Limited ability of coronal alignment parameters to assess rotational alignment in patients with varus knee osteoarthritis: a cross-sectional study

Abstract

Purpose

This study aimed to investigate whether rotational alignment parameters vary across different severities of varus deformity and to examine the correlation between coronal and rotational alignment parameters in end-stage varus knee osteoarthritis (KOA).

Methods

We performed a retrospective study on 315 lower extremities with end-stage KOA. The lower extremities were categorized into the mild varus group (n = 141) and the severe varus (n = 174) group according to mechanical hip-knee-ankle angle (mHKA). Coronal alignment parameters were measured on standing full-limb radiographs, including the mHKA, lateral distal femoral angle (LDFA), and medial proximal tibial angle (MPTA). Rotational alignment parameters were measured on computed tomography (CT), including the femoral anteversion angle (FAA), condylar twist angle (CTA), and tibial torsion angle (TTA).

Results

We found no significant differences in any lower-limb rotational parameters between the mild varus and severe varus groups. Correlation analysis revealed that only TTA was significantly correlated with coronal alignment parameters, which was associated with MPTA in the mild varus group (r = 0.200, p = 0.020), mHKA in the severe varus group (r = 0.239, p = 0.001), and MPTA in the entire varus cohort (r = 0.118, p = 0.036).

Conclusion

The assessment of lower limb coronal alignment has limited ability to evaluate rotational alignment in patients with end-stage varus knee osteoarthritis. Therefore, surgeons cannot rely on coronal alignment to estimate rotational parameters in patients with varus deformity, and CT examination is recommended for accurate assessment of rotational parameters.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12891-026-10093-8.

Introduction

Rotational alignment is a challenge for total knee arthroplasty (TKA). During the progression of knee osteoarthritis (KOA), torsional relationships between the hip and ankle relative to the knee may change [1–3]. Previous studies have demonstrated that rotational deformities of the femur and tibia can result in knee pain, patellar instability, and abnormal gait patterns [4–8]. Furthermore, rotational deformities of the lower extremity may contribute to component malrotation. Malrotation of femoral and tibial components often impairs tibiofemoral kinematics, thereby reducing functional scores and patient satisfaction [9–12]. Therefore, accurate assessment of rotational alignment parameters is essential for achieving optimal TKA outcomes.

Rotation alignment parameters have traditionally been evaluated using computed tomography (CT) imaging. However, the absence of preoperative CT assessment in most TKA cases and the lack of accurate intraoperative measurement techniques render the identification of individuals with rotational abnormalities challenging. Coronal radiographs offer a simple, accurate, and cost-effective method for assessing the severity of varus and valgus deformities in patients with knee osteoarthritis. Previous studies have demonstrated significant correlations between rotational parameters, including the femoral anteversion angle (FAA), condylar twist angle (CTA), and tibial torsion angle (TTA), and changes in coronal alignment from varus to valgus [13, 14]. These findings provide a basis for preoperative identification of patients with abnormal rotational parameters and for intraoperative adjustment of component rotation, which is of considerable importance for improving surgical precision, optimizing implant positioning, reducing postoperative complications, and enhancing functional recovery and patient satisfaction.

These rotational parameters differ between patients with varus and valgus knees [13, 14]. However, Asian patients with KOA predominantly present with varus deformity [15, 16]. Given the distinct pathological mechanisms underlying varus and valgus knees in KOA, the relationship between rotational and coronal alignment parameters may not be applicable to patients with varus deformity. Moreover, the correlations reported in previous studies between rotational alignment parameters and varus-to-valgus changes may be influenced by extreme cases of severe valgus and severe varus, and these effects may not be applicable to a varus-only cohort. Patients with varying degrees of varus deformity may exhibit similar rotational parameters due to shared pathological processes, which could compromise the identification of individuals with abnormal lower limb rotation. We hypothesized that coronal alignment parameters would not reflect variations in rotational alignment among patients with varus KOA, implying that surgeons cannot rely on coronal alignment to estimate rotational parameters in these patients. The present study investigates whether rotational alignment parameters differ across patients with varying severities of varus deformity. Additionally, we examined whether coronal alignment parameters, including the mechanical hip–knee–ankle angle (mHKA), lateral distal femoral angle (LDFA), and medial proximal tibial angle (MPTA), correlate with rotational alignment parameters, including FAA, CTA, and TTA, in patients with varus KOA.

