Study on the correlation between early three-dimensional gait analysis and clinical efficacy after robot-assisted total knee arthroplasty

Abstract

Purpose

Robot-assisted technology is a forefront of surgical innovation that improves the accuracy of total knee arthroplasty (TKA). But whether the accuracy of surgery can improve the clinical efficacy still needs further research. The purpose of this study is to perform three-dimensional (3D) analysis in the early postoperative period of patients who received robot-assisted total knee arthroplasty (RATKA), and to study the trend of changes in gait parameters after RATKA and the correlation with the early clinical efficacy.

Methods

Patients who received RATKA in the Center of Joint Surgery, the First Hospital Affiliated to Army Military Medical University from October 2020 to January 2021 were included. The imaging parameters, i.e., hip-knee-ankle angle, lateral distal femoral angle, medial proximal tibial angle, posterior condylar angle were measured 3 months post-TKA. The 3D gait analysis and clinical efficacy by Western Ontario Mac Master University Index (WOMAC) score were performed pre-TKA, 3 and 6 months post-TKA. The differences in spatiotemporal parameters of gait, kinetic parameters, and kinematic parameters of the operated limb and the contralateral limb were compared. The correlation between gait parameters and WOMAC scores was analyzed. Paired sample t-test and Wilcoxon rank-sum test were used to analyze the difference between groups, and Spearman correlation coefficient was used to analyze the correlation.

Results

There were 31 patients included in this study, and the imaging indexes showed that all of them returned to normal post-TKA. The WOMAC score at 3 months post-TKA was significantly lower than that pre-TKA, and there was no significant difference between at 3 and 6 months. The 3D gait analysis results showed that the double support time of the operated limb reduced at 3 and 6 months (all p < 0.05), the maximum extension and maximum external rotation of the knee joint increased at stance phase, and the maximum flexion angle, the range of motion and the maximum external rotation increased at swing phase. Compared with the preoperative data, there were significant improvements (all p < 0.05). Compared with the contralateral knee joint, the maximum external rotation of the knee joint at swing phase was smaller than that of the contralateral side, and the maximum flexion and extension moment was greater than that of the contralateral knee. The maximum external rotation moment of the joint was greater than that of the contralateral knee joint (p < 0.05). There was a negative correlation between the single support time pre-TKA and the WOMAC score at 3 months (p = 0.017), and the single support time at 3 months was negatively correlated with the WOMAC score at 6 months (p = 0.043). The cadence at 6 months was negatively correlated with the WOMAC score at 6 months (p = 0.031). The maximum knee extension at stance phase at 6 months was negatively correlated with the WOMAC score at 6 month (p = 0.048). The maximum external rotation at stance phase at 6 months was negatively correlated with the WOMAC score at 6 months (p = 0.024).

Conclusion

The 3D gait analysis of RATKA patients is more sensitive than WOMAC score in evaluating the clinical efficacy. Trend of changes in gait parameters shows that the knee joint support, flexion and extension function, range of motion, external rotation and varus deformity moment of the patient were significantly improved at 3 months after surgery, and continued to 6 months after surgery. Compared with the contralateral knee, the gait parameters of the operated limb still has significant gaps in functionality, such as the external rotation and flexion and extension. The single support time, cadence, knee extension, and knee external rotation of the operated limb have a greater correlation with the postoperative WOMAC score. Postoperative rehabilitation exercises should be emphasized, which is of great value for improving the early efficacy of RATKA.

Introduction

Osteoarthritis (OA) is a type of disease characterized by destruction of articular cartilage and subchondral bone, progressive narrowing of the joint space, and hyperplasia of osteophytes around the joints. Advanced OA causes damage to the articular surface and loss of joint function. Among the weight-bearing joints of the lower limbs, the knee joints have the highest incidence of OA.^1,2 Total knee arthroplasty (TKA) is an important treatment of advanced knee OA. Judging from the results of long-term follow-up of a large number of cases, the clinical efficacy of TKA was reliable, and the survival rate of prosthesis in the medium and long-term follow-up was 89%–95%. In the past 30 years, the curative effect of this technique has been recognized and promoted on a large scale.^3,4 However, due to postoperative pain, decreased quadriceps muscle strength, poor prosthesis position and deviation of lower limb alignment, 20% of patients were dissatisfied with the results.⁵

