Introduction

In the treatment of isolated medial compartment osteoarthritis, medial unicondylar knee arthroplasty (UKA) has been used successfully with around 95% of 10-year implant survival rate, which is similar to that of total knee arthroplasty (TKA) [5, 18, 20, 22, 23, 28, 30]. Besides its advantages in terms of morbidity, mortality, rehabilitation, and functional outcomes [12, 13, 31], mobile-bearing systems have a greater risk of complications in less experienced hands, some of which are directly related to incorrect surgical technique, mostly in the tibial cutting stage [11, 12, 21, 24, 27]. Therefore, tibial component malpositioning in the coronal, sagittal, or axial plane may cause various complications, such as insert dislocation [14], medial tibial plateau fracture [4, 26], impingement [3, 29], and loosening [2].

Tibial plateau resection for the Oxford partial knee system is performed through a saw guide, which determines varus/valgus and cut depth in the coronal plane, and posterior tibial slope, but does not provide guidance for the vertical cut in the sagittal plane, which determines the rotation of the tibial component. The designers suggest pointing the tip of the blade toward the anterior superior iliac spine (ASIS) or using the flexion plane while making the vertical cut; however, ASIS may not be a good bony landmark because it is far from the surgical field, difficult to palpate for the surgeon and draped. The current literature shows a wide variability among the series in the results of tibial component rotation based on computed tomography (CT) images obtained after UKA [2, 6, 8, 15, 25].

The anteroposterior (AP) axis of the tibia is on the flexion plane of the knee. Because the tibial saw guide is placed parallel to the tibial crest on the flexion plane of the knee, the AP axis of the tibia, which is easy to palpate, observe, and correct intraoperatively, is an appropriate landmark for the vertical cut. We hypothesized that a more accurate vertical tibial cut could be made by placing the saw blade on the flexion plane of the knee, which is created by the tibial crest and saw guide. This study compared tibial component rotation, and radiological and functional results, in patients who underwent Oxford UKA performed using the ASIS or AP axis as the landmark to determine the direction of the vertical cut.

Material and methods

The study was conducted according to the Declaration of Helsinki and a priori institutional review board approval was obtained. Prospectively collected data of 86 consecutive patients who underwent Oxford medial UKA between March 2013 and August 2014, with a diagnosis of anteromedial osteoarthritis and an at least 5‑year postoperative follow-up, were analyzed retrospectively. All patients were operated on by two high volume arthroplasty surgeons who have expertise in Oxford UKA and prefer to use either AP axis (group I) or ASIS (group II) as landmarks for the vertical cut.

Surgical technique

All patients were operated on under a tourniquet, through a mini-medial parapatellar approach using a thigh support and the Oxford microplasty instrumentation system. The surgeons routinely placed the tibial saw guide parallel to the tibial crest in both planes. To place the cutting guide on the AP axis, its distal end was positioned anterior to the midpoint of the bimalleolar line and the tibial crest, while its proximal end was placed just medial to the patellar tendon so that the tibial crest and guide created the flexion plane of the knee. For the vertical tibial cut, the surgeon placed the saw blade parallel to the AP axis of the tibia, just medial to the apex of the medial tibial spine (group I; Fig. 1a–c). In group II, a senior resident palpated the ASIS and placed his fingers on it to show its position to the surgeon, who in turn placed the saw blade just medial to the tip of the medial tibial eminence and pointed it towards the resident’s hand while making the vertical tibial cut (Fig. 2). The medial tibial plateau resection was completed with a transverse cut through the guide. Femoral resection, gap balancing in flexion and extension, and tibial baseplate sizing were performed. Appropriately sized cementless femoral and tibial components were implanted in all patients.

Fig. 1
figure 1

Tibial cutting guide is placed on the AP axis, anterior to the tibial crest, medial to the patellar tendon in proximal and on the midpoint of the bimalleolar axis in distal (a). For the vertical cut, saw blade is placed on the plane created by the guide and tibial crest (b,c)

Full size image
Fig. 2
figure 2

The saw blade is placed medial to the eminence and it points to the hand of the resident, who shows ASIS for the vertical tibial cut

Full size image

All patients were allowed full weight-bearing as tolerated on the day of or day after surgery and were discharged on the first or second postoperative day. They were followed up at 2 weeks, 6 weeks, and 6 months, and annually thereafter.

Radiological evaluation

A senior resident, who did not participate in the surgeries and was blinded to the groups made all measurements on the PACS system and repeated 2 weeks later for intraobserver reliability, therefore the average values were used in the statistical analysis.

