CLINICAL OUTCOMES AFTER UNICOMPARTMENTAL KNEE ARTHROPLASTY FOR OSTEONECROSIS OF THE KNEE.

OBJECTIVE: Although the mobile-bearing Oxford unicompartmental knee arthroplasty (OUKA) seems an appropriate procedure to treat spontaneous osteonecrosis of the knee (SONK), aseptic tibial component loosening was the leading cause for medial UKA failure. This study aimed to observe short-term and midterm clinical outcomes following OUKA and determine whether tibial lesion affects the procedure clinical and radiographic outcomes.
METHODS: Sixty patients (mean age 73.1 ± 6.6 years) diagnosed with SONK in the medial femoral condyle and treated with OUKA were separated into two groups using T1-weighted preoperative magnetic resonance imaging (MRI): group F (necrotic lesion confined to the femur) and group T (necrotic lesion spread to the tibia). The Oxford Knee Score (OKS), maximum flexion angle (MFA), and radiographic findings (radiolucent line and subsidence) were compared between the two groups using unpaired t-test.
RESULTS: Both groups showed significant improvement in OKS and MFA values at the final follow-up, but without significant differences in the clinical and radiographic outcomes.
CONCLUSION: OUKA is a reliable treatment procedure for SONK in the short and midterm. The presence of tibial lesions on preoperative MRI does not affect postoperative radiographic and clinical outcomes. Level of Evidence IV, Case Series.

INTRODUCTION

Unicompartmental knee arthroplasty (UKA) seems an appropriate procedure for treating spontaneous osteonecrosis of the knee (SONK) with unaffected lateral and patellofemoral compartments. Less invasive than other surgical procedures and with positive clinical outcomes, UKA replaces only the affected condyle and preserves all major ligaments, including the anterior cruciate ligament (ACL). )- ( However, a recent systematic review on the use of UKA for the treatment of SONK verified a wide variation in survival rates within the literature, stressing the importance of appropriately selecting patients. (

Despite extensive knowledge on SONK affecting the femoral condyle, studies approaching SONK in the tibial plateau are still scarce due to its lower prevalence when compared to the femoral component, comprising only approximately 2% of the cases. ( Recent studies found tibial component aseptic mechanical loosening and subsidence to be the leading causes of UKA failure. ( )- ( A recent research reported the use of Oxford UKA for the treatment of medial tibial plateau osteonecrosis to achieve positive clinical outcomes. ( However, we found no reports in the literature comparing UKA outcomes for SONK affecting the femoral condyle and the tibial plateau due to the major difference in their incidences. Our hypothesis is that the presence of lesion on the tibial component, verified by preoperative magnetic resonance imaging (MRI), can affect UKA clinical and radiographic outcomes for SONK treatment.

This study sought to observe the short and midterm clinical outcomes of mobile-bearing Oxford UKA for treating SONK and determine whether the presence of tibial lesions in preoperative MRI would affect the clinical and radiographic outcomes of the procedure.

MATERIALS AND METHODS

Subjects, clinical data, and surgery

The study protocol was approved by the Ethics Committee of our hospital (No. 2018-79). All patients included in this study signed an informed consent form. Our study group comprised 60 patients diagnosed with SONK affecting the medial femoral condyle, treated in our hospital between 2012 and 2014 with UKA using an Oxford mobile-bearing knee implant (The Oxford Partial Knee, Zimmer Biomet Ltd), and followed up for at least 3 years after surgery. In case of persistent pain and joint collapse with failure of conservative treatment after at least 3 months, surgical treatment was performed. Patients without multicompartment necrotizing involvement, soft tissue imbalance, ligament involvement, and coronal malalignment greater than 15° underwent UKA. Patients whose outcomes, radiographic evaluation, and knee condition could be measured at the outpatient clinic during the final follow-up were included in the study. Exclusion criteria were knees with fixed flexion greater than 15°, active knee joint infection, and bilateral UKA. Among the 60 patients that participated in the study, 47 were women and 13 men (mean age, 73.1 ± 6.6 years; body mass index, 23.1 ± 3.6 kg/m2) (Table 1).

Table 1

Patient demographic data.

	Group F (N=34)	Group T (N=26)	p-value
OKS
preoperatively	23.4±9.2	22.8±9.4	0.65
final follow-up	37.8±8.8	38.3±9.1	0.76
improvement	14.4±6.7	15.5±6.1	0.51
MFA
preoperatively	109.8±12.9	107.9±12.0	0.88
final follow-up	132.2±11.5	130.7±11.9	0.71
improvement	22.3±6.1	22.8±5.7	0.85
Complications
Superficial infection	0	1	N.S
Deep infection	1	0	N.S
Tibial plateau fracture	0	1	N.S
Revision cases	1	0	N.S
	Group F (n=34)	Group T (n=26)	p-value
Sex	male 8, female 26	male 5, female 21	NSw
Age	73.4 ± 6.4 y.o	72.7 ± 7.6 y.o	P=0.86
Body mass index	22.7 ± 3.5 kg/m2	23.5 ± 3.7 kg/m2	P-0.42
Varus deformity	7.6° ± 5.3°	7.1° ± 4.7°	P=0.77
Period from onset to surgery	61.0±9.8 weeks	68.6±11.1weeks	P=0.66
Follow-up	60.5±9.3 months	56.7±10.4 months	P=0.49

Group F: necrotic lesion confined to the femur; Group T: necrotic lesion spread to the tibia; N.S: not significant.

