1 Introduction

Unicompartmental knee arthroplasty (UKA) has emerged as an effective surgical treatment for anteromedial compartment osteoarthritis.^1,2 In UKA, there is no release performed, so it is valuable to know the alignment achieved postoperatively merely by ligament re-tensioning and resurfacing the worn medial compartment.³ In theory, it should lead to very close restoration of the pre-arthritic or constitutional alignment, the joint line obliquity, and the manner in which the weightbearing mechanical axis traverses the tibial plateau. Due to significant cartilage and bone loss, adjunct extra-articular deformity and possible lateral ligament stretching, it may be difficult to assess what the pre-arthritic alignment may have been when confronted with a radiograph of an arthritic knee and hence challenging to gauge from postsurgical radiographs whether or not it has been restored. In any case, considerable ambiguity remains as to the optimal alignment a surgeon should aim for in UKA.

Aligning the knee to its pre-arthritic position is known as personalised alignment and in recent times, arithmetic HKA (aHKA) has been proposed as a measurement on the preoperative radiograph which suggests what the pre-arthritic alignment may have been. There have been several studies of TKA, and a few of UKAs using this measurement but these have been of small deformities and small sample sizes.^4–7 The value in larger deformities is yet to be established as the reliability of aHKA in predicting constitutional alignment decreases when the deformity is greater than 8°.^8,9

Lately, robotically-assisted fixed-bearing UKAs (FB UKA) have generated data regarding the accuracy with which alignment has been restored.¹⁰ Robotic assistance enables a few degrees of undercorrection to be achieved accurately and consistently, and this has led to promoting the use of a robot for FB UKA.^11,12 The aims of alignment in FB UKA may be somewhat different from MB UKA. Notwithstanding, a large series giving results of conventionally-performed mobile-bearing UKAs is lacking, largely because the Oxford group did not use full-length radiographs to describe the alignment achieved in their multiple publications.¹³

The aims of the study therefore were to determine, in one of the largest single-surgeon series of Oxford UKAs, (i) what is the alignment achieved with a conventional technique for mobile-bearing UKA, (ii) whether mobile-bearing UKA delivers personalised alignment by restoring the constitutional or pre-arthritic alignment, (iii) whether there is any correlation between preoperative and postoperative alignment, and (iv) whether preoperative aHKA is a reliable predictor of postoperative alignment in mobile-bearing UKA.

2 Material and methods

This study is a retrospective analysis of prospectively collected data of 2818 consecutive cemented Oxford phase 3 medial UKAs performed by a single surgeon. All were performed for primary medial compartment osteoarthritis between January 2015 and December 2024. Patients were divided into 2 groups for analysis (Fig. 1). The inclusion criterion for group 1 was primary medial OA in the operated knee with a clinically asymptomatic and unaffected contralateral knee and a radiographic grading of less than Kellgren-Lawrence grade 1 medial OA.¹⁴ The inclusion criteria for group 2 were those patients undergoing unilateral UKA for primary medial OA, with contralateral knee either affected with OA, or previously operated by UKR or TKR, and patients undergoing bilateral UKA. The exclusion criteria for both groups were: incomplete clinical or radiologic records, history of fractures in either lower limb, and hip arthroplasty.

Close

Fig. 1 The distribution of total surgeries into two different groups.

Export to PPT

Group 1 (UKA in one knee, contralateral knee asymptomatic) had 370 patients; we excluded 35 with osteoarthritis of KL grade 2 or 3; 63 with poor quality radiographs or with suspicion of joint space narrowing (KL grades 1). This left 272 patients in this group with UKA in one knee and no evidence of joint space narrowing in the other knee.

Group 2 (UKA in both knees and those who underwent unilateral UKA with the other knee being arthritic or having had previous knee surgery) comprised 2448 knees. Of these, 248 knees were eliminated due to inappropriate x-rays and records. Thus, we had 2200 knees in this group. Of these, 1106 knees underwent bilateral UKAs (in 553 patients) and 1094 knees underwent unilateral UKA (Fig. 1).

The mean age of the patients was 62.4 (45–89) years. The final number of knees included in the two groups was 2472. All patients were operated upon by the senior surgeon. Spinal anaesthesia was administered and a tourniquet used. We followed the technique recommended by the Oxford group. All patients received Oxford phase 3 UKA implants (Biomet, Bridgend, UK). The thigh was supported in a thigh holder and surgery was performed in the “hanging-leg” position with the knee flexed. A minimally invasive quadriceps-sparing approach was used. No soft tissue releases were performed. After excision of osteophytes, tibial resection was first performed, and flexion gap assessed. The posterior femoral condylar resection was performed. An equal extension gap was created by graduated milling of the distal femur. Cemented components were implanted after trialling.

