Postarthroscopy Osteonecrosis of the Knee: Current Concepts

Abstract

Knee osteonecrosis is a severe disease rapidly leading to end-stage osteoarthritis, which was classified into three categories: spontaneous, secondary, and postarthroscopy. To understand postarthroscopy osteonecrosis of the knee, all the three types of knee osteonecrosis have to be deepened. This article reviewed spontaneous and secondary osteonecroses of the knee, with special focus upon postarthroscopy osteonecrosis, which is a rare form, affecting patients operated for arthroscopic knee surgery, most commonly for meniscectomy. Due to its rarity, patients and surgeons are often unprepared for this complication. A correct diagnosis is essential for appropriate treatment, and also to determine if a preexisting osteonecrosis was present, avoiding medicolegal sequelae, although many authors agree that osteonecrosis (both spontaneous and postarthroscopy) represent unpreventable and unpredictable conditions. In spontaneous osteonecrosis, the treatment is defined according to the size and the degree of the lesion, whereas in postarthroscopy osteonecrosis, the size of the lesion has no prognostic value, and therefore, the choice of the correct treatment is based more on the timing of the diagnosis. A diagnostic and therapeutic algorithm was outlined on the basis of the actual knowledge.

Introduction

Osteonecrosis of the knee is a severe disease rapidly leading to end-stage osteoarthritis, which was first described by Ahlbäck et al in 1968. 1 The knee is the second most common affected site after the hip. 2

Knee osteonecrosis was later delineated into three categories: spontaneous, secondary, and postarthroscopy. A good knowledge of the first two is essential to understand the third, which is the most rare and unexpected form. Many arthroscopic surgeons in fact are not aware of its existence until a case does occur.

Spontaneous osteonecrosis of the knee (SPONK) (also called primary osteonecrosis) is the most common type, with an incidence of 3.4% in patients older than 50 years and 9.4% in patients older than 65 years. 3

Secondary osteonecrosis affects more commonly younger patients, frequently have multiple and bilateral locations, and is associated with recognized conditions, such as alcohol abuse, corticosteroids, tobacco, sickle cell disease, and myeloproliferative disorders.

Postarthroscopy osteonecrosis is the rarest form, affecting patients operated for arthroscopic knee surgery, most commonly for meniscectomy.

This article reviewed the three types of knee osteonecrosis, summarizing the actual knowledge about demographics, etiology, diagnosis, and treatment. Special focus was made upon postarthroscopy osteonecrosis, delineating risk factors, differential diagnosis, medicolegal implications, and treatment.

Spontaneous Osteonecrosis of the Knee

SPONK is most common in patients older than 65 years, with a male-to-female ratio of 1:5, 3 and is typically unilateral. 4 The medial femoral condyle is affected in 94% of the cases. 5 Its real prevalence may be underestimated because many cases of end-stage osteoarthritis could be due to occult osteonecrosis. 6

Vascular and mechanical factors are involved. While a vascular etiology continues to be the dominant theory for osteonecrosis of the femoral head, these predisposing factors have yet to be consistently demonstrated in patients with SPONK. Nevertheless, a cadaveric study demonstrated limited vascular supply to the medial femoral condyle compared with the lateral. 7

Recent studies suggested subchondral insufficiency fractures in osteopenic bone, leading to synovial fluid accumulation in the bone marrow, with subsequent edema, focal ischemia, and osteonecrosis. 8 An association was demonstrated between low bone mineral density and the incidence of SPONK in women more than 60 years of age. 9

Case series have identified medial meniscal tears in 50 to 78% of patients with SPONK. 10 11 12 In elderly patients with osteoporotic bone, discontinuity of the medial meniscus results in loss of hoop stress distribution in the medial compartment, thus increasing the load experienced in the femoral condyle and potentially predisposing patients to the development of subchondral insufficiency fracture.

Patients with SPONK usually present with acute onset of medial knee pain, just proximal to the joint line. X-rays are usually negative. 13 Magnetic resonance imaging (MRI) is recommended for the diagnosis. 14

SPONK is commonly classified in four stages following the Koshino et al classification: 15 (1) knee symptoms with normal radiographic finding, (2) flattening in the weight-bearing area and subchondral radiolucency surrounded by osteosclerosis, (3) extended radiolucency and subchondral collapse, and (4) osteoarthritis.

