The natural course of spontaneous osteonecrosis of the knee: A retrospective 3-year follow-up study.

Takayuki Kawasaki^a, Hisashi Kurosawa ^a, Hiroshi Ikeda^a, Yuji Takazawa ^a, Muneaki Ishijima^a, Mitsuaki Kubota^a,  Tokuhide Doi^b

Abstract:

Objective: This study investigated the characteristics of spontaneous osteonecrosis of the knee (SONK) in order to establish appropriate surgical indications.
Materials and Methods:This is a retrospective study of 70 knees with SONK observed in our hospital, with an average follow-up of 37.9 months after conservative treatment.
Results: The pain and functional scores of the subjects with SONK significantly improved over the period. X-ray findings worsened slightly, whereas limb alignment, size of necrosis, and knee range-of-motion were stable. Of the cases that progressed to higher X-rays stage during the follow-up period, 71.4% of knees changed within 12 months of disease onset. Of the cases that demonstrated stage III to IV at the first visit, 81.1% of knees maintained the same stage at the final examination. Conclusions            According to both observations, it seems reasonable to choose conservative treatment for SONK, even if X-rays of the affected knee reveal more than stage III severity.  

J.Orthopaedics 2009;6(1)e6

Keywords:

knee; osteonecrosis; conservative treatment

Introduction:

Spontaneous osteonecrosis of the knee (SONK) was first described in 1968 [1].  The typical SONK patient is a woman over sixty years old who has sudden onset of severe pain on the medial side of the knee which often is related to a specific activity or a minor trauma. In clinical situations, it is sometimes difficult to decide the timing and indication of surgical treatment because SONK shows variable prognoses [2-6]. Some cases need surgery because of severe discomfort and joint deformity, while in others the symptoms disappear a few months later and no joint deformity occurs.

We aimed to clarify the characteristics of this disease and establish an appropriate surgical indication. We hypothesize that the progression of joint deformity is quite slow or in due time reaches a plateau if the patient is responsive to conservative treatment. In this study, we assessed retrospectively the characteristics of subjects who successfully continued conservative treatment for SONK.

Materials and Methods:

We retrospectively evaluated the natural history of 70 knees in 64 subjects (48 female, 16 male) with SONK, with an average follow-up of 37.9 months (range 8 to 96). They were selected from the patient list of our hospital from January 1998 through December 2007 at their first visit. The average age was 68.9±8.5 years old and the average body mass index was 24.9±3.7 kg/m². All subjects initially demonstrated positive magnetic resonance imaging with a well-localized low signal intensity lesion on T1-weighted images over the medial femoral condyle that was diagnosed according to past reports [5,7] by expert radiologists. None of the patients had a history of cortisone consumption or any disease known to cause secondary osteonecrosis. The time interval between the onset of symptoms and the first visit was 13.0 months on average (range 0 to 60 months). All patients were encouraged to exercise the affected knee at home (description in detail as below) and used a lateral wedged sole to avoid mechanical overload to the medial compartment of the knee. Only non-steroidal anti-inflammatory drugs were used as pain rescue analgesics. The characteristics of our subjects are summarized in Table 1.       

^† Yates chi-square test

^†† unpaired t-test

Table 1: Subject characteristics

We recommended subjects perform the exercises as follows: (1) isometric muscle exercises of the bilateral lower limbs, 1 set (each exercise done 20 times) of straight leg rising training, and hip abduction and adduction exercises twice a day, in the morning and evening. SLR exercise: Lying on the back, raise the entire involved leg straight. Be sure to keep the involved ankle 10 cm above the floor. Hold for 5 seconds, then slowly lower. Hip abduction exercise: Lying on the uninvolved side, slowly lift the involved leg straight out to the side to the horizontal level. Hold for 5 seconds, then slowly lower. Hip adduction exercise: Sitting on the edge of a table or a high chair with the knees bent, squeeze an appropriately sized ball between bended knees. Hold for 5 seconds, then slowly open.  (2) As a ROM exercise, maximum flexion and maximum extension were performed twice a day in the morning and evening after the knee was warmed (in a bath or shower). After decreasing the pain and improvement in symptoms, we recommended normal speed walking as much as possible to the subjects.

