1 Introduction

In the total knee arthroplasty (TKA), patient dissatisfaction following an uncomplicated primary procedure is well documented, with an average rate of 10–20 %.¹ To decrease this dissatisfaction, concepts of personalized alignment targets has been developed and implemented in practice.^2–4 Indeed only up to 17 % of native knees have a neutral mechanical alignment (MA) i.e. a straight leg where the line joining the centre of the femoral head to the centre of the ankle passes through the centre of the intercondylar notch.^5,6

To reproduce the native knee's pre-arthritic alignment, various classification systems have been developed which label each knee's phenotype, thereby providing the surgeon targets to aim for to try and reproduce the native knee anatomy. In 2018, Lin proposed a classification system comprising 27 potential knee phenotypes, of which only five were considered clinically relevant.⁷ Variations on this framework have since been introduced, including the Functional Knee Phenotype classification,² the Coronal Plane Alignment of the Knee (CPAK) classification,⁵ mCPAK classification⁸ and the Coronal Extraarticular Deformity Phenotype (CEDP) classification.⁹ A key limitation of these classification systems is that they categorize knees based solely on their coronal plane alignment. In reality, a replaced knee rarely fails due to coronal malalignment alone, and several studies have shown equal or even better implant survival in patients with coronal plane malalignment to those with well-aligned TKA.^10,11

To date, all classification systems have relied on plain radiographs, which are two-dimensional (2D) representations of three-dimensional (3D) anatomical structures. However, these 2D images can be misleading due to observer variability, technical limitations (radiographs are not always perfectly standardized), subtle anatomical differences between individuals, and rotational or projectional variations. Recent reviews have further highlighted the substantial variability in methods used for 3D leg alignment analysis, underscoring the lack of consensus on how to derive axes and joint orientations from 3D bone models.¹² This variability prevents the establishment of universal reference values and makes it difficult to compare alignment parameters across studies.^13–15 Importantly, to achieve a comprehensive, biplanar assessment of knee alignment, sagittal plane alignment must also be evaluated in addition to coronal plane alignment. However, it remains unclear whether there is any correlation between sagittal and coronal plane alignment in native knees.The primary aim of this research is to investigate whether a correlation exists between coronal and sagittal knee alignment. The secondary aim is to analyse and understand differences in posterior tibial slope (PTS) among various racial groups.

2 Materials and Methods

2.2

2.2 Radiograph measurements

2.2.1

2.2.1 Coronal alignment

For coronal alignment, the coronal plane alignment of the knee (CPAK) classification was selected and the angles required for this classification were measured accordingly.⁵The mechanical hip-knee-ankle (mHKA) angle was defined as the angle formed by the intersection of the mechanical axes of the femur and tibia. The mechanical lateral distal femoral angle (mLDFA) was defined as the lateral angle between the femoral mechanical axis and the joint line of the distal femur. The mechanical medial proximal tibial angle (mMPTA) was defined as the medial angle between the tibial mechanical axis and the joint line of the proximal tibia.⁵ The arithmetic hip-knee-ankle angle (aHKA), as described by Macdessi¹⁸ is calculated from the mLDFA and the mMPTA. Calculate aHKA and the knee joint line obliquity (JLO) based on formulas; aHKA = mMPTA - mLDFA, and JLO = mMPTA + mLDFA(5).

2.2.2

2.2.2 Sagittal alignment

The medial proximal tibial joint orientation line was defined as the tangent to the deepest point of the medial plateau concavity¹⁹ for the measurement of medial PTS. The proximal circle was positioned just distal to the tibial tuberosity, and the distal circle was positioned just proximal to the distal tibial diaphysis.²⁰ The anatomical posterior proximal tibial angle (aPPTA) was measured as the acute angle formed between the perpendicular to the lateral tibial central anatomical axis and the proximal tibial joint orientation line. PTS was defined as 90°- aPPTA (Fig. 1).

Close

Fig. 1 Posterior tibial slope measurement in long limb radiographs Proximal tibial joint orientation line. The cross marks the deepest point of the medial plateau concavity. The line is tangent to the curve at the deepest point. Measurements of anatomical posterior tibial slope.

Export to PPT

Patients were categorized into classifications based on their PTS which were divided into intervals of four degrees as follows: Group A: less than 4.0°, Group B: 4.1°–8.0°, Group C: 8.1°–12.0°, Group D: 12.1°–16.0°, Group E: 16.1°–20.0° and Group F: greater than 20.0°. These intervals were determined arbitrarily as part of the study design, as no established precedent for PTS grouping exists in the current literature.