The Ficat classification 16 , originally described for femoral head osteonecrosis, includes four stages based on joint space narrowing, subchondral collapse, and trabecular pattern: –(1) knee symptoms with normal radiographic findings, (2) sclerosis or cysts with normal joint line, (3) normal or slightly narrowed joint space with subchondral collapse and a sequestered appearance of the trabeculae (“crescent sign”), and (4) joint space narrowing and subchondral collapse.

The size of the lesion measured upon X-rays in case of SPONK represents a prognostic factor and gives indication for treatment, as lesions <3.5 cm ² usually regress with nonsurgical management, lesions sizing 3.5 to 5.0 cm ² may or may not regress, while large lesions >5 cm ² usually lead to condyle collapse. 6 17

Based on these findings, a treatment protocol for SPONK is summarized in Fig. 1 .

Fig. 1.

Treatment algorithm for SPONK. NSAIDS, nonsteroidal anti-inflammatory drugs; SPONK, spontaneous osteonecrosis of the knee; TKA, total knee arthroplasty; UKA, unicompartmental knee arthroplasty.

Secondary Osteonecrosis of the Knee

Secondary osteonecrosis of the knee (SONK) usually affects patients younger than 45 years, frequently with multiple lesions in multiple joints. SONK preferably affects women and hip joints, and is bilateral in 80% of cases. 2

Corticosteroid and alcohol abuse are the most common risk factors, being involved in 90% of cases. 2 Other risk factors include tobacco, sickle cell disease, Caisson's disease, and coagulopathies are responsible of vaso-occlusive effects ( Table 1 ).

Table 1. Secondary causes of osteonecrosis.

Alcoholism

Coagulopathies

Caisson's disease

Chemotherapy

Corticosteroids

Cushing's syndrome

Diabetes

Familial thrombophilia

Gaucher's disease

Gout

Hyperlipidemia

Inflammatory bowel disease

Liver disease

Organ transplantation

Pancreatitis

Pregnancy

Radiation

Renal disease

Sickle cell disease (and other hemoglobinopathies)

Smoking

Systemic lupus erythematosus (and other connective tissue disorders)

Tumors

In contrast to SPONK, SONK usually presents with gradual onset of pain. Pain may be located in multiple joints, and the knee may be affected in various sites. Similarly to SPONK, MRI is the recommended exam, while X-rays are often negative.

Secondary osteonecrosis showed different outcomes; in fact, symptomatic patients will almost uniformly need surgery. Therefore, nonsurgical management is recommended only for those patients who are asymptomatic. 2 It was found that only 19% of the symptomatic patients had satisfactory outcomes when managed nonoperatively, while 70% went on to eventually require total knee arthroplasty (TKA). Asymptomatic lesions treated nonoperatively were successful in 80%, avoiding arthroplasty and with no radiographic progression.

Postarthroscopy Osteonecrosis of the Knee

This condition was first described by Brahme et al in 1991. 18 Later, it was referred as “postarthroscopy,” 19 “postmeniscectomy,” 20 21 22 or “osteonecrosis in the postoperative knee” (ONPK). 23

Considering the high number of arthroscopic meniscectomies worldwide performed, the prevalence of ONPK seems to be very low. Santori et al 24 reported that of more than 2,000 knee arthroscopies performed at their hospital over 10 years, there were only two documented cases of osteonecrosis (0.2%) that occurred early, and mentioned that other cases could have been misdiagnosed. Pruès-Latour et al 25 reported nine cases among 585 arthroscopic meniscectomies on patients older than 50 years (1.5%).

A wide review published in 2007 by Pape et al 23 collected 47 cases of ONPK described in the literature after arthroscopic meniscectomy. The cited review did not consider the use of radiofrequencies (RFs) or laser procedures, in which thermal or photoacoustic effects could play a role.

Men and women were equally affected by ONPK after arthroscopic meniscectomy, with a mean age of 58 years. 23 This was in contrast with SPONK, which affects predominantly elderly females.