Tibial component rotation was measured on axial CT images (16-MDCT Aquilion 16 system; Toshiba Medical Systems, Japan) obtained at the 2nd week, postoperatively. For CT, the patient was positioned with the lower extremity in full extension and the hips in neutral rotation, with the patella pointing upward. Tibial component rotation (α-angle) was defined as the angle between the line passing through the lateral wall of the tibial component and Akagi’s line, i.e. the line between the center of the tibial attachment site of the posterior cruciate ligament and medial edge of the patellar tendon insertion (Fig. 3; [1]). Based on the reference line, the α‑angle was defined as positive (+) if the tibial component was placed in external rotation (ER) and negative (−) if it was in internal rotation (IR). The instant bearing position (IBP) was evaluated by measuring the angle between the line passing through the lateral wall of the tibial component and a line drawn perpendicular to the radiopaque linear line placed in the anterior part of the bearing (Fig. 4; [15]). Similarly, insert ER relative to the lateral side of the tibial component was considered (+), while IR was considered (−). Posterior overhang of the tibial component into posterior cruciate ligament (PCL) fossa was also assessed on axial CT scans.

Fig. 3
figure 3

Tibial component rotation (α-angle) was measured on axial CT images as the angle between the line passing through the lateral wall of the tibial component (blue line) and Akagi’s line (orange line), i.e. the line between the center of the tibial attachment site of the posterior cruciate ligament and medial edge of the patellar tendon insertion

Full size image
Fig. 4
figure 4

Position of the mobile bearing was evaluated by measuring the angle between the line passing through the lateral wall of the tibial component and a line drawn perpendicular to the radiopaque linear line placed in the anterior part of the bearing

Full size image

The varus (−)/valgus (+) angulation of the tibial and femoral components, posterior tibial slope, and femoral component flexion angle were measured on AP and lateral full-length radiographs (Figs. 5 and 6). Posterior tibial slope was considered as (+) and reverse tibial slope as (−). Flexion of the femoral component was considered as (+) and extension as (−).

Fig. 5
figure 5

The varus (−)/valgus (+) angulation of the tibial and femoral components, posterior tibial slope, and femoral component flexion angle were measured on AP and lateral full-length radiographs

Full size image
Fig. 6
figure 6

Posterior tibial slope was considered (+) and reverse tibial slope (−). Flexion of the femoral component was considered as (+) and extension as (−)

Full size image

Functional evaluation

Patients were evaluated using the Oxford knee score (OKS) and Knee Society score (KSS), preoperatively and at the last follow-up. Complications, reoperations, and revisions during follow-up were also recorded.

Outcome parameters

The primary outcome of this study was the α‑angle. Secondary outcomes were varus or valgus of the femoral and tibial components in the coronal plane, posterior tibial slope, flexion of the femoral component in the sagittal plane, rate of posterior overhang of the tibial component into the PCL fossa, and position of the mobile bearing relative to the tibial component (instant bearing position, IBP) in the axial plane [15]. Functional outcomes, i.e. the OKS, KSS and complications, were also analyzed.

Statistical analyses

The statistical analyses were performed using SPSS, version 22 (IBM SPSS Statistics for Windows, Armonk, NY, USA; IBM Corp., Released 2013). In order to find the difference of between the two groups statistically significant in terms of rotational angle, the minimum number of subjects required in each group was determined to be 43 [15] (power of 80% and significance level at 0.05). The statistical significance level was set to p < 0.05.

Kolmogorov-Smirnov test was used to determine if the data were normally distributed. Nonparametric tests of Mann-Whitney U, Wilcoxon signed ranks and Kruskal-Wallis were used to compare continuous variables across the groups. The χ2-test and Fisher’s exact test were used to compare categorical variables between two groups. The median (Q1 (1st quartile)-Q3 (3rd quartile)), mean ± standard deviation, frequency and percentage were reported as descriptive statistics. Intraobserver reliability of the measurements were assessed using the intraclass correlation coefficient and the 95% confidence interval (CI).

Results

Each group consisted of 43 patients with a similar gender distribution. The mean ages of the patients were 65.1 ± 7.2 years in group I and 63.6 ± 8.7 years in group II. Table 1 summarizes the demographic data.

Table 1 Patient demographic data
Full size table

Radiological results

The median α‑angle was significantly smaller in group I than in group II (2.5°, range: −4–5.5° vs. −6 range: −13–0.5), indicating that the tibial component rotation was closer to the neutral in group I (p < 0.001). The median flexion angles of the femoral components (7 vs. 10.5) and posterior tibial slopes (6 vs. 8) were significantly lower in group I than group II (p = 0.001). All other radiological parameters were similar in both groups (Table 2). Radiological parameters did not differ by tibial component size (Table 3). While PCL fossa involvement rate was 32.6% in group, in group 2, this ratio was 39.5%. The rates of PCL fossa involvement were 14 (32.6%) and 17 (39.5) (p = 0.7). The median α‑angle was significantly greater in patients with versus without PCL fossa involvement (−5, range: −13–2 vs. 0.5, range: −5–5.5; p = 0.02).

Table 2 Radiologic parameters
Full size table
Table 3 Comparison of the radiological parameters between different tibial component sizes. A-D Tibial component sizes
Full size table

Functional outcomes

The preoperative and final OKS and KSS were similar in both groups (Table 4). Both groups had significant functional improvements (p < 0.001).

