OA pathophysiology. The pathophysiology of OA involves a variety of systemic and local factors, including trauma, aging, obesity, epigenetic, and genetic factors. Chronic inflammation plays a strong role, including inflammatory signals from local (cartilage breakdown products stimulating the innate immune system) and distant (adipose tissue-derived adipokines) sources. The end result is chronic tissue damage which does not undergo appropriate wound healing
Increased cellular senescence
Increased systemic inflammation
Reduced ability of periarticular structures to absorb stress
Unclear. Potential defects in growth and remodeling, potential defects in vitamin A metabolism, cartilage calcification
Create a gene regulatory environment permissible for overexpression of cartilage-destroying enzymes, inflammatory cytokines. Downregulation of collagen. May mediate, in part, aging risk
Produces localized cartilage defects, increased extracellular matrix breakdown products, increased stress on subchondral bone, musculature, ligaments
Chronic systemic inflammation stimulated by cartilage breakdown products. Synovitis further stimulates immune responses, increases vascular permeability, and allows additional inflammatory cells to migrate to joint tissues. ECM breakdown also assembles complement and increases destruction of cartilage cells
Increases systemic inflammation, leading to paracrine effects, further worsening chronic joint inflammation. Increases joint load
Among the first changes in OA joints are inflammation within the synovial lining (synovitis), focal changes within bone marrow underlying the joint (bone marrow lesions, characterized by fibrosis, necrosis, and trabecular abnormalities), and matrix changes within the cartilage itself [7, 8]. The end-stage pathology of OA consists of erosion of articular cartilage, subchondral bone change, and loss of function of the joint “unit” in diarthrodial joints. Grossly, OA joints exhibit joint space narrowing, subchondral sclerosis of underlying bone, hypertrophic osteophyte formation of neighboring bone, and subchondral cyst formation. Moderate-to-severe cases of OA may demonstrate fibrillated cartilage, especially in areas of maximal loading, an irregular and disordered attempt at cartilage regrowth . We will briefly discuss each of these joint components individually.
Articular cartilage undergoes substantial changes during the development of OA. Although the majority of the physical load of a joint is borne by extra-articular structures (musculature, menisci, subchondral bone), normal cartilage provides a remarkably low-friction surface for smooth movement of a joint throughout its range of motion. Cartilage itself is made up of relatively scant, long-lived, metabolically inactive chondrocytes embedded within a tightly woven extracellular matrix (ECM) consisting of collagen fibers and proteoglycans coated by lubricin, a protein which reduces friction . As OA develops, chondrocytes begin to proliferate and aggregate into nests, and switch from an anabolic gene transcription program to a catabolic one. Counterintuitively, this switch results in the production of matrix metalloproteinases and other enzymes which begin actively breaking down neighboring ECM, led principally by the actions of ADAMTS5 and the matrix metalloproteinases MMP-9 and MMP-13 . Large shifts in epigenetic regulation occur within OA chondrocytes, providing a gene transcription regulatory pattern permissive for these changes [6, 12]. Remarkably, this transcriptomic shift closely resembles cellular changes seen in the senescence-associated secretory phenotype (SASP) of senescent cells seen in other body tissues, leading some to speculate that the predilection for OA development later in life is due, at least in part, to age-related senescence .
Although the inciting event(s) remain unclear and are most likely multifactorial, a proinflammatory environment is created once cartilage breakdown begins which propagates and further accelerates OA joint damage. A variety of ECM components stimulate the innate immune receptors of macrophages and other antigen-presenting cells via toll-like receptors (TLRs) recognizing danger-associated molecular patterns (DAMPs) [14, 15]. This immune stimulation, and the cytokine production that results from it, further stimulates the production of catabolic enzymes. The complement cascade is important as well, as cartilage ECM components also catalyze the assembly of various complement proteins, further disrupting cartilage homeostasis. Elevations of inflammatory markers, including tumor necrosis factor-alpha (TNF alpha) and interleukin-6 are seen in synovial fluid of patients with progressive OA . Furthermore, low-level systemic inflammation is also seen in patients with OA. For example, baseline prostaglandin E2 synthase levels in peripheral blood leukocytes can easily distinguish OA patients from controls, and a variety of cytokines, including interleukin-1 beta, TNF alpha, and cyclooxygenase 2 are increased in OA peripheral blood leukocytes and predict future rapid radiographic progression .
