Unrestricted Kinematic Alignment TKA in Asian Knees: Restoring Native Alignment Without Crossing the Safety Line

Ken Okazaki, MD, PhD, Prof., JAPAN

ISAKOS eNewsletters   Current Perspective 2026   rating

Predicting Constitutional Alignment for Kinematic Alignment TKA

Kinematic alignment (KA) total knee arthroplasty (TKA) is based on a simple idea: to restore the patient’s pre-arthritic joint surfaces, constitutional coronal alignment, and soft-tissue envelope, rather than correcting every knee toward a neutral mechanical axis. If that anatomy can be reproduced, the reconstructed knee may also restore more physiologic kinematics and feel more natural to the patient.

The challenge is that the patient presenting for TKA may already have severe knee deformity. Cartilage loss, bone wear, joint-line convergence, and soft-tissue laxity all influence standing alignment. The surgeon therefore needs a way to estimate the alignment that existed before osteoarthritis altered the joint.

The coronal plane alignment of the knee (CPAK) classification provides a practical language for this purpose. CPAK combines the arithmetic hip-knee-ankle angle (aHKA = MPTA − LDFA), an estimate of constitutional limb alignment, with joint-line obliquity (JLO = MPTA + LDFA), and classifies the knee into nine phenotypes (with MPTA standing for medial proximal tibial angle and LDFA standing for lateral distal femoral angle). CPAK is not itself an alignment prescription, but it helps surgeons distinguish phenotype from deformity and compare populations using a common framework.

Why Is Unrestricted KA More Controversial in Asia?

A growing body of literature using the CPAK classification suggests that coronal phenotypes are not distributed uniformly around the world. A recent systematic review of 29 studies and 21,632 arthritic knees found that CPAK Type I—the varus aHKA/apex-distal JLO phenotype—accounted for 51.3% of arthritic knees in Asia, compared with 32.8% in Europe and North America (Figure 1). Across published arthritic cohorts, mean aHKA is often approximately 3°–6° varus in Asian series, whereas many Western cohorts are closer to neutral, commonly around 0°–1° varus¹.

The tibial anatomy is particularly relevant. Several East Asian series have reported relatively low MPTAs, reflecting greater proximal tibial varus. In the geographic review, mean MPTA values in Japanese, Korean, Chinese, and Indian arthritic cohorts were often around 83°–86°, whereas many Western arthritic cohorts were closer to 86°–87°. Therefore, constitutional tibial varus is common in Asian TKA practice.

For surgeons working in Asia, this naturally raises a concern. If unrestricted KA (unKA) reproduces the native joint line without predefined coronal limits, will a patient with pronounced tibial varus end up with a markedly varus tibial component and excessive overall limb varus? Varus placement of the tibial component has historically been associated with edge-loading, tibial subsidence, and aseptic loosening. Therefore, restricted KA (rKA) and functional alignment, which preserve patient-specific anatomy while avoiding predefined extremes, have gained attention in Asia.

Even so, the basic idea behind unKA remains attractive. Moving the joint line to stay within a predefined boundary may alter ligament tension and knee kinematics in patients whose native anatomy lies outside that boundary. In some knees, doing so may require ligament releases to achieve balance or may compromise native soft-tissue balance. The key question, therefore, is whether well-executed unKA across the full spectrum of preoperative coronal alignment actually leads to unsafe postoperative alignment or implant failure in Asian patients.

Figure 1. Pooled prevalence of CPAK Type I in arthritic knees by geographic region. Asian arthritic knees show the highest prevalence of the varus/apex-distal phenotype. Data from Zhao et al.

What Do the Asian Clinical Data Show?

The most mature Asian evidence comes from a Korean series of 96 caliper-verified unKA TKAs, with mean clinical and radiographic follow-up of approximately 9 years². The cohort had a mean preoperative limb alignment of 8.3° varus and a mean aHKA of −3.8°, with postoperative aHKA averaging −2.9°. Mean tibial component alignment was 4.6° varus, and more than half of the limbs remained in >3° varus. Despite these findings, estimated implant survival at 10 years was 98.8%. The only revision was for a periprosthetic femoral fracture rather than loosening. Among knees with radiographic follow-up, there was no aseptic loosening or polyethylene wear; one tibial radiolucent line remained nonprogressive and asymptomatic for 9.7 years. CPAK phenotype was restored in 84.4% of knees.

A Japanese prospective multicenter study subsequently evaluated 1,950 unKA TKAs performed with a medial-pivot implant³. The population was highly varus: the mean standing HKA angle was 170.5° preoperatively, equivalent to approximately 9.5° varus, and the mean MPTA was 83.1°. At 1 year, the mean HKA angle was 177.8° (approximately 2.2° varus) and the mean MPTA was 86.2°. Although 1,074 knees (55.1%) fell outside the commonly cited restricted-KA envelope for HKA angle, MPTA, or LDFA, no knee required reoperation for a tibiofemoral complication, loosening, or subsidence. Three knees (0.15%) required reoperation for patellofemoral problems.