Among patients, the average preoperative coronal plane alignment on standard weight-bearing anteroposterior (AP) radiographs was 7.4° ± 5.1° in varus. All surgeries were performed by the senior author (TH) or by surgeons directly instructed by him.

Surgical procedures were conducted as described in the literature. ( Before performing UKA, all intact cruciate ligaments and healthy cartilage in unaffected compartments were confirmed. Cemented implants were used in both tibial and femoral sides for all cases. Osteonecrotic lesion was removed as much as possible so that normal bone could be used as base for cement impregnation while avoiding deeper bone cut, thus prioritizing joint-line preservation and soft tissue balance over complete removal of necrotic lesion. After bone cut, eventual large craters were filled with autologous bone graft, harvested from the bone removed during surgery according to a previously reported method. (

Preoperative MRI classification

Patients were separated into two groups using T1-weighted preoperative magnetic resonance imaging (MRI). Whereas in group F the necrotic lesion was confined to the femur, in group T it was spread to the tibia (Figure 1).

Figure 1

Preoperative T1-weighted magnetic resonance imaging classification. (A): Group F - necrotic lesion confined to the femur; (B) Group T - necrotic lesion spread to the tibia.

Clinical outcome

Clinical outcome was measured using the Oxford Knee Score (OKS) - a patient-based questionnaire that allows patients to report functional activity levels and clinical symptoms. OKS has been validated for use in degenerative arthrosis of the knee and received a Japanese translation, which was also validated. (

The maximum flexion angle (MFA) was measured using a goniometer, both preoperatively and postoperatively (at the final follow-up), to assess clinical outcome.

Both parameters (OKS and MFA) were compared 2 weeks preoperatively and postoperatively (at the final follow-up) using the paired t-test (p < 0.05).

Radiolucent line assessment

Radiolucent line (RLL), that is, the radiolucent interval (measured in millimeters) between the cement and the bone, ( was assessed using AP radiography at the final follow-up. Radiolucency was evaluated by adjusting the X-ray beam direction parallel to the tibial component and dividing the bone-cement interface into four zones, according to a previously reported method (Figure 2). (

Figure 2

Standardized analysis of the tibial bone-cement interface in four zones, according to a previously reported method.

Those with RLLs in 1-3 zones were considered partial RLL, whereas those with RLLs in all 4 zones were considered completed RLL.

Assessment of tibial component subsidence

The subsidence of tibial component was evaluated according to changes in radiographic parameters from 2 weeks postoperatively to the final follow-up.

Subsidence distance was measured by the height difference between the center of the lateral compartment and the line in contact with the tibial prosthesis lower surface at 2 weeks postoperatively and at the final follow-up. Subsidence angle was measured by the angular difference between the lateral compartment and the line in contact with the tibial prosthesis lower surface at 2 weeks postoperatively and final follow-up (Figure 3).

Figure 3

Assessment of tibial component subsidence. The subsidence of the tibial component was evaluated according to changes in radiographic parameters from 2 weeks postoperatively (A); to final follow-up (B). Subsidence distance (mm) = b − a. Subsidence angle (°) = β − α.

Statistical analysis

Results were analyzed using StatView 5.0 (Abacus Concepts Inc., CA, USA). RLL ratios between the groups were compared using Pearson’s chi-square test (p < 0.05). Clinical outcome (OKS and MFA) and subsidence (distance and angle) at the final follow- were also compared using unpaired t-test (p < 0.05). All values are presented as mean ± standard deviation (SD).

Considering a prespecified significance level at α < 0.05 and assuming a medium effect size (effect size = 0.5) using G power 3, the statistical power analysis performed before the study expected a 0.8 power. ( The estimated sample size was 54 patients, and p < 0.05 was considered statistically significant.

RESULTS

Clinical data

Table 2 shows the clinical data of both groups (group F, n = 32; group T, n = 28). We found no significant differences regarding age, sex, body mass index (BMI), and follow-up period between groups.

Table 2

Clinical outcome for both groups.

	Group F (N=34)	Group T (N=26)	p-value
OKS
preoperatively	23.4±9.2	22.8±9.4	0.65
final follow-up	37.8±8.8	38.3±9.1	0.76
improvement	14.4±6.7	15.5±6.1	0.51
MFA
preoperatively	109.8±12.9	107.9±12.0	0.88
final follow-up	132.2±11.5	130.7±11.9	0.71
improvement	22.3±6.1	22.8±5.7	0.85
Complications
Superficial infection	0	1	N.S
Deep infection	1	0	N.S
Tibial plateau fracture	0	1	N.S
Revision cases	1	0	N.S

Group F: necrotic lesion confined to the femur; Group T: necrotic lesion spread to the tibia; OKS: Oxford knee score; MFA: Maximum flexion angle; N.S: not significant.

OKS and MFA significantly improved in both groups from preoperative to final follow-up, according to paired t-test. However, we verified no significant differences between both groups.

Radiographic evaluation (radiolucent line and subsidence)

Table 3 shows radiographic evaluation results. Radiolucent line (RLL) ratios, partial RLL, and complete RLL were not significantly different between groups. Group T shows slightly larger subsidence distance and angle than Group F, but without significant differences.

Table 3

Radiographic evaluation for both groups.

	Group F (N=34)	Group T (N=26)	p-value
RLL
No RLL	18/34 (52.9%)	14/26 (53.8%)	0.79
Partial RLL	8/34 (23.5%)	7/26 (26.9%)	0.53
Completed RLL	6/34 (17.6%)	4/26 (15.3 %)	0.55
Subsidence
Subsidence distance (mm)	0.98±0.08	1.09±0.10	0.28
Subsidence angle (°)	0.85±0.11	1.01±0.12	0.24

Group F: necrotic lesion confined to the femur; Group T: necrotic lesion spread to the tibia; RLL: radiolucent line.

DISCUSSION

Oxford mobile-bearing unicompartmental knee arthroplasty (UKA) seems an appropriate procedure to treat spontaneous osteonecrosis of the knee (SONK) regardless of the presence of necrotic lesion on the tibial side besides the femoral side, verified by preoperative magnetic resonance imaging (MRI). To the best of our knowledge, this is the first study to describe the midterm clinical and radiographic outcomes of Oxford UKA in Japanese populations, as well as the impact of tibial lesion on clinical outcomes.

Several authors reported positive midterm outcomes of Oxford mobile-bearing UKA for SONK treatment. A study conducted by Langdown et al. ( with 29 patients reported a 100% survival rate at a mean follow-up of 5 years. Guo et al. ( and Zhang et al. ( also verified a 100% survival rate at short follow-up using a minimally invasive approach. In our study, the survival rate was 96.7%, showing that Oxford medial UKA is an effective technique for the treatment of Japanese patients with SONK in the short and midterm. Despite the positive results, further studies should continue to pursue long-term follow-up, given that several studies conducted with follow-up of over 10 years reported a slightly high revision rate, of approximately 10%. ), (

To the detriment of plain radiographs, MRI provides more detailed information for SONK evaluation. Early-stage SONK shows low signal intensity at T1-weighted images, but subchondral areas with high-intensity signals surrounded by low-intensity band-like signals at T2-weighted imaging. ( ), ( In this study, we compared the clinical and radiographic outcomes verified by preoperative T1-weighed MRI of two groups: one with condylar necrosis localized on the femoral side (Group F) and one with lesion spreading on the tibial side (Group T). According to the literature, lesions at the femoral condyle and tibial plateau (Group T) co-occur in around 30-40% of all SONK patients ), ( ), ( following two patterns. The first occurs in SONK natural course in the femoral condyle, as explained by Koshino et al. ( After osteonecrosis onset in the femoral condyle with a typical oval shadow in the subchondral area at the weight-bearing portion, the shadow expands and a surrounding sclerotic halo is formed. Then, these changes lead to ipsilateral lesions of the subchondral bone and articular cartilage in tibial plateau and consequently to osteoarthritis (OA) progression, including osteophytes, osteosclerosis, and joint-space narrowing. The second pattern refers to the occurrence after minor trauma, such as meniscal injury. Mechanical environmental changes after medial meniscus injury may increase contact stresses across the joint, resulting in focal subchondral overload and fracture, and possibly leading to osteonecrosis development on both the femur and tibial side simultaneously. ( )- ( Still regarding UKA for Group T, surgeons should also be concerned about tibial component loosening caused by incomplete seating due to poor bone quality, given that aseptic tibial component loosening was reported as the leading cause of failure in medial UKA. ( ), ( We found no significant differences in the clinical and radiographic evaluations between groups F and T, suggesting that the presence of necrotic lesion on the tibial plateau in preoperative MRI should not be considered a contraindication for the use of medial UKA for the treatment of SONK. We found no reports in the literature correlating clinical outcome to the extent of the surrounding bone marrow lesions after medial UKA. ( Yet, to our best knowledge, this is the first study to describe the minimal influence of tibial plateau lesion in preoperative MRI on clinical and radiographic outcomes after the use of Oxford UKA for treating SONK affecting the femoral condyle.

Our study had some limitations, such as the limited sample size and short follow-up period, indicating the need for further long-term research with larger sample size. We also did not perform a histological examination of osteonecrosis, which would be important for differentiate osteonecrosis from insufficiency fracture. Moreover, imaging results are influenced by the onset period in view of its diversity until MRI evaluation. Regardless of the limitations, this study suggests that the presence of tibial side lesions in preoperative MRI is not a predictor of poor outcome in the use of UKA for SONK.

CONCLUSION

The Oxford medial UKA is a reliable option for treating SONK in the short and midterm. The presence of lesions spread to the tibia in preoperative MRI does not affect the procedure radiographic and clinical outcomes.