In order to further improve the accuracy of TKA and postoperative efficacy, robot-assisted total knee arthroplasty (RATKA) technology is gradually applied to the treatment of advanced knee OA. Assistance of active robots (CASPAR and ROBODOC) and semi-active robot (MAKO RIO), many studies reported that the postoperative lower limb alignment of RATKA were more accurate, and the deviation in the sagittal, coronal and transverse plane dropped significantly compared with conventional TKA.6, 7, 8, 9 However, the knee society score or the Western Ontario Mac Master University Index (WOMAC) score used for the postoperative functional evaluation showed no significant difference compared with the conventional technique.¹⁰ These results have made it difficult to evaluate the clinical efficacy of RATKA. Some scholars have proposed that due to the flaws in the design of the scoring system and the insufficient accuracy of the scoring standard, the quantitative evaluation of the curative effect after TKA cannot meet the requirements of the increasingly precise technological progress.11, 12, 13

The three-dimensional (3D) gait analysis can collect the motion trajectory of the knee joint in multiple gait cycles from multiple angles through cameras placed in different positions. Gait analysis can accurately present the small changes in knee joint motion, achieve precise quantification, avoid the interference of subjective factors, and provide an objective, real and accurate method for solving the problem of roughness in the evaluation system of knee joint function. This accurate and quantifiable solution is gradually being applied in the efficacy evaluation after TKA.¹⁴ The 3D patterns of knee joint loading during gait have been shown in the past to move toward asymptomatic patterns post-TKA. Gait patterns of RATKA are expected to be closer to normal gaits, which is more accurate and minimally invasive than conventional TKA.

In previous studies, gait analysis was usually performed at 1 year post-TKA, with the purpose of reducing the adverse effects of pain and lameness on gait analysis during postoperative rehabilitation. With the continuous improvement of TKA technique and the application of rapid rehabilitation technology, most patients could restore knee function within 6 months after TKA. At this time, gait analysis can help surgeons find deficiencies and defects during the functional rehabilitation of the knee joint, guide patients to conduct individualized functional training, and further improve the rehabilitation level post-TKA. For RATKA, its accuracy is significantly improved compared with conventional technique, and more necessary to adopt accurate and quantifiable efficacy evaluation methods are needed to help patients obtain rapid and effective rehabilitation treatment. However, to our knowledge, there is no such report of 3D gait analysis after RATKA.

This study aims to conduct 3 consecutive gait analyses of RATKA patients in the early postoperative period (pre-TKA, 3 months and 6 months post-TKA) to quantitatively evaluate the patient's gait from 3 aspects (spatiotemporal parameters, kinetic parameters and kinematic parameters), and to compare it with the contralateral lower limb to investigate the degree and trend of knee function recovery after using the robotic technology. Combined with the knee joint function WOMAC score, the correlation between the function score and gait parameters is studied and analyzed.

Methods

Participants

Patients who underwent RATKA for severe knee OA in the hospital from October 2020 to February 2021 were included. The inclusion criteria were: (1) Age ≤ 80 years old; (2) Those who needed unilateral knee arthroplasty and had no contraindications; (3) Able to complete follow-up and gait analysis tests, with good compliance; (4) Volunteered to participate in this study and signed informed consent in writing.

The exclusion criteria were: (1) Knee varus or knee valgus deformity >15°; (2) Ankylosing/ankylosing deformity of the hip or ankle joint in the lower extremities, or having undergone joint arthroplasty before; (3) Known or suspected of being allergic to polyethylene, titanium, cobalt, chromium or materials containing iron elements; (4) Unable to support and/or fix the prosthesis due to disease; (5) Those who have participated in clinical trials of other investigational drugs or devices within the past 3 months before the start of the study; (6) Suffering from autoimmune diseases leading to inflammatory joint disease such as rheumatoid arthritis, Charcot's arthritis, systemic lupus erythematosus, ankylosing spondylitis and accumulated multiple joints throughout the body (including the contralateral knee joint, bilateral hip joints), which might cause errors in the gait analysis results.

