Ethics approval

This study was conducted in accordance with the principles of the Declaration of Helsinki. Our Institute's Bioethics Committee for Medical Research approved the study (approval ID:24–060) and waived the requirement for informed consent from individual participants given the retrospective study design. All patients received standard treatment.

2.1 Participants

This retrospective study initially identified 116 consecutive patients with mass KOA who underwent CR-TKA using the ATTUNE Knee System (DePuySynthes, Warsaw, IN, USA) between July 2018 and December 2021. Of these, 16 patients were excluded due to insufficient follow-up (<2 years), resulting in a final cohort of 100 patients included in the analysis (Fig. 1). Experienced knee surgeons determined the indication for TKA based on clinical (e.g., loss of ROM) and radiological findings according to Kellgren and Lawrence classification grade 3 or 4.

Close

Fig. 1 Patient cohort flowchart.

Export to PPT

Patients were categorised into two groups according to their knee flexion position: Group G (flexion range of motion (ROM) over 120° and a Numerical Rating Scale (NRS) score of 1.0 or less; 34 knees [34 %]) and Group NG (flexion ROM less than 120° or an NRS score over 1.0; 66 knees [66 %]). The cut-off value for knee joint flexion angle was set at 120 because previous studies have reported that the cut-off value for the postoperative flexion angle indicating good knee joint function is 120° in Asian populations.¹¹ The postoperative flexion angle significantly affects knee joint function, and an improved flexion angle significantly affects patient satisfaction.⁹

2.2 Measurements

The patient characteristics collected were age; height; weight; BMI; sex; surgical procedure (measured resection (MR) or tibia first gap navigation [TFGN] technique); preoperative knee flexion ROM and numerical rating scale (NRS) score; Oxford Knee Score (OKS); 1-year postoperative knee flexion ROM and NRS; and 2 years postoperative knee flexion ROM. ROM was measured using a double-arm goniometer.

2.3 Surgical procedure

All surgical procedures were performed by a senior board-certified orthopedic surgeon. TKA was performed using the midvastus approach, cemented fixed prosthesis, pneumatic tourniquet, and the same type of CR-TKA implant. The anterior cruciate ligament was dissected, while the posterior cruciate ligament was retained. No patients underwent patellar resurfacing. Mechanical alignment TKA was performed to achieve the neutral alignment of the mechanical coronal limb by cutting the femur and tibia to ensure that the rectangular flexion and extension gap were perpendicular to the mechanical axis. Surgical techniques were performed using either femur-first MR or TFGN technique, as described by Iguchi et al..¹⁵ In MR, an intramedullary alignment system was used to cut the distal femur. An anterior reference guide was used in making a posterior femoral cut to ensure that the osteotomy line was parallel to the surgical epicondyle line and perpendicular to the Whiteside line. Proximal tibial cuts were made using an extramedullary alignment system. The collateral and reticular ligaments were not released to balance the flexion and extension gap. In TFGN, tibial osteotomy was performed first to ensure that the tibial osteotomy was perpendicular to the anatomical tibial axis. Computer-assisted surgery was used to obtain anatomical landmarks and individual knee biomechanical data.¹⁵ After tibial osteotomy, the medial and lateral gap sizes were measured dynamically using a tibial trial component. No release of the medial collateral ligament was observed. The extents of the distal femoral and posterior osteotomy and the mass–lateral–mass balance were determined, and the position of the femoral component was planned by navigation. Appropriate medial and lateral identical gaps were observed throughout the ROM.

2.4 Postoperative physiotherapy

As in our previous report,¹⁶ all patients in this study began a physical therapist-assisted rehabilitation protocol 3–4 weeks after surgery. This protocol consisted of three exercises: seated pedalling, knee extension stretches and heel-to-toe walking exercises. Patients were recommended to perform these exercises at least three times per day, provided they felt comfortable doing so. Patients received individually supervised outpatient physical therapy after discharge. This was patient-centred, with the supervising physical therapist determining the type of intervention, duration and frequency of sessions.

