Introduction

High tibial osteotomy (HTO) is an effective treatment for medial compartmental osteoarthritis (OA) of the knee with good clinical outcomes [5, 8, 9, 11, 13, 14, 16, 18, 22, 23, 25, 27, 29,30,31, 33, 35,36,37,38,39,40,41,42,43,44]. In particular, the two most commonly used surgical techniques of HTO are medial open wedge HTO (OWHTO) and lateral closed wedge HTO (CWHTO) [5, 8, 9, 11, 13, 14, 16, 18, 22, 23, 25, 27, 29,30,31, 33, 35,36,37,38,39,40,41,42,43,44]. However, OWHTO has been reported to have a negative effect on the patellofemoral (PF) joint postoperatively due to a decrease in patellar height and an increase in PF contact pressure, in contrast to CWHTO [3, 10, 13, 22,23,24, 39]. Biomechanical studies indicated that the alteration in patellar height can be attributed to distalization of the tibial tuberosity following OWHTO, which leads to the patella being pulled down toward the joint lines [3, 13, 44]. However, in radiographic studies, the alteration in patellar height after OWHTO has remained controversial because the results have differed depending upon the various methods used for measurement [18, 25]. To avoid an influence on patellar height, descending OWHTO leaving the tuberosity attached to the proximal tibia have been reported [14, 25]. In a radiographic study, descending OWHTO did not alter the patellar height postoperatively [25]. Kloos et al. [24] indicated that this procedure could significantly decrease the PF contact pressure, based on a biomechanical study. In CWHTO, the tibial tuberosity was reported to move proximally in a biomechanical study [13]. Therefore, CWHTO may lead to improvement of PF joint congruity compared with descending OWHTO. Meanwhile, CWHTO has also been reported to have several disadvantages, such as loss of the large bone block below the lateral tibial plateau, increase in lateral offset due to the horizontal osteotomy, and late full weight-bearing walking [17, 41]. To improve these disadvantages, the hybrid closed wedge HTO (HCWHTO) technique, a kind of CWHTO, was developed [41]. The points of difference for HCWHTO compared with conventional CWHTO are an oblique osteotomy to achieve a large contact area, a contact for the lateral cortices in the proximal and distal fragments without lateral offset for early full weight-bearing, and a removal of a smaller bone wedge to maintain leg length [17, 41]. Furthermore, this procedure has good effects on PF joint congruity through improvement in tibial tuberosity–trochlear groove distance and the proximalization of the tibial tuberosity when compared with OWHTO [18, 41]. If a patient has PF-OA preoperatively, HCWHTO may be chosen to prevent progression of PF-OA postoperatively. However, HCWHTO may be performed in patients with severe varus deformity of the knee combined with PF-OA, because progression of PF-OA is correlated with varus knee [36]. Thus, HCWHTO may introduce a risk of worsening the mid-term and long-term clinical results when compared with OWHTO, which is most effective for knees with early or middle stages of varus deformity [41].

Regarding clinical results for OWHTO and CWHTO, van Egmond et al. [43] compared the two procedures with a mean 7.9-year follow-up. The clinical results were significantly better for CWHTO than those for OWHTO, and the authors concluded that alterations in the PF joint after OWHTO may lead to worse results. Otsuki et al. [35] compared the Kujala scores for PF joint congruity between OWHTO and HCWHTO. Although the preoperative Kujala score was lower in HCWHTO than in OWHTO, the postoperative scores did not differ between the two groups at a mean follow-up of 31 months, because of improved PF joint congruity after HCWHTO [35]. Thus, in the future, it will be important to establish the correlations between clinical results and PF joint congruity following HTO with mid-term follow-up. However, to our knowledge, there are currently no reports in the literature that have compared clinical outcomes and PF joint congruity between HCWHTO and OWHTO with a minimum 5-year follow-up. To provide an update for the previous report, the present study aimed to clarify the mid-term clinical and radiographic outcomes for PF joint congruity following HCWHTO and OWHTO based on the degree of preoperative PF-OA. The study hypothesis was that the outcomes for the two techniques would not be different because HCWHTO improves PF joint congruity in patients with varus knee combined with PF-OA compared with OWHTO.