Cartilage is not the only tissue which undergoes extensive alteration during the development of OA. The subchondral bone plate is a thin cortical lamella directly underlying calcified cartilage. Although it is not a trabecular structure, it nonetheless has quite high porosity and contains channels for arteries, veins, and nerves, which can reach up to the cartilage surface . During OA development, stress on the subchondral bone plate underlying damaged cartilage regions increases substantially, resulting in reactive thickening  and leading to the radiographic appearance of subchondral sclerosis. Inflammatory markers released during this bone remodeling process can reach the overlying articular cartilage . Osteophytes, another hallmark feature of OA, originate in the periosteum  next to the bone/cartilage interface. They are a reactionary feature thought to develop in response to joint instability; one key player in the development of osteophytes is bone morphogenic protein-2 (BMP2) . Interestingly enough, the inflammatory cytokine (and target of rheumatoid arthritis drugs) TNF alpha also plays a role in osteophyte formation . Bone marrow lesions are present in symptomatic OA joints as well. Sometimes erroneously referred to simply as bone marrow “edema,” these are defined as discrete regions of hyperintense marrow signal in fat-suppressed magnetic resonance imaging sequences. Gene transcription analysis of these lesions demonstrates substantial upregulation of genes involved in pain sensitization, extracellular matrix, and proinflammatory gene signaling . The baseline volume of these lesions in OA patients is highly correlated with both joint pain and future radiographic narrowing of OA-affected joints .
Unlike cartilage, synovium is richly innervated and highly vascularized. Early OA is characterized almost universally by a degree of synovial inflammation, or synovitis. This is characterized by distinct histological findings, including synovial hypertrophy and hyperplasia, infiltration by mononuclear cells (T and B lymphocytes, tissue macrophages, monocytes), and increased angiogenesis. Inflamed synovial tissue itself releases a variety of proinflammatory cytokines and catabolic factors in OA, including interleukins, TNF alpha, matrix metalloproteinases, bone morphogenic proteins, and pain-producing neurotransmitters (i.e., nerve growth factor, substance P) [26, 27]. Although it may not be as clinically apparent as the florid synovitis seen in autoimmune forms of arthritis, MRI-detectable synovitis is strongly correlated with knee radiographic OA severity . This is not limited only to large-joint OA; the interphalangeal joints of hand OA patients also demonstrate increases in synovitis compared to non-OA controls, which correlates with both pain and radiographic severity . As one might expect, patients with the erosive hand OA subtype exhibit additional increases in joint synovitis scores .
Risk Factors for OA Development
Genetics certainly play a role in the development of OA. The overall genetic contribution to OA can be estimated through the use of twin studies, where a comparison is made between the “shared” genetic risk of identical, monozygotic twins and compared to the “half-shared” genetic risk of fraternal, dizygotic twins. Older twin studies of hip OA studies, including the UK Adult Twin Registry, have estimated genetic contributions to hand OA at around 50% and hip OA at around 60% . A newer study published in 2018 used more advanced data modeling techniques to adjust for modifiable risk factors and included a large number (n = 18,058) of twins from Norway . Their model suggested that 73% of hip and 45% of knee variance was genetically determined.
Investigations into individual genetic risk alleles (single nucleotide polymorphisms, SNPs) in OA have been somewhat less fruitful, and are quite specific to joint site (hand vs. hip vs. knee). The largest genome-wide association studies (GWAS) have been performed in knee and hip OA. The only risk alleles that have been independently confirmed for knee OA include mutations in the collagen gene COL11A1 and vascular endothelial growth factor VEGF . Another gene, growth differentiation factor 5 (GDF5), deserves special mention. This gene and the rs143383 SNP located within it have been strongly associated with both hip and knee OA in both humans and mice; furthermore, the risk allele causes reduced gene expression in joint tissues [34–37]. GDF5 is also under epigenetic control via changes in DNA methylation, and this conspires with underlying genetic changes to modulate gene expression further . Relatively fewer genetic studies have been performed in hand OA; in fact, only two large GWAS studies have been completed to date. The first study, in 2014, found an association with the retinaldehyde dehydrogenase gene ALDH1A2, involved in vitamin A metabolism . The second study, published in 2018, identified changes within the matrix GLA protein (MGP) gene, involved in cartilage calcification .