Of the 47 cases, 41 had a medial meniscus tear (87%) and 6 a lateral (13%). The medial femoral condyle was affected in 39 cases (83%), the lateral femoral condyle in 4 (8.5%), the lateral tibial plateau in 2 (4.3%), and the medial tibial plateau in 2 (4.3%). The affected area always coincided with the site of the arthroscopic procedure. 23 Sixty-five percent of the cases had a preexisting chondral lesion of varying degree.

The etiology of ONPK is debated. Altered knee biomechanics after meniscectomy may be a predisposing factor for osteonecrosis. 26 Increased tibiofemoral contact pressure might result in insufficiency fracture of the cartilage and subchondral bone with an intraosseous leak of synovial fluid and subsequent osteonecrosis. 27 28 A preexisting cartilage damage may have increased permeability for the arthroscopy fluid, which might lead to subchondral edema and consequent osteonecrosis. 25 29 In the same way, inadvertent contact of the arthroscopic instruments with the femoral condyle during meniscectomy may result in iatrogenic chondral damage, predisposing to ONPK. 30

Aggressive postoperative rehabilitation may contribute to the development of ONPK. It is possible that if aggressive therapy is resumed prior to bony remodeling in response to the altered load distribution that occurs postmeniscectomy, insufficiency fractures may develop.

Other authors have described the lesion as being in fact a subchondral fracture, and not pure osteonecrosis as traditionally described. 31 32 The arthroscopic procedure itself was discussed as a cause of osteonecrosis. 21 25 In addition, the meniscal tear itself should be a cause of osteonecrosis without arthroscopic surgery, especially in patient older than 60 years. 11 33 Other causes such as the use of an irrigation pump or tourniquet did not show an association with ONPK. 11 19 20

Other studies investigated the role of RFs procedures in the occurrence of ONPK. 29 34 35 The heat effect on the fluid medium and the direct transfer of energy to subchondral bone were thought to be the explanatory mechanism of ONPK after RF treatment. 34 36 37 38 39 40 41 Nevertheless, other studies did not confirm this hypothesis, and also found that RF chondroplasty resulted in a decrease in the number of patients developing ONPK. 36 42 43 They suggested that RF chondroplasty can be safely used for shrinking, smoothing of the joint surface, sealing, and debridement by controlled application of heat for the treatment of grades II and III chondromalacia. 44 Compared with mechanical shaving, RF debridement resulted in less chondrocyte death, 45 shortened operative time, 40 46 and created a smooth joint surface preventing irregular surface contact, with patients reporting fewer clinical symptoms in the postoperative period. On the contrary, mechanical shavers result in a 300- to 800-µm chondrocyte loss due to tissue removal, with the cartilage layer still remaining rough. 36

Some studies did not demonstrate significant chondrocyte death after RF treatment. 45 47 Others demonstrated chondrocyte termination extending in to the treatment area despite a well-controlled debridement. 48 49 According to Lu et al, 50 bipolar RF systems penetrates to 78 to 92% of the total depth of the cartilage, and causes more chondrocyte death and a larger thermal lesion than the monopolar systems. The majority of studies concluded that the relation between RF and ONPK is not meaningful, while meniscal tears and partial meniscectomy are thought to be the major responsible of ONPK. 36

Osteonecrosis lesions have also been noted following other arthroscopic procedures including anterior cruciate ligament reconstruction. 51 52 53

Diagnosis

MRI obtained in the early stages of ONPK will demonstrate a nonspecific, large area of bone marrow edema (BME) in the femoral condyle, ipsilateral to the prior meniscectomy with heterogenous signal present on T2 imaging ( Fig. 2 ). By 3 months postoperatively, the extent of edema typically decreases and MRI findings in cases of ONPK are similar to those seen in cases of SPONK with T1 imaging showing a discrete low signal area surrounded by an area of intermediate signal intensity. A line of low signal is often present at the margin of the lesion, delineating the necrotic area from the adjacent area of BME. T2 images will typically show a high signal intensity at lesion edge in the region of the BME ( Fig. 3 ). As the lesion progresses to its final stages, bone sequestration may be present with a surrounding high signal rim along with condylar flattening and the possibility of loose body development. 54

Fig. 2.