2.1

2.1 Radiographic evaluation

A standard technique was followed for obtaining full-length hip-to-ankle radiographs both pre and post-surgery: patellae facing forward and both feet placed at a fixed distance from each other. For alignment we measured the hip-knee-ankle (HKA) angle, knee joint line obliquity (KJLO) and mechanical axis deviation (MAD). The HKA angle was measured in routine fashion (Fig. 2a). The KJLO (Fig. 2b) was measured as an angle between the knee joint line and the perpendicular to the line joining the centre of the tibial plafond of the lower limbs. The knee joint line on the unaffected side was plotted by joining the mid points of the medial and lateral joint spaces at the level of the most prominent femoral condylar points. The knee joint line on the operated side was plotted by drawing a tangent from the lower surface of the femoral component to the midpoint of the lateral joint space at the level of the most prominent femoral condylar point. The mechanical axis deviation (MAD) was evaluated on the basis of Kennedy and White zonal classification (Fig. 2c).¹⁵ This system divides tibial plateau into 7 zones (medial to lateral: zones 0, 1, 2, C, 3, 4, 5). Zone 0 lies medial to the medial tibial plateau. The area between medial and lateral tibial eminences is zone “C”. The medial plateau, medial to the medial eminence, is equally divided into zones 1 and 2. The lateral plateau, lateral to the lateral eminence, is equally divided into zones 3 and 4. Zone 5 referred to the area lateral to the lateral tibial plateau. The zone through which the mechanical axis (MA) passes is noted to be the MAD.

Close

Fig. 2 The method of measuring variables on standing scanogram and determining a) HKA (Hip Knee axis angle), b) KJLO (Knee joint line obliquity) and c) MAD (mechanical axis deviation).

Export to PPT

The arithmetic HKA (aHKA) was calculated by subtracting the medial proximal tibia angle (MPTA) from the lateral distal femoral angle (mLDFA) calculated as per the method described.¹⁶

3 Results

The preop and postop HKAs of the entire cohort are shown in Table 1 and depicted in the histograms (Fig. 3a and b).

Close

Fig. 3 Histogram showing distribution of HKA angle in all 2472 knees-a) preoperatively and b) postoperatively.

Export to PPT

3.1

3.1 Group 1

The mean postop HKA angle of 175.7° (SD 2.8, CI 175:176, range 164–186) was not significantly different from the mean HKA angle of 175.4° (SD 3.2, CI 175:176, range 167–186) on the contralateral unaffected side (p = 0.106) (Fig. 4a & b). We calculated the difference between the HKAs of each pair of knees to determine the mean of the differences in HKA angles between the 2 sides – and this was 2.29° (SD 2.06, CI 1.91:2.40, range 0–17). There was a highly significant and moderately strong positive correlation (r = 0.523, p < 0.001) between postop HKA angle and the HKA angle of the contralateral limb (Fig. 5a). Postop HKA angle was restored to ±3° of that of the contralateral limb HKA angle in 246 of 272 patients (91 %) and to ±2° of that of the contralateral limb HKA angle in 177 of 272 patients (65 %). Postoperatively, the HKA angle was in <177° varus in 184 (67.6 %) of the operated limbs, in 177–180° varus in 77 (28.3 %) of the operated limbs, and in >180° varus in 11 (4 %) of the operated limbs. We found only 44 of the 272 knees to be aligned as per±3 aHKA and there was weak correlation (r = 0.355, p < 0.001) between the aHKA and post op HKA (Fig. 5b).

Close

Fig. 4 Histogram showing distribution of HKA angle in 272 knees of group 1- a) preoperatively and b) postoperatively.

Export to PPT

Close

Fig. 5 The scatter plot showing distribution of post op HKA angle of knees in group 1with-a) contralateral limb HKA and b) pre op aHKA of same limb.

Export to PPT

The mean post op KJLO in operated limb was 91.6° (SD 2.6, CI 91.3:91.9, range 84–96) and the value of mean KJLO was 90.2° (SD 2.8, CI 89.8–90.5, range 84–97) in the contralateral limbs (p < 0.001). The mean difference in KJLO between the operated and unoperated limbs was 1.4° (SD 3.19, CI 1.05:1.81, range −8.8 to 10.8). The KJLO in operated limbs was within ±3° of that of the contralateral limbs in 235 of the 272 patients (86.4 %). Further analysis based on the HKA angle showed that there was no significant difference in mean KJLO in the contralateral limbs (p = 0.342) compared to the operated limbs (p = 0.665) when the post HKA angle was within mild varus range (i.e. HKA 177–180°), when the HKA angle was in moderate varus range (i.e. HKA 170–176°) and when the HKA angle was in valgus range (i.e. HKA >180°).