For the clinical and physical evaluation, three scales that are usually used for osteoarthritis of the knee were used. With the visual analogue scale (VAS) for pain, patients can indicate their actual pain level using a 100-mm straight line as a scale of pain. The JKOM (Japanese Knee Osteoarthritis Measure) consists of 25 self-completion questionnaires with 4 subcategories: Pain and stiffness during activities of daily living, social activities, and general health with 100 points as the maximum score [8], incorporating the concepts of the World Health Organization’s International Classification of Functioning, Disability and Health (ICF 2004) [9]. The (Japan Orthopaedics Association) JOA score reflects the treatment criteria of the Japanese Orthopaedic Association, which is used by physicians for subjective evaluation of sub-categories: pain during walking, pain when using stairs, range of motion, and swelling, with 100 points set as the full score [10]. Clinical findings, as active range of motion (ROM; maximum extension and flexion in degrees) measured by an orthopaedic goniometer and swelling of the affected knee in three degrees, non=0, mild=1, moderate or more=2, were also recorded at the same time.

The radiographic evaluation was also assessed. Weight-bearing anteroposterior X-rays of the tibiofemoral joint using the standing extended view (also known as the standing anteroposterior view) were taken to measure the femorotibial joint space width (JSW) as the degree of joint deformity [11]. JSW was determined at the center point of the medial femorotibial compartment on a radiograph using a 0.1-mm graduated magnifying lens. In order to obtain acceptable X-rays for accurate reading, we decided to take X-rays according to the Buckland-Wright criteria [12] in which inter-margin distance (IMD) of the medial tibial plateau fell within 1 mm. If the IMD was over 1 mm, another X-ray was taken up to three times. If we failed to obtain an X-ray within 1 mm of IMD from the three X-rays, we adopted the X-ray of the least IMD. The femorotibial angle (FTA) as a parameter of limb alignment was also measured according to the method of Moreland et al. [13]. In the plane X-ray examination, we determined X-ray stage according to Agliatti et al. [2], and also simply assessed the occupied ratio (OR) of the osteonecrotic lesion in the standing anteroposterior view because exact measurement of the lesion was sometimes difficult, especially in the lateral view. In cases of early SONK with stage I that represented no radiographic abnormality, magnetic resonance imaging of the affected knee was also used to measure this ratio instead. All radiographs were obtained by an experienced expert technician and quantified by a single reader who was blinded to the features of the subjects.

Statistical analysis
Seventy knees in 64 SONK subjects were included in the analyses. Changes in factors including JKOM, VAS, JOA, ROM, JSW, FTA, X-ray stage, and OR from the first visit to the final status of follow-up were assessed using a paired t-test for parametric data, or Wilcoxon’s signed ranks test for other data.

It could not be ignored that our subjects had various time intervals from the onset to the first visit, and this appeared to affect their clinical course. Therefore, for analysis we divided subjects into two groups so as to divide the population close to equal distribution as follows: One was the acute or subacute group (n=40) of subjects who visited within 12 months from the disease onset, and the other was the chronic group (n=30) of subjects who visited later than 12 months. The characteristics of all subjects and each group are summarized in Table 1. The gender difference was estimated by chi-square test, and other factors were compared using the unpaired t-test. There was no difference between the both groups except for time intervals from disease onset and follow-up period.

Linear mixed model analysis was also devised to address the effect of inter-group and longitudinal course on the eight above factors. This method provides the most reliable results as it uses all available data to estimate the dependent variables. Subjects with at least 8 months’ follow-up were selected for analysis from the complete, longitudinal dataset. Time intervals from the first visit through the examination were divided into five categories as follows; the first visit, less than 12 months from the first visit, 12 to less than 36 months, 36 to less than 54 months, 54 months or more. The estimations were made by using 267 observations from 70 subjects. A P value of <.05 was considered significant and all tests were two sided. All data analyses were conducted with SPSS for Windows, version 16.0J (SPSS Inc., Chicago, IL, USA).

