Although loose bodies within a joint were first described by Paget,¹ König later suggested three methods by which loose bodies could be created: (1) direct trauma with acute fracture, (2) minimal trauma that develops into osteonecrosis and subsequent fragmentation, or (3) no trauma with spontaneous fragmentation. The latter variety he coined osteochondritis dissecans (OCD).²,³ Although it should be pointed out that the exact pathophysiology remains unknown, it is agreed that OCD is likely an acquired lesion of subchondral bone.

Beyond this characterization, it is less clear. There are degrees of osseous resorption, collapse, and sequestrum formation with possible involvement of the articular cartilage through delamination unrelated to an acute osteochondral fracture of normal cartilage (Fig. 21.1).⁴,⁵ This understanding of the end point of the disease process has led to many etiologies of OCD being postulated (particularly concerning the knee) including trauma,⁶,⁷ inflammation,²,⁸ genetics,⁹ vascular abnormalities,¹⁰,¹¹ and constitutional factors.¹² However, the etiology remains unknown, even though our veterinary medicine colleagues have made some leaps in understanding over recent years.¹³,¹⁴

Historically, there has been a distinction between juvenileonset OCD and adult-onset OCD. Many surgeons have suggested that skeletally immature patients (juvenile onset) have a better prognosis that has been inconsistently defined in the literature as either radiographic healing or merely resolution of pain.⁴,⁵,⁹,¹¹,¹²,¹⁵ Despite the lack of an open physis at the time of diagnosis in an adult OCD, however, many authors suggest that the only true difference between juvenile- and adult-onset OCD is purely a reflection of patient age at the time of diagnosis.

A recent definition of human OCD lesions, proposed by the Research in Osteochondritis of the Knee (ROCK) study group, highlights the fact that these are (1) focal, (2) idiopathic, (3) involve subchondral bone, and (4) risk instability and disruption of articular cartilage with potential long-term consequences, such as premature osteoarthritis.¹⁶ This definition is the summary of epidemiology, etiology, classification, and assessment of OCD.

There are only three true epidemiology papers regarding knee OCD.¹⁰,¹⁷,¹⁸ The first was performed by Marsden and Wiernik¹⁸ in a review of 18,405 radiographs at a military hospital. They found an incidence of symptomatic OCD of 2.3% of the radiographs and an overall incidence (including incidental discovery) of 4% OCD in their cohort.

A classic study by Linden¹⁰ in 1977 from Malmo, Sweden demonstrated an incidence of 29 per 100,000 boys and 19 per 100,000 girls. More importantly, after he reviewed radiographs and obtained follow-up with many of these patients 33 years later, he discovered that there was an “accumulated risk of roentgenographic gonarthrosis in patients who have had osteochondritis dissecans.” If he assumed the risk of arthritis to be 0% at the start of life, this was unchanged at age 40 years with an OCD; but the risk increased to 70% at age 48 years and continued to increase to 95% at 70 years of age. However, in those initially discovered with juvenile-onset OCD, he could not directly correlate any pathology such as gonarthritis with their OCD directly.

The only other study was published in 2014 and included a review of just over 1 million children aged 2 to 19 years within a closed health system.¹⁷ These authors found 192 children with 206 OCD lesions of the knee. The majority (64%) of the lesions involved the medial femoral condyle, and the overall incidence seen in the cohort was 18.1 per 100,000 boys and 3.9 per 100,000 girls. The incidence of knee OCD varied by ethnicity: non-Hispanic white was 10.3 per 100,000 overall (17.3 and 3.0 per 100,000 for boys and girls, respectively), non-Hispanic black was 10.3 per 100,000 overall (17.3 and 3.0 per 100,000 for boys and girls, respectively), Hispanic was 8.6 per
100,000 overall (14.3 and 2.8 per 100,000 for boys and girls, respectively), and Asian was 4.7 per 100,000 overall (9.1 and 0.0 per 100,000 for boys and girls, respectively). These authors performed multivariable logistic regression analysis that revealed a 3.3-fold increased risk of OCD of the knee in children aged 12 to 19 years compared with those aged 6 to 11 years. Moreover, boys had 3.8 times greater risk of OCD than girls.

Finally, there is the discrete possibility that knee OCD may occur in both knees. In the literature, there is a relatively wide range of bilaterality noted with about a 3% to 30% chance of discovering it in both knees by x-ray.⁵,¹⁸,¹⁹,²⁰,²¹,²²,²³

No definitive etiology has yet been determined for the origin of knee OCD. There are, of course, many hypotheses that have been presented and tested primarily via ex vivo histology. The potential etiologies include inflammation, spontaneous osteonecrosis and vascular deficiency, genetic predisposition, and repetitive trauma. Each of these will be discussed.

As suggested by the name “osteochondritis,” the first description in 1888 by König,² by definition was atraumatic and it was postulated to be a result of inflammation. However,
histologic studies have not supported this etiology⁷,²⁴; instead, these works appear to highlight findings of necrosis within the OCD lesions rather than inflammation.

Based on their histology findings, Green and Banks¹²,¹⁹ proposed that ischemia was the primary etiology, and Milgram,²² identifying revascularization in partially detached OCD lesion, further promoted this concept of poor vascularity. Yet, Yonetani and colleagues²⁵ found no evidence of necrosis on biopsies. Uozumi and colleagues²⁶ discovered a discrete absence of subchondral bone in many of their biopsy samples and, in those who had an osseous component present, only two demonstrated no viable osteocytes. The difference between the first two studies and the second two is state of the OCD lesion prior to biopsy. When the lesion was unstable (or even a loose body) at the time of biopsy, then necrosis was found by microscopic evaluation. However, if the lesion was fully intact and in situ, then there was less definitive evidence for avascular necrosis.

Another hypothesis of the etiology is familial inheritance of OCD lesions. A mild form of skeletal dysplasia with associated short stature has even been proposed.⁹,²³,²⁷,²⁸,²⁹ Two different authors, via different family tree assessments, have identified what they believe to be an autosomal dominant inheritance pattern.⁹,²⁸ In contrast, Petrie³⁰ reported on a radiographic examination of first-degree relatives of those patients with known OCD and discovered only 1.2% with OCD themselves. This suggests that the majority of knee OCD does not follow a predictable inheritance pattern. However, it does not exclude the possibility that genetics may still play a role in OCD etiology.