Materials and methods

Study population

We conducted a retrospective, observational, and single-center study of 315 total knee arthroplasties (TKAs) in 290 patients (Fig. 1). Patients who underwent TKA between December 2023 and June 2025 were assessed for eligibility. The inclusion criteria were cases that presented with the end-stage of KOA, and had preoperative standing full-limb anteroposterior radiographs along with axial CT scans. The exclusion criteria were preoperative valgus alignment of the lower extremity, rheumatoid arthritis, flexion contracture > 10°, and a previous history of lower extremity fracture or surgery. All cases were divided into two groups based on the degree of preoperative varus deformity. The cases with 0° ≤ varus < 10° were assigned to the mild varus group, and the cases with 10° ≤ varus were assigned to the severe varus group [17]. The demographic data and preoperative angles of these subjects are summarised in Table 1.

Fig. 1.

Flowchart of the patient selection process in the study. TKAs = total knee arthroplasties

Table 1.

Demographic characteristics and lower limb alignment parameters of patients stratified by varus severity

Parameters	Mild varus group (n = 141)	Severe varus group (n = 174)	P value
	0° ≤ varus < 10°	10° ≤ varus
Age (years)	67.6 ± 5.8	67.5 ± 5.6	0.703
Sex (men : women)	33 : 108	41 : 133	0.974
Height (cm)	162.0 ± 7.1	162.1 ± 7.5	0.861
Weight (kg)	72.2 ± 11.2	73.9 ± 11.4	0.176
BMI (kg/m2)	27.5 ± 3.6	28.0 ± 3.4	0.280
mHKA (°)	173.4 ± 2.4	166.1 ± 3.0	< 0.001
LDFA (°)	87.9 ± 2.2	90.1 ± 2.3	< 0.001
MPTA (°)	85.7 ± 2.0	83.2 ± 2.3	< 0.001
FAA (°)	15.1 ± 8.8	16.8 ± 8.9	0.081
CTA (°)	5.8 ± 2.2	5.6 ± 2.1	0.493
TTA (°)	25.0 ± 7.7	25.4 ± 7.4	0.593

BMI body mass index, mHKA mechanical hip-knee-ankle angle, LDFA lateral distal femoral angle, MPTA medial proximal tibial angle, FAA femoral anteversion angle, CTA condylar twisting angle, TTA tibial torsion angle

Imaging techniques

Axial CT (SOMATOM^® Perspective, Siemens Healthcare GmbH, Erlangen, Germany) was used to acquire simultaneous axial images of the hip, knee, and ankle joints. The lower extremities of patients were fixed with a cylindrical metal rod in a state of neutral rotation in a supine position to obtain the CT scan [18]. Multislice CT images (130 keV, 170 mA) of the whole lower extremity with multiplanar reconstruction (slice thickness of 1.00 mm; reconstruction interval of 1.00 mm) were also obtained.

Standing full-limb anteroposterior radiographs were obtained using an X-ray machine (Philips SRO 33100, Philips, Eindhoven, The Netherlands) at a source-to-image distance of 180 cm. To ensure the authenticity of the radiographs, all the imaging procedures were performed in strict accordance with the following standardized protocol. For standing full-limb anteroposterior radiographs, patients were positioned upright with the back against the detector, knees fully extended, feet shoulder-width apart, and internally rotated approximately 15°, ensuring one-third overlap of the fibular head with the tibia and anterior orientation of the patellae. The standing radiographs and CT scans were acquired on the same day.

Measurements

All radiographs were digitally obtained by the Centricity™ universal viewer zero footprint. The coronal parameters, including mHKA, LDFA, and MPTA, were shown in Fig. 2. All the coronal parameters were measured in standing full-limb anteroposterior radiographs. The mHKA was defined as a medial angle formed between femoral mechanical axis and tibial mechanical axis. The LDFA was defined as a lateral angle formed between femoral mechanical axis and the distal articular surface of the femoral condyles. The MPTA was defined as a medial angle formed between tibial mechanical axis and the proximal articular surface of the tibial platform. The varus deformity angle was calculated as 180° minus the mHKA.