Demographics of patients

Forty-two cases met the inclusion criteria; 4 of rheumatoid arthritis and 5 with a history of TKA on the contralateral knee were excluded; 2 did not complete the gait analysis and lacked of gait data. A total of 31 cases (31 knees) were included in this study. Among them, 6 were male and 25 were female, and 11 were left knee and 20 were right knee. They were 55–80 years old, with an average of 66.1 ± 8.6 years old. The height was 153.7 ± 7.6 cm (range of 140–168 cm), the weight was 62.9 ± 10.1 kg (range of 46.2–80.6 kg), and the body mass index (BMI) was 25.9 ± 4.1 (range of 19.1–34.7). The research protocol was approved by the Ethics Committee of the First Hospital Affiliated to Army Military Medical University (No. KY2019163). All patients gave informed consent and signed an informed consent form. The study was registered on http://www.chictr.org.cn/(ChiCTR2100054391).

Study design

This is a retrospective cohort study, with an experimental group (operated limb of RATKA) and a control group (the contralateral limb). The difference of the alignment parameters of the operated limb pre- and post-TKA, the change trend of the gait analysis parameters before and 3, 6 months after surgery in the experimental group were compared. The differences of the gait parameters between the two groups were compared. Combined with the WOMAC score pre- and post-TKA, the early clinical efficacy of RATKA was comprehensively evaluated, and the correlation between gait parameters and the WOMAC score was analyzed.

Surgical procedure

All surgeries were performed by the same team of senior doctors. The anesthesia method was subarachnoid block combined with epidural anesthesia. Surgery was performed under the tourniquet. A median anterior knee incision was used with the medial patellar approach. Advanced MP (Microport Orthopedics, USA) prosthesis was used, and no patellar resurfacing was conducted.

The surgery was performed with the aid of the robotic system SkyWalker™ (Suzhou MicroPort Changxing Robot Co., Ltd., China). This system includes 4 parts: navigation system, robotic arm, surgical design planning system and instant image feedback system.

Surgical planning and design

CT thin-slice scans (slice thickness 1 mm) of the lower extremity were used to establish a 3D skeletal model of the affected lower limb. The surgeon completed the surgical design and determined the prosthesis installation parameters in the surgical planning system. The knee osteotomy angle, osteotomy amount, alignment and rotation of the femoral and tibial prosthesis were calculated (Fig. 1).

Fig. 1.

Adjusting the installation parameters of the femoral and tibial prostheses in the surgical planning system. (A) Adjust the installation parameters of the femoral prosthesis in the surgical planning system. The femoral prosthesis is perpendicular to the lower limb alignment, and the size of the prosthesis is appropriate. (B) The tibial prosthesis is planned and designed with 0° inversion and 3° retroversion, reference for rotation positioning is medial 1/3 anatomical orientation of tibial tubercle.

The femoral and tibial horizontal osteotomy in the coronal plane was set to be perpendicular to the mechanical axis of the tibia; the femoral osteotomy in the sagittal plane was 3° anteverted to the anatomical axis of the femur; the tibial osteotomy was 3° retroverted relative to the anatomical axis of the tibia. On the transverse section, the femoral anterior and posterior condyle osteotomy line was parallel to the surgical condyle axis, and the tibial prosthesis rotation alignment was perpendicular to the projected line of the femoral prosthesis' transverse axis and aligned with the medial 1/3 anatomical orientation of tibial tubercle. The target value of the lower limb alignment was set to 180°.

Navigation system registration and accuracy test

The optical targets installed on the robotic arm, femur and tibia coordinated with the navigation system for spatial positioning and registration; the lower limb moved in multiple directions to locate the center points of the hip, knee and ankle joints to position the lower limb alignment (Fig. 2).

Fig. 2.

Complete robot registration and positioning in the navigation system. (A) Complete the femoral articular surface positioning and registration through articular surface registration technology. (B) Guided by the navigation system, the robotic arm completes the identification and spatial positioning of the tibial articular surface.

Robot-assisted osteotomy

The robotic arm reached the designated osteotomy area according to the navigation guidance and aligned the osteotomy guide connected to the robotic arm with the pre-designed osteotomy line. The instant image feedback system would display the error between the preset position and the real-time position, when the error was less than 0.5 mm, the robotic arm would be locked, and the osteotomy could be performed (Fig. 3).

Fig. 3.

Osteotomy is completed with the assistance of the robotic arm. (A) The robotic arm reaches the tibial osteotomy area according to the navigation guidance. (B) After the robotic arm is locked, the tibial osteotomy is completed through the osteotomy guide connected to the robotic arm.