Post op MA passed through zone 2 in 54 %, zone 1 in 30 % and zone C in 13 %. This distribution was almost similar to the contralateral lower limb in which the MA went through zone 2 in 42 %, zone 1 in 32 %, and zone C in 17 %. It was observed that if the MA traversed in contralateral limb in zone 0 then postop the MA went through zones 1 or 2. Similarly, the MA passed through zones 1 or 2 when contralateral limb is in zone 1 and zone 2; through zone C or 2 when contralateral limb is in zone C, and zone C or 3 if the contralateral limb is in zone 3 (Fig. 6).

Close

Fig. 6 Comparative stacked column chart of proportions of post op MAD zones in each Kennedy & White zone of the contralateral unaffected knee in group 1(The Y-axis depicts the percentages of knees in each zone).

Export to PPT

3.2

3.2 Group 2

The mean preop HKA was 170.7 (SD 3.86, CI 170.5:170.8, range 153–181) and postop HKA was 176.2 (SD 2.8, CI 176.1:176.3, range 169–186) with a statistically significant difference between two groups (p < 0.001). The mean difference between pre and post HKA (ΔHKA) was 5.62 (SD 3.34, CI 5.5:5.8, range 0–18.6). The values of ΔHKA were then correlated with the preop and postop HKA. We found strong negative correlation of ΔHKA with preop HKA (r = 0.695, p < 0.001) (Fig. 7a) and a weak correlation between the preop aHKA and postop HKA (r = 0.345, p < 0.001) (Fig. 7b). KJLO measurements could not be correctly made owing to bone defects, angular deformity and lateral joint laxity in a substantial proportion of cases; hence to avoid erroneous statistical analysis, these were excluded in group 2.

Close

Fig. 7 Scatter plot showing distribution of post op HKA angle of knees in group 2 with-a) ΔHKA (post op HKA-pre op HKA) and b) pre op aHKA.

Export to PPT

Preop and postop MAD was evaluated based on Kennedy and White zones. MA passed through zones 0 in 38 %, 1 in 43 % and 2 in 17 % pre-surgery. Postoperatively, MA went through 1 in 22 %, zone 2 in 52 %, C in 23 %. It was observed that MAD improved post-surgery and the MA went from zone 0 to zones 1 and 2. From zone 1 a shift was seen towards zone 2 in most; the others went to zone C or remained in zone 1. If MA was in zone 2, postoperatively it remained in the same zone or shifted to zone C. Zone C cases remained in zone C, with very few moving to zone 3 (Fig. 8).

Close

Fig. 8 Comparative stacked column chart of proportions of post op MAD zones in each pre op Kennedy & White zone for all knees in group 2 (The Y-axis depicts the percentages of knees in each zone).

Export to PPT

4 Discussion

The main results of the study show that the mean preoperative alignment of 9 degrees of varus gets partially corrected postoperatively to a mean of about 4 degrees of varus after Oxford mobile-bearing UKA. The final alignment is very closely correlated with the alignment of the contralateral non-arthritic and asymptomatic limb if that is available for comparison. The joint line obliquity with respect to the ground is also restored very close to that of the unaffected limb. MAD also shows that the weight-bearing axis after UKA traverses the tibial plateau in zones very akin to that of the contralateral normal knee. Using the opposite limb is acceptable for angular measurements as there is little difference in the femoral and tibial angular measurements on the two sides.¹⁷

In patients in whom unilateral UKA was performed and the opposite limb was also arthritic, or had undergone surgery, or in whom bilateral UKA surgery had been performed, the postoperative alignment was corrected by a mean of 5.6°. However, there is a very strong correlation between the extent of varus deformity and the quantum of correction achieved; minimal correction is observed in minor deformity, and larger correction is seen in severe deformities. There is a high degree of correlation with the preop HKA but only a weak association with aHKA.

The Oxford mobile-bearing UKA can achieve alignment very close to the native alignment.^6,18,19 This has been the goal of advocates of kinematic alignment for TKA.^20,21 That it can be achieved with conventional techniques and instruments has been demonstrated in this large series in the vast majority of cases. There is a fair amount of data to support the superior outcomes that ensue when the patient's limb alignment has been restored close to its pre-arthritic alignment in TKAs and one can easily understand why there are excellent and superior PROMs reported with UKAs.^2,22,23

Restoring pre-arthritic/constitutional alignment is referred to as personalised alignment. What is the normal alignment of the general population remains undefined. The general acceptability of normal alignment of the lower extremity is considered 2 to 3^o of varus but there are many deviations from this range. Bellemans et al. found, in 250 healthy adult volunteers (aged 20–27), mechanical alignment to be > 3^o of varus in 32 % of men and in 17 % of women.²⁴ Similar findings have been reported by others.^25,26 The variation in native alignment from 167 to 186° has also been observed in our study with mean HKA angle of 175.4° (SD 3.2, CI 175:176) in the contralateral leg in Group 1.