Fig. 2.

The measurement of the coronal alignment parameters on standing full-limb anteroposterior radiographs. a The hip–knee–ankle angle (mHKA) was shown as a medial angle formed between the femoral mechanical axis and the tibial mechanical axis. b The lateral distal femoral angle (LDFA) was shown as a lateral angle formed between the femoral mechanical axis and the distal articular surface of the femoral condyles. The medial proximal tibial angle (MPTA) was shown as a medial angle formed between the tibial mechanical axis and the proximal articular surface of the tibial platform

The rotational parameters, including femoral anteversion angle (FAA), condylar twisting angle (CTA), and tibial torsion angle (TTA), were measured on axial CT images and are illustrated in Fig. 3. FAA was measured as the medial angle between the femoral neck axis (A, Fig. 3a) and the posterior condylar line (PCL) (B2, Fig. 3b) on axial CT slices through the femoral neck and distal femoral condyles [19]. A positive value indicated anteversion. CTA was defined as the medial angle between the clinical transepicondylar axis (cTEA) (B1, Fig. 3b) and the PCL (B2, Fig. 3b), also measured on axial slices of the distal femur [19]. Relative internal rotation of the PCL with respect to the cTEA was considered positive. TTA was measured as the lateral angle between the proximal tibial axis (C, Fig. 3c), defined along a tangential line of the tibial posterior cortex just proximal to the fibular head, and the distal tibial axis (D, Fig. 3d), defined through the centers of the tibia and fibula along the bimalleolar axis on axial CT slices [20]. Relative internal rotation of the proximal tibia with respect to the distal tibia was defined as a positive value.

Fig. 3.

The measurement of the rotational alignment parameters on axial computed tomographic scan images. a The femoral neck axis (A). b The clinical transepicondylar axis (B1) and the posterior condylar line (B2). c The proximal tibia axis (C). d The distal tibia axis (D)

Statistical analysis

All statistical analyses were conducted using RStudio software (version 4.4.1), with p < 0.05 considered statistically significant. The Shapiro-Wilk test was conducted for continuous variables to evaluate the normality of the data. For two independent groups, a t-test was used to compare means when the data were normally distributed, whereas the Wilcoxon rank-sum test was applied when the data did not follow a normal distribution. The chi-square test was used to compare categorical variables between two independent groups. Spearman correlation analysis was utilized to evaluate the statistical significance of the association between the coronal parameters and the rotational parameters of the lower extremity in different varus groups. Intraclass correlation coefficients (ICCs) were employed to assess the intra-observer and inter-observer reliabilities of the radiographic measurement techniques. Two independent observers repeated the measurements twice, separated by a two-week interval.

Results

A total of 315 TKAs were enrolled and eligible for the study. According to the varus deformity angle, 141 TKAs and 174 TKAs were classified into the mild varus and severe varus groups, respectively. The mean age of the total varus group was 67.5 ± 5.7 years, the mean height was 162.1 ± 7.3 cm, and the mean weight was 73.1 ± 11.3 kg. The cases included 241 (76.5%) women and 74 (23.5%) men. There was a significant difference (p < 0.001) in all the coronal parameters between the mild varus and severe varus groups (Table 1). The mean mHKA was 173.4° ± 2.4° in the mild varus group, and 166.1° ± 3.0° in the severe varus group. The mean LDFA was 87.9° ± 2.2° in the mild varus group, and 90.1° ± 2.3° in the severe varus group. The mean MPTA was 85.7° ± 2.0° in the mild varus group, and 83.2° ± 2.3° in the severe varus group. However, no significant differences (p > 0.05) in any rotational parameters were observed between the mild varus and severe varus groups (Table 1). The mean FAA was 15.1° ± 8.8° in the mild varus group, and 16.8° ± 8.9° in the severe varus group. The mean CTA was 5.8° ± 2.2° in the mild varus group, and 5.6° ± 2.1° in the severe varus group. The mean TTA was 25.0° ± 7.7° in the mild varus group, and 25.4° ± 7.4° in the severe varus group.

