and largely no better than placebo over extended use. Duloxetine (SNRI) is conditionally recommended for chronic OA pain with a small but real effect. Opioids are conditionally recommended against. Glucosamine and chondroitin have been tested in multiple large trials including GAIT (n=1,583); the combination did not outperform placebo on the primary endpoint, and ACR 2019 strongly recommends against their use in knee OA Clegg et al. 2006 Kolasinski et al. 2020.

Evidence — arthroscopic debridement and meniscectomy

Two NEJM trials closed the chapter on arthroscopy for OA. Moseley 2002 (n=180) randomised arthroscopic debridement, lavage, and sham surgery; no significant pain or function difference at 2 years Moseley et al. 2002. Sihvonen 2013 / FIDELITY (n=146) compared arthroscopic partial meniscectomy with sham surgery in degenerative meniscal tears (most of which occur in the context of underlying OA); no benefit at 12 months, confirmed at 5-year MRI follow-up Sihvonen et al. 2013. AAOS, ACR, and NICE all recommend against arthroscopic debridement for OA AAOS 2021 Kolasinski et al. 2020 NICE 2022. Medicare and many private payers have ceased routine reimbursement. The intervention persists in some practices.

Misconceptions

(a) "Don't exercise — you'll wear it out faster." The opposite is true: progressive loading is the single best-supported non-pharmacological intervention; sedentary behavior accelerates deconditioning, weight gain, and functional decline Fransen et al. 2015. (b) "Cartilage grows back with the right supplement." Glucosamine, chondroitin, and collagen peptides have failed to demonstrate disease modification in adequately powered trials Clegg et al. 2006. (c) "X-ray findings dictate symptoms." Radiographic-symptom correlation is weak; severe imaging can be asymptomatic, mild imaging can be debilitating. (d) "Joint replacement is a last resort to delay as long as possible." Modern arthroplasty has excellent long-term outcomes; significant pre-operative quadriceps atrophy and functional decline can worsen post-op recovery, so timing matters Evans et al. 2019. (e) "OA only affects the elderly." Onset is typically 40s–50s for risk factors like prior ACL or meniscus injury, where ~50% develop radiographic OA within 10–20 years Lohmander et al. 2007.

Contraindications and red flags

Exercise: essentially no absolute contraindications; relative ones (acute joint flare with effusion, unstable cardiovascular disease) are addressed by program modification, not avoidance. NSAIDs: caution with CKD, peptic ulcer history, anticoagulants, uncontrolled hypertension, established cardiovascular disease. Steroid injections: caution near infection, in poorly controlled diabetes (transient glycaemic excursion). Red flags to distinguish OA from other diagnoses: hot, red, severely swollen joint (rule out septic arthritis or crystal disease); prolonged morning stiffness >1 hour (suggests inflammatory arthritis); systemic symptoms (fever, weight loss, multi-joint symmetric involvement); neurologic deficit (suggests radiculopathy or referred pain). Suspected hip OA presenting only as knee pain is a classic referred pattern.

Practicalities

The bottleneck for evidence-based OA care is access to structured exercise programs delivered by trained physiotherapists. GLA:D-style 8-week programs are widely available in Scandinavia, Australia, parts of Canada, and increasingly the UK and US; in the US, insurance may cover physical therapy as PT visits rather than as a structured group program Skou & Roos 2017. Cost spectrum: physical therapy ($50–$200/session, often 6–24 sessions); topical diclofenac ($20–$50/month OTC); oral NSAIDs ($5–$30/month); corticosteroid injection ($100–$500 per injection in US, typically covered); HA injection ($500–$1,500 per series, often covered); PRP ($500–$2,500 per injection, usually not covered); total knee or hip replacement ($30,000–$50,000 in US, typically covered with deductible). Weight loss via lifestyle is cheap; via GLP-1 receptor agonists like semaglutide is $200–$1,300/month if paying cash Bliddal et al. 2024.

Failure modes

The dominant failure mode is intensity and consistency of exercise. Patients prescribed exercise but not supervised drop off; pain-avoidant patients reduce load below the threshold for adaptation; flare-up phobia leads to deconditioning. The fix is structured supervision, progressive loading (most patients undertrain), and the explicit reframe that some discomfort during loading is expected and does not indicate damage. The second failure mode is the injection-treadmill: cycles of corticosteroid every 3 months instead of structural intervention (exercise, weight, surgery). The third is delayed surgery: patients told to "exhaust all conservative options" deconditioning so severely that surgical recovery is compromised. The fourth is unfounded faith in supplements and biologics, displacing higher-value interventions.

Stakes and payoff

The stakes side: progressive loss of walking distance, stair-climb capacity, sleep quality (pain-induced waking), and mood; cascade into weight gain (less activity), sarcopenia (less loading), cardiometabolic decline, and the mortality signal associated with walking disability Hawker et al. 2014. The payoff side: structured exercise produces 25–35% pain reduction and meaningful function gains within 8–12 weeks, sustained months to years with maintained activity Skou & Roos 2017 Fransen et al. 2015. For 10% body weight loss, WOMAC scores improve roughly 20% Panunzi et al. 2021. For end-stage disease, joint replacement returns walking, stair-climbing, sleep, and daily activities at >90% durability over 15 years Evans et al. 2019 Evans et al. 2019.

The credibility range

The optimist case

For an under-50 patient with early-to-moderate OA, the evidence base supports an active treatment plan that genuinely changes the disease trajectory. Structured progressive exercise + 10% weight loss + topical NSAIDs as needed produces pain reduction and functional improvement on par with the best pharmacological options, with no serious adverse-event profile, and the gains are durable as long as the activity is maintained. The IDEA trial's effect sizes at 18 months are large enough to redefine what "conservative management" can deliver. For late-stage disease, modern arthroplasty is one of the most effective procedures in surgery: ~90%+ survivorship at 15–20 years, transformative pain and function gains, and recovery measured in months. The future pipeline (GLP-1s for weight-mediated effect, possible disease-modifying agents) further raises the ceiling. The honest read on OA is closer to "an actively managed chronic disease with multiple effective interventions" than to the older "wear-and-tear, eventually replace" framing.

The skeptic case

Trials of exercise and weight loss enroll motivated patients in supervised programs with intensive support — the gap between trial efficacy and real-world effectiveness is large. Adherence is the binding constraint, and most patients are not given structured supervised programs (in the US in particular). Topical NSAIDs have small effect sizes; oral NSAIDs carry real risks at scale; acetaminophen is functionally placebo at chronic doses; opioids are net harmful. Most injectables either don't work (PRP, HA) or work briefly with possible structural cost (corticosteroids). Even joint replacement leaves 15–20% of patients with persistent pain or dissatisfaction, and revision rates accumulate with time. The disease has no disease-modifying pharmaceutical despite decades of investment and several high-profile failures. For many patients, OA is a progressive, function-stealing chronic disease that nudges them toward surgery on a timeline they don't control.

The author's call

The evidence base for the core interventions — exercise, weight loss, and (when needed) joint replacement — is strong, consistent, and replicated across multiple guidelines (OARSI 2019, ACR 2019, NICE 2022, AAOS 2021), placing the entry at evidence 5. The controversy is moderate and concentrated in the middle of the protocol: corticosteroid frequency, HA, PRP, the timing of arthroplasty, and the role of arthroscopy in subgroups. Controversy 2. The substance carries real felt-experience effects on pain, mobility, sleep, and quality of life, and acting on the evidence (loading exercise + weight + appropriate use of medication + arthroplasty when warranted) genuinely changes the trajectory. The clinical bar isn't perfection — it's beating the natural history of an undertreated case, which the protocol does, reliably.

Stakeholder and incentive map

Orthopaedic surgery — appropriately incentivised toward arthroplasty, which is genuinely effective for end-stage disease; misalignment can push toward earlier surgery than needed or toward retained low-value procedures (arthroscopy for OA persists in some practices despite guideline reversal).
Injection providers (orthopaedics, sports medicine, pain medicine) — fee-for-service environment rewards repeat injection (corticosteroid every 3 months, HA series annually, PRP cycles every 6–12 months). Reimbursement does not track evidence quality.
Pharmaceutical industry — substantial pipelines for disease-modifying OA drugs (DMOADs); no approved agent to date. Recent semaglutide trial (STEP-9) opens a credible weight-mediated pharmacological pathway. NSAID makers have stable mature markets.
Supplement industry — glucosamine/chondroitin and collagen peptides are billion-dollar categories despite negative or null evidence. Marketing claims continue.
Physical therapy / structured exercise programs — under-resourced in many systems despite carrying the strongest evidence; reimbursement for supervised programs is patchy. GLA:D in Scandinavia is the proof-of-concept that scale delivery works.
Public health and rheumatology guideline bodies — increasingly aligned on exercise + weight as first-line, with growing willingness to recommend against low-value interventions.
Patient advocacy — strongly aligned with active management; tension with the "cartilage regenerator" supplement marketing space.

Population variability

Sex — women carry roughly 60% of global OA burden; hand and knee OA in particular skew female, with a clear post-menopausal acceleration. Hip OA is more sex-balanced.
Age — symptom onset typically 40s–50s with risk-factor exposure (joint injury, obesity, family history); accumulates through 60s–70s; rare under 30 outside of post-traumatic disease.
Prior joint injury — ACL and meniscus injuries produce radiographic OA in ~50% of cases within 10–20 years regardless of treatment choice, with similar prevalence after ACL reconstruction Lohmander et al. 2007.
Obesity — RR ~2.8 (men) and ~4.4 (women) for knee OA; mechanism is biomechanical (load) and metabolic (adipose-driven inflammation). Hip OA shows weaker but real association.
Occupation and high-impact sport — repetitive squatting, kneeling, heavy lifting, and elite running/contact sport correlate with knee and hip OA. Recreational running does not.
Genetics — moderate heritability (~40–60% for hand and hip OA, ~25% for knee), with multiple GWAS hits but no single dominant locus.
Joint-specific responsiveness — hand OA responds less to weight loss and to most injectables than knee or hip OA; splinting plus targeted exercise is the hand-OA mainstay Kolasinski et al. 2020.

Knowledge gaps

No approved disease-modifying drug. Multiple candidates (anti-NGF — efficacy with safety signal; cathepsin K inhibitors; Wnt-pathway modulators) remain in trials; sprifermin and lorecivivint have shown structural but inconsistent symptom signals.
Whether GLP-1 weight loss in non-obese OA patients produces independent benefit beyond the weight-mediated effect is untested.
Best-in-class biomarkers for early disease (before radiographic change), to identify candidates for prevention.
Personalised matching of patient subgroup to intervention — the heterogeneity of OA is well-recognised but clinical trials largely enroll lumped populations.
Long-term effects of repeated corticosteroid injection on cartilage are signalled but not definitive — the McAlindon 2017 trial used 8 injections over 2 years; the safety profile of less-frequent injection (1–2x per year) is unclear.
Hand OA broadly under-studied relative to knee and hip; trials are smaller and outcome measures less standardised.

Scope vs brief. The brief named pain, stiffness, mobility, function, weight management, progressive loading exercise, injections, and joint replacement. All eight are covered. The brief listed knee, hip, and hand explicitly; the article's centre of gravity is knee (where the burden is largest and the trial base deepest), with a dedicated audience section addressing hip-specific notes (smaller weight-loss effect, referred pain, the most reliable arthroplasty) and hand/thumb-base specifics (female predominance, splinting + targeted exercise mainstay, weight argument largely absent).

Score-coverage check. Every non-zero meta dimension has a home in the body: health_short_term 4 in evidence/protocol/payoff; longevity 3 in stakes; energy, sleep, mood in stakes and payoff; focus 2 lightly in stakes (cognitive bandwidth pain steals); beauty_cumulative 1 indirectly via maintained activity/posture; cost_burden in practicalities; effort_burden in protocol's honest catch and dek.

Hard calls. Action coded as respond rather than do because the substance is a condition the reader manages, not a behaviour they initiate. Cadence daily because the loading-exercise piece is daily-to-most-days and that's the binding ask. Audience ages set to 40-59 and 60+ (most symptomatic burden lands there) but explicitly NOT excluding younger readers because the post-injury risk window pulls people in their thirties and early forties into the protocol; the article calls this out in audience and misconceptions.

Controversy 2 rationale. Core (exercise, weight, arthroplasty) is uncontested across OARSI/ACR/NICE/AAOS. The disagreement that does exist sits cleanly in the middle of the protocol — HA injections (AAOS/ACR/NICE against, OARSI conditional, US practice continues), PRP (RESTORE negative, practice continues), corticosteroid injection frequency (McAlindon signal vs short-term symptomatic value), arthroscopy (guideline-rejected, practice patchy), and arthroplasty timing.

Excluded from this entry. Rheumatoid and other inflammatory arthritides (genuinely different disease — mentioned only in red flags so the reader can self-distinguish). Detailed surgical technique comparisons (medial vs total knee, anterior vs posterior hip approach — not actionable for a non-clinician audience). Disease-modifying OA drug pipeline (no approved agent; would be speculation). Detailed nutritional/supplement weeds beyond the glucosamine/chondroitin canonical negative.

Future-link candidates. Strength training, structured walking, weight loss, sleep, joint replacement (as its own entry covering pre-op prep and recovery), GLP-1 receptor agonists, ACL injury and post-traumatic OA prevention, semaglutide/tirzepatide for chronic conditions — the out-of-scope closing points the reader toward these without making promises about catalogue contents.

Separate-entry candidates. Joint replacement (pre-op optimisation + post-op recovery + revision timeline is its own substantial substance). Post-traumatic OA after ACL/meniscus injury could become its own entry given the younger audience and the specific prevention window. Pharmacological weight loss for OA specifically, if STEP-9 leads to a label expansion.

Dream tier. Overall ≈ 49 (above the 40 threshold). Lever chosen: reclaim rather than pure aspiration — the honest hook is getting the active life back from a disease that takes it slowly, not building a new life. The dek and tagline lean in: cartilage-isn't-a-tyre reframe, get-the-decade-back close.