The alignment targets for UKA have been debated with proponents for undercorrection, neutral and overcorrection. Undercorrection has seen to be giving good results for many decades as mentioned in studies by Mullaji et al.,²⁷ Fisher et al.²⁸ and Hernigou et al.²⁹ Overcorrection (valgus alignment) was associated with poor results in these studies. This leads to lateral compartment osteoarthritis and is the main reason of revision after medial UKA (Price et al.).³⁰ In contrast Collier et al. and Whiteside et al. stated that risk of revision was not associated with neutral or overcorrected alignment.^31,32 Neutral or mild varus alignment has been advocated in some kinematic simulation studies. Innocenti et al. recommended neutral tibial alignment or a slight varus alignment (3°), based on collateral ligament strain and bone and polyethylene insert stress distribution.³³ Sekiguchi et al. observed the least mediolateral translation within one cycle of the deep knee bend and gait motions with neutral to 2° varus alignment.³⁴

Therefore, the alignment philosophy for the UKA surgery has been controversial over the past few decades with support of undercorrection, overcorrection and neutral alignment. This decade has now seen the preferred choice being restoration of native status or pre-arthritic alignment. The shift of TKA alignment techniques from mechanical to kinematic led to surgeons also performing UKA with the target of restoring pre-arthritic alignment. Ever since the concept of aHKA has been introduced, many authors have retrospectively evaluated their results and attributed getting good results on the basis of restoring aHKA (Jiang Y et al., Plancher et al., and Zhao Y et al.).^6,35,36 However, Bayoumi et al. found good results with both restoring and overcorrecting the aHKA.³⁷ The limitation of aHKA is that it is not applicable across a wider range of deformities as has also been seen in our study.

Our study demonstrates that the postoperative HKA angle was successfully restored to within ±3° of that of the unaffected lower limb in 91 % patients, merely by optimum balancing of the soft tissues, with no attempt being made to assess alignment during surgery. There was no statistically significant difference in the mean HKA angle between the operated side and the contralateral unaffected side. The MA passed through zone 2 or zone C in 81 % of our patients, as compared to 76 % (Emerson and Higgins et al.) and 87 % (Kim et al.).^1,38 The MAD was restored to within ±1 zone of the contralateral unaffected lower limb in 95 % of our patients (16–17). In the second scenario in which the other knee is affected, the values of ΔHKA were moderately correlated with the pre HKA.

One contentious issue is whether target alignments for both mobile-bearing and fixed-bearing UKAs ought to be similar. Perhaps FB UKAs may be more sensitive to alignment. Kleeblad et al. defined the optimal postoperative alignment to be 4 degrees of varus, and acceptable alignment as 5 to 7^o of varus.³⁹ Hernigou and Deschamps et al. in their series of 58 knees with a mean follow-up of 15 years, found that postoperative alignment of >7^o of varus increased the risk of early polyethylene wear and aseptic loosening.²⁹ Overcorrection was associated with increased risk of osteoarthritis in the lateral compartment, whereas, severe undercorrection (i.e., >10^o of varus) was associated with increased tibial component wear. Conversely, Zuiderbaan et al. reported no revisions at a mean of 2.3 years in a consecutive series of 104 patients who underwent medial UKAs.⁴⁰ In that study, 25 % of patients had a preoperative HKAA of 10–15° of varus, and 32 % had a postoperative HKAA of >4° of varus. Large varus deformities of >10° and <15° were not found to be a contraindication for UKA in that study, and no patient with these large varus deformities underwent conversion to TKA. This remains an area in need of exploration.

Limitations of our study: We did not compare the long-term functional outcomes with the alignment and nor did we assess the outcomes between the 2 groups. We have also not listed postoperative complications in this study as that was not the aim of this study. However, the patients are all being monitored and followed up regularly. Future research should focus on prediction of postoperative alignment and also defining the safe limits for personalised alignment in UKAs.

5 Conclusions

Conventional mobile-bearing UKA consistently delivers personalised alignment, restoring the mechanical axis and joint line obliquity close to the patient's native state, most evident in cases with an unaffected contralateral limb. This large series supports that optimal alignment can be achieved without robotic assistance, so as to confer the potential benefits of alignment restoration on patient outcomes.

Ethical Statement

Not applicable as this is retrospective evaluation of collected records.

Credit author contribution statement

ABM: Conceptualization, Supervision, Writing – review & editing. AG: Data Collection, Formal analysis, Writing – original draft, Writing – review & editing. SC: Data Collection, Writing – review & editing.

Consent for surgery

Not applicable.

Funding statement

None.