Epigenetics also play a role in OA development, as mentioned previously. A number of epigenome-wide association studies have been performed in both hip and knee OA, and have both identified and confirmed a number of genes and genetic pathways as strongly dysregulated in OA cartilage and subchondral bone [6, 12, 41–44], including a number of inflammatory pathways and inflammation-related transcription factors. As in genetics, epigenetic patterns are distinctly geographic (different in knee OA samples compared to hips).
The strongest environmental risk factor for OA is advanced age, although how exactly age contributes to OA is still somewhat unclear. Aging increases levels of c-reactive protein (C-RP), interleukin 6 (IL6), and tumor necrosis factor alpha (TNFα), a phenomenon known as “inflammaging” . Levels of systemic inflammatory mediators correlate with knee pain and decreased functional capacity in older adults with knee OA [35, 36]; furthermore, elevated levels of C-RP and IL6 are found in patients with knee OA and the level of these markers are related to the risk of OA progression [37, 38]. A dysregulated epigenome appears at least partly responsible for this phenomenon . For example, using a DNA methylation-based age estimator, OA cartilage has been shown to be epigenetically “older” than control cartilage . Autophagy, the process by which normal cells “clean up” old proteins, is defective in both aging and in OA articular cartilage and has been proposed as another possible explanation for the increased risk of OA associated with aging . Supporting this theory, aged mice also demonstrate a reduced autophagy phenotype in cartilage, and this defect precedes the development of OA-like cartilage damage .
Another quite important risk factor for OA development is obesity. Notably, like epigenetic and genetic risk, the risk conferred by obesity varies by joint. In knees, those with the highest body mass index (BMI) have an approximate 8.5-fold increased risk for OA compared to individuals with a “normal” BMI . Even more striking, each 2-unit increase in BMI equates to an increase in OA risk by 1.36-fold. A recent meta-analysis of 14 studies confirmed this finding that being overweight increased the risk of knee OA by 2.45-fold and obesity carried an increased risk of 4.55-fold . In hips, the risk is somewhat lower, with increased risk of around 1.1-fold . Hand OA is also associated with obesity, with increased risk in the 1.1-fold range . How obesity contributes to OA pathogenesis is complex and is not simply related to increased stress on the joint itself; rather, it likely also involves the increased basal systemic inflammation related to obesity, as well as an increased production of “adipokines,” inflammatory signaling molecules originating in adipose tissue [52, 53].
Trauma and “traumatic” occupations also increase the risk of OA. Dock workers, agricultural workers, carpenters, and cleaners have an increased risk of OA [54, 55]. Perhaps counterintuitively, running does not carry an increased risk for OA , nor does running worsen OA when it already is present . However, “elite” athletes, mainly those with a history of high impact activities, do have a higher chance of developing OA as they age . A history of previous injury is strongly associated with OA; this is perhaps best seen in the military population, where soldiers are more than five times more likely to develop PTOA compared to the general population. Soldiers with a history of knee joint trauma during active duty have a 5.7-fold increased risk of knee OA compared to those without a history of trauma .
Early: pain on strenuous activity, walking up or down stairs
Intermediate: pain in everyday activity
Late: constant rest pain
Early: tibial spine sharpening, subchondral sclerosis, subchondral cyst formation
Intermediate: joint space loss (usually medial>lateral), marginal osteophyte formation
Late: complete cartilage loss, bone-on-bone appearance, joint deformity
Early: occasional pain on activity, referred to groin or to knee; pain on internal rotation or full flexion
Intermediate: pain with activity, walking
Late: constant rest pain
Early: asymmetrical joint space narrowing
Intermediate: diffuse joint space loss, subchondral sclerosis
Late: marginal osteophyte formation, bone-on-bone appearance
Early: occasional stiffness and pain on repetitive motion
Intermediate: predictable pain with certain movements, stiffness daily, Heberden’s and Bouchard’s nodes may develop
Late: pain with minimal movement, perceived loss of hand “strength”
Early: DIP and thumb 1st CMC joint space narrowing
Intermediate: Substantial joint space narrowing, marginal osteophyte formation
Late: Fixed flexion deformities develop, marginal osteophyte formation may cause lateral or medial distal phalanx deviation
Hand: erosive OA subtype
Rapidly progressive PIP, DIP joint pain and stiffness with significant synovitis
Characteristic “gull-wing” and “sawtooth” appearance of DIP, PIP, respectively. Substantial soft tissue swelling. Spontaneous joint fusion may occur. Significant marginal osteophyte formation and bony proliferation
Other frequently-occurring signs and symptoms include bony hypertrophy, reflecting osteophyte formation (see section “Pathophysiology”), which tends to occur on marginal surfaces of OA-affected joints. Osteophyte formation and/or cartilage degradation can lead to frank joint deformities, which subsequently lead to joint instability. In fact, joint “buckling” or “giving out” is a very common symptom, particularly of knee OA. Over a quarter of patients with physician-diagnosed knee osteoarthritis will report knee instability symptoms, and a substantial number of these also report falls. Frequent falls in elderly OA patients can lead not only to fractures and other sequelae, but perhaps even more damaging, to fear of falling and poor balance confidence which can reduce physical activity further and worsen pre-existing deconditioning .
Hand OA generally affects the distal interphalangeal joints (DIP), first carpometacarpal joint (CMC, base of the thumb), proximal interphalangeal (PIP) joints, and occasionally the index and long finger metacarpophalangeal (MCP) joint, especially in cases associated with calcium pyrophosphate deposition disease. Like large-joint OA, hand OA is generally characterized by pain with activity. Some patients may complain mainly of stiffness, although this is generally less prolonged than autoimmune causes of hand arthritis. A frequent finding in hand OA are Heberden’s (DIP) and Bouchard’s (PIP) nodes. The appearance of these nodules is the result of early inflammation at the insertion of ligaments on the phalanges , further reinforcing the role of inflammation in OA. A less common but more aggressive variant of hand OA, known as erosive OA , is characterized by synovitis of the DIP joints with more extensive pain, erythema, and tenderness than one would expect of typical hand OA . Erosive OA tends to progress more rapidly than traditional hand OA. Cartilage and joint capsule erosion lead to lateral DIP instability and sclerosis, causing characteristic “twisting” and lateral deviation of the distal phalanges, with eventual and spontaneous DIP fusion a common finding. As one might expect, this erosive form of OA carries with it worse functional outcomes .
Classification criteria (using history and physical examination)
Pain in the knee and at least 3 of:
>50 years of age
Less than 30 minutes of am stiffness
Crepitus on active range of motion
No palpable warmth of synovium
Pain in the hip and:
>50 years of age
Internal hip rotation ≥15 degrees
Pain associated with internal hip
Morning stiffness of the hip less than 60 minutes
Internal hip rotation <15 degrees
Hip flexion ≤115 degrees
Pain, aching, or stiffness in the hand and at least 3 of:
Bony enlargement of 2 or more of:
2nd and 3rd distal interphalangeal (DIP)
2nd and 3rd proximal interphalangeal (PIP)
1st carpometacarpal joint of the thumb (CMC)
Bony enlargement of 2 or more distal interphalangeal (DIP)
Less than 3 swollen MCP joints
Deformity of at least one of:
2nd and 3rd distal interphalangeal (DIP)
2nd and 3rd proximal interphalangeal (PIP)
1st carpometacarpal joint of the thumb (CMC)
When the suspicion for OA is high based on clinical symptoms, there is not generally an indication for additional testing, and empiric treatment can commence. The presence of atypical symptoms may, however, indicate the need for additional workup; these include rapid progression of pain (imaging may be necessary here), a clear-cut periodicity of symptoms (periodic symptoms self-resolving after just a few days to weeks is suggestive of a crystal arthritis), or other constitutional symptoms such as weight loss, fevers, recent or current infections, etc. Testing for autoantibodies associated with rheumatoid arthritis (rheumatoid factor and anti-cyclic citrullinated peptide), along with systemic inflammatory markers (erythrocyte sedimentation rate or c-reactive peptide) can be useful in ruling out an autoimmune cause of arthritis symptoms in patients with a more inflammatory presentation.