Results:

Pain and functional scores of the affected knee with JKOM, JOA, and VAS initially were 34.1, 70.0, and 50.5 on average, respectively (Table 2). The scores at the final follow-up examination were 22.7, 82.1, and 32.3, and the differences were statistically significant (P<.01) from the first visit. For radiographic evaluations, there were also significant differences regarding JSW and X-ray stage between the first visit and the last follow-up examination, while OR, FTA and ROM demonstrated no differences. In the acute group, there were significant differences between the first and last examinations regarding JKOM, JOA, VAS, and X-ray stage, whereas no such differences were seen in the chronic group regarding JOA, VAS, or X-ray stage.

† paired t-test
†† Wilcoxon’s signed rank test
ROM: range of motion, VAS: the visual analogue scale, JOA: the Japan Orthopaedics Association score, JKOM: the Japanese Knee Osteoarthritis Measure, JSW: joint space width of the affected condyle, FTA: the femorotibial angle

Table 2: Differences in variables between the first visit and the final follow-up

The changes in the X-ray stage from the first visit through the final observation are summarized in Table 3, which includes the cases for whom follow-up periods were at least 12 months. Of 21 cases that progressed to the higher X-ray stage during the period, 15 knees (71.4%) changed within 12 months of disease onset. Of 37 cases that demonstrated stage III to IV on the X-ray at the first visit, 30 knees (81.1%) maintained their stage level at the final examination.

Sixty subjects with at least 12months’ observation period are summarized.
* 38.5 months on average (range 12-96).

Table 3:Changes in X-ray staging between the first visit and the last examination.

There was no difference in the inter-group analysis regarding all dependent variables in linear mixed model analysis (Figure 1, Table 4). These analyses also revealed that VAS, JKOM, JOA, JSW and X-ray stage changed significantly with time, even in the chronic group. This means that time intervals are significant to improve of the functional scores of VAS, JKOM, and JOA. JSW and X-ray stage worsened slightly, whereas FTA, OR and ROM were unchanged in each group. 

^† Univariate Test. This test is based on linearly independent pairwise comparisons among the estimated marginal means.
ROM: range of motion, VAS: the visual analogue scale, JOA: the Japan Orthopaedics Association score, JKOM: the Japanese Knee Osteoarthritis Measure, JSW: joint space width of the affected condyle, FTA: the femorotibial angle

Table 4 Fixed effects in the linear mixed model analysis

The effect of longitudinal time course is significant in JKOM, JOA, VAS, JSW and X-ray stage,  representing the difference between the value at the first visit and that of follow-up periods. ** p< .01,  * p < .05 squares) white opened; at the baseline, oblique lined; less than 12 months from the first visit, dotted; 12 to less than 36 months, checked; 36 to less than 54 months, striped; 54 months or more. Figure 1. Linear mixed model analysis.

Typical cases of SONK with conservative treatment are shown in Figure 2 and 3. A 72 y.o. woman had severe knee pain with acute onset, scoring 45 on VAS, 54 on JKOM, and 70 on JOA (Figure 2). She visited our hospital 5 months after the disease onset. X-rays revealed typical characteristics of stage III SONK and she was treated conservatively with home exercise and protected from excess weight bearing. Three months after the first visit, she complained of intense knee pain with a VAS score of 90 and requested surgical treatment. We planned surgical treatment for the affected knee, but she could not undergo surgical treatment immediately due to her family reasons. Nine months after the first visit, her pain decreased gradually, although X-rays showed progression to stage IV. At the last observation, symptoms had almost resolved and X-rays showed no progression. She could climb a slope, fold her legs under herself in the Japanese style ‘seiza’, and even run a short distance without pain. Scores at the last examination were 20 on VAS, 11 on JKOM, and 90 on JOA. The OR was 37.4% and JSW was 2.1 mm at the last observation (-0.8mm compared to that of the first visit).

