Posttraumatic Bone Marrow Lesions

Posttraumatic bone marrow lesions (BMLs, also called bone bruises or contusions) are observed on magnetic resonance imaging as regions of diffuse signal abnormality in the subchondral bone marrow. These contusions are a result of the impaction forces between the anterolateral femur and the posterolateral tibia that occur during the initial trauma (kissing lesions), and are present in virtually all knees with complete ACL rupture. Most of these BMLs occur in the lateral compartment, most notably on the lateral femoral condyle and the posterior lateral tibial plateau as a result of the valgus distribution of force usually experienced during the injury ( Fig. 1 ). A recent study by Boks and colleagues suggests that, contrary to expectations, reticular posttraumatic BMLs are not associated with increased pain severity in posttraumatic knees. Simple posttraumatic BMLs without involvement of the articular surface are thus likely to be benign occurrences. Posttraumatic BMLs generally resolve without sequelae within 6 to 12 months after the injury, although new BMLs have been shown to develop in approximately one-third of ACL-injured knees over the first 2 years after injury.

Characteristic traumatic BMLs without associated osteochondral injury after rotational trauma. A 28-year-old patient suffered complete ACL disruption. ( A ) The most common locations of traumatic bone changes in conjunction with ACL disruption are the central lateral femoral condyle ( white arrows ) and the posterior lateral tibial plateau ( arrowheads ). Bone contusions are characterized by signal hyperintensity on fat-suppressed water-sensitive sequences. In addition, there is a subchondral fracture of the posterior lateral tibial plateau ( black arrow ). ( B ) At 4-month follow-up, there is almost complete resolution of BMLs. Only minimal residual hyperintensity is still observed in the central lateral femur. Note susceptibility artifact caused by femoral metallic screw after ACL reconstruction, which may be mistaken as a posttraumatic bone marrow edemalike lesion ( arrowhead ). ( C ) At 12-month follow-up, there is complete resolution of subchondral bone changes. Cortical depression of lateral tibial plateau is persistently observed.

In contrast, BMLs that are accompanied by disruption to the articular surface are predictive of long-term osteochondral sequelae. Johnson and colleagues found significant proteoglycan loss, chondrocyte injury, and matrix degeneration in the articular cartilage adjacent to a geographic BML, as well as osteocyte necrosis within the affected bone marrow. A separate follow-up study by Theologis and colleagues found that the matrix composition in cartilage overlying bruises in the lateral tibia was still abnormal 1 year after injury, despite most of the original osteochondral lesions healing almost completely within 2 weeks to 6 months. This finding indicates that the initial cartilage injury accompanying geographic bone bruises, osteochondral defects, and cortical impactions may lead to sustained cartilage trauma, and could therefore play a role in long-term osteoarthritic changes.

Large BML volumes have also been shown to be associated with the presence of cortical depression fractures, which are likely to be of greater short-term clinical relevance than the presence of a simple BML. A recent study published by Kijowski and colleagues reported that patients with cortical depression fractures had lower International Knee Documentation Committee clinical outcome scores 1 year after injury and higher rates of meniscal tears.

Meniscal Injury

Damage to the menisci is observed in approximately 65% to 75% of ACL-ruptured knees during arthroscopy. Traumatic longitudinal tears in the posterior and middle one-third of the medial menisci account for most lesions, although damage to the posterior-middle portion of the lateral meniscus is also common ( Fig. 2 ). It is still unclear whether this meniscal damage occurs primarily as a result of trauma during the initial injury or is secondary to the initial trauma and occurs between ACL rupture and arthroscopy. Retrospective observational studies have suggested that increased time between ACL injury and ligament reconstruction may result in higher rates of meniscal tears, but the fact that most of these reports are confounded by indication makes it difficult to interpret their findings.

Sagittal dual echo at steady state image shows a longitudinal meniscal tear of the posterior horn of the lateral meniscus in conjunction with a complete ACL disruption. Tear is characterized by a longitudinal hyperintense line extending from the meniscal upper surface to the undersurface ( arrows ).

Numerous studies have shown that meniscal damage in ACL-insufficient knees is associated with cartilage damage. Murrel and colleagues found that patients with meniscal injury had a 3-fold increase in cartilage damage 2 years after injury, and partial or complete menisectomies have long been linked to cartilage damage and earlier-onset osteoarthritic changes. This finding may indicate that the role of the menisci in reducing contact stresses and friction within the joint is protective of articular cartilage, and therefore the development of osteoarthritis. However, it is as yet unclear whether the loss of the meniscal function causes articular cartilage damage, or is merely a concurrent destructive occurrence.