Magnetic resonance imaging T2 scan showing bone marrow edema in the early phase of osteonecrosis of the medial femoral condyle.

Fig. 3.

Magnetic resonance imaging T1 scan showing the delineated necrotic area surrounded by bone marrow edema.

In cases of persisting or worsening pain after knee arthroscopy, the surgeon must distinguish among a preexisting SPONK, ONPK, transient BME, and a recurrent meniscal tear.

To establish the diagnosis of ONPK, it is necessary the absence of BME on a preoperative MRI performed 4 to 6 weeks after the onset of preoperative symptoms, and its presence on the postoperative MRI. Johnson et al 19 have chosen 6 weeks as the minimal time interval between the onset of symptoms and MRI examination. This “window period” was based on an animal study by Nakamura et al, 55 who surgically induced osteonecrosis in the femoral head in a canine model and showed that it may take up to 4 weeks between surgery and the MRI findings to become positive.

In the 47 cases collected by Pape et al, 23 a preoperative MRI was performed in 44 of 47 patients, but the exact onset of clinical symptoms before preoperative MRI was not always mentioned. In total, up to 28 of the 47 ONPK cases (59.5%) could represent a preexisting undiagnosed SPONK.

Kobayashi et al 21 found postoperative BME on MRI in 34% of patients after partial meniscectomy within 8 months after surgery, with no preoperative signal changes. Postoperative changes were restricted to the meniscectomized compartment, both on the proximal tibia and on the distal femur. There was no correlation between the incidence or extent of BME and age, gender, or degree of chondromalacia. In addition, they did not observe progression of the disease to ONPK. 21 56 Muscolo et al 11 and Pruès-Latour et al 25 suspected that the chance of progression to ONPK after partial meniscectomy seems to increase in patients aged older than 50 years. To assume the presence of ONPK, early-stage SPONK must have been excluded before arthroscopy, which is not a current or cost-effective standard of care. Based on the actual knowledge, a diagnostic algorithm was created and illustrated in Fig. 4 .

Fig. 4.

Diagnostic algorithm for ONPK. BME, bone marrow edema; ONPK, osteonecrosis in the postoperative knee; MRI, magnetic resonance imaging; SPONK, spontaneous osteonecrosis of the knee.

Treatment

Of the 47 patients diagnosed with ONPK in the review by Pape et al, 23 44 (93.6%) had either permanent MRI lesions or showed a progression to irreversible stages. In 17 of these 47 patients (36%), further surgery was needed. Knee arthroplasty has been performed in nine, high tibial osteotomy (HTO) in two, and repeat arthroscopy in six.

It was well documented that the size of the lesion as percentage of the diameter of the medial femoral condyle is a prognostic factor and guides the treatment in SPONK. 6 17 57 58 59 60 In ONPK, the size of the lesion has rarely been correlated with outcome. 23 It seems that even relatively small bone marrow changes on postoperative MRI may lead to osteonecrosis. Nevertheless, the treatment protocol should be uniform for SPONK and ONPK.

A nonsurgical treatment protocol is recommended and more prudent than quickly performing another surgical intervention with the potential for accelerated joint destruction. 23 Protected weight bearing with crutches for a 4- to 8-week period is recommended, coupled with nonsteroid anti-inflammatory medications. As the symptoms improve, gradual resumption of normal activities is allowed, in conjunction with physical therapies. An 89% success rate was reported with conservative management of grade I SPONK. 57 In contrast to SPONK, the nonoperative treatment of ONPK is less successful, with only 3 of 47 cases improving in the series of Pape et al. 23

Bisphosphonates showed to be beneficial in the nonsurgical management of both SPONK 61 and ONPK, 62 by preventing resorption of the necrotic region. 63 These data were not confirmed by other studies. 64 Available medical treatments also include vasodilators, statins, and anticoagulants. Pulsed electromagnetic fields therapy demonstrated an improvement in pain and a reduction of the necrotic area at MRI in early-stage SPONK. 65

The use of the hyperbaric chamber in the treatment of SPONK for young patients was described, 66 even if further reports and studies are needed to demonstrate the validity of this treatment.

