Prevalence and incidence of OA

The Framingham Osteoarthritis Study found that 6.8% and 19% of adults exhibited radiographic hand and knee OA , respectively. Similarly, both hip and knee OA had a prevalence of 28% of African–American and Caucasian men and women in the Johnston County Osteoarthritis Project. This potentially may have been related to genetic, anatomic or occupational differences .

Symptomatic OA prevalence estimates are lower as its presence is defined by a combination of symptoms such as pain, aching and stiffness as well as radiographic features. Framingham reported the prevalence of symptomatic hand OA to be 26% and 13% in women and men, respectively, and knee OA to be 7% . The Johnston County cohort reported prevalence rates of 17% for symptomatic knee OA and 10% for symptomatic hip OA . However, not all individuals with radiographic OA have concomitant symptoms. Furthermore, risk factors for structural disease and symptomatic disease may not be the same.

Defining OA

OA can be defined pathologically, radiographically or clinically. Due to the ease of standardisation and acquisition, radiography is often used as the standard for defining the presence and severity of OA using the Kellgren and Lawrence (KL) grading system . This system scales OA severity on a scale of 0–4 with >2 defining radiographic OA. The KL grading system has been used for hand and hip OA, but for the knee it can only be used to define tibiofemoral OA with the distinct radiographic feature of X-rays defining OA of the patellofemoral joint. Currently, there is no standard magnetic resonance imaging (MRI) definition of OA although a preliminary definition including cartilage lesions, osteophytes, bone marrow lesions (BMLs), synovitis and effusion has been described .

Age and gender

Age is one of the strongest predictors of OA ; however, the exact mechanism/s behind the increased prevalence and incidence of OA with age is poorly understood. A combination of changes including the capacity for joint tissues to adapt to biomechanical insults, biological changes such as cellular senescence as well as having a reduced capacity to adjust to biomechanical challenges as a consequence of age-related sarcopenia and increased bone turnover are likely contributing factors.

Females are associated with a higher prevalence and severity of OA and are more often affected with hand, foot and knee OA than men . In addition, women are more likely to suffer more severe radiographic knee OA then men, particularly following menopause . The increase in incidence of OA at the time of menopause has led to hypotheses regarding the role of oestrogen in OA. Oestrogen may unmask the symptoms of OA by enhancing pain sensitivity; however, results from observational studies and clinical trials have been conflicting . Gender disparities may also be caused by differences in bone strength, alignment, ligament laxity, pregnancy and neuromuscular strength. Women may also have a reduced volume of knee cartilage than men, but it is not clear if this could contribute to accelerated cartilage loss.

Genetics

OA in all of its forms appears to be strongly genetically determined with genetic factors accounting for at least 60% of hip and hand OA and up to 40% of knee OA . Whilst many studies focus on OA prevalence, many genes have been identified in playing a role in OA pathophysiologic pathways and thus may contribute to OA risk. Genes for vitamin D receptors, insulin-like growth factor 1 , type 2 collagen and growth differentiation factor 5 (GDF5) may provide targets for future pharmaceutical approaches. Linkages have been made between an area of ‘chromosome 2q’ and nodal OA , hip OA in women and chromosome 11q , ‘chromosome 7q22’ and knee OA and GDF5 (a bone morphogenetic protein expressed in articular and skeletal structures) and knee and hip OA in meta-analytic studies . Other signals such as DIO2, SMAD3 and ASPN may also be involved in OA susceptibility. Many of these findings are yet to be confirmed in all ethnicities and populations possibly due to inadequate phenotypic assessment of controls and social stratifications.

Obesity

Obesity has become a global problem leading to excess morbidity and mortality. There is considerable evidence indicating that obesity represents one of the most important risk factors for OA at peripheral joints such as the knee and hip . As obesity is increasing in prevalence and is also a risk factor for OA development, it is thus likely that more individuals will be affected by knee OA in the future.

Being overweight not only antedates the development of disease but also increases the risk of radiographic progression . A recent meta-analysis found that a dose–response relationship exists between obesity and the risk of knee OA. For every 5-unit increase in body mass index (BMI), there is an associated 35% increased risk of knee OA. The magnitude of the association was significantly stronger in women than that found in men. Furthermore, Lohmander et al. found a relative risk of knee OA to be 8.1 for patients with a BMI > 30 in a Swedish population.

Conversely, the Framingham Study showed that weight reduction by 5 kg provides a decreased risk for the development of knee OA by 50% , the findings also supported by a more recent meta-analysis . In a separate study, the population attributable risk for knee OA due to obesity was 29% and was higher in populations in which the prevalence of obesity was even higher .

The relationship between body weight and hip OA, however, is inconsistent and weaker than that for knee OA . Obesity is also associated with hand OA conferring the possibility that obesity may also provide some metabolic and inflammatory systemic effects. Greater weight has been associated with both incident radiographic and incident symptomatic hand OA .

Diet

Reactive oxygen species can accumulate with age and can be generated in joints, damaging articular tissues . Furthermore, animal studies have emphasised the importance of early life nutrition as a contributor to OA susceptibility but these effects have yet to be established in humans .

As vitamin D is associated with several aspects of articular cartilage and bone metabolism, it has been postulated that low levels of vitamin D may increase the incidence and progression of knee and hip OA . However, the relationship between vitamin D and OA has been conflicting. Previous studies have found that subjects with low to moderate levels of vitamin D (3–347 IU/d) were at an increased risk of incident hip OA and further progression of knee OA . In addition, low serum levels of vitamin D also predicted loss of joint space and increased osteophyte growth in knee OA . Despite this, recent randomised controlled trials concluded that there was no demonstrable cartilage loss on MRI in subjects with low levels of vitamin D .

Antioxidant vitamins may also play a role in OA. Low intake of vitamin C has been associated with an increased risk of knee OA progression and, similarly, high vitamin C intake has been shown to reduce the progression of radiographic knee OA threefold as well as reducing the risk of developing knee pain .

Vitamin K is an important regulator of bone and cartilage mineralisation. Low levels of plasma phylloquinone (<0.5 nmoles/l) have been shown to increase the prevalence ratios for OA, osteophytes and joint space narrowing in the hand and osteophytes in the knee . Kashin–Beck osteoarthropathy has been associated with iodine and selenium deficiency .

Conversely, no evidence has been found that a high intake of any of the aforementioned antioxidant nutrients reduces the incidence of knee OA or affects the volume of cartilage. Further studies are required to better refine the association between OA and dietary factors.

Occupation and physical activity

Repetitive joint use has been associated with an increased risk of OA. Studies have found that individuals whose occupations require squatting or kneeling have twice the risk of developing knee OA than occupations that do not require physical activity, particularly among those who are overweight or whose jobs required carrying or lifting . Hip OA has also been associated with prolonged standing and lifting . Occupations that require increased manual dexterity have been associated with features of hand OA .

The issue of repetitive joint use may also be relevant for physical activity. Physical activity may potentially be detrimental if it places undue load on the joint despite the observed benefits of strengthening periarticular muscles to help stabilise the joint. A recent review by Hansen et al. failed to find evidence to support an association or causal relationship between low- and moderate-distance running and hip and knee OA in general population studies defined as incident radiographic or symptomatic OA. The review also produced inconclusive evidence regarding high-volume running and OA development suggesting that, in the absence of joint injury, the risk of OA development due to running and exercise is minimal. Limited information exists regarding the effect of running and the development of OA at the ankle and the lumbar spine.

Conversely, there does appear to be an association between elite-level athletes and OA development due to the highly repetitive, intense and high-impact nature associated with sports such as tennis, squash and team sports. This cohort confers an increased risk of developing radiographic hip and knee OA when compared to an age-matched, non-elite cohort . It is unclear, however, whether this association is solely due to sports participation or as a result of injury. In two studies of athletes, the increased risk of OA appeared to be related to knee injury amongst soccer players rather than due to training loads .

Post-traumatic knee injuries

Multiple intrinsic factors affect a joint’s ability to withstand destructive forces including the thickness of the articular cartilage, the strength of the bone adjacent to the joint, ligament and muscle strength and neuromuscular control of the joint. Yet, despite these factors and the introduction of biomechanical training programmes initiated into schools and elite sporting organisations, the knee remains one of the most commonly injured joints. In the context of OA, the most important injuries are those resulting in the rupture of the anterior cruciate ligament (ACL), which is often accompanied by damage to the articular cartilage, subchondral bone and collateral ligaments and, importantly, damage to the menisci is observed in approximately 65–75% of ACL-injured knees . The risk of ACL rupture is higher in adolescents and up to 70% higher in high-risk sports than in the general population .

ACL rupture, traumatic meniscal tears and direct articular cartilage damage sustained during injury are strongly linked to the subsequent development of OA, with a substantial percentage of patients showing OA changes and functional disability as early as 10 years after the initial injury . Furthermore, direct damage to the articular cartilage and the development of BMLs have been associated with matrix disruption, chondrocyte necrosis and proteoglycan loss which may not be reversible . Thus, individuals who suffer a knee injury are at an increased risk of early-onset knee OA.

The precise pathogenesis behind why ACL ruptures lead to an increased risk of developing OA and why OA development can be accelerated in injured joints is not known. It has been postulated that the majority of the tissue damage is related to the large forces required to injure the ACL . Occult osteochondral lesions occur in 80–90% of patients with an acute ACL injury as seen on MRI suggesting that articular cartilage sustains a considerable mechanical impact at the time of injury. Most commonly, these lesions occur on the posterolateral tibial plateau and the anterolateral femoral condyle . In addition, intra-articular bleeding commonly occurs with the initial injury, as well as the surgical repair, causing both an acute and sustained release of inflammatory cytokines and proteases from joint tissues , which may lead to further damage of the type 2 collagen network. A more recent study demonstrated that the lack of a functional ACL changes the static and dynamic loading of the injured knee resulting in a change to the region of cartilage that is in contact during weight bearing. This may cause increased loading of areas that were not conditioned to constant load prior to injury .

It is difficult to attain precise figures on the prevalence of knee OA after ACL injuries, with estimates ranging between 10% and 90% at 10–15 years after injury as studies attempting to assess the relationship between ACL rupture and OA have made use of heterogeneous populations and inconsistent radiologic classification methods. Thus, Oiestad et al. suggested that post-traumatic knee OA resulting from an isolated ACL injury may be greatly overestimated with the highest-rated studies reporting a prevalence of 13% after isolated ACL rupture as compared to the 50–70% prevalence often quoted in the literature. However, when combined with damage to the meniscus, they found a higher prevalence of OA, between 21% and 40%.

Treatment of post-traumatic knee injuries

Surgery is the most common form of treatment for injuries to the ACL and the menisci. ACL reconstructions are commonly advocated on the basis that they protect against secondary injuries to the meniscus and ligaments and therefore reduce the risk of OA development. Studies investigating the two treatment arms of surgical and conservative management for ACL rupture or meniscal tears have concluded that both treatment groups confer the same prevalence of radiographic knee OA 10–15 years after injury with the only difference being that the rate of secondary meniscal surgery was higher after nonsurgical treatment for ACL-injured knees . A systematic review also produced similar findings .

Hoffelner et al. found that athletes who had an isolated ACL injury and underwent surgical reconstruction conferred no increased risk of OA development in that knee when compared to the contralateral knee. This suggests that ACL reconstruction for an isolated ACL injury may prevent secondary meniscal and chondral injuries, which advance OA. Furthermore, a study of European handball players found that 22% of those who returned to their sport post ACL reconstruction would later re-injure their ACL . This has led a Cochrane review to declare that there is insufficient evidence to determine which method of treatment is best for ACL injuries, as it is unclear whether ACL reconstruction decreases the incidence of OA development over the long term. Similarly, it is unclear whether ACL reconstruction or arthroscopy for treatment of meniscal tears decreases the incidence of OA . Meniscal replacement surgeries including the use of allogeneic, xenogeneic and artificial menisci have been tried in younger patients but the transplant survival is variable and long-term results are lacking .

Thus, whilst surgery may repair an injured knee in the short term, it may not prevent the development of knee OA in the long term nor will it protect the knee from re-injury or restore normal knee kinematics. If long-term joint health is the primary concern, this raises questions as to whether returning to sports that involve pivoting is really in the athlete’s long-term interest with regard to joint health. Neuman et al. reported low rates of OA in patients who injure their ACL and agree to moderate their level of activity to avoid re-injury.

Local pathologic changes

OA was once considered a degenerative, primary disorder of the articular cartilage but now it is widely appreciated that multiple structures are involved and affected in the development of OA. Early osteoarthritic cartilage may be thicker and swollen with water due to disruption of the collagen network. Evidence also suggests that there is increased cell proliferation and an up-regulation of synthetic activity resulting in clusters of chondrocytes . Despite this, these cells are not able to maintain the integrity of the cartilage matrix, mainly due to their inability to respond to growth factors, and thus further contribute to the increased matrix degradation and the destruction of type II collagen .

Damage to articular cartilage and bone may lead to adjacent synovial inflammation causing further release of proteinases, inflammatory cytokines, matrix metalloproteinases (MMPs) and aggrecanase, which acts to accelerate cartilage degradation . The release of these inflammatory mediators, as well as the formation of osteophytes may act to irritate sensory nerve endings within the synovium causing pain. MRI analysis of subjects with knee OA showed that synovial thickening was greater amongst subjects who experienced knee pain than in asymptomatic subjects .

In patients with symptomatic OA, progression of cartilage defects over 30 months was found in 46% and 22% for the medial and lateral tibiofemoral compartments, respectively. MRI is able to capture these initial structural changes in the earliest phases of the disease, whilst changes such as joint space narrowing as detected by radiographs emerge at a much later stage . Furthermore, cartilage defects are also associated with bone expansion, BMLs, meniscal injuries and ACL rupture, suggesting that they have multiple causes .

Normally organised menisci are rarely found in patients with knee OA, suggesting that there is a strong disorder of the meniscus involved with the development of OA . Degenerative meniscal lesions often occur in middle-aged and elderly individuals with knees that have already been compromised by OA . The Framingham Study found that 82% of subjects who displayed radiographic knee OA had meniscal damage with the majority suffering from degenerative lesions . Bhattacharyya et al. found that 91% of subjects who had symptomatic knee OA had a meniscal tear . The long-term radiographic outcome for those subjects with a degenerative lesion has been found to be worse than those with traumatic lesions including an increased risk for early-onset OA when compared to subjects who had normal menisci . This suggests that damage to the meniscus antedates radiographic cartilage changes. Cartilage destruction due to the pathological processes that are active during the early stages of OA could also affect meniscus and ligament integrity as well. Thus, knee OA may also cause meniscal lesions and act to further accelerate the disease .

Bone cells are more able to self-repair and modify their surrounding extracellular matrix than articular cartilage . Subchondral bone undergoes adaptations during the development of OA including an increase in subchondral plate thickness, sclerosis, joint space narrowing, reduced matrix mineralisation, increased cancellous bone volume, formation of osteophytes at the joint margins, development of bone cysts and advancement of the tidemark associated with vascular invasion of the calcified cartilage. These changes may cause alterations in the adjacent joint surfaces, which in turn will change the joint congruity and hence progress the disease .

It is the adaptive capacity of bone that underlies the more rapid appearance of detectable skeletal changes, especially after joint injuries or with altered mechanics. The presence of BML correlates with the severity of pain as well as the areas of the greatest cartilage loss . BMLs were present in 77.5% of subjects who experienced painful knees compared with only 30% of subjects who reported no knee pain . Furthermore, BMLs have been found to be compartment specific for cartilage progression. Subjects who were varus in alignment developed medial lesions whilst those who were valgus developed lateral lesions. In terms of bone abnormalities, BML is the only effective risk factor for predicting knee OA progression .

Muscle strength

Quadriceps femoris is the primary antigravity muscle of the lower limb and serves to decelerate the lower limb during ambulation, absorb limb loading as well as to provide dynamic joint stability. Thus, it has been postulated that quadriceps femoris weakness could play an important role in the genesis of knee OA.

Deficits in muscle strength, activation and proprioception are common in patients with knee OA and can occur as a consequence of OA related to disuse due to pain avoidance. Furthermore, a recent literature review concluded that there is some evidence that muscle weakness may predispose to the onset and potentially the progression of knee OA .

Ikeda and colleagues found that the quadriceps cross-sectional area was significantly reduced in women with incident asymptomatic radiological OA, compared with women matched for age and body mass. Signs of muscle fibre atrophy have been reported in later disease states with one study finding the quadriceps lean muscle cross-sectional area to be 12% lower in the affected limb compared with the contralateral limb in patients prior to knee replacement . Thus, it seems likely that loss of strength associated with OA may be associated with the loss of muscle cross-sectional area.

Additionally, quadriceps weakness may also increase the risk of structural damage. For every 5-kg increase in extensor strength, Slemenda et al. found an associated 20% and 29% reduction in the odds of developing radiographic knee OA and symptomatic knee OA, respectively . Improvement in muscle function, especially strength, through exercise has been associated with reduced pain and improved function in people with knee OA .

Alignment

A shift from neutral alignment will alter load distribution across the knee; thus, malalignment may contribute to abnormal mechanical forces. Knee malalignment is one of the strongest predictors of knee OA progression with a prospective cohort study showing that abnormal alignment was strongly associated with increased structural degradation in the compartment that was under the greatest compressive stress . Medial progression of knee OA was four times more likely in individuals with varus alignment, whilst lateral progression was five times more likely in individuals with valgus alignment . BML as well as rapid cartilage loss displayed on MRI have also been associated with knee malalignment . It is important to note that no study as yet has documented the slowing of disease progression when the alignment is corrected.

The association between incident knee OA and malalignment is less apparent. The Rotterdam Study produced an odds ratio of 2.06 and 1.54 of developing radiographic knee OA in individuals with varus and valgus knee alignments, respectively . These results were not supported by the Framingham Study, which found no association between knee joint alignment and an increased risk of incident radiographic knee OA .

Implications for disease prevention

Presently, therapeutic interventions for OA are palliative and are primarily analgesia and surgical intervention. In the absence of pharmacologic agents that can modify disease, we need to instead focus on the modifiable risk factors mentioned in this review, namely obesity, alignment and injury prevention, for pain incidence and disease progression.

The majority of people with OA are overweight or obese, and there is good evidence for the efficacy of weight management for OA . For each kilogram of weight lost, the knee will experience a fourfold reduction in load during daily activities . In practice, however, weight management is not frequently implemented . If someone is overweight or obese, they should be engaged in a combination diet and exercise programme aimed at a weight reduction of >5% of body weight . Exercise is often the forgotten aspect of a conservative treatment programme. Not only can exercise increase muscle strength and aerobic capacity but it also helps to facilitate weight loss. For this reason, all patients should be encouraged to participate in low-impact aerobic exercise programmes such as walking, bike riding or swimming.

It has also been suggested that targeting the pathomechanics of OA, such as correcting knee joint alignment, is effective in preventing OA progression . Knee braces, orthotics, patella taping and knee osteotomies are some of the therapeutic options available to help modify joint forces in an attempt to reduce both symptoms and structural changes.

Preventing knee injuries, particularly to the ACL, meniscus and chondral bone surfaces, would, in turn, help prevent a large proportion of young adults who injured their knee playing sports from developing early-onset OA. The efficacy of knee bracing to prevent knee injury has been studied, most recently in American football players, with many studies reporting relative risk reductions in injury incidence of 10–50% . However, compliance is a major issue with many athletes fearing impaired athletic performance and discomfort. There is strong evidence for the role of intrinsic risk factors in the aetiology of knee injury. Studies investigating the implementation of neuromuscular and proprioceptive training programmes in young athletes have shown that these programmes may reduce ACL injuries . These programmes are designed to train athletes to land and decelerate in a more controlled fashion with reduced valgus collapse, increased knee flexion and improved trunk control, balance and proprioception.

One large and as yet unmet treatment goal is modification of the underlying joint structure. Some studies suggest that disease progression may be modified by glucosamine sulphate, chondroitin sulphate, sodium hyaluronan, doxycycline, MMP inhibitors, bisphosphonates and calcitonin . Disappointingly, it may be a while before a disease-modifying drug is available as cartilage remains the major focus of drug development, and yet cartilage has been shown to not be the direct cause of symptoms .