Direct Articular Cartilage Damage

Nearly half of knee injuries that result in an ACL rupture also cause direct articular cartilage damage, particularly on the medial (41%–43%) and lateral (20%) femoral condyles ( Fig. 3 ). Direct cartilage damage is associated with short-term matrix disruption, chondrocyte necrosis, and proteoglycan loss. Although it is not yet known whether these changes are reversible, or become irreversible if a certain amount of damage is sustained, it is possible that the initial trauma plays a role in instigating the well-described progressive cartilage loss that is characteristic of osteoarthritis.

Traumatic articular cartilage damage in conjunction with ACL disruption. ( A ) Coronal T2-weighted fat-suppressed image shows a traumatic BML in the lateral femoral condyle ( arrowheads ) and an osteochondral depression ( arrow ). ( B ) Corresponding sagittal image shows depression and disruption of the articular surface ( arrow ). ( C ) Coronal short-tau inversion recovery image of different patient shows an example of a traumatic focal cartilage defect in the posterior lateral tibial plateau ( arrow ). Note subchondral traumatic BML adjacent to defect reflected as hyperintensity in the subchondral bone marrow.

A recent study by Frobell using data from the longitudinal Knee Anterior cruciate ligament NON-operative vs. operative treatment (KANON) trial reported that 2 years after injury, significant cartilage thickening was observed in the central medial aspect of the femur, whereas marked thinning had occurred in the femoral trochlea and the posterior aspects of both the medial and lateral aspects of the femur. These findings were particularly interesting in the context of osteoarthritis, given that osteoarthritis occurs predominantly in the medial compartment and that animal models have shown that cartilage hypertrophy precedes the characteristic cartilage breakdown.

How Strong is the Link?

As noted by Oiestad and colleagues in a systematic review in 2009, most studies assessing the long-term link between ACL rupture and osteoarthritis made use of inconsistent radiologic classification methods and heterogeneous populations with respect to treatment, previous activity levels, and the presence of concurrent injuries. It is therefore difficult to draw firm figures from the literature on the prevalence of osteoarthritis after ACL injuries, with reported rates ranging from 10% to 90% at 10 to 15 years after injury.

In their 2009 review, Oiestad and colleagues suggested that the lack of a consistent radiologic classification system (7 distinct classification systems were identified in the articles included in the analysis) has resulted in the prevalence of knee osteoarthritis after isolated ACL ruptures being greatly overestimated ( Fig. 4 ). Oiestad and colleagues conducted a methodological quality assessment of 31 studies and found that the highest-rated studies reported a prevalence for knee osteoarthritis of 0% to 13% after isolated rupture of the ACL; significantly less than the 50% to 70% prevalence rate often quoted in the literature. However, combined injuries involving ACL rupture and meniscal damage resulted in a higher prevalence of osteoarthritis of 21% to 48%. As discussed earlier, both meniscal injury and direct articular cartilage trauma are linked to long-term cartilage damage after a knee injury and are predictive of long-term tibiofemoral and patellofemoral osteoarthritis. Given that isolated rupture of the ACL is rare, and most ACL ruptures are accompanied by meniscal and chondral damage, the overall rate of osteoarthritis after an injury resulting in an ACL rupture is likely to be closer to the quoted combined injury rate than that reported for isolated injuries.

Radiographic osteoarthritis development over 5 years. ( A ) Anteroposterior radiograph at baseline obtained directly after trauma shows no signs of radiographic osteoarthritis. Normal medial and lateral joint space width and absence of osteophytes are observed. ( B ) Two-year follow-up image shows metallic screws in the femur and tibia after ACL reconstruction. Normal joint space width is observed. There is a tiny equivocal marginal osteophyte at the lateral tibial plateau. ( C ) At 5-year follow-up, definite osteophytes are observed at the lateral femoral condyle ( arrowheads ) and tibial plateau ( large arrow ), representing radiographic osteoarthritis grade 2 according to the Kellgren-Lawrence classification scheme. In addition, there is a prominent notch osteophyte at the lateral femoral condyle, potentially causing ligament impingement ( small arrow ).

Why the Increased Prevalence of Osteoarthritis After ACL Rupture?

It has long been suggested that osteochondral damage and intra-articular bleeding experienced during the initial trauma may induce a cascade of biochemical events within the joint that result in the development of osteoarthritis. Recent studies seem to support this idea, with Sward and colleagues reporting that an acute knee injury is associated with an immediate local biochemical response, potentially affecting the adjacent cartilage and bone in addition to inducing inflammation. However, little is known about the relationship between the immediate release of traumatic factors and subsequent osteoarthritis development, although the area is gaining increasing interest.