Installation trial and stability test

After completing the osteotomy of the extension gap, the interposition block was placed to measure the alignment at extension position, and the soft tissue was released according to the prompt of the navigation system. The flexion gap osteotomy was adjusted according to the extension gap. The balance of the flexion and extension gap was measured after the osteotomy of the flexion and extension gap was completed. After installing the trial model, the tension balance of the lateral and medial soft tissues in the navigation system was observed (Fig. 4).

Fig. 4.

The balance and stability test of the flexion-extension gap is completed with the assistance of the robotic arm. (A) The extension gap and the tension balance of the lateral and medial soft tissues are tested. (B) The flexion gap and the tension balance of the lateral and medial soft tissues are tested.

After the test, pulsed irrigation was applied to the osteotomy surface and bone cement was used to complete the installation of the prosthesis. The knee joint flexion and extension function, patella trajectory, medial and lateral tension balance were tested again in the navigation system. Negative pressure drainage tube was placed, and the incision was sutured.

Postoperative treatment and rehabilitation

Postoperative treatments included infection prevention, preemptive analgesia, ice compress to reduce swelling and thrombosis prevention. At 12 h post-TKA, the drainage tube could be removed, and the walking aid could be used to perform weight-bearing walking. Raise of extended leg in bed could be conducted to exercise quadriceps strength. Active and passive knee flexion and extension exercises can prevent joint stiffness, and achieve the goal of flexion not less than 90° within 48 h post-TKA. The patient was discharged from the hospital 5–7 days post-TKA.

3D gait analysis

Gait analysis was conducted within 3 days prior to their surgery, and at approximately 3 and 6 months postoperatively. The tests were all done by the same researcher to reduce some patterns of loading, motion and neuromuscular control over the gait cycle. Vicon 3D gait capture system (Oxford Metrics, UK) was used to complete the gait analysis. The steps include: (1) Spherical reflective markers were placed around the walking trail for camera testing, and the abnormal reflective points were covered. (2) Proof of the coordinate system was established by the 3D gait capture system, and the original coordinates were located. (3) The gait data collection interface was entered to input the patient's height, weight, leg length and other basic information, and the force measuring platform was reset to 0. (4) Reflective markers were pasted on the patient's anatomical landmarks, including bilateral anterior superior iliac spine, posterior superior iliac crest, femoral lateral condyle, femoral trochanter and femoral lateral condyle midpoint, lateral malleolus, between the 2nd to 3rd phalanx, the midpoint of the lateral calf, and the midpoint of the heel, and tracked during the walking trials. Individual markers were placed based on standardized protocols to define anatomical and joint coordinate systems (Fig. 5). (5) Patients stood on the force measuring platform to calibrate the static coordinates and data were collected. (6) Gait test started and selected all the reflective points at the anatomical landmarks to display the gait data. All gait data were averaged from 10 complete gait cycle measurements and the spatiotemporal parameters (walking velocity, cadence, step length, single and double support time), kinematic parameters (knee flexion and extension at stance phase/swing phase, varus and valgus angle, internal and external rotation), and kinetic parameters (flexion and extension moment at stance phase/swing phase, rotation moment, varus and valgus moment) were recorded for analysis. Patients walked at their self-selected speeds along a 6 m walkway. The 3D motion of the bilateral lower limbs and external ground reaction forces and moments were captured during gait.

Fig. 5.

The patient underwent a gait analysis test post-TKA, and the movement trajectory of the lower limbs was captured by the reflective target pasted on the corresponding anatomical mark.

Clinical efficacy evaluation

All follow-ups were implemented by 2 investigators. The investigator was a member of an independent third-party organization and did not participate in the operation or related clinical treatments.

Imaging measurement

Three months post-TKA, the measurement was carried out by full-length weight-bearing X-ray (anteroposterior view and lateral view), as well as CT of the lower extremity. In the radiograph of anteroposterior view, the hip-knee-ankle angle (HKA) was determined in the coronal plane by measuring the angle between a line connecting the center of the femoral head and the center of the knee and a line connecting the center of the knee to the center of the ankle, and the value should be less than 180° when the knee was at varus deformity. The lateral distal femoral angle (LDFA) was the lateral angle between the mechanical axis and the joint line of the femur. The medial proximal tibia angle (MPTA) was the medial angle between the mechanical axis and the joint line of the tibia. In the radiograph of lateral view, the sagittal tibial component angle (sTCA) was the posterior angle formed by the anatomical axis of the proximal tibia and the joint line of the tibial plateau. The posterior condylar angle (PCA) was the angle between the surgical transepicondylar axis and the posterior condylar line of the femoral component (Fig. 6).