A.	B.
C.	D.
Figure 2. A 72 y.o. woman had severe knee pain with acute onset, scoring 45 on VAS, 54 on JKOM, and 70 on JOA. A. She visited our hospital 5 months after the disease onset. X-rays revealed typical characteristics of stage III SONK. A radiolucent area (arrowhead) located in the subchondral bone. B. After 3 months from the first visit, she complained of intense knee pain with a VAS score of 90. C. 9 months from the first visit, her pain decreased gradually although X-rays showed progression to stage IV. D. At the last observation, symptoms had almost resolved and X-rays showed no progression and slight sclerotic change. Scores were 20 on VAS, 11 on JKOM, and 90 on JOA. The OR was 37.4% and JSW was 2.1 mm at the last observation (-0.8mm compared to that of the first visit).

Another case of SONK belonged to the chronic group. An 83 y.o. woman who had severe pain with sudden onset 18 months earlier and was treated at another hospital followed by our hospital with continued pain, scored 40 on VAS, 30 on JKOM, and 85 on JOA at her first visit (Figure 3). X-rays revealed typical characteristics of stage III SONK and she was treated conservatively with home exercise and protected from excess weight bearing. Twenty-four months after the first visit at the last observation, symptoms almost resolved and X-rays showed no progression and slight sclerotic changes. She could climb a slope and even run a short distance without pain. Scores at the last examination were 10 on VAS, 9 on JKOM, and 100 on JOA. The OR was 42.5% and JSW was 4.0 mm at the last observation (no change from that of the first visit).

A.

B.

Figure 3. An 83-year-old woman had sudden onset of medial pain in the right knee. A. Plain radiographs after 18 months demonstrated a radiolucent area of the medial femoral condyle that appeared to be stage III SONK, scoring 40 on VAS, 30 on JKOM, and 85 on JOA. B. After 24 months: Symptoms had almost resolved and X-rays showed no progression and slight sclerotic change. Scores at the last examination were 10 on VAS, 9 on JKOM, and 100 on JOA. The OR was 42.5% and JSW was 4.0 mm at the last observation.

Discussion:

Since Ahlbach’s first description of SONK in 1968 [1], there have been many published data concerning its clinical prognosis and it is well known that the clinical course of this disease varies from remission with conservative treatment to severe dysfunction requiring surgical treatment [14].

In the cases of early SONK of stage I that shows no radiographic abnormalities, there are several reports of a benign prognosis with conservative treatment [7,15,16]. Yates et al. [16] described 20 sequential cases of early SONK diagnosed by magnetic resonance imaging (MRI) in which non-operative management led to the spontaneous resolution of symptoms and MRI changes at an average of 4.8 months after disease onset. In our study, most of the stage I cases changed to a higher stage at the last examination. It is possible that we missed early stage SONK by omitting of magnetic resonance image or radioisotope examination in our outpatient department. There were some attempts to predict the prognosis of SONK in past reports [7, 17]. It is worth noting that Locouvet et al. [7] demonstrated that predictive data for irreversible osteonecrosis by means of MR images was a subchondral low signal on T2 of >4 mm depth or >14 mm long, focal epiphyseal contour depressions and low signal lines deep in the affected condyle.

Regarding stage II or higher stages when these radiographic findings are visible on X-rays, the clinical course and prognosis were reported to depend on the radiographic size of the lesion [2,18,19,20]. In these reports, several authors noted that when a more than 50% of the occupied ratio of the lesion, and more than 5 cm square of the lesion area, the patient had a poor clinical and radiographic prognosis with rapid development of osteoarthritis and surgical treatment was recommended. In this study, we retrospectively evaluated SONK subjects who were successfully treated without surgical treatment over the past ten years. There was no discrepancy between the past reports and our findings that revealed a 32.5% lesion occupied ratio and 181.7 degrees of FTA on average at the final evaluation (37.9 months’ observation period on average). Aglietti et al., suggested other factors that worsened prognosis. They were flexion contracture and extent of local swelling [2], and our results were also consistent with these findings. In seventy knees with SONK, non of the knees represented severe contracture, and only 4 knees had swelling, scoring 1 in three knees and 2 in one knee at the final observation. Of these knees, the case with severe swelling (score=2) had worse symptoms and ROM, but not in terms of X-ray stage, which consistently showed stage 4 (data not shown in the results).