Radiography is not generally indicated for the diagnosis of OA but can be useful in ruling out alternative causes for arthritis, making a diagnosis of erosive OA, and in monitoring the degree of cartilage loss if one is considering joint replacement. Moderately to severely affected OA joints are characterized radiographically by joint space narrowing (generally asymmetrical), subchondral sclerosis, marginal osteophyte formation, and the presence of subchondral cysts. Hand and knee radiographs are frequently obtained in patients with OA-like symptoms to rule out cartilage calcification, which is suggestive of concomitant calcium pyrophosphate deposition disease. Erosive OA of the hands is associated with a particular appearance of DIP joints, including cartilage erosion leading to a “gullwing” pattern in DIP joints and/or “sawtooth”-type pattern in PIP joints . Magnetic resonance imaging (MRI) can allow for direct quantitation both of synovitis, cartilage thickness, and screen for the presence of chondral lesions. Although MRI screening and monitoring of OA is not routinely done, it can predict patients who will have subsequent clinical OA progression . Finally, synovial fluid analysis is not generally indicated to diagnose OA; however, it can be quite useful in ruling out alternative diagnoses, particularly the crystalline arthropathies .
Treatment intervention or drug
OA subtype where specific treatment is appropriate (knee-only vs. multi-joint, with vs. without comorbidities)
Intra-articular steroid injection
Oral nonselective NSAIDs
Knee-only and multijoint OA without comorbidities
Oral COX2-selective NSAIDs
Knee-only and multijoint OA without comorbidities, or with up to moderate comorbidity risk
Knee-only OA both with and without comorbidities
Appropriate for knee-only and multijoint OA without comorbidities. (∗Note: more recent data suggest benefit no greater than placebo)
Uncertain for knee-only OA, not appropriate for multijoint OA
Uncertain for all forms of OA
Not recommended for any form of OA
First, modifiable risk factors should always be addressed. Weight loss should be discussed with every OA patient, and dietary changes made (including referral to a dietician if necessary) to achieve sustained weight loss. Several studies have indicated that even as little as 10% weight loss has substantial benefits in reducing OA-related pain and decreasing functional disability in OA patients. For example, a recent large study combined dietary and exercise interventions in knee OA patients and resulted in a mean weight loss of 11%. In the intervention group, significant reductions in pain and improvements in function were noted, along with better physical health-related quality of life scores and even reductions in serum levels of the inflammatory cytokine interleukin 6 (IL6) . Bariatric surgery, both for the treatment of OA and as an adjunct to total joint replacement, has been the focus of recent interest. Although studies have shown that massive weight loss induced by bariatric surgery does improve both pain and serum inflammatory markers in knee OA , several studies have also shown that bariatric surgery before joint replacement does not improve postarthroplasty functional or pain outcomes . There have not been definitive studies to indicate that one particular diet is any better than another for the treatment of OA symptoms, and no dietary supplements have been shown effective for OA.
Physical therapy should be a part of every OA prescription. Exercise in essentially any form is beneficial in OA and should be part of every OA treatment plan. There do not appear to be differences between land-based and water-based exercise from an efficacy standpoint, and the beneficial effects of exercise last for up to 6 months after cessation (although patients should be encouraged to continue an exercise regimen indefinitely) . Tai Chi, a Chinese martial art practiced with slow, methodical movements and an emphasis on balance, has a similar benefit in improving OA pain, physical function, and quality of life when compared to physical therapy regimens, with the added benefit of improving depression in OA patients .
Pharmacologic treatment in OA consists of a stepwise approach to analgesia. The best practice guidelines for the treatment of knee and hip OA come from several national and international societies, including the Osteoarthritis Research Society International (OARSI), the American College of Rheumatology (ACR), and the American Academy of Orthopedic Research (AAOS). The most recently updated of these are the OARSI guidelines for the management of knee osteoarthritis , and will be the basis for the following recommendations. Contrary to popular practice, acetaminophen has little place in the modern treatment of OA, as it has been demonstrated in multiple meta-analyses to be no better than placebo at pain relief in OA . The first question when treating a patient with OA regards their comorbidities. These include comorbidities which place the patient at moderate risk, including diabetes, advanced age, hypertension, cardiovascular disease (CVD), acute renal failure, history of gastrointestinal (GI) complications, depression, or physical impairment resulting in severe limitation of activity or exercise, including obesity. High-risk comorbidities include a history of a GI bleed, history of myocardial infarction, and chronic renal failure. Patients are then subdivided into knee-only OA or multijoint OA.