Table 4 The comparison of functional outcomes between groups
Full size table

Complications

In group I, two complications were observed and all required reoperation without revision TKA. In group II, there were three complications, with one reoperation and two conversions to TKA.

Discussion

The designers of the Oxford partial knee recommend using the ASIS or flexion plane of the knee as landmarks to ensure an accurate vertical tibial cut for the appropriate tibial baseplate rotation; however, the literature does not clearly demonstrate that one of these two techniques is superior to another. In addition, there are limited data about the effect of tibial component rotation on the functional outcomes and complications. This study found that the tibial AP axis is a better landmark than the ASIS because it enables more accurate tibial component rotation; however, this difference did not influence the clinical outcome, and good and excellent results were obtained using both methods. Different studies have used the ASIS [8, 15], femoral head, medial wall of the intercondylar notch [10], medial femoral condyle, medial intercondylar ridge [29], and tibia AP line [6, 29] as landmarks. Strengths of our study include its comparative design, CT measurements, and evaluation of the effect of the methods on the mid-term clinical outcome. In addition, all measurements were made by a blinded observer.

Making a vertical tibial cut with near-neutral rotation is important. Excessive ER of the tibial component reduces bone support, increases polyethylene wear, and adversely affects the clinical outcome [6,7,8, 16]. Kamenaga et al. showed that the mobile insert hits the lateral wall of the tibial component and tends to escape posteriorly after 60° of knee flexion if it is placed in ER [9]. Excessive IR increases the space available for the baseplate in the mediolateral direction, but this may cause PCL damage [7]. Compared to the ASIS, the AP axis allowed us to achieve significantly more neutral tibial baseplate rotation. This could be due to its location and proximity to the surgeon, which promotes easy palpation and visibility. Although the mean α‑angle was within 6° with both methods, and the Oxford mobile bearing may tolerate this difference, excessive IR or ER was seen in some patients. If up to 10° of malrotation is considered normal, 18 knees in group II had excessive baseplate rotation versus only 3 patients in group I. This shows that the AP axis provides a more reliable landmark and less variation in component rotation than the ASIS. In a magnetic resonance study of healthy knees, Kawahara et al. stated that the medial wall of the intercondylar notch is almost parallel to Akagi’s line [10]; however, its reliability is questionable in cases with osteophytes and advanced degeneration. Comparing the tibial AP line with the medial intercondylar ridge line as a guide, Tsukamoto et al. demonstrated that rotation of the vertical cut is closer to neutral, where neutral rotation provides more surface area for the baseplate when the tibial AP line is used [29]. Theoretically, using the AP axis may help prevent plateau tibia fractures and medial overhang of the baseplate by providing a larger medial tibial surface area.

The mean posterior tibial slope and flexion of the femoral component were higher in group II, which is considered to be unrelated to the method; however, the difference was small and did not affect the OKS or KSS in our series. As we show here, excessive baseplate IR increases the possibility of overhang into the PCL fossa [15]. The designers recommend less than 2 mm of posterior overhang, but we could not demonstrate the clinical importance of this threshold.

The relationship between tibial component rotation and pain and function scores is unclear. Kamenaga et al. reported that increased external tibial rotation was related to decreased OKS and KSS [8], while Iriberri et al. found no significant relationship between tibial rotation and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), KSS, or visual analogue scale (VAS) scores, although the groups with good scores tended to have smaller rotation angles [6]. Liow et al. showed that patients with UKA femoral and/or tibial component rotation angles within 3° external rotation to 3° internal rotation of neutral component alignment reported better functional outcomes [17]. They emphasized that surgeons should be cognizant of the high variability noted in UKA component axial rotation and its potential correlation with functional scores. Additionally internal rotational error of the tibial component could be a major cause of pain after TKA [19]. In our study we found no significant association between tibial rotation and the postoperative OKS or KSS; the degree of improvement in the scores also showed no association with tibial rotation.

This study had some limitations, and the results should thus be interpreted cautiously. First, the patients were not randomized into the groups and surgeons. Therefore, results of this study need to be validated by multicenter, randomized, prospective studies. The sample size was relatively small, although we performed an a priori power analysis based on previous studies [15] and used the minimum number of patients to avoid unnecessary radiation exposure. We did not assess the interobserver reliability of the methods and measurements. Another limitation of the study is that the outcomes in patients with a high body mass index, for whom palpation of the ASIS or tibial crest may be difficult have not been studied. Lastly, the study was not powered to detect a relationship between tibial rotation and complications; however, our primary aim was to compare rotation angles between the methods.

In conclusion, the AP tibial axis is a better landmark for accurate rotation of the vertical tibial cut in Oxford medial partial knee replacement compared to the ASIS; however, the clinical outcomes of the two landmarks are comparable.