The difference between the first visit to the last observation regarding the eight examined factors in the acute group tended to be much clearer compared with that of the chronic group (table 2). The two-comparison analyses (paired t-test and Wilcoxon signed rank test) revealed that changes in VAS, JOA, ROM and X-ray stage were significant only in the acute group.

In our conservatively treated subjects with SONK, however, linear mixed model analyses revealed that pain and function gradually improved, JSW and X-ray stage barely progressed, and no malalignment or joint contracture was seen, regardless of the time interval from disease onset to the first visit. In addition, it appears that the improvements in pain and functional scores were rapid within a year from the disease onset, although the differences were not significant between the groups. After that period, the symptoms became stable and joint deformity tended to be unchanged or remained at a plateau in our selected population.

Among 60 subjects with SONK who were followed up for at least 12 months, the change in the X-ray stage from the first visit through the final observation reveals that 71.4% of knees (15/21) progressed to the higher stage within 12 months from disease onset (Table 3). Also 81.1% (30/37) of knees with stage III to IV on X-rays at the first visit maintained the same stage until the final examination. According to both observations, it seems reasonable to choose conservative treatment for SONK, even if X-ray of the affected knee reveals greater than stage III severity.  

In clinical situations, the timing of surgical treatment for SONK remains controversial. For younger SONK patients with severe malalignment, proximal tibial osteotomy is preferred in most cases, while for elderly SONK patients with secondary osteoarthritis, total knee arthroplasty may be the first choice. Unicompartmental replacement can be considered for patients with single compartment involvement and no degenerative changes in the opposite compartment in the femorotibial joint.

In the early stage, Koshino [14] concluded that bone grafting or drilling into the necrotic lesion are effective in promoting osteonecrosis healing, and surgery should be performed in the early stages of the disease (before the onset of degenerative changes) to obtain maximum clinical and radiographic improvement. It remains unclear when surgical treatment is indicated; the timing depends on whether degenerative changes have progressed to represent varus knee, an earlier stage of SONK, or severe impediment of activity of daily living regardless of these stages. 

Although it is clear that there are cases who need surgical intervention and cases without progression of joint deformity. Therefore, we think it important to consider the natural course of SONK under conservative treatment to avoid unnecessary surgery. Our findings suggest that surgical treatment may not be the first choice even if symptoms are severe at the first visit, especially within 12 months, or radiographic examination demonstrates stage III or later X-ray characteristics. Our recommended home-based exercise is effective for SONK to maintain good condition.

There are several limitations to this study as follows. First, surgical treatment cases were excluded from this study and were not compared with the present subjects, so we could not show the cumulative survival rate of SONK with conservative treatment.

Second, a 3-year follow-up period on average may not be enough to evaluate the prognosis of SONK, and some of our subjects may undergo surgery in the future.

Third, the term ‘12-months’ from the disease onset in this study was merely a convenient way to divide subjects into 2 groups according to equal deviation. It is necessary to calculate an accurate ‘safety’ period cut-off value so the ROC curve or survival rate can be computed together with cases of surgical intervention. Another study is required to explore these issues.

Although there are several limitations as described above, this study reveals relevance of conservative treatment for SONK. Further study is needed to explore the accurate timing of surgical treatment.

Conclusion:

The 70 knees receiving conservative treatment for SONK was retrospectively assessed. Most cases demonstrated that changes in the functional scores were correlated with the time interval from the disease onset, while the radiological aspects did not progress in the 3 year follow-up period. It appears that conservative treatment in useful to alleviate SONK symptoms, if it is over 12-months from disease onset, even if X-rays of the affected knee reveal greater than stage III severity.

Reference :

1. Ahlbäck S, Bauer GCH, Bohne WH. Spontaneous osteonecrosis of the knee. Arthritis Rheum 1968; 11: 705-33.

2. Aglietti P, Insall JN, Buzzi R, Deschamps G. Idiopathic osteonecrosis of the knee. Aetiology, prognosis and treatment. J Bone Joint Surg Br 1983; 65: 588-97.

3. Lotke PA, Ecker ML. Current concept review. Osteonecrosis of the knee. J Bone Joint Surg Am 1988; 70: 470-3.

4. Björkengren AG, Al-Rowaih A, Anders L, Thorngren K, Pettersson H. Spontaneous osteonecrosis of the knee: Value of MR imaging in determining prognosis. AJR 1990; 154: 331-6.