For knee-only OA without comorbidities, pharmacologic treatment may include nonsteroidal anti-inflammatory medications (NSAIDs), either “traditional” nonselective NSAIDs (i.e., naproxen), or COX-2-selective NSAIDs (i.e., celecoxib), or via topical application (i.e., diclofenac gel), or intra-articular (IA) therapies. Knee-only OA with comorbidities should avoid systemic nonselective NSAIDs and use instead IA treatments and topical NSAIDs. Multijoint OA benefits from systemic NSAIDs and IA treatments; generally, topical NSAIDs are not recommended, as the maximum dose may be inadequate to treat all involved joints. Multijoint OA in patients with comorbidities represents a challenge, with IA therapy and COX-2-selective NSAIDs being the preferred pharmacologic agents.
There have been surprisingly few head-to-head studies comparing the efficacy of various individual NSAIDs. One recent large meta-analysis including 76 individual trials suggested that the most effective oral NSAID for pain relief in OA was diclofenac at a dose of 150 mg total daily dose, followed by ibuprofen at 2400 mg total daily dose and naproxen at 1000 mg total daily dose . This should be interpreted with caution, however, given the lack of direct comparison in published data. There was some concern recently over the cardiovascular safety of COX-2-selective NSAIDs (the one in the US market being celecoxib) when compared to nonselective NSAIDs; however, the Prospective Randomized Evaluation of Celecoxib Integrated Safety versus Ibuprofen or Naproxen (PRECISION) trial, published in 2016, did not find evidence for an increased risk of celecoxib compared to either ibuprofen or naproxen . One area where there is a clear difference based on COX selectivity is in the risk of GI bleed, where COX-2-selective NSAIDs are strongly superior to nonselective NSAIDs; in patients with a past history of GI bleed, topical NSAIDs should be used if at all possible, with COX-2-selective oral NSAIDs used cautiously, and consideration given to concomitant proton pump inhibitor therapy. Nonselective NSAID use should be avoided in these patients. Duloxetine, an oral serotonin-norepinephrine reuptake inhibitor (SNRI), is a nonopiate, non-NSAID alternative appropriate for OA treatment and has good evidence for its efficacy; it may be an appropriate choice in patients with contraindications for NSAID or IA therapy or to be used in combination with NSAIDs . Other oral pharmacologic therapies with dubious evidence for efficacy (and not recommended) include glucosamine/chondroitin, diacerein, both oral and transdermal opiates, and risedronate.
IA corticosteroid injections should also be considered. IA steroids have strong evidence for pain improvement in the short term, although longer-term data are lacking. A trial published in 2017 also demonstrated a statistically significant increase in the rate of cartilage loss after 2 years of every-three-month short-acting steroid injection, although the incremental cartilage loss was not likely to be clinically significant . An extended-release steroid preparation of triamcinolone for IA injection has been recently approved which may offer both extended symptom improvement and a reduction in systemic side effects owing to a reduction in acute diffusion of steroid out of the joint . The other intra-articular therapy frequently used for knee OA, hyaluronic acid, does not have robust support in the literature and has received either an “inappropriate” or “not certain” recommendation from OARSI [77, 84].
Surgical treatment is the only definitive therapy available to physicians for knee and hip OA at the present time and demonstrates substantial pain relief and improvement of physical function that are better than the best pharmacologic management. The benefits of joint replacement for OA should not be overstated, however, as a measurable percentage (up to one-third in some studies) of patients have persistent symptoms following arthroplasty . The morbidity and mortality associated with joint replacement is low but should always be carefully considered before a decision is made to go to the operating room, with patient age and the presence of medical comorbidities (diabetes, obesity, and cardiovascular risks) being the strongest predictors of poor outcomes .
A 65 year-old Caucasian woman (BMI 35, sedentary, history of GERD with a treated gastric ulcer 2 months ago) presents to your clinic complaining of a 1-year history of steadily worsening bilateral medial knee pain and hip pain, worse with exercise, better with rest. When asked, she has morning stiffness lasting less than 30 minutes in both of her knees. She does not complain of any periodicity (i.e., no “flares” lasting for days to weeks) in any of her joints. She has tried over-the-counter acetaminophen, up to 1000 mg three times daily, without any benefit.