Restoring tibial obliquity for kinematic alignment in total knee arthroplasty: conventional versus patient-specific instrumentation

For decades, surgeons tried to achieve neutral tibial alignment with a medial proximal tibial angle (MPTA) of 90° and, therefore, were not trained to align the tibia according to this physiological MTPA equivalent to a mean of 2.9° with a range from 20.5° of valgus to 20.5° of varus [17]. The theoretical advantage of neutral mechanical alignment—with the mechanical axis passing mid-line—is the even load distribution in the medial and lateral femorotibial compartment with consecutively reduced risk for implant loosening and wear [3‐5, 13, 16]. In literature, however, this is discussed controversially [1]. In addition, mechanical alignment does not necessarily depict patients’ individual knee anatomy. In fact, only 0.1% of TKA patients actually have neutral femoral and tibial mechanical axes, whilst the vast majority presents with some degree of variation [2, 8]. Even more, according to a study by Bellemans et al. a constitutional varus knee, defined as natural mechanical alignment of at least 3° varus, can be found in 17% and 32% of asymptomatic adult women and men, respectively [3]. Upon TKA, neutral mechanical alignment in these patients would overcorrect the pre-existing varus, and eventually require some sort of medial soft tissue release [3]. Based on these observations, the philosophy of kinematic alignment (KA) was first proposed by Howell et al. aiming at a more patient-tailored reconstruction of the mechanical axis [8, 10]. First long-term results on KA are encouraging, with implant survivorship of 97.4% at 10 years [11].

Methods to achieve KA in TKA include caliper verification (as proposed by Howell et al.), robotics, computer-aided surgery, or patient-specific instrumentation (PSI) [8, 15]. In comparison to conventional instrumentation with standard blocks, PSI uses customised cutting blocks based on 3D models deriving from preoperative magnetic resonance imaging (MRI) or computed tomography (CT) scans [14], to achieve a more individualized reconstruction of preoperative knee anatomy [17]. However, the definite decision which instrumentation should be used to achieve KA still has to be made.

The aim of the current study was, therefore, to evaluate the accuracy of conventional instrumentation (CI) versus patient-specific instrumentation (PSI) with respect to restoration of tibial obliquity (based on mechanical medial proximal tibial angle [MPTA]) measured on conventional X-rays. It was hypothesized that PSI would be superior to conventional instrumentation with respect to accuracy of restoring tibial obliquity.

This study was approved by the institutional review board (blinded for review). Of 386 patients undergoing TKA at a single high-volume orthopaedic surgery centre due to end-stage osteoarthritis (grade IV according to Kellgren and Lawrence) of 2 out of 3 compartments of their joint, one-hundred patients were prospectively included between May 2020 and May 2022 and randomized into CI (n = 50) or PSI (n = 50) groups. Further 100 patients from the initial cohort directly underwent TKA with CI (Fig. 1). Mean age at surgery of the entire cohort (n = 200) was 70.7 (range: 48–90) years, and 127 were women (63.5%). All procedures were performed by one experienced senior knee surgeon.

There were no significant differences in terms of gender distribution (p = 0.932), age at surgery (p = 0.290) or BMI (p = 0.141) between the two treatment groups (Table 1).

Overall, no change in tibial obliquity from pre- to postoperative larger than 2° was found. One patient presented with change in tibial obliquity between 1° and 2° (0.5%), and 62 with a change between 0° and 1° (31.0%). In the remaining 137 patients, preoperative tibial obliquity had been restored, with no deviation from preoperative measurement (68.5%).

Notably, in a significantly higher proportion of patients with CI, preoperative tibial obliquity had been restored (78.0%; n = 117) as compared to PSI (40.0%; n = 20; p < 0.001). Furthermore, 21.3% (n = 32) and 60.0% (n = 30) of patients with CI and PSI, respectively, presented with a tibial obliquity change between 0° and 1°. One change in tibial obliquity between 1° and 2° was observed with the conventional technique (0.7%). In detail, PSI rather resulted in valgus deviation from preoperative tibial obliquity than CI, i.e., a mean increase in tibial obliquity of 0.1° (range: − 1°–0.5°) was observed for PSI, compared to a mean decrease of − 0.1° (range: − 1.5°–1°) for CI. Moreover, change in tibial obliquity was comparable between men and women (p = 0.642), and did not alter significantly with increasing age (p = 0.169), or BMI (p = 0.320).

ICCs between reviewers for pre- (ICC = 0.988) and postoperative measurements (ICC = 0.999) showed excellent agreement.

According to the present study, both PSI and conventional instrumentation achieved adequate results with respect to restoring tibial obliquity in kinematically aligned TKA, and that CI was superior to PSI.

Till now, no study comparing KA with PSI and CI regarding change in tibial obliquity has been published in literature. However, some studies have focused on change in tibial obliquity comparing PSI-based kinematic with conventional mechanical alignment [6, 7, 18]. Radiological results of the present study (no change in tibial obliquity > 2°) are, for example, better than the ones observed by Waterson et al. in a in a prospective randomized controlled trial comparing KA with PSI (n = 71) vs. conventional mechanical alignment (n = 71) using the Stryker Triathlon (Stryker Navigation, Kalamazoo, Michigan, US) system [18]. They reported on 78% and 77% of kinematically and mechanically aligned TKAs to be within 3° of preoperative alignment, but did not provide statistical tests [18]. Other than in the present study, Waterson et al. referred to alignment measurements based on preoperative MRI scans that had been performed in the PSI group [18], whereas radiological alignment measurements of the present study were based on x-rays.

Calliess et al. performed a similar prospective randomized controlled trial in 100 patients with PSI and conventional mechanical instrumentation [6]. Comparable to our observations, they reported on larger deviations in postoperative limb alignment following PSI than conventional mechanical instrumentation, although not providing statistical tests [6].

Dossett et al. likewise carried out a prospective randomized controlled trial comparing KA with PSI (n = 44) and conventional mechanical alignment (n = 44) using the Vanguard (Biomet Inc., Warsaw, Indiana, US) system [7]. In their study, knee and limb alignment was comparable between groups, whilst knee joint alignment of patients in the kinematic group was significantly more valgus (by 1.9°), the tibial component significantly more varus (by 2.1°), and the femoral component significantly more valgus (by 2.2°) compared to the mechanically aligned group [7]. This corroborates our findings, with PSI leading to tibial obliquity in more valgus than CI, although we used CI with kinematic rather than mechanical alignment.

A limitation of the current study is the lack of clinical outcome data (e.g., WOMAC, KSS) to correlate with radiographic findings made. Furthermore, 3 patients had to be excluded secondarily due to missing postoperative knee images. However, the resulting drop-out rate of 0.5% may be considered negligible. Also, tibial obliquity was measured on a.p. x-rays only, whereas the sagittal alignment (i.e., tibial slope) had not been considered in the current study.

Furthermore, 50 patients were randomized into both groups. This process was further randomized due to visiting surgeons requesting a specific technique. In addition, 100 patients were operated without randomization on using the CI instrumentation due to lack of available CT scans for PSI.

Both conventional and patient-specific instrumentation revealed adequate results with respect to restoring tibial obliquity in kinematically aligned TKA, with conventional instrumentation presenting superior results. In the future, the conventional technique may be preferred to PSI due to fewer radiation exposure and reduced costs. Nevertheless, further studies on larger cohorts might reveal the effect on potential outlier limitation using PSI.