5. Motohashi M, Morii T, Koshino T. Clinical course and roentgenographic changes of osteonecrosis in the femoral condyle under conservative treatment. Clin Orthop Related Res. 1991; 266: 156-161.

6. Levcouvet FE, van der Breg BC, Maldague BE, Lebon CJ, Jamart J, Saleh M, et al. Early irreversible osteonecrosis versus transient lesions of the femoral condyles: prognostic value of subchondral bone and marrow changes on MR imaging. Am J Roentgenol 1998; 170: 71–7.

7. Akai M, Doi T, Fujino K, Iwaya T, Kurosawa H, Nasu T. An Outcome Measure for Japanese People with Knee Osteoarthritis. J Rheumatol 2005; 32: 1524-32.

8. World Health Organization. International classification of functioning, disability and health. Geneva: World Health Organization 2004. http://www.who.int/classifications/icf/en/

9. Koshino T, Saito T, Wada J, Akamatsu Y. New design of TKA; unicompartmental arthroplasty for osteoarthritis of the knee using the ceramic YMCK model. In: Niwa S, Yoshino S, Kurosawa M, Shino K, Yamamoto S eds. Reconstruction of the Knee Joint. Berlin: Springer, 1997: 200–206.

10. Ravaud P, Auleley GR, Chastang C, Rousselin B, Paolozzi L, Amor B, et al. Knee joint space width measurement: an experimental study of the infl uence of radiographic procedure and joint positioning. Br J Rheumatol 1996;35:761–6.

11. Buckland-Wright JC, Macfarlance DG, Williams SA, Ward RJ. Accuracy and precision of joint space width measurements in standard and macroradiographs of osteoarthritic knees. Ann Rheum Dis 1995;54:872-880

12. Moreland JR, Bassett LW, Hanker GJ. Radiographic analysis of the axial alignment of the lower extremity. J Bone Joint Surg Am 1987; 69: 745-9.

13. Koshino T. The treatment of spontaneous osteonecrosis of the knee by high tibial osteotomy with and without bone-grafting or drilling of the lesion. J Bone Joint Surg Am 1982; 64: 47-58.

14. Lotke PA, Ecker ML, Alavi A. Painful knees in older patients. Radionuclide diagnosis of possible osteonecrosis with spontaneous resolution. J Bone Joint Surg Am 1977; 59: 617-21.

15. Yates PJ, Calder JD, Stranks GJ, Conn KS, Peppercorn D, Thomas NP. Early MRI diagnosis and non-surgical management of spontaneous osteonecrosis of the knee. Knee 2007; 14: 112–116.

16. Mont MA, Baumgarten KM, Rifai A, Bluemke DA, Jones LC, Hungerford DS. Atraumatic osteonecrosis of the knee. J Bone Joint Surg Am 2000; 82–9:1279–90.

17. Muheim G, Bohne WH. Prognosis in spontaneous osteonecrosis of the knee: investigation by radionuclide scintimetry and radiography. J Bone Joint Surg Br 1970; 52: 605-12.

18. Rozing PM, Insall J, Bohne WH. Spontaneous osteonecrosis of the knee. J Bone Joint Surg [Am] 1980; 62: 2–7.

19. Lotke PA, Abend JA, Ecker M L. The treatment of osteonecrosis of the medial femoral condyle. Clin Orthop 1982; 171: 109-16.

This is a peer reviewed paper 

Please cite as: Takayuki Kawasaki: The natural course of spontaneous osteonecrosis of the knee: A retrospective 3-year follow-up study.

J.Orthopaedics 2009;6(1)e6

URL: http://www.jortho.org/2009/6/1/e6

ANNOUNCEMENTS

Arthrocon 2011


Refresher Course in Hip Arthroplasty

13th March,  2011

At Malabar Palace,
Calicut, Kerala, India

Download Registration Form

For Details
Dr Anwar Marthya,
Ph:+91 9961303044

E-Mail:
anwarmh@gmail.com

Powered by
VirtualMedOnline