2.5 Statistical analysis

Numerical variables are expressed as mean ± standard deviation, while categorical variables are expressed as number and percentage. All statistical analyses were performed using EZR software.¹⁷ Comparisons between the two patient groups were based on knee flexion ROM and NRS at 2 years postoperatively. Numerical and categorical variables were assessed using Student's t-test. The significance threshold for the statistical test was set at P < 0.05. To exclude beta errors in the present results, the sample size calculation for the unpaired t-test for the primary outcome was performed a priori, with the significance level at P < 0.05. The minimum sample size with alpha error 0.05, beta error 0.20, and Cohen's effect size 0.8 was 52 patients (G Power 3.1, using Franz Paul, Kiel, Germany).¹⁸ In total, 34 knees were in the G group, and 66 knees were in the NG group. Therefore, the post hoc power analysis calculated an alpha error of 0.05 and an effect size of 0.8, with a power of 96.3 %. Thus, the sample size of the study (N = 100) was sufficient for determining statistical significance. An initial model was constructed using univariate analysis between groups. Variables with a P-value <0.05 in the univariate analysis were included in the multivariate model. The cut-off values were analyzed using receiver operating characteristic (ROC) curves and areas under the curve.

In addition, a subgroup analysis was performed to compare outcomes between MR and TFGN surgical techniques. Statistical significance was evaluated using a Student's t-test, with a P value < 0.05 being considered significant.

3.1 Multivariate analysis

The results of the logistic regression analysis are shown in Table 3. In univariate analysis, BMI, flexion ROM at 1 year postoperatively, and NRS were associated with good postoperative knee flexion and pain-free recovery. Variables that showed statistical significance in the univariate analysis were included in the multivariate analysis. In multivariate analysis, BMI (odds ratio [OR]: 0.90, 95 % confidence interval [CI]: 0.77–1.04, P = 0.16) was not associated with the desired outcomes. However, 1-year postoperative NRS (OR: 0.49, 95 % CI: 0.29–0.83, P < 0.001) and 1-year postoperative flexion ROM (OR: 1.17, 95 % CI: 1.03–1.27, P < 0.001) were significant predictors of postoperative flexion ≥120° and pain-free recovery. The ROC cut-off values for group G were flexion ROM ≥125° and NRS≦1.0 at 1 year postoperatively.

The positive predictive value for achieving both 1 year postoperative flexion ROM ≥125° and NRS ≦1 was 82.1 %. The positive predictive value for achieving NRS ≦ 1 alone was 50.8 %, while that for achieving a flexion ROM ≥125° alone 61.5 %. The negative predictive value for not achieving either was 84.7 %.

3.2 Subgroup analysis

In the subgroup analysis based on surgical technique, no significant differences were observed between the MR and TFGN groups in terms of postoperative 1-year flexion ROM and NRS. Furthermore, the proportion of patients achieving group G criteria (flexion ROM ≥120° and NRS ≤1.0 at 2 years) did not differ significantly between the two surgical techniques. These findings indicate that the choice of surgical technique did not significantly influence the likelihood of achieving favorable outcomes. Details are presented in Table 4.

4.1 Limitations

The results of this study should be considered within certain limitations. First, this study was a cohort study conducted by a single surgeon at a single institution, and therefore its generalizability is limited. Second, this study included only Asian participants, and there may be differences from other ethnic groups in terms of expectations for treatment outcomes, physical characteristics, and implant sizing. Third, psychological factors such as depression and anxiety, preoperative use of analgesics, adherence to rehabilitation protocols, and socioeconomic status are known to be major predictors of postoperative outcomes but were not assessed in this study. Additionally, implant positioning, alignment, and radiographic parameters—such as posterior condylar offset and tibial slope—were not evaluated, although they may significantly influence postoperative knee flexion. In addition, a selection bias may have influenced the cases that could be followed up. Patients with complaints may have continued to attend the clinic and it is possible that cases with both residual ROM limitation and pain were more likely to be included. However, in our study, one experienced surgeon carefully selected the indications for CR-TKA, performed the procedures, and conducted the postoperative follow-up. Thus, our study identified for the first time the predictive factors for achieving good postoperative knee flexion ROM and pain-free knee flexion.