Correlation analysis was performed to evaluate the relationships between the coronal parameters and the rotational parameters. Neither FAA nor CTA was significantly associated with mHKA or LDFA in any of the varus groups. In the total varus group, only TTA was significantly associated with MPTA (r = 0.118, p = 0.036) (Table 2). In the mild varus group, only TTA was significantly associated with MPTA (r = 0.200, p = 0.020) (Table 3). In the severe varus group, only TTA was significantly associated with mHKA (r = 0.239, p = 0.001) (Table 3). The statistically significant correlations were weak (r = 0.118–0.239), indicating that coronal parameters explained less than 6% of the variance in TTA. The scatter plots and regression lines for the relationship between coronal and rotational alignment parameters are presented in Fig. 4. Table 4 summarizes the ICCs and corresponding 95% confidence intervals for each parameter. Inter-observer ICCs ranged from 0.844 to 0.931, whereas intra-observer ICCs ranged from 0.885 to 0.949 for observer 1 and from 0.845 to 0.918 for observer 2, indicating excellent measurement reliability.

Table 2.

Correlation analysis between coronal alignment parameters and rotational alignment parameters across the entire varus cohort

Rotational parameter	Coronal parameter	Correlation coefficient	P value
FAA	mHKA	−0.094	0.095
CTA	mHKA	0.053	0.348
TTA	mHKA	0.014	0.804
FAA	LDFA	0.015	0.797
CTA	LDFA	−0.090	0.111
TTA	MPTA	0.118	0.036

FAA femoral anteversion angle, CTA condylar twisting angle, TTA tibial torsion angle, mHKA mechanical hip-knee-ankle angle, LDFA lateral distal femoral angle, MPTA medial proximal tibial angle

Table 3.

Subgroup correlation analysis between coronal and rotational alignment parameters stratified by varus severity

Rotational parameter	Coronal parameter	Mild varus group (0° ≤ varus < 10°, n = 141)		Severe varus group (10° ≤ varus, n = 174)
		Correlation coefficient	P value	Correlation coefficient	P value
FAA	mHKA	0.116	0.171	−0.128	0.091
CTA	mHKA	−0.095	0.265	0.133	0.080
TTA	mHKA	−0.078	0.356	0.239	0.001
FAA	LDFA	−0.020	0.818	−0.029	0.705
CTA	LDFA	−0.058	0.493	−0.134	0.079
TTA	MPTA	0.200	0.020	0.125	0.100

FAA femoral anteversion angle, CTA condylar twisting angle, TTA tibial torsion angle, mHKA mechanical hip-knee-ankle angle, LDFA lateral distal femoral angle, MPTA medial proximal tibial angle

Fig. 4.

Correlation between coronal and rotational alignment parameters. FAA = femoral anteversion angle; CTA = condylar twisting angle; TTA = tibial torsion angle; mHKA = mechanical hip-knee-ankle angle; LDFA = lateral distal femoral angle; MPTA = medial proximal tibial angle

Table 4.

Intraclass correlation coefficients for inter-observer and intra-observer measurement reliability (95% confidence intervals)

Parameters	Inter-observer	Intra-observer
		Observer 1	Observer 2
mHKA	0.923 (0.909–0.935)	0.949 (0.937–0.959)	0.900 (0.877–0.919)
LDFA	0.921 (0.906–0.933)	0.939 (0.924–0.951)	0.893 (0.864–0.916)
MPTA	0.881 (0.859–0.901)	0.917 (0.897–0.933)	0.887 (0.826–0.923)
FAA	0.844 (0.818–0.867)	0.885 (0.859–0.907)	0.845 (0.811–0.874)
CTA	0.931 (0.919–0.942)	0.930 (0.913–0.943)	0.918 (0.898–0.934)
TTA	0.909 (0.892–0.923)	0.913 (0.890–0.931)	0.906 (0.884–0.924)

mHKA mechanical hip-knee-ankle angle, LDFA lateral distal femoral angle, MPTA medial proximal tibial angle, FAA femoral anteversion angle, CTA condylar twisting angle, TTA tibial torsion angle

Discussion

This study investigated the relationship between coronal alignment parameters and rotational alignment parameters in patients with varus knee osteoarthritis. Among the 315 TKAs included in the present study, the varus knee deformity measured a mean of 6.6° in the mild varus group and 13.9° in severe cases (p < 0.001). However, there were no statistically significant differences in the rotational angles (FAA, CTA, and TTA) between patients with mild varus deformity and those with severe varus deformity. Furthermore, femoral rotational alignment parameters (FAA and CTA) had no significant correlations with mHKA and LDFA. Notably, we found significant correlations between both mHKA and MPTA and the tibial rotational alignment parameter (TTA).