Fig. 6.

Postoperative imaging measurement indicators.

Clinical efficacy score measurement

WOMAC score was measured pre-TKA, 3 and 6 months post-TKA. The score evaluated the severity of knee OA and the treatment efficacy based on the symptoms and signs of the patient's knee joint, and was divided into 3 subscales: pain, stiffness, and physical function, with a total of 24 items.

Statistical analysis

SPSS 26.0 (IBM, the USA) was used for statistical analysis. Kolmogorov-smirnov test was used to verify whether the data conform to the normal distribution. If it satisfies the normal distribution, the quantitative data was described by mean ± standard deviation, otherwise, using the median. Paired-sample t-test was used for the imaging indicators; Wilcoxon rank-sum test was used for WOMAC scores; analysis of variance was used for the differences of the 3D gait analysis parameters intra-group and between groups. Spearman correlation coefficient was used to analyze the correlation between WOMAC score and gait parameters, and the test level was α = 0.05.

Results

Comparison of alignment parameters pre- and post-TKA in the experimental group

The experimental group had knee varus deformity pre-TKA, and the HKA angle ranged 165.06°–176.26°, with the mean value of 170.33° ± 2.77°, and the postoperative data reached the standard of neutral alignment. The LDFA ranged 87.4°–92.7°, which was not statistically different from that pre-TKA. The MPTA, PCA and sTCA were significantly improved postoperatively (Table 1).

Table 1.

Comparison of alignment parameters pre- and post-TKA in the experimental group.

Lower limb alignment parameters	Pre-TKA	Post-TKA	t value	p value
HKA (°)	170.33 ± 2.77	180.71 ± 1.30	−18.854	< 0.001
LDFA (°)	88.88 ± 3.14	89.88 ± 1.26	−1.769	0.087
MPTA (°)	79.57 ± 4.37	89.70 ± 1.26	−11.616	< 0.000
PCA (°)	3.14 ± 1.15	0.64 ± 0.27	11.309	< 0.000
sTCA (°)	/	3.49 ± 2.21	/	/

HKA: hip-knee-ankle angle; LDFA: lateral distal femoral angle; MPTA: medial proximal tibia angle; PCA: posterior condylar angle; sTCA: sagittal tibial component angle.

Comparison of preoperative and postoperative WOMAC scores in the experimental group

Comparison of the total WOMAC score

The preoperative total WOMAC score was 35.45 ± 15.44 (range of 9–71), and 10.16 ± 6.12 (range of 2–27) at 3 months, 6.74 ± 7.47 (range of 0–26) at 6 months postoperatively. There was a statistical difference between the preoperative total WOMAC score and the 3-month score (p < 0.001) and the 6-month score (p < 0.001). There was no significant difference between the 3-month score and the 6-month score postoperatively (p = 0.204).

Comparison of the subscale WOMAC score

The WOMAC score were expressed as median (Q1, Q3). The preoperative WOMAC pain score was 6 (5, 10), 3 (1, 4) at 3 months, and 1 (0, 3) at 6 months postoperatively. There was a statistical difference between the preoperative score and the 3-month score (p < 0.001) and the 6-month score (p < 0.001), and there were statistical differences between 3 months and 6 months postoperatively (p = 0.040).

The preoperative WOMAC stiffness score was 2 (0, 3), 1 (1, 2) at 3 months, and 1 (1, 2) at 6 months postoperatively. There was no statistical difference between the preoperative score and the 3-month score (p = 0.086) and the 6-month score postoperatively (p = 0.05). There was no statistical difference between 3 months and 6 months postoperatively (p = 0.263).

The preoperative WOMAC functional score was 24 (20, 30), 5 (2, 8) at 3 months, and 2 (0, 6) at 6 months postoperatively. There were statistical differences between the preoperative score and the 3-month score (p < 0.001) and the 6-month score postoperatively (p < 0.001). There was no statistical difference between 3 months and 6 months postoperatively (p = 0.407) (Fig. 7).