Recurrent episodes of instability may also play a role in initiating the pathologic changes to the articular cartilage observed after injury. It has previously been postulated that frequent episodes of instability or pivot shifting could result in sustained damage to both the articular cartilage and menisci that results in loss of the cartilage. The extent to which cartilage must be damaged in the initial trauma before structural damage becomes irreversible is not yet known, but it is possible that the regular occurrence of instability-related trauma and altered biomechanical loading could overwhelm the limited restorative capacity of the joint and lead to longer-term osteoarthritic changes. Concurrent injury to the menisci and the corresponding loss of its protective function would merely serve to exacerbate the damage to articular cartilage.

Although reconstructive surgery can partially restore joint stability after ACL rupture, it is unlikely that surgery fully restores normal biomechanical loading across the knee. An altered loading pattern causes a shift in compressive and tension load bearing to unconditioned regions, and reduces loads in conditioned regions. Numerous studies have described adaptations by cartilage to altered loading: chondrocyte metabolism and volume/aspect ratio, proteoglycan production, collagen fiber orientation, and matrix metalloproteinase expression are all altered during the cartilage response. It has therefore been suggested that early changes in cartilage may be partially explained by the altered biomechanics of the knee after injury.

Surgical Versus Nonsurgical Treatment

Despite a paucity of evidence that ACL reconstruction is the most effective treatment of an ACL rupture, more than 200,000 procedures are performed each year in the United States alone. Allografts, ipsilateral bone-patellar tendon-bone autografts and quadruple hamstring tendon (HT) autografts are the most commonly used procedures. Recent meta-analyses have suggested that although the 3 procedures produce similar long-term functional outcomes, allografts and HT autografts may be associated with lower rates of anterior knee pain.

The short-term benefits of surgical intervention in relation to nonsurgical treatments are still unclear, and several recent studies have reported that surgery and rehabilitation alone may produce comparable functionality. Frobell and colleagues conducted a randomized trial in which patients were assigned to receive either structured rehabilitation and early reconstruction or structured rehabilitation alone (with the option of delayed ACL reconstruction). These investigators reported that although early surgical treatment was associated with greater measured stability in Lachman and pivot-shift tests, after 2 years there were no significant differences between the treatment groups with respect to patient-relevant outcomes, knee-related adverse events, or return to preinjury activity levels. Similarly, a prospective cohort study by Moknes and Risberg found no difference in performance-based outcomes and the number of patients returning to preinjury activity levels between nonoperatively and operatively treated groups. In a case-control study, Meuffels and colleagues also found no statistical difference in activity levels or subjective and objective functional outcomes at 10 years after injury between patients treated conservatively or operatively.

ACL reconstructions are commonly advocated on the basis that they are protective against secondary meniscal injury and thereby reduce the risk of osteoarthritis development. Numerous retrospective studies have suggested that an increased time between injury and reconstruction is associated with higher rates of chondral and meniscal injuries. However, the studies are largely confounded by indication, because the fact that patients have symptomatic meniscal or cartilage injuries means they are more likely to undergo surgery. A long-term follow-up of a previous randomized controlled trial showed that, although the rate of secondary meniscal surgery was higher after nonsurgical treatment, there was no statistical difference in terms of radiographic osteoarthritis. A systematic review in 2007 had similar findings, reporting that no treatment-related differences in osteoarthritis could be found within the literature.

It is similarly unclear whether ACL reconstruction decreases the incidence of osteoarthritis development over the long-term. That osteoarthritis still develops in a substantial portion of patients treated with surgical repair is beyond doubt, but virtually no high-quality randomized studies comparing nonoperative treatment and reconstructive surgery have been conducted. A case-control study by Meuffels and colleagues found no statistical difference in prevalence of radiographic osteoarthritis between the operatively and nonsurgically treated groups. In 2007, Meunier and colleagues published the results of a 15-year-long trial in which 2 groups were allocated to receive either surgical treatment or conservative nonsurgical treatment by their year of birth (odd or even), and no statistically significant differences in osteoarthritis development were identified between the 2 groups. However, the investigators did note that there were several major problems with the randomization procedures used when the study was initiated in the early 1980s, most notably the exclusion of some patients from the surgical group because their injuries were not deemed to be amenable to surgical treatment, and the markedly different rehabilitation protocols used across the groups. This finding is symptomatic of the difficulties faced in attempting a meta-analysis, with the available studies all being of poor methodological quality or of insufficient length to allow a proper evaluation of osteoarthritis development. As noted in a recent Cochrane review, there is therefore a need for long-term randomized trials comparing surgical reconstruction and nonsurgical treatment to establish the efficacy of surgical repair in reducing the incidence of osteoarthritis.

However, from the limited evidence available, there is little to suggest that surgical intervention is significantly superior to conservative rehabilitation in terms of decreasing the rate of osteoarthritis.