2.2.3

2.2.3 Data analyses

Data are presented as mean values with standard deviation (SD). Group differences were evaluated using one-way analysis of variance (ANOVA) and Fisher's exact test with Bonferroni post hoc analysis. All statistical analyses were performed using EZR software (http://www.jichi.ac.jp/saitama-sct/SaitamaHP.files/statmed.html).²¹

The necessary sample size for inter-observer and intra-observer reliability calculations was based on Zou's method, using an intraclass correlation coefficient (ICC) effect size of 0.8, a two-tailed significance level (α) of 0.05, and a power (β) of 0.8(19). Inter- and intra-observer reliability was assessed using a random two-way, single-measure ICC (ICC(2,1)) to evaluate agreement between observations.²² According to Koo and Li's guidelines.²² ICC values above 0.9 indicate excellent reproducibility, values between 0.75 and 0.9 indicate good reproducibility, and values between 0.5 and 0.75 indicate moderate reproducibility.²² We confirmed the ICC for the Japanese and Chinese groups, which had a larger number of cases. The relationship between CPAK and PTS was evaluated for each ethnicity using Spearman's rank correlation coefficient and Pearson's product-moment correlation coefficient.

3 Results

We included in 147 Japanese patients (288 legs: 148 men and 140 women), with a mean age of 43.9 ± 19.6 years, 200 Chinese patients (324 legs: 71 men and 253 women), with a mean age of 63.8 ± 10.3 years and 73 Indian patients (135 legs: 60 men and 75 women), with a mean age of 29.2 ± 4.0 years. There were significant differences among the three ethnic groups in terms of PTS: Japanese: 7.9 ± 3.6°, Chinese: 10.3 ± 4.9°, Indian: 13.8 ± 4.1°. Post hoc analyses further revealed significant differences between each ethnicity (Table .1).

One-way ANOVA demonstrated significant differences among the three groups in age, mLDFA, mMPTA, aHKA, JLO and PTS. Post hoc analyses showed no significant differences between Chinese and Indian in mLDFA, mMPTA, aHKA and JLO.

Fisher's exact test identified significant differences among the three ethnicities in terms of sex, CPAK, and PTS classification. Post hoc analyses showed significant differences between each pair of countries in CPAK and PTS classification. However, regarding sex, there was no significant difference between Japanese and Indian groups.

Inter- and intra-observer reliability for PTS were 0.91(95 % Confidence interval (CI): 0.85–0.93, P < 0.001) in PTS of Japanese patients. Inter- and intra-observer reliability for PTS were 0.85(95 % CI: 0.50–0.96, P < 0.001) and 0.76(95 % CI: 0.65–0.84, P < 0.001), respectively in PTS of Chinese patients. Both ICC values indicated good reproducibility.

We analyzed the distribution of PTS across CPAK types in all three ethnic groups (Table .2).

For all patients combined, the Spearman correlation coefficient between CPAK and PTS was 0.056 (p = 0.114), and the Pearson correlation coefficient was 0.038 (p = 0.283) indicating very low and non-significant correlations.

In Japanese patients, Spearman correlation coefficient was −0.153 (p = 0.076), and the Pearson correlation coefficient was −0.176 (p = 0.041). While both indicated a very weak negative correlation, only the Pearson correlation reached statistical significance. In Chinese patients, Spearman correlation coefficient was −0.086 (p = 0.123), and the Pearson correlation coefficient was −0.092 (p = 0.100), both extremely low and not statistically significant.

In Indian patients, Spearman correlation coefficient was −0.153 (p = 0.076), and the Pearson correlation coefficient was −0.176 (p = 0.041). Again, the Spearman result suggested a weak, non-significant negative correlation, whereas the Pearson result showed a weak negative correlation with statistical significance.

Mean PTS values and PTS classification rates are presented in Table 3, and detailed PTS distributions across CPAK categories are shown in Table 4.

4 Discussion

This study demonstrated significant differences in PTS across different ethnicities and found no meaningful correlation between coronal and sagittal plane alignment. These findings suggest that evaluating coronal plane alignment alone as a measure of knee phenotype is insufficient and, in some cases, potentially misleading.

Accurate and reproducible assessment of both coronal and sagittal alignment is essential in TKA, including during preoperative planning, surgical execution, and postoperative avaluation.²³

Historically, surgical strategies have focused predominantly on coronal alignment, guided by phenotyping classifications such as the CPAK system.⁵ While CPAK offers a structured framework, its two-dimensional nature limits its ability to capture the complexity of three-dimensional alignment and segmental deformities. Previous reports have highlighted CPAK's limited accuracy in identifying joint line apex²⁴ and inability to differentiate femoral from tibial deformities,²⁵ raising concerns about its clinical applicability in surgical planning.²⁶

To address these limitations, never systems such as the Functional Knee Phenotype Classification²⁷ have been developed, incorporating coronal variables like HKA, femoral mechanical angle (FMA), and tibial mechanical angle (TMA). This system allows for up to 125 possible phenotypes² and has shown better correlation with anatomical variability across diverse populations, offering greater reproducibility and clinical relevance.^28–30 Similarly, the CEDP classification⁹ builds upon CPAK by accounting for segmental coronal derformities, emphasizing the importance of individual femoral and tibial alignment in surgical decision making.