Fig. 7.

Comparison of total WOMAC score and subscales of pain, stiffness and function scores.

∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ns: p > 0.05, no statistical difference.

WOMAC: Western Ontario Mac Master University Index.

Comparison of preoperative and postoperative gait parameters in the experimental group

Compared with pre-TKA, the double support time of the experimental group post-TKA gradually decreased, the maximum knee extension gradually decreased, the maximum knee flexion angle gradually increased, the knee joint range of motion (ROM) increased, and the maximum external rotation of the knee joint increased significantly at both the stance phase and the swing phase. There were several knee joint kinetic differences between pre- and post-TKA during gait. During the stance phase of gait, the maximum varus moment (peak t) gradually decreased, and the difference was statistically significant (Table 2).

Table 2.

Comparison of 3D gait parameters in the experimental group.

Groups	Double support time (s)	Max knee extension (°)	Max knee flexion (°)	Knee ROM	Max knee rotation at swing phase (°)	Max knee rotation at stance phase (°)	Max knee valgus/varus moment at stance phase (N·m/Kg)
Pre-TKA	0.34 ± 0.09	8.38 ± 6.00	44.86 ± 13.36	41.38 ± 13.19	−3.95 ± 2.18	7.38 ± 1.19	0.68 ± 0.31
3 months post-TKA	0.29 ± 0.09	7.79 ± 7.22	55.18 ± 10.77	47.38 ± 10.70	6.48 ± 3.31	11.01 ± 1.814	0.57 ± 0.33
6 months post-TKA	0.29 ± 0.07	3.48 ± 7.99	58.76 ± 7.05	50.38 ± 6.87	7.67 ± 4.13	17.16 ± 1.75	0.49 ± 0.31
Statistical value	F = 5.870 p = 0.005	F = 10.909 p = 0.000	F = 15.418 p = 0.000	F = 7.086 p = 0.002	F = 9.953 p = 0.000	F = 3.721 p = 0.038	F = 5.783 p = 0.005

TKA: total knee arthroplasty, ROM: range of motion.

Comparison of postoperative gait parameters between the experimental group and the control group (mean ± standard error)

Comparison between the experimental group and the control group showed that the maximum knee external rotation at swing phase of the experimental group increased continuously post-TKA, but it was still smaller than that of the control group (F = 4.070, p = 0.048); the maximum knee flexion/extension moment at swing phase of the experimental group increased continuously post-TKA and was greater than that of the control group (F = 4.403, p = 0.040); the maximum knee flexion/extension moment at stance phase of the experimental group continued to increase post-TKA and was greater than the control group (F = 4.110, p = 0.047) (Table 3).

Table 3.

. Comparison of gait parameters between the experimental group and the control group

Groups	Maximum knee external rotation at swing phase (°)	Maximum knee flexion/extension moment at swing phase (N·m/kg)	Maximum knee flexion/extension moment at stance phase (N·m/kg)
Pre-TKA
Experimental Group	-3.95 ± 2.18	0.19 ± 0.019	0.24 ± 0.04
Control group	8.83 ± 2.59	0.15 ± 0.012	0.20 ± 0.01
3 months post-TKA
Experimental group	6.48 ± 3.31	0.24 ± 0.057	0.26 ± 0.04
Control group	11.08 ± 1.84	0.15 ± 0.007	0.19 ± 0.02
6 months post-TKA
Experimental group	7.67 ± 4.13	0.29 ± 0.069	0.27 ± 0.05
Control group	12.46 ± 2.28	0.13 ± 0.001	0.15 ± 0.01
Statistical value	F = 4.070	F = 4.403	F = 4.110
	p = 0.048	p = 0.040	p = 0.047

TKA: total knee arthroplasty

Correlation between gait parameters and WOMAC score of the experimental group

The cadence at 6 months was negatively correlated with the WOMAC score at 6 months postoperatively (r = −0.388, p = 0.031); the preoperative single support time was negatively correlated with the WOMAC score at 3 months (r = −0.425, p = 0.017); the single support time at 3 months was negatively correlated with the WOMAC score at 6 months (r = −0.366, p = 0.043); the maximum knee extension at stance phase at 6 months was negatively correlated with the WOMAC score at 6 months (r = −0.359, p = 0.048); the maximum external rotation at stance phase at 6 months was negatively correlated with the WOMAC score at 6 months (r = −0.406, p = 0.024) (Fig. 8).

Fig. 8.

(A) The cadence at 6 months postoperatively was negatively correlated with the Western Ontario Mac Master University Index (WOMAC) score at 6 months. (B) Preoperative single support was negatively correlated with the WOMAC score at 3 months. (C) Single support at 3 months was negatively correlated with the WOMAC score at 6 months. (D) Maximum knee extension at 6 months was negatively correlated with the WOMAC score at 6 months. (E) Maximum rotation at 6 months was negatively correlated with the WOMAC score at 6 months.

Discussion

With the continuous popularization of TKA technology and the continuous development of patient education, more and more patients have higher expectations for the function recovery after TKA. However, the design and operation of TKA prosthesis still do not have enough breakthroughs, so more and more patients are not satisfied with the effect of operation. Some scholars analyzed and studied the postoperative patient satisfaction and found that 15%–23% of patients were dissatisfied with the results of surgery, but doctors often question this conclusion.^15,16 The main reason for this phenomenon is that there are differences in the evaluation of the postoperative efficacy between the doctor and the patient. From a doctor's perspective, TKA can be called successful realizing correction of the lower limb alignment, alleviation of pain, and restoration of most of the flexion and extension activities post-TKA; however, the patient always hopes to return to sports or a barrier-free life. Therefore, at present, we need to reduce the weight of subjective evaluation in the scoring system after TKA, and increase the objective evaluation, in order to correctly guide both doctors and patients to evaluate the curative effect of TKA. And gait analysis is one of the important objective indicators. Slight asymmetry is always undetectable in a standard physical examination but may generate many negative long-term consequences. Adequate and sensitive gait analysis could help to reveal this asymmetry.^17,18

Sensitivity of WOMAC scoring system to evaluate curative effect after TKA

In this study, we found that WOMAC score was more valuable in assessing knee joint function in the early postoperative period than in the late stage. The greatest improvement in WOMAC scores including pain subscore, was within the first 3 months postoperatively.¹⁹ At 3 months, WOMAC score reflected patients' satisfaction post-TKA and they were in accordance with the major goal of TKA: relieving pain and stiffness as well as restoring function.²⁰ We found that the WOMAC scores at 3 months post-TKA were significantly improved compared with those pre-TKA, but there was no significant difference between the WOMAC scores at 6 months and 3 months post-TKA. The change trend showed that after experiencing early postoperative functional exercise, WOMAC score could reflect the significant improvement in patient's function compared with pre-TKA. The pain score was 7.23 ± 3.04 pre-TKA, decreased to 2.94 ± 2.56 at 3 months post-TKA, and 1.61 ± 1.70 at 6 months post-TKA. There was a significant difference between 3 months and 6 months post-TKA. The total stiffness score was 8, 2.00 ± 2.08 pre-TKA, 1.35 ± 0.91 at 3 months post-TKA, and 0.94 ± 1.12 at 6 months post-TKA. Because of the low scores of these 2 subscales, they had a low impact on the total WOMAC score. In function subscale, the total score of 17 items was 68. The patient's preoperative score was 26.61 ± 12.54, which rapidly dropped to 5.94 ± 4.27 at 3 months post-TKA and 4.19 ± 5.29 at 6 months post-TKA with no significant difference. This subscale had a high proportion of scores, so the overall total WOMAC score no longer had a statistical difference after 3 months post-TKA. Therefore, the evaluation of functional recovery using WOMAC score after 3 months would have a certain degree of limitation. To be specific, this limitations are: (1) The set scoring items and corresponding scores have a large span, and some scenes of the scoring items are not suitable for the life scenes of Chinese patients, which affects the accuracy of the scoring to a certain extent. (2) Among the scoring items, there is still insufficient response to the needs of patients at different stages of rehabilitation. At 3 months post-TKA, the pain and stiffness of the patients have been greatly relieved, and the improvement of the scores of such items is limited, making scores during the follow-up period difficult to show statistical differences. At 3 months, the pain caused by stretching soft tissue was still obvious during knