Abstract
Post-traumatic osteoarthritis (PTOA) subsequent to joint injury accounts for over 12% of the overall disease burden of OA, and higher in the most at-risk ankle and knee joints. Evidence suggests that the pathogenesis of PTOA may be related to inflammatory processes and alterations to the articular cartilage, menisci, muscle and subchondral bone that are initiated in the acute post-injury phase. Imaging of these early changes, as well as a number of biochemical markers, demonstrates the potential for use as predictors of future disease, and may help stratify patients on the likelihood of their developing clinical disease. This will be important in guiding future interventions, which will likely target elements of the inflammatory response within the joint, molecular abnormalities related to cartilage matrix degradation, chondrocyte function and subchondral bone remodelling. Until significant improvements are made, however, in identifying patients most at risk for developing PTOA – and therefore those who are candidates for therapy – primary prevention programmes will remain the most effective current management tools.
Introduction
Osteoarthritis (OA) is the most common of the arthropathies, with the prevalence of clinically defined OA of at least one joint estimated to be approximately 9% . Despite this, a consensus has still not been reached on the exact nature of OA as a disease, as it remains unclear whether OA represents a single disease entity, or is rather a clinical end point with characteristic features (pain, functional disability and evidence of cartilage loss, osteophyte formation, synovitis and subchondral bone remodelling) resulting from a number of different disease subtypes with distinct aetiologies . Stratifying OA into multiple disease entities could provide an explanation for its heterogeneity in terms of joint tissue pathology and rate of disease progression. Subtyping may also have particular clinical relevance, partly explaining the failure of current therapies to achieve substantive response rates, and raising the prospect that future therapies could be targeted towards individual disease processes .
Of particular interest within the potential subtypes of OA – especially in the context of targeted treatment – is post-traumatic osteoarthritis (PTOA), the form of OA that develops following a known injurious event to the affected joint . PTOA accounts for over 12% of OA, and most commonly occurs in the ankle and knee joints . The fact that primary injuries are more likely to be sustained by younger individuals indicates that PTOA develops earlier than other forms of OA, with a recent study finding that individuals with ankle and knee PTOA were approximately 14 and 10.4 years younger, respectively, than their counterparts with other forms of OA . PTOA therefore represents a cause of functional disability in a disproportionately young population group, leading to substantial economic implications in the form of direct health-care costs (estimated to be over US$ 3 billion annually in the USA), disability-affected life years (DALYs) and lost work productivity . Given the increased rates of participation in sports with high rates of joint injury – notably basketball and soccer in the USA with their predilection for anterior cruciate ligament (ACL) injuries – and the tendency towards heavier players within a number of sports, it is likely that the importance of this subtype of the OA disease process will only increase in the future .
PTOA also represents an appealing opportunity for targeted therapy, as it commonly has a known ‘start point’ after which interventions could theoretically be initiated at an early stage to prevent the future progression of the disease. An understanding of the mechanisms by which the primary joint injury initiates the subsequent development of OA, and the way in which the disease process is maintained, would therefore be central to the development of targeted treatments.
This narrative review outlines the pathologic changes that take place in PTOA following joint injury, and the potential for these changes to be used as biomarkers and predictors of future disease progression. The possibilities for future interventions targeted towards specific disease processes in PTOA are also reviewed.
Mechanisms of joint injury
The ankle is the joint most commonly injured during sporting activities, with over 300,000 injuries occurring annually in the USA alone, at an estimated injury rate of 52.3 ankle injuries per 1000 athlete exposures . Although sprains of the lateral ligament complex account for the majority of injuries (83.4%), fractures are perhaps the most clinically relevant ankle injuries in the context of PTOA, and occur at a rate of approximately 17,000 per year . The typical mechanism of ankle injury is thought to be a rapid change of direction upon landing, producing excessive inversion, plantar flexion and internal rotation . Sports that involve cutting manoeuvres and jumping in close proximity to other players therefore exhibit the highest risk, with basketball and football representing the greatest sources owing to their high overall injury rates (77.4 and 54.1 per 1000 athlete exposures, respectively) and population participation rates .
Injuries of the knee remain the most prevalent within the wider community, with an estimated 700,000 occurring in the USA each year at a rate of 2.29 knee injuries per 1000 individuals . Of particular note are knee injuries involving rupture of the ACL. ACL tears are believed to result from postural readjustments that produce concurrent valgus force and internal or external rotation . This dynamic loading creates sufficient tension to rupture the ACL, and is usually associated with sudden deceleration and changes in direction . Most ACL injuries therefore occur in the absence of direct contact in sports such as basketball (0.29 tears per 1000 exposures for women and 0.08 for men) and football/soccer (0.32 for women and 0.12 for men) and skiing (0.40), rather than heavy-contact sports like American football (0.08) . By far, the largest numbers of knee injuries in the USA occur in basketball and soccer due to their higher participation rates . The significantly higher rate of ACL tears in females is consistently observed across a multitude of studies and sports, with an average female:male tear rate of approximately 3:1 . Although the exact mechanisms underlying this disparity have not been identified, a number of factors – including differences in quadriceps activation, movement patterns, hormone-dependent knee laxity and muscle stiffness – have been suggested to be likely contributors .
Unlike injuries to the ankle, there is significant heterogeneity in terms of the exact pathology within the acutely injured knee. Even when considering only knee injuries involving ACL rupture, a wide range of associated structures are frequently compromised. Direct articular cartilage damage, post-traumatic bone lesions, meniscal damage and collateral ligament injuries frequently occur concomitantly with ACL ruptures (see Fig. 1 ), and have been linked to long-term structural sequelae, and so will be discussed in the following section.
Although PTOA has been shown to occur in a number of other joints (most notably the hip), the predominance of ankle and knee is such that only those joints will henceforth be considered in detail during the course of this chapter.
Mechanisms of joint injury
The ankle is the joint most commonly injured during sporting activities, with over 300,000 injuries occurring annually in the USA alone, at an estimated injury rate of 52.3 ankle injuries per 1000 athlete exposures . Although sprains of the lateral ligament complex account for the majority of injuries (83.4%), fractures are perhaps the most clinically relevant ankle injuries in the context of PTOA, and occur at a rate of approximately 17,000 per year . The typical mechanism of ankle injury is thought to be a rapid change of direction upon landing, producing excessive inversion, plantar flexion and internal rotation . Sports that involve cutting manoeuvres and jumping in close proximity to other players therefore exhibit the highest risk, with basketball and football representing the greatest sources owing to their high overall injury rates (77.4 and 54.1 per 1000 athlete exposures, respectively) and population participation rates .
Injuries of the knee remain the most prevalent within the wider community, with an estimated 700,000 occurring in the USA each year at a rate of 2.29 knee injuries per 1000 individuals . Of particular note are knee injuries involving rupture of the ACL. ACL tears are believed to result from postural readjustments that produce concurrent valgus force and internal or external rotation . This dynamic loading creates sufficient tension to rupture the ACL, and is usually associated with sudden deceleration and changes in direction . Most ACL injuries therefore occur in the absence of direct contact in sports such as basketball (0.29 tears per 1000 exposures for women and 0.08 for men) and football/soccer (0.32 for women and 0.12 for men) and skiing (0.40), rather than heavy-contact sports like American football (0.08) . By far, the largest numbers of knee injuries in the USA occur in basketball and soccer due to their higher participation rates . The significantly higher rate of ACL tears in females is consistently observed across a multitude of studies and sports, with an average female:male tear rate of approximately 3:1 . Although the exact mechanisms underlying this disparity have not been identified, a number of factors – including differences in quadriceps activation, movement patterns, hormone-dependent knee laxity and muscle stiffness – have been suggested to be likely contributors .
Unlike injuries to the ankle, there is significant heterogeneity in terms of the exact pathology within the acutely injured knee. Even when considering only knee injuries involving ACL rupture, a wide range of associated structures are frequently compromised. Direct articular cartilage damage, post-traumatic bone lesions, meniscal damage and collateral ligament injuries frequently occur concomitantly with ACL ruptures (see Fig. 1 ), and have been linked to long-term structural sequelae, and so will be discussed in the following section.
Although PTOA has been shown to occur in a number of other joints (most notably the hip), the predominance of ankle and knee is such that only those joints will henceforth be considered in detail during the course of this chapter.
Pathogenesis – acute injury phase
The fact that a single traumatic incident strongly predisposes a joint to PTOA lends itself to the idea that PTOA is simply the clinical end point resulting from the persistence and progression of a pathogenetic processes initiated at the time of the initial joint trauma. It has, however, proven extremely difficult to separate the processes in the acute post-traumatic phase that predispose to PTOA from the myriad of pathological, inflammatory and regenerative processes that always take place within injured tissue. This difficulty is exacerbated by the fact that the disease is not typically diagnosed until 15–20 years after the injury, making it hard to link minute changes in the post-injury phase to the gross histopathological changes that are characteristic of clinical OA .
Despite this, a number of studies have reported observations that would appear to support the idea that a number of processes may play a role in the long-term development of PTOA.
Articular cartilage damage
Articular cartilage damage resulting from impaction is common following joint injury, with the cartilage of the medial and lateral femoral condyles exhibiting signs of direct structural damage in almost half of ACL tears . Direct cartilage impaction >15–20 MPa (MPa) has previously been shown to be associated with chondrocyte necrosis, with forces >40 MPa causing virtually complete cell death .
Chondrocyte death beyond the directly impacted area is probably brought about by subsequent apoptosis, however, rather than necrosis . Evidence suggests that disruption of the extracellular matrix (ECM) through direct damage, and the disproportionate increase in chondrocyte expression of matrix-degrading enzymes relative to their inhibitors (namely matrix metalloproteinase 3 (MMP-3), a disintegrin and metalloproteinase with thrombosponin repeat (ADAMTS)-5 and tissue inhibitor of metalloproteinase (TIMP)-1, respectively) and inflammatory cytokines (tumour necrosis factor (TNF)α and interleukin (IL)-1) following mechanical stress is associated with chondrocyte apoptosis, potentially mediated by the caspase pathway . As suggested by Kramer et al. in their recent review, it is possible that the loss of the ECM surrounding a chondrocyte following its apoptosis places greater mechanical and metabolic stress on the remaining chondrocytes . This would produce a positive-feedback system, raising the possibility that once a ‘critical number’ of chondrocytes undergo apoptosis, the damage to the cartilage would be ongoing.
It should also be noted that injuries that cause severe short- to medium-term functional disability are likely to involve instability and altered loading patterns within the joint. Direct impaction-related cartilage damage due to frequent episodes of instability or pivot shifting has previously been postulated as a possible cause of articular cartilage loss subsequent to ACL injury . Similarly, altered biomechanical loading following ACL rupture may produce changes in the articular cartilage, as cartilage loading patterns have previously been shown to influence chondrocyte metabolism, proteoglycan production, collagen fibre orientation and MMP expression . It is therefore possible that the functional alteration and disability associated with an injury are involved in long-term osteoarthritic changes to articular cartilage.
Inflammatory processes
Numerous inflammatory cytokines are found at increased levels in joint tissues during the acute post-injury phase, including IL-1, IL-6, IL-17 and TNFα . Lee et al. reported that this increase in cytokines may be mediated by resistin, an inflammatory mediator produced by macrophages . IL-1 has previously been shown to down-regulate ECM synthesis and up-regulate the synthesis of metalloproteinases via the induction of nitric oxide synthesis in chondrocytes . IL-6 and IL-17 have also previously been associated with cartilage damage, largely through their capacity to act synergistically with IL-1 in stimulating chondrocyte-driven destruction of the ECM .
While TNFα also has pro-degradative effects on chondrocytes, it has been implicated as the primary driver of synovial inflammation . The increase in TNFα post injury is also inversely proportional to the decrease in the levels of lubricin, a key glycoprotein involved in lubrication of the articular cartilage . This causes a decrease in boundary lubrication, a finding that correlates well with the level of cartilage damage . TNFα has also been shown to facilitate the release of IL-6 by synovial cells, exacerbating the damage to the adjacent cartilage . This is in accordance with recent evidence that suggests that synoviocytes may be key regulators of the joint inflammation that occurs in response to mechanical stimuli. Morisugi et al. demonstrated that the expression of nuclear factor kappa B (NFκB), cyclooxygenase-2 (COX-2), inducible nitric oxide synthase (iNOS) and poly adenosine triphosphate (ADP)-ribose by synoviocytes was increased following mechanical stress, and levels of IL-1, IL-6 and TNFα have all been found at elevated levels in synovial fluid following ACL tears .
There is growing evidence that the innate and adaptive immune systems may play roles in the post-traumatic joint, with increases in activated macrophages, CD4+ T cells, complement (C3a) and cytokines (IL-1, IL-15 and TNFα) reported in injured knee joints . The importance of soluble factors released from the synovium in the progression of load-induced cartilage damage has been demonstrated in co-culture studies . Similarly, the expression of IL-1 and iNOS by meniscal cells is increased in response to excessive compressive loading relative to physiological loading in vitro . Abnormal loading of the menisci may therefore explain the elevated expression of a number of catabolic factors (including IL-1, TNFα and MMP-13) by the meniscus in knees with an ACL tear compared to those with an intact ligament . These studies highlight the importance of the interplay between different joint tissues in regulating the response to mechanical injury, not only as a result of changing joint biomechanics but also indirectly through the secretion of soluble mediators that can regulate cell viability and ECM turnover.
Subchondral bone
Injury to the subchondral bone – as indicated by the presence of post-traumatic bone marrow lesions (BMLs) – appears to be exceedingly common in acutely injured joints, with recent studies reporting that BMLs are virtually ubiquitous on the lateral femoral condyle and the posterior lateral tibial plateau following ACL tears . Simple BMLs without involvement of the articular surface are thought to be benign occurrences, but geographic BMLs (characterized by their contiguity to adjacent cortical bone and their association with disruption to the articular surface) are associated with osteocyte necrosis within the affected bone marrow, as well as significant proteoglycan loss, chondrocyte injury and matrix degeneration in the overlying cartilage . It remains unclear whether the pathophysiological mechanisms that link geographic BMLs with overlying cartilage damage are biomechanical, trophic/soluble or a combination of the two. Nevertheless, Theologis et al. reported that the matrix composition of the overlying cartilage was still abnormal 1 year post injury, and was indicative of longer-term cartilage damage .
BMLs with particularly large volumes have also demonstrated an association with cortical depression fractures. These are likely to be of greater short-term clinical relevance than BMLs, with a recent study reporting that patients with cortical depression fractures had lower International Knee Documentation Committee (IKDC) clinical outcome scores and higher rates of meniscal tears at 1 year post injury .
There is also emerging evidence that structural changes in the shape of subchondral bone may occur significantly earlier than previously thought, and therefore may play a role in the alteration of cartilage loading that serves to hasten cartilage destruction. Hunter et al. recently reported that flattening of the femur and increased depression of the tibial surface was detectable within 3 months of ACL injury .
Preclinical identification of PTOA
Despite this increasingly detailed knowledge of the potential pathogenic mechanisms occurring in the injured joint, there is still no clear understanding of the reasons underlying the observation that only 20–50% of patients with joint injuries eventually develop clinically significant PTOA, whereas 50–80% of their injured colleagues do not . This represents a major obstacle to the future implementation of targeted therapies for PTOA, as the early stratification of patients on the basis of risk factors that determine their likelihood of developing clinical disease would be key to guiding treatment. The identification of biomarkers and predictors of future disease is therefore of great importance, and a number of possible biomarkers are currently under investigation.
Direct articular cartilage damage
There is evidence to suggest that direct articular damage sustained in the initial joint injury may correlate with cartilage loss at least 2 years post injury . This would certainly be in accordance with expectations, given the extensive chondrocyte death associated with direct impaction. A recent study by Frobell using data from the Longitudinal Knee Anterior Cruciate Ligament Non-operative vs. Operative Treatment (KANON) trial reported a number of significant findings at 2 years after ACL injury; cartilage thickening was described in the central medial aspect of the femur, whereas thinning had occurred in the femoral trochlear and the medial and lateral aspects of the femur . The increase in the thickness of cartilage in the acute injury phase has previously been suggested to be a result of the loss of small proteoglycans (such as decorin and fibromodulin) and other matrix components (cartilage oligomeric matrix protein (COMP) and type IX collagen, for example) that act to cross-link the fibrillar collagen matrix and resist the swelling pressure of the large proteoglycans, namely aggrecan . It is therefore possible that the central medial cartilage thickening represents some of the earliest macroscopic osteoarthritic changes, and so may serve as a marker of preliminary disease progression. Further studies with longer follow-up times are required, however, in order to establish whether patients with these early changes do in fact have an increased tendency to progress to clinical disease.
As noted by Su et al., the development of new magnetic resonance imaging (MRI) techniques such as T 1 ρ , T 2 and delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) have enabled much more sophisticated analysis of cartilage to be carried out than was previously possible using non-invasive techniques . The fact that it is now possible to quantify biochemical changes in the cartilage matrix, detect very early cartilage matrix loss and analyse cartilage matrix composition would suggest that future studies will be able to use more sensitive measures that better lend themselves to use as biomarkers that can feasibly discriminate between cellular hypertrophy, swelling and loss of cartilage, as well as track these changes longitudinally.
Meniscal injury
Approximately 65–75% of ACL-ruptured knees also show evidence of meniscal damage, with longitudinal tears in the posterior and middle one-third of the medial menisci accounting for the majority of lesions . A number of studies have reported that baseline meniscal damage is associated with cartilage loss at follow-up. Su et al. recently published their findings that baseline meniscal lesions were correlated with higher T 1 ρ values – reflective of degenerative biochemical changes in the cartilage – in the medial femoral cartilage at 2 years post injury . Similarly, Murrel et al. found that meniscal injury was associated with a threefold increase in cartilage loss 2 years after the initial injury . The meniscus increases the effective surface area for the transmission of force through the joint, and therefore decreases the contact stress experienced per unit of area . The increased risk of cartilage pathology with meniscal injury may reflect a loss of the capacity of the meniscus to distribute contact forces, a function that would appear to be protective of the articular cartilage. The meniscus is, however, a biologically active tissue that has been shown to respond to traumatic and trophic stimuli . As previously discussed, this response involves the synthesis of a variety of soluble enzymes and inflammatory mediators that may lead to damage to the adjacent cartilage. The well-documented link between partial or complete meniscectomies and earlier-onset osteoarthritic changes would appear to suggest that it is also protective against long-term OA, and therefore detecting meniscal trauma may serve as a useful biomarker for future disease .
Subchondral bone
As previously mentioned, geographic post-traumatic BMLs are associated with damage to the overlying articular cartilage and cortical depression fractures, both of which appear to be predictive of sustained chondral damage . Given that BMLs are easily identifiable as regions of diffuse signal abnormality in the subchondral bone marrow on MRI, they may serve as useful indicators of areas of articular cartilage that may be predisposed to long-term degradation, and hence future disease .
Molecular biomarkers
Given the pivotal role that a number of previously discussed molecular processes are thought to play in the pathogenesis of PTOA, it would be expected that molecules in the acute injury phase may provide the earliest indication of the direction of the future disease process. Lubricin has demonstrated particular promise, as it has been shown to decrease in proportion to alterations to the activity of synoviocytes and adjacent chondrocytes, possibly serving as an indicator of functional cartilage injury . Stromal cell-derived factor (SDF-1), cartilage ECM fragments (such as small and large proteoglycans, COMP, cartilage intermediate layer protein (CILP), fibril-associated collagens with interrupted triple helices (FACIT) and fibrillar collagens) and various cytokines and MMPs have been shown to be associated with acute joint injury, but it remains to be seen whether they are predictive of long-term structural damage .
Genetic biomarkers
Specific genetic abnormalities, such as those seen in type II collagen, have been associated with an increased risk of familial OA, but are comparatively uncommon . A number of large-scale genome-wide association scan (GWAS) studies have examined the minor genetic variations such as single nucleotide polymorphisms (SNPs) in human OA . These studies have identified a number of candidate genes for involvement in OA pathogenesis, but replication of significant alleles across different study populations has proven to be difficult. The effect size of individual SNPs is quite small, and is dependent upon a number of factors (gender, ethnicity and the affected joint, for example), limiting the potential utility of the genetic variations as predictive biomarkers for OA susceptibility. It is possible that the improving understanding of the importance of epigenetic phenomena such as DNA methylation and micro-RNA expression in regulating gene expression, and their evaluation in future genome-wide studies, may provide further information . No GWAS or epigenetic associated studies conducted to date have reported susceptibilities for PTOA or other OA subtypes, and the stratification of future studies may be important.
As discussed in a recent review, the use of genetically engineered mice (most notably “global or tissue-specific, constitutive or inducible gene knockouts, knockdowns and knockin mutations, as well as transgenic overexpression”) to model the molecular pathophysiology of PTOA has enabled the identification of a number of genetic modifications that protect against or worsen the erosion of articular cartilage . Of those that provided protection against the development of PTOA, the majority directly targeted cartilage matrix degradation, the differentiation and apoptosis of hypertrophic chondrocytes, or inflammation within the joint . Modifications that appeared to worsen the breakdown of articular cartilage in PTOA affected the same pathogenic processes . These studies clearly demonstrated the potential of specific genetic mutations that alter the synthesis of a variety of molecules – present both within the joint and in systemic metabolic processes – to alter the susceptibility to PTOA following joint injury. Although the discrepancies in terms of joint structure and physiology between mice and humans make direct extrapolation from animal models problematic, it is possible that the heterogeneity in PTOA disease progression and presentation in humans could be partly explained by differing expression of the genes identified in the mouse studies. Identification of such genetic associations would provide novel targets for use as predictors of disease.
Interventions
Primary prevention
The current paucity of treatments that help to prevent the eventual development of PTOA means that prevention of the initial injury is the most effective management tool.
A number of studies have attempted to assess the effect of preventative strategies on the rate of ankle injuries in sports, with the most common approaches being proprioceptive training and external support . Both taping and brace support mechanisms appear to provide reductions in the rate of ankle injuries. McGuine et al. recently reported that injury rates in college basketball players in a randomized controlled trial were significantly lower in athletes using a lace-up ankle brace than in the control group, and Sitler et al. found injury rates to be decreased roughly threefold (1.6 vs. 5.2 injuries per 1000 exposures) . There is evidence to suggest, however, that both support methods are markedly more effective in preventing re-injury than first-time injury, though the exact reasons for this finding have yet to be elucidated .
Proprioceptive training regimes have generally consisted of a range of balance exercises – such as balance board and single-leg and dynamic jumping exercises – performed with the aim of improving neuromuscular control of the joint . Multiple studies have reported decreases of up to 60% in the rates of ankle injury following proprioceptive training, though the heterogeneity in terms of the type, duration and frequency of the training programmes has made it difficult to identify an optimal protocol . A number of novel, sport-specific interventions have also been successful. Perhaps the best example is provided by the use of breakaway bases in baseball and softball, with an 80% reduction in sliding-related injuries seen following the their introduction .
Analysis of the results from studies carried out for the prevention of ACL injury is similarly problematic, with the complexity of the training programmes and the failure to adequately identify the intervention target (players or coaching staff) making it virtually impossible to determine which particular components of the programmes were effective. Despite this, there is a growing body of evidence to indicate that interventions involving neuromuscular training, aerobic conditioning, resistance training and plyometrics reduce the incidence of ACL ruptures . A recent systematic review by Gagnier et al. of six randomized controlled trials and eight cohort studies found a 50% decrease in the rate of ACL ruptures . Meta-analyses conducted by Hewett et al. and Grindstaff et al. reported similarly promising statistics, with fixed-effect estimates of 0.40 and 0.30, respectively . Major obstacles to the completion of large, randomized studies are the monetary and time cost of the programmes. Most intervention programmes are relatively expensive to run – with an estimated minimum cost of US$ 50–400 per athlete – and involve a substantial additional training load of approximately 3 h per week, making it likely that in-season programmes incorporating the training as an element of the warm-up would be the most widely implemented . It is worth noting that although preventative programmes are expensive, the cost of surgical reconstruction and rehabilitation is estimated to be approximately US$ 17,000 per patient, meaning that the prevention of a significant percentage of ACL tears would yield considerable long-term savings . This is quite apart from the economic costs; the fact that roughly 12% of OA is attributable to joint injury means that a decrease in the rate of joint injury would likely bring about a major decrease in overall disease burden ( Table 1 ).
Related posts:
The epidemiology of osteoarthritis
The role of imaging in osteoarthritis
Biomarkers for osteoarthritis: Current position and steps towards further validation
Patellofemoral joint osteoarthritis: An individualised pathomechanical approach to management
The role of arthroscopy in the management of knee osteoarthritis
Chronic disease management: Improving care for people with osteoarthritis
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