Coronal alignment parameters have a limited ability to predict TTA. Although significant associations were observed between TTA and both mHKA and MPTA in varus knees, the correlation coefficients ranged from 0.118 to 0.239, indicating a weak correlation [21]. Furthermore, a total of six distinct parameter combinations (mHKA-FAA, mHKA-CTA, mHKA-TTA, LDFA-FAA, LDFA-CTA, and MPTA-TTA) were tested for correlation analysis in each subgroup and the overall cohort. Accordingly, Bonferroni multiple comparison correction was performed with α = 0.05/6 ≈ 0.008. After correction, the original significant correlation of MPTA-TTA in the overall cohort (r = 0.118, p = 0.036) and mild varus group (r = 0.200, p = 0.020) failed to meet the corrected significance threshold, whereas the mHKA-TTA correlation in severe varus (r = 0.239, p = 0.001) remained statistically significant. Therefore, these findings suggest that mHKA and MPTA in varus knees may not be interpreted as clinically meaningful predictors of tibial torsion.

There are differences in rotational alignment between patients with varus and valgus deformities. Chang et al. [13] reported that the mean rotational alignment parameters in patients with valgus deformities (FAA = 16.7°, CTA = 10.2°, TTA = 32.6°) were higher than those in patients with varus deformities (FAA = 10.9°, CTA = 6.6°, TTA = 22.6°). Similarly, León-Muñoz et al. [14] reported that the mean rotational alignment parameters in valgus deformities (FAA = 12.1°; CTA = 6.4°; TTA = 24.6°) were higher than those in patients with varus deformities (FAA = 10.8°; CTA = 5.4°; TTA = 18.6°). Adjusting the prosthetic rotational alignment based on the preoperative deformity type can result in satisfactory clinical outcomes in TKA [22]. However, there is no consensus on whether prosthetic rotation should be adjusted based on the severity of varus deformity. Yoon et al. [23] suggest that prosthetic rotation should be adjusted to the degree of varus deformity in knee varus patients. However, previous studies demonstrated no significant differences in postoperative functional scores or complication rates between mild and severe varus deformity patients [24, 25]. Furthermore, Oh et al. [26] confirmed that the long-term survival in TKAs had no significant difference in different degrees of preoperative varus deformity. Our study indicated that no statistical difference was found between the mild and severe varus groups in terms of rotational profiles. Consequently, we do not recommend using the severity of varus deformity to guide intraoperative prosthetic rotational alignment.

Malrotation of femoral components remains a major cause of early TKA revision [27]. It has been shown that axial malrotation of femoral components can lead to postoperative pain, patellofemoral maltracking or instability, accelerated wear, or implant loosening [28, 29]. Luyckx et al. [30] found that for each 1° increase in coronal alignment from varus to valgus, the posterior condylar angle increased by 0.1°. However, we did not find a significant correlation between CTA and mHKA or LDFA in varus patients. Although our study shows no significant correlation between FAA and mHKA or LDFA, decreased femoral anteversion increases hip external rotation and toe-out gait [31]. Additionally, the distal femur shows a considerable individual variability in the rotational alignment [32]. External rotational positioning of the femoral component relative to the PCL may reduce the femoral anteversion. This could potentially increase the risk of gait abnormalities in patients with low FAA, but further studies are needed to confirm the association. To improve femoral prosthetic rotation accuracy, surgeons should consider anatomical variations and perform additional measurements preoperatively and intraoperatively.

Excessive internal rotation of the tibial component in TKA is a widely recognized risk factor for postoperative pain and inferior functional outcomes. Nicoll and Rowley [33] reported that more than half of the 740 consecutive TKA cases who experienced postoperative pain had excessive internal rotation of the tibial component. Similarly, Barrack et al. [34] reported that patients who presented with anterior knee pain exhibited an average internal rotation of the tibial component of 6.2°, compared with only 0.4° in the asymptomatic control group. Moreover, increased internal rotation of the tibial component may result in relative external rotation of the distal tibia. In addition, excessive tibial torsion is associated with anterior knee pain, patellar instability, and gait dysfunction [6, 35–37]. Increased tibial component internal rotation may be associated with poorer postoperative outcomes in patients with larger preoperative TTA, although further studies are needed to confirm this relationship. Therefore, in patients with varus knee osteoarthritis, surgeons need to carefully assess tibial rotation and avoid excessive internal rotation of the tibial component. Intraoperatively, excessive internal rotation can be prevented by identifying the Akagi line and the posterior condylar line of the tibia [38, 39].

This study had several limitations. First, our study included only Asian patients from a single center in China, and the cohort was predominantly female (76.5%). These factors may limit the generalizability of our results to other populations, ethnic groups, or more balanced sex distributions. While racial differences in rotational parameters have been reported in prior study [1], further multicenter and multiethnic studies are warranted to validate our findings and determine their applicability to broader patient populations. Second, this study evaluated lower-limb rotational profiles using CT, which provides a static rather than a dynamic assessment. As a result, soft-tissue conditions that may influence dynamic limb alignment were not captured. Third, articular cartilage erosion can affect rotational parameters in KOA patients undergoing TKA. While cartilage status was not explicitly analyzed, all participants were in the end-stage of KOA requiring TKA, indicating that the risk of bias was likely minimal. Fourth, this retrospective study included 25 patients who underwent bilateral TKA. Previous study has reported asymmetry between left and right lower limbs [40]. Because the left and right knee surgeries were performed at different times, each lower limb exhibited different rotational alignment and varus deformity, providing relative independence between the measurements. We also conducted a sensitivity analysis excluding one randomly selected limb from bilateral patients, and the results remained essentially unchanged (Supplementary material 1). Fifth, the cross-sectional design precludes causal inference. Consequently, it is not possible to determine whether coronal deformity influences rotational alignment over time, or whether pre-existing rotational abnormalities contribute to the development or progression of varus deformity. Sixth, this study only focused on patients with varus osteoarthritis. The coronal and axial alignment relationships in patients with varying degrees of valgus deformity require further investigation in future studies. In addition, a 10° threshold was used to classify varus deformity in the present study; however, a 15° threshold was used in another report [24]. Therefore, different cut-off values may lead to variations in patient classification and could influence the observed results. Future prospective studies can adopt multiple classification thresholds to verify the reliability of our research results. Finally, our study lacked clinical outcome data. Therefore, it remains unclear whether patients with mild versus severe varus deformity, who showed no significant differences in rotational parameters, experience similar postoperative functional outcomes. In addition, whether patients with preoperative abnormal rotational alignment have worse functional results after TKA requires further investigation. Further studies with clinical and functional outcome are needed to clarify these relationships.

Conclusion

The assessment of lower limb coronal alignment has limited ability to evaluate rotational alignment in patients with end-stage varus knee osteoarthritis. Therefore, surgeons cannot rely on coronal alignment to estimate rotational parameters in patients with varus deformity, and CT examination is recommended for accurate assessment of rotational parameters.

Supplementary information

Supplementary Material 1. ^{(18.1KB, docx)}

Abbreviations

KOA

knee osteoarthritis

mHKA

mechanical hip–knee–ankle angle

LDFA

lateral distal femoral angle

MPTA

medial proximal tibial angle

TKA

total knee arthroplasty

CT

computed tomography

FAA

femoral anteversion angle

CTA

condylar twist angle

TTA

tibial torsion angle

TKAs

total knee arthroplasties

PCL

posterior condylar line

cTEA

clinical transepicondylar axis

ICCs

intraclass correlation coefficients

Authors' contributions

H.W. and Z. Z. wrote the paper and edited the manuscript. N.P. and F.M. collected the data. F.C. and W.W. measured the imaging parameters. C.Y. performed all the statistical analyses. H.Z. designed the whole study and revised the paper. All authors have read and approved the final manuscript.

Funding

The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.

Data availability

The datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of the Affiliated Hospital of Qingdao University (Project ID: QYFY WZLL 30614). Informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.