Surgical management should be considered when patients do not improve clinically or radiographically after 3 months of nonoperative treatment, as well as in patients who present with osteonecrotic lesions larger than 5 cm. 6 17

Joint-preserving surgical techniques may successfully postpone the need for joint arthroplasty. 67 The use of arthroscopic debridement in the management of SPONK and ONPK has limited applications. As the primary pathology is intraosseous, arthroscopic debridement has little likelihood of altering the course of the disease process; however, it may lead to symptomatic improvement when mechanical symptoms are present due to unstable chondral fragments or loose bodies. While retrograde drilling or microfractures may stimulate revascularization within the lesion, the potential for damage to the intact articular surface and the difficulty associated with accurately localizing the focus of the lesion makes antegrade drilling or core decompression more attractive treatment options.

Core decompression ( Fig. 5 ), first described by Jacobs et al, 68 demonstrated successful healing and pain relief in stages I and II SPONK. 69 Bone grafting could be associated with decompression, using autogenic or allogenic cancellous bone or demineralized bone matrix. 70 71

Fig. 5.

X-rays showing a core decompression procedure in for osteonecrosis of the medial femoral condyle.

Joint-preserving techniques include HTO, which may also be considered in young, active individuals who have failed core decompression and still have early-stage disease. HTO can offload the affected femoral condyle by shifting the weight-bearing axis laterally ( Fig. 6 ). HTO demonstrated better improvement of symptoms and MRI features of the lesions, compared with nonoperative treatment. 72

Fig. 6.

Osteonecrosis of the medial femoral condyle treated by high tibial osteotomy.

However, patients who have progressed to subchondral collapse may benefit more from osteochondral autograft due to restoration of the cartilage surface. 67 73

Osteochondral allograft transplantation is useful even in large or complex lesions, and has the advantage of restoring mature hyaline cartilage to the affected area. 74

For patients in whom joint-preserving treatments fail to provide symptomatic improvement and in those with large or advanced lesions, knee arthroplasty is the treatment of choice. Depending on patient factors, lesion characteristics, and the condition of the remainder of the joint, unicompartmental arthroplasty or standard TKA may be utilized. A treatment protocol for ONPK is outlined in Fig. 7 .

Fig. 7.

Treatment algorithm for ONPK. NSAIDS, nonsteroidal anti-inflammatory drugs; ONPK, osteonecrosis in the postoperative knee; TKA, total knee arthroplasty; UKA, unicompartmental knee arthroplasty.

Conclusion

ONPK is a rare form of osteonecrosis that often comes unexpected for both the patient and the surgeon. Its knowledge is important for the arthroscopic surgeons, given the number of knee arthroscopies worldwide performed.

In the case of ONPK, the surgeon might be involved in medicolegal sequelae. However, many authors agree that both SPONK and ONPK represent unpreventable and unpredictable conditions. Also, the diagnosis of ONPK is not always sure, as it is not possible to exclude a preexisting SPONK if the preoperative MRI was performed in the so-called window period. An original diagnostic algorithm has been described to assist the surgeon to orientate in the ONPK diagnosis. Basically, it is important to know that elderly patients with meniscal tears and chondral lesions should be alerted that there is a rare risk of osteonecrosis developing after knee arthroscopy. 19 The meniscal tear itself may be a cause of osteonecrosis, especially in patient older than 60 years. 11 The majority of studies agree that risk factors exist for the development of ONPK, but a certain causal relationship with arthroscopy is lacking. 21 25 75

The prognosis and treatment of ONPK are not in relation to the size of the lesions, but rather to the timing of the diagnosis. An early diagnosis is required to have chances of success with conservative treatment. The latter is always indicated as first treatment. If, however, more than 6 months from the previous intervention have passed, surgical treatment can be considered directly, depending on the degree of the lesion as for SPONK. On the basis of current knowledge, an original treatment algorithm of ONPK has been described, inspired by the treatment for SPONK, but differing in the indications, which are not dictated by the size of the lesion, that in case of ONPK have not demonstrated prognostic value, but by the timing of treatment.