An interesting feature of the Japanese multicenter data is that unKA did not simply reproduce the severe varus seen on the preoperative radiograph. CPAK Type I decreased from 68.5% preoperatively to 36.4% postoperatively, whereas Type II increased from 16.7% to 30.0%; the most common change was from Type I to Type II. Many knees therefore moved toward a less-varus phenotype, but without being forced into mechanical neutrality.

More recently, a Japanese 5-year study evaluated 184 unKA TKAs radiographically, with 100 knees completing clinical follow-up⁴. The mean HKA angle improved from 11° varus to 3° varus, and the mean MPTA increased from 83° to 86°. Even in the subgroup with at least 15° of preoperative varus, the mean postoperative alignment was 6° varus rather than a reproduction of the full deformity. Clinical outcomes at 5 years were similar across preoperative varus-severity groups, with no significant differences in Knee Society Scores. There were no revisions for infection, fracture, or aseptic loosening, and only one nonprogressive tibial radiolucent line was observed.

Not all Asian data favor an unrestricted approach. A Singapore study reviewed 232 KA-TKAs and compared restricted and unrestricted techniques among CPAK Type-I knees, which accounted for 47.8% of the cohort⁵. Among 111 Type-I knees, 81 underwent rKA and 30 underwent unKA. The mean postoperative HKA (hip-knee-ankle) angle was 2.4° varus after unKA and 0.9° varus after rKA. Restricted KA produced higher objective KSS and satisfaction scores at 6 months and better Oxford Knee Scores at 2 years, although the differences in KSS and Oxford Knee Score did not reach the reported minimal clinically important difference, and some early advantages were no longer significant at 2 years. Thus, unrestricted restoration may not be optimal for every CPAK Type-I knee.

So what can we reasonably conclude? The Korean and Japanese cohorts provide reassuring safety signals for unKA in populations with a high prevalence of tibial varus. However, there is no clear evidence that unKA provides a clinical benefit over other alignment strategies. The studies differed in technique, implant design, and follow-up, so no universal target can yet be defined.

Table 1. Comparison of Asian unKA-TKA Cohorts

*Angles are presented as mean values.

Why Is Postoperative Varus Less Extreme Than Expected?

At first sight, this seems puzzling. The arithmetic hip-knee-ankle angle (aHKA) used in the CPAK classification is calculated as MPTA minus LDFA and, unlike the standing HKA angle, does not include the joint-line convergence angle. If unKA exactly reproduced the preoperative LDFA and MPTA, one would expect the postoperative HKA angle to be close to the preoperative aHKA. In the Japanese series, however, postoperative alignment was on average about 2°–3° closer to neutral, mostly due to an increase of MPTA. There are two likely reasons for this.

The first is inaccuracy in the preoperative aHKA, particularly in advanced varus knees. Flexion contracture and external rotation can prevent a true anteroposterior projection on standing long-leg radiographs. With external rotation, the posterior tibial slope is projected into the coronal plane and can make the MPTA appear more varus; femoral curvature can similarly affect the LDFA. Posteromedial tibial wear or bone loss may also place the measured articular surface below the original joint line, underestimating the native MPTA. Thus, the true pre-arthritic MPTA may be greater—and the constitutional aHKA less varus—than the radiograph suggests.

The second explanation comes from the surgical technique. In the Japanese cohorts, tibial resection was not determined from tibial morphology alone. After femoral reconstruction and osteophyte removal, longitudinal traction was applied with the MCL under appropriate tension, and the tibial resection was referenced parallel to the distal femoral cut. Complete osteophyte removal is accompanied by minimal release of the deep MCL, which slightly reduces medial soft-tissue tension. Referencing the cut to this ligament envelope can therefore reduce medial bone resection and shift the postoperative MPTA by roughly 2°–3° toward neutral compared with a cut based only on the worn tibial surface.

This distinction is important. In these Japanese reports, unrestricted KA did not mean reproducing every degree of varus visible on the preoperative standing radiograph. Although this tibial resection is not purely morphology-based, it can be categorized as a soft-tissue-respecting form of unrestricted KA. In this approach, the femoral anatomy is restored by resurfacing, whereas the tibial resection is determined from the reconstructed femur under appropriate ligament tension, without applying predefined limits to the HKA angle, MPTA, or LDFA. The femur was resurfaced accurately, while the tibial resection was guided by the reconstructed femur and the soft-tissue envelope. Together, more reliable estimation of the pre-arthritic bony anatomy and the controlled influence of medial ligament tension help explain why postoperative alignment was less varus than the preoperative aHKA.

Soft-Tissue-Guided Tibial Resection Using Manual In-Line Traction

In the Japanese series, unrestricted KA was performed using a soft-tissue-guided tibial resection technique based on manual in-line traction (Figure 2). After the distal femoral cut had been completed using a caliper-verified technique and accessible osteophytes around the femur and tibia had been removed, the knee was kept in slight flexion, with the degree of flexion adjusted to account for the posterior tibial slope. An assistant then applied steady longitudinal traction with both hands at the ankle, keeping the lower leg in line with the limb and bringing the MCL and the remaining capsuloligamentous structures under appropriate tension.