Materials and methods

This retrospective study evaluated patients who underwent OWHTO and HCWHTO in our hospital between May 2011 and April 2013. During the study period, OWHTO and HCWHTO were performed for 78 knees in 65 consecutive patients with medial compartment OA of the knee and varus knee deformities.

The inclusion criteria were medial compartment OA of the knee, varus alignment, no inflammatory arthritis, no neuromuscular disease, intact anterior or posterior cruciate ligaments, and follow-up of at least 5 years. The exclusion criteria included follow-up of less than 5 years and incomplete or absent preoperative or postoperative skyline views, anteroposterior weight-bearing whole-leg radiographs, or outcome scores, including the Japanese version of the Knee injury and Osteoarthritis Outcome Score (J-KOOS) and the Japanese version of the Oxford Knee Score (J-OKS) [15, 34]. As a result, two knees in patients with Parkinson’s disease that occurred postoperatively, nine knees with less than 5 years of follow-up, three knees with incomplete radiographs, and seven knees with incomplete outcome scores were excluded. Therefore, 57 knees in 48 patients comprised the final study population, and these patients underwent examination by magnetic resonance imaging (MRI) to confirm cartilage degeneration at the patella and trochlear groove. Based on Iwano’s radiographic classification of PF-OA [20], OWHTO was selected when the severity of PF-OA was less than stage II, and HCWHTO was indicated when the severity of PF-OA was more than stage III or the patient had cartilage degeneration of the patella and/or trochlear groove on MRI.

Radiological evaluation

The degree of OA at the femorotibial joint was evaluated using the Kellgren–Lawrence classification. The following data were measured using digital software (Plissimo: Konica, Tokyo, Japan). Using an anteroposterior weight-bearing whole-leg radiograph with the knee joint in extension, the hip–knee–ankle angle (HKA; defined as the angle between the femoral mechanical axis and the tibial mechanical axis), the weight-bearing line ratio (WBL ratio; calculated as the horizontal distance from the WBL to the medial edge of the tibial plateau (d), divided by the width of the tibial plateau (W); d/W × 100%), and the medial proximal tibial angle (MPTA; defined as the angle between the tibial mechanical axis and the articular surface of the proximal tibia) were calculated (Fig. 1a–c). The denominator was the width of the tibia measured using a ruler, and the numerator was the tibial intersection of the WBL (with the medial tibial edge at 0% and the lateral tibial edge at 100%) [28]. In addition, the mechanical lateral distal femoral angle (mLDFA; defined as the angle between the femoral mechanical axis and the articular surface of the distal femur), the joint convergence angle (JLCA; defined as the angle between the line connecting the distal femur and the proximal tibial articular surfaces), and the femoral patellar height index (FPHI; defined as distance A’/B’, where distance A’ is between the medial and lateral margins of the femoral epicondyles and distance B’ is between the base of the patella and the joint surface of the femoral condyles) were measured (Fig. 1d–f) [18]. If the apex of the JLCA was medial, it was recorded as positive (+) and denoted as varus, and if it was lateral, it was recorded as negative (−) and denoted as valgus. The tibial posterior slope (TPS) was measured on a standing lateral view of the knee flexed at 30° (Fig. 1g) [2]. TPS was defined as the angle between the line perpendicular to the mid-diaphysis of the tibia and the posterior inclination of the tibial plateau. In a skyline view with the knee flexed at 30°, as determined by a dedicated radiology technician using a goniometer under the instruction of the orthopedist, the lateral patellar tilt (LPT) and lateral patellar shift (LPS) were calculated (Fig. 2a) [5, 26, 35]. LPT was defined as the angle between the line intersecting the widest bony structure of the patella and the line crossing the anterior surfaces of the femoral condyles tangentially, and LPS was defined as the hh’/cc’ ratio [5, 26, 35]. For the latter, cc’ was defined as the distance between the summits of the medial and lateral femoral condyles of the femur, while hh’ was defined as the distance between the summit of the lateral femoral condyle and the point at which a line from the lateral edge of the patella perpendicular to the line passing through the summits of the femoral condyles crosses that line. The medial and lateral joint spaces of the PF joint were measured from the centre of the joint surface to the patella by drawing a perpendicular line (Fig. 2b) [35]. The congruence angle was calculated as the angle between the zero-reference line (bisecting the sulcus angle) and the line that links the lowest point of the intercondylar sulcus to the lowest point on the articular ridge of the patella. If the lowest point on the articular ridge of the patella was medial, it was recorded as minus (−), and if it was lateral, it was recorded as plus (+) (Fig. 2c) [32]. Preoperative and final follow-up values were compared for the outcome analyses.

Fig. 1
figure 1

Radiographic assessments. a Hip–knee–ankle angle (HKA). HKA was defined as the angle between the femoral mechanical axis and the tibial mechanical axis. b Weight-bearing line ratio (WBL ratio). WBL ratio was calculated as the horizontal distance from the WBL to the medial edge of the tibial plateau (d), divided by the width of the tibial plateau (W); d/W × 100%. c Medial proximal tibial angle (MPTA). MPTA was defined as the angle between the tibial mechanical axis and the articular surface of the proximal tibia. d Mechanical lateral distal femoral angle (mLDFA). mLDFA was defined as the angle between the femoral mechanical axis and the articular surface of the distal femur. e Joint convergence angle (JLCA). JLCA was defined as the angle between the line connecting the distal femur and the proximal tibial articular surfaces. If the apex of the JLCA was medial, it was recorded as positive (+) and denoted as varus, and if it was lateral, it was recorded as negative (−) and denoted as valgus. f Femoral patella height index (FPHI). FPHI was defined as distance A’/B’, where distance A’ is between the medial and lateral margins of the femoral epicondyles and distance B’ is between the base of the patella and the joint surface of the femoral condyles. g Tibial posterior slope (TPS). TPS was defined as the angle between the line perpendicular to the mid-diaphysis of the tibia and the posterior inclination of the tibial plateau

Full size image
Fig. 2
figure 2

Patellofemoral joint measurements. a Lateral patellar tilt (LPT). LPT was defined as the angle between the line intersecting the widest bony structure of the patella and the line tangentially passing the anterior surface of the femoral condyles. Lateral patellar shift (LPS). LPS was defined as the ratio of distances hh’/cc’, where cc’ was defined as the distance between the summits of the medial and lateral femoral condyles and hh’ was defined as the distance between the summit of the lateral femoral condyle and the point where a line from the lateral edge of the patella perpendicular to the line passing through the summits of the femoral condyles crosses that line. b The medial and lateral joint spaces were measured from the centre of the joint surface to the patella by drawing a perpendicular line. c The congruence angle was calculated. If the lowest point on the articular ridge of the patella was medial, it was recorded as minus (−), and if it was lateral, it was recorded as plus (+)

Full size image

Clinical evaluation

The J-KOOS and J-OKS were used to evaluate the clinical outcomes preoperatively and at the final follow-up. The maximum score for the J-KOOS is 168 points and that for the J-OKS is 48 points [15, 34]. At the final follow-up, patient satisfaction was determined using a 5-point Likert scale [21]. Patients were asked whether they were satisfied with the procedure, and scored five points for highly satisfied, four points for satisfied, three points for somewhat satisfied, two points for not satisfied, and one point for strongly dissatisfied. In addition, postoperative anterior knee pain for stairs or slope descent and adverse events including peroneal nerve palsy, paresthesia of the infrapatellar branch of the saphenous nerve, and tenderness of the pes anserinus were evaluated. The study was approved by the Medical Ethics Committee (Yokosuka Municipal Hospital, ID-number: 30-2), and all patients provided informed consent for participation in the study.

Surgical techniques

OWHTO was performed according to Staubli et al. [38] and Lobenhoffer et al. [30] using a modified method. Two wedges of β-tricalcium phosphate (Osferion 60, Olympus Biomaterial, Tokyo, Japan) were formed in a triangle pole equivalent to the size of the opening and then inserted into the osteotomy gap. To match the shape of the tibia, TomoFix MHT small and TomoFix MHT standard (Synthes Inc., Bettlach, Switzerland) were inserted into a subcutaneous tunnel formed on the medial side of the tibia in 29 knees and 7 knees, respectively, and fixed in place with eight locking screws in a minimally invasive fashion [40] (Fig. 3).

Fig. 3
figure 3

Medial compartment osteoarthritis of the left knee. a Anteroposterior radiograph before OWHTO. b Anteroposterior radiograph after OWHTO. c Skyline view before OWHTO. d Skyline view at 6 years after OWHTO. e Anteroposterior radiograph before HCWHTO. f Anteroposterior radiograph after HCWHTO. g Skyline view before HCWHTO. h Skyline view at 6 years after HCWHTO

Full size image

HCWHTO was performed according to the method described by Takeuchi et al. [41]. Under knee arthroscopy, lateral retinacular release (LRR) of the PF joint was performed. Before the tibial osteotomy, the fibular osteotomy at the mid-portion and segmental resection were performed. The tibia underwent a biplanar osteotomy. The hinge point that divided the proximal tibial osteotomy line by approximately 2–1 was then determined. After removal of the lateral closed wedge bone block, the medial side was opened, and the lateral side was closed. A proximal-lateral tibial plate (five distal screw holes; Synthes, Solothurn, Switzerland) was inserted into a subcutaneous tunnel formed on the lateral side of the tibia after plate bending for the proximal screws to become parallel to the tibial plateau, and fixed in place with nine locking screws with minimal invasiveness [41] (Fig. 3).

Postoperative rehabilitation

The same postoperative rehabilitation regimen was performed in both groups. On the day after surgery, active and passive range-of-motion exercises using continuous passive motion and isotonic muscle strengthening were started. Range-of-motion training with continuous passive motion was continued until 2 weeks after surgery. Standing exercises were initiated as soon as possible. At 1 week after surgery, patients were permitted to begin partial weight-bearing or full weight-bearing walking depending on pain with walker equipment, and full weight-bearing walking was started at 2 weeks after surgery.

Statistical analysis

Statistical analyses were performed using SPSS software (version 23.0 for Windows; IBM Japan). The radiographic parameters for HTO were measured by two authors who were blinded to the clinical results. The same observers reviewed the radiographs three times for each subject on different days, and the mean values were calculated. Intraobserver and interobserver reliabilities were assessed by calculating the intraclass correlation coefficients (ICCs) with a 95% confidence interval using a two-way mixed model for each parameter. Data for OWHTO and HCWHTO patients were evaluated separately. The preoperative and final follow-up values of the radiographic parameters, J-KOOS, and J-OKS were compared by a paired t test. Differences among demographic parameters were analysed by the Mann–Whitney U test, and those among categorical variables were analysed by the Chi-square test. The differences in radiographic parameters between OWHTO and HCWHTO patients were analysed by Student’s t test for a normal distribution. Statistical significance was assumed for p values less than 0.05. A power analysis was performed for the primary comparison of interest (postoperative congruence angle between OWHTO and HCWHTO). The power analysis revealed that a total sample size of n = 57 for the study was sufficient to detect a minimum effect size (d) of 0.78 at a power of 0.80 and an α-error of 5%, based on G power software (version 3.1.9.2).

Results

Radiologic evaluation

The demographic parameters and the results of the radiographic measurements are listed in Tables 1, 2 and 3. There were more cases of grade 4 OA than grade 2 or 3 OA in the HCWHTO group (57% vs. 43%) compared with the OWHTO group (22% vs. 78%) (p = 0.008). Furthermore, PF-OA was more severe in the HCWHTO group compared with the OWHTO group (p = 0.0001). The mean postoperative values for HKA and WBL ratio were significantly higher in the OWHTO group compared with the HCWHTO group (p = 0.007, p = 0.003). For the mean FPHI, the HCWHTO group showed a significant decrease at the final follow-up compared with the preoperative value (p = 0.01), while there was no significant change in the OWHTO group (n.s.). The mean postoperative FPHI was significantly lower in the HCWHTO group than in the OWHTO group (p = 0.04). Regarding changes in the skyline view, the mean LPT values were significantly decreased at the final follow-up compared with the preoperative values in both groups (OWHTO; p = 0.0001, HCWHTO; p = 0.003). For the mean LPS, the HCWHTO group showed significant medial translation at the final follow-up compared with the preoperative value (p = 0.004), while there was no significant change in the OWHTO group (n.s.). The mean postoperative LPT and LPS values showed no significant differences between the two groups (n.s., n.s.). In the PF joint, the medial and lateral joint spaces were both significantly decreased at the final follow-up in the OWHTO group (p = 0.0001, p = 0.018), while the medial joint space showed no significant difference (n.s.) and the lateral joint space was significantly increased at the final follow-up in the HCWHTO group (p = 0.0001). Although the congruence angle did not differ significantly in the OWHTO group (n.s.), it was significantly improved at the final follow-up in the HCWHTO group (p = 0.005). The ICCs of the radiographic parameters indicated high intraobserver and interobserver reliabilities (Table 4).

Table 1 Demographic parameters of open and hybrid closed wedge high tibial osteotomy
Full size table
Table 2 Pre- and Post-operative radiographic measurements of open and hybrid closed wedge high tibial osteotomy
Full size table
Table 3 Change in patellofemoral joint congruity with open and hybrid closed wedge high tibial osteotomy
Full size table
Table 4 Intra- and Inter-observer reliability of radiographic measurement
Full size table

Clinical evaluation

The mean preoperative and postoperative J-KOOS and J-OKS were significantly improved at the final follow-up in both groups (OWHTO: p = 0.0001, p = 0.0001; HCWHTO: p = 0.0001, p = 0.0001) (Table 5). However, the postoperative J-KOOS and J-OKS did not differ significantly between the two groups (n.s., n.s.). In the HCWHTO group, deep peroneal nerve palsy was seen in one patient. Paresthesia of the infrapatellar branch of the saphenous nerve and tenderness of the pes anserinus were each seen in 10 of 36 knees in the OWHTO group, and in no knees in the HCWHTO group at the final follow-up. Significantly more patients experienced adverse events in the OWHTO group compared with the HCWHTO group (p = 0.002). Anterior knee pain during stairs or slope descent was seen in 8 of 36 knees in the OWHTO group and 3 of 21 knees in the HCWHTO group (n.s.). For the mean patient satisfaction scale, there were no significant differences between the groups (n.s.). However, in the OWHTO group, patient satisfaction for 16 knees with adverse events (3.7 points; range: 1–5 points) was significantly lower than that for 20 knees without adverse events (4.7 points; range: 3–5 points) (p = 0.002).

Table 5 Change in clinical parameters with open and hybrid closed wedge high tibial osteotomy
Full size table

Discussion

There were four principal findings of the present study. First, patellar height was significantly higher in the HCWHTO group compared with the OWHTO group at the final follow-up. Second, regarding postoperative PF joint congruity in the HCWHTO group, LPT was significantly decreased and LPS showed significant translation in a medial direction. Furthermore, the lateral joint space of the PF joint was significantly increased and the congruence angle was significantly improved. Third, in the OWHTO group, LPT was significantly decreased, while LPS was not significantly altered. Although the congruence angle did not differ significantly, the medial and lateral joint spaces of the PF joint were significantly decreased. Fourth, the clinical outcomes in the two groups were equivalent at a minimum 5-year follow-up.

There are several methods for measurement of patellar height using radiography, including the Caton–Deschamps index, Blackburne–Peel index, Insall–Salvati index, and FPHI [6, 7, 18, 19]. Ihle et al. [18] reported that the Insall–Salvati index and FPHI are more exact for measuring patellar height in OWHTO than the Caton–Deschamps index and Blackburne–Peel index, because the latter indices are affected by the TPS following OWHTO. Use of the FPHI to measure patellar height after OWHTO in the present study revealed that, although TPS was significantly altered at the final follow-up, patellar height was not influenced, similar to the previous study [18]. In addition, for PF contact pressure based on a biomechanical study, Kloos et al. [24] indicated that descending OWHTO could significantly decrease PF contact pressure compared with OWHTO. Although patellar height was reported to be uninfluenced in descending OWHTO compared with that in OWHTO by leaving the tuberosity attached to the proximal tibia [25], this procedure seemed to decrease the PF contact pressure [24]. In our study, patellar height in HCWHTO was moved significantly proximally based on the FPHI compared with OWHTO. Therefore, HCWHTO may have superior effects on patellar height and PF contact pressure when compared with OWHTO and descending OWHTO.

With respect to PF joint congruency, a systematic review and meta-analysis revealed that LPT was significantly decreased, but LPS did not differ significantly after OWHTO, while there was a lack of evidence to confirm the changes in LPT and LPS after CWHTO [27]. In our study, the LPT and LPS after OWHTO were similar to those in previous studies [5, 16, 27, 29]. Meanwhile, the LPT was significantly decreased and LPS showed significant translation in a medial direction after HCWHTO. Regarding the cause of these changes in HCWHTO, LRR may have a good influence on PF joint congruency. Several studies have reported the efficacy of LRR for PF-OA [1, 12, 45]. Aderinto et al. [1] evaluated 53 patients who underwent arthroscopic LRR with a mean 31-month follow-up. Except for four patients who underwent total knee arthroplasty after LRR, 80% of patients had a reduction in pain with improvement of the clinical score compared with the preoperative state [1]. Christodoulou et al. [8] compared the radiographic measurements and clinical scores of patients with PF symptoms who underwent CWHTO with or without LRR with a minimum 5-year follow-up. These authors strongly recommended the combination of LRR with CWHTO, because the combined procedure improved the LPT and clinical scores compared with CWHTO alone [8]. Therefore, in our study, LRR at the time of HCWHTO was performed for improvement of the PF joint congruency and clinical score. However, although the LPT and LPS were significantly improved after HCWHTO, there were no significant differences between HCWHTO and OWHTO postoperatively. Therefore, regarding the LPT and LPS, HCWHTO with LRR may not be more effective than OWHTO. However, the congruence angle following HCWHTO was significantly improved in contrast to the findings for OWHTO, while the clinical scores were significantly improved in both groups. Regarding the contribution of LRR to PF-OA grade, Wu [45] indicated that LRR was effective for patients with stage II PF-OA by Iwano’s radiographic classification, and the rate of improvement of both the clinical score and congruence angle was reported as 88.3%. Meanwhile, the corresponding rate in patients with stage III PF-OA was only 34.2% [45]. Therefore, in the present study, HCWHTO may contribute more than LRR to improvement of the congruence angle and clinical score, because all patients with stage III PF-OA underwent HCWHTO.

Regarding the cartilage status of the PF joint after OWHTO, Goshima et al. [16] reported that progressive cartilage degeneration was observed on second-look arthroscopy at the lateral facet in 27% of patients, at the medial facet in 17% of patients, and at the trochlear groove in 30% of patients at a mean follow-up of 19 months after OWHTO. In our study using radiography, the medial and lateral joint spaces of the PF joint were significantly decreased after OWHTO with a minimum 5-year follow-up. Therefore, the cartilage status of the PF joint may be worsened by narrowing of the joint space. Meanwhile, HCWHTO was able to improve the lateral joint space of the PF joint, which was narrowed by PF-OA, and maintain the medial joint space. Therefore, HCWHTO provided significantly better postoperative PF joint congruency compared with OWHTO.

Several studies have reported adverse event rates after OWHTO and CWHTO [9, 37, 42]. However, the findings remain controversial, because the results varied among the studies. In our study, significantly more patients experienced adverse events in the OWHTO group compared with the HCWHTO group, but the rate of anterior knee pain on the PF joint did not differ significantly between the two groups. Meanwhile, regarding the clinical score, some studies reported that the PF joint congruity after HTO can affect the clinical score postoperatively [35, 43]. Otsuki et al. [35] recommended that HCWHTO, rather than OWHTO, should be the preferred technique for treatment of varus knees combined with PF-OA, based on improved clinical scores and PF joint congruity following HCWHTO. However, although HCWHTO may be performed for treatment of varus knees combined with PF-OA in the future, there are currently no reports comparing clinical outcomes and PF joint congruity between HCWHTO and OWHTO with more than 5 years of follow-up. In the present study, patients with severe PF-OA were treated with HCWHTO to prevent postoperative deterioration of the PF joint congruency. Although the grades of OA, PF-OA, and varus knee in the HCWHTO group were more severe than those in the OWHTO group, these two techniques did not lead to significantly different clinical outcomes based on the J-KOOS and J-OKS at a minimum 5-year follow-up. Therefore, HCWHTO is considered an effective treatment for varus knee combined with PF-OA at mid-term follow-up.

Several limitations of this study should be considered. First, the preoperative demographic parameters of the patients differed between the two groups because the indications for the procedures were different based on the grade of PF-OA, and this was not a prospective study. Second, LRR was only performed in HCWHTO. Although the combination of HCWHTO with LRR was performed for improvement of varus knee and PF-OA, it was difficult to achieve an accurate distinction between the osteotomy and the LRR regarding effects on the PF joint. Meanwhile, the efficacy of the combination of OWHTO with LRR remains controversial, because there are currently no reports in the literature comparing clinical outcomes and PF joint congruity between OWHTO with or without LRR at mid-term follow-up. Third, preoperative mLDFA of more than 90° was seen in one of 36 knees in the OWHTO group and 3 of 21 knees in the HCWHTO group. These patients may have had not only an indication for HTO, but also an indication for double-level osteotomy. Fourth, there was no evaluation of the PF joint using MRI at the final follow-up. In this study, the PF joint congruency showed significant differences before and after surgery in both groups. Thus, the postoperative status of the cartilage in the PF joint needs to be evaluated. Finally, overcorrection (HKA of more than 6°) was seen in 11 of 36 knees in the OWHTO group and 2 of 21 knees in the HCWHTO group. Overcorrection would be a risk factor for development of lateral patellar compression and progression of PF-OA [8]. In OWHTO, the postoperative lateral joint space of the PF joint in 11 knees with overcorrection (4.2 mm; range: 1.9–6.3 mm) was significantly smaller than that in the remaining 25 knees without overcorrection (5.0 mm; range: 2.6–6.7 mm) (p = 0.028), while the other parameters of the PF joint showed no significant difference (n.s.). To prevent overcorrection, a precise surgical technique is important, because a difference of 1.1° in the mechanical femorotibial angle leads to a change of 5% in the WBL ratio and severe varus deformity alters the tension of the medial and lateral soft tissues [4, 11]. In addition, long-term follow-up studies are needed to confirm PF joint congruity after HTO.

HCWHTO and OWHTO did not lead to significantly different clinical outcomes based on the J-KOOS and J-OKS at a minimum 5-year follow-up because HCWHTO was better able to correct deterioration of PF joint congruency than OWHTO. Therefore, in patients with varus knee combined with PF-OA preoperatively, HCWHTO is a more effective treatment than OWHTO.

Conclusion

Although the degrees of varus knee and PF-OA in HCWHTO were more severe than those in OWHTO, HCWHTO led to improved PF joint congruity and its mid-term clinical outcomes were equivalent to those of OWHTO. HCWHTO is considered an appropriate treatment for varus knee combined with PF-OA.