Despite these advances, sagittal alignment has received comparatively little attention, and currently, no well-established measurement methods or classification systems exist for sagittal plane assessment. To help address this gap, we employed a standardized measurement method for PTS, ensuring inter- and intra-observer reliability by using full-length radiographs and normalizing tibial length. Prior work by Garra¹⁷ demostrated that PTS measurements can vary with tibial length, with the most consistent results achieved using half-tibial images.

5 Ethnic differences

This study revealed significant differences among the three ethnic groups—Japanese, Chinese, and Indian—in both CPAK and PTS classifications. Post hoc analyses showed that these differences were significant between each ethnic pair. Our findings are consistent with a previous systematic review, which reported notable variation in CPAK classification prevalence across geographic regions, including Europe, North America, Asia, and Australia.³¹ Furthermore, a separate study examining PTS in 250 cadaveric specimens reported that African Americans/Blacks and Asian Americans generally have steeper PTS compared to Whites, with nearly 25 % of individuals demonstrating clinically significant slopes outside the commonly referenced 6°–12° range, regardless of sex or age.³² imilarly, our research confirmed marked differences in PTS angles and classifications among Asian countries, with one-way ANOVA showing significant differences in mLDFA, mMPTA, aHKA, JLO, and PTS among the three countries. Notably, however, no significant differences were found between Chinese and Indian patients in mLDFA, mMPTA, aHKA, and JLO.

6 Relationship between PTS and CPAK

Although a weak negative Pearson correlation between PTS and CPAK was observed in Indian patients, the practical impact appears minimal, and overall, no clinically meaningful correlation between PTS and CPAK was identified in any group. These findings suggest that PTS and CPAK represent independent variables influenced by ethnicity but unrelated to each other in terms of alignment behavior. Importantly, this independence highlights the need to evaluate sagittal plane alignment separately from coronal plane classifications when characterizing knee phenotypes or planning surgical interventions.

7 Clinical implications, limitations, and future directions

Taken together, these results underscore the importance of incorporating sagittal plane parameters, particularly posterior tibial slope (PTS), into knee phenotyping systems to improve surgical planning and optimise clinical outcomes. As personalized alignment strategies in total knee arthroplasty (TKA) continue to evolve, future research should aim to develop comprehensive classification systems that integrate both coronal and sagittal alignment to more accurately reflect the three-dimensional complexity of the knee. Such integrated systems may ultimately allow for more precise preoperative planning, better restoration of native knee kinematics, and improved patient satisfaction.

However, several limitations of this study should be acknowledged. First, this was a retrospective study, which may carry inherent selection bias and limit the ability to establish causal relationships. Second, the study population was restricted to three Asian ethnic groups (Japanese, Chinese, and Indian), which may limit the generalisability of the findings to other racial or geographic populations. Third, although inter- and intra-observer reliability was assessed and found to be good, radiographic measurement of PTS can still be influenced by subtle variations in imaging technique, positioning, or anatomical landmarks, which may introduce measurement error. Fourth, the study did not assess the clinical outcomes or functional impact of the observed anatomical variations; thus, the clinical relevance of the differences in PTS and their interaction with CPAK remains speculative. Finally, the cross-sectional design precludes evaluation of temporal changes or longitudinal effects, such as how these alignment patterns might evolve over time or influence the risk of osteoarthritis or implant survival.

Addressing these limitations in future research, including prospective, multi-ethnic, and longitudinal studies with functional outcome measures, will be crucial to refining knee phenotyping systems and advancing personalized approaches in TKA.

8 Conclusions

PTS and CPAK exhibit variations across different races in Asia; however, no significant correlation was identified between these two parameters. To validate these findings and further explore potential racial differences, a larger study involving data from additional racial groups is necessary.

CRediT authorship contribution statement

Shuhei Hiyama: The, Conceptualization, and design of this study were performed, Funding acquisition, of data was done, Formal analysis, and/or interpretation of, Data curation, was carried out, Writing – original draft, of the article was done. Reuben P. Rao: Funding acquisition, of data was done, Formal analysis, and/or interpretation of, Data curation, was carried out. Feng Xie: Funding acquisition, of data was done. Tsuneari Takahashi: Funding acquisition, of data was done. Vivek Shetty: Funding acquisition, of data was done. Jignesh Tandel: Funding acquisition, of data was done. Sajeev Shekhar: Funding acquisition, of data was done. Aniketh Wagh: Funding acquisition, of data was done. Jeya Palan: The Conceptualization, and design of this study were performed, Editing – original draft, of the article was done. Katsushi Takeshita: Funding acquisition, of data was done. Hemant Pandit: The, Conceptualization, and design of this study were performed, Writing – original draft, of the article was done, All authors have contributed significantly to the study approved the article and agreed with the submission.

Guardian/patient's consent

This study was conducted in compliance with the ethical principles outlined in the Declaration of Helsinki. Written informed consent was obtained from all patients prior to their participation in the study.

Ethics statement

This study was conducted in accordance with the principles of the Declaration of Helsinki. Our Institute's Bioethics Committee for Medical Research approved the study and waived the requirement for informed consent from individual participants due to the study's retrospective nature and all patients received standard treatment (Approval ID: 24–066).

Funding statement

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.