With traction maintained, a spacer block corresponding to the intended implant thickness was placed against the distal femoral cut and held parallel to the resection surface. The level at which its distal surface meets the anterior tibia was marked. The knee was then flexed, and the tibial cutting guide was aligned to this mark. After the planned posterior slope was set and the resection depth and coronal orientation were reconfirmed, the tibial cut was made. Thus, the tibial resection was referenced from the reconstructed femur while incorporating soft-tissue tension.

The key point is that this is not a purely morphology-based tibial resection: the reference comes from both the reconstructed femur and the tensioned soft-tissue envelope. In a varus knee, this helps compensate for medial osteochondral wear and the small decrease in medial tension after minimal deep MCL release. Less bone may therefore be removed from the worn medial side than the apparent preoperative MPTA would suggest. This is particularly useful in knees with severe varus deformity or medial tibial bone loss, where the worn surface poorly reflects the native resection level.

Two details are worth emphasizing. Traction should be applied in line with the limb with the knee held a few degrees short of full extension; hyperextension can lead to an unnecessarily deep tibial resection. Also, the goal is appropriate MCL tension, not maximal tension. Manual in-line traction allows a competent ligament envelope to guide the tibial resection; it is not a maneuver to force the limb toward a predetermined mechanical axis.

Figure 2. Manual in-line traction technique for determining the tibial resection level. A: An assistant applies steady longitudinal traction with both hands at the ankle and lower leg, keeping the lower leg in line with the limb and bringing the MCL and the remaining capsuloligamentous structures under appropriate tension. B: With traction maintained, a spacer block corresponding to the intended implant thickness is placed against the distal femoral cut and held parallel to the resection surface. The level at which its distal surface meets the anterior tibia is marked.

Clinical Pearls and Pitfalls

In practice, unKA should be regarded as a precise reconstructive technique. It depends on reliable imaging, accurate femoral reconstruction, correct interpretation of bone loss, a competent ligament envelope, and controlled soft-tissue-guided tibial resection. It does not justify accepting any component position simply because it is patient-specific. When the anatomy, ligament competence, implant support, or patellofemoral mechanics are unfavorable, or when there is pathological extra-articular deformity, a restricted or functional strategy may be the better personalized option.

Future Perspectives

The key unanswered question is which phenotype-implant-technique combinations remain durable over time. In published Asian cohorts, postoperative varus was generally 2°–3° less severe than preoperative aHKA would have predicted, especially in severely varus knees, yet a meaningful number of knees still remained in >6° varus. Early and mid-term implant failure was uncommon, but whether these residual outliers carry a long-term risk of polyethylene wear, tibial migration, or loosening remains unknown. Multicenter cohorts and registry studies with 10- to 15-year follow-up should report outcomes according to actual component angles and CPAK phenotype.

Another important question is what to do when an unrestricted reconstruction is expected to leave substantial residual varus. Future studies should not pool all varus knees together. Pre-arthritic phenotype, tibial bone quality and support, collateral ligament competence, and implant design are all likely to matter. A knee with constitutional tibial varus, preserved collateral competence, and an apex-distal joint line may behave very differently from one with acquired deformity, severe bone loss, or extra-articular malalignment. Comparisons of unKA, rKA, and functional alignment within these higher-risk phenotypes would be more useful than another all-comers comparison and may help define when a boundary should be introduced.

Implant design should be part of the same discussion: tibial baseplate support, articular conformity, medial-pivot behavior, and trochlear geometry may all modify the effect of residual varus. Rather than a single universal limit, the more realistic goal may be a phenotype-specific safety range that preserves native kinematics without compromising fixation or patellofemoral function.

Disclosures

The author is a paid consultant for Medacta International and Smith+Nephew, Inc.

References

1. Zhao G, Ma C, Luo Z, Ma J, Wang J. A systematic review of geographic differences in knee phenotypes based on the coronal plane alignment of the knee (CPAK) classification. Arthroplasty. 2025;7:26.
2. Won SH, Eim SH, Shen QH, Kim KK, Won YY. Caliper-verified unrestricted kinematically aligned total knee arthroplasty in Asian patients showed efficacious mid- to long-term results regardless of postoperative alignment categories. Knee Surg Sports Traumatol Arthrosc. 2024;32:941-952.
3. Nakamura M, Soda Y, Yamamoto T, et al. Short-term outcomes of unrestricted kinematic alignment total knee arthroplasty with a medial pivot implant: a Japanese multicenter study. Arthroplasty Today. 2026;37:101882.
4. Toyono S, Fukushima S, Yamamoto T, et al. Unrestricted kinematic alignment total knee arthroplasty provides favourable five-year outcomes regardless of varus severity in Japanese patients. J Exp Orthop. 2026;13:e70802.
5. Loh JY, Liow MHL, Purnomo G, et al. Restricted kinematic total knee arthroplasty provided better functional outcomes and higher satisfaction rates for Asians of genu varum with apex distal joint line over unrestricted kinematic total knee arthroplasty. Arthroplasty. 2025;7:33.

Please note: ISAKOS Newsletter Current Perspectives are not peer-reviewed articles. For peer-reviewed articles, please visit the Journal of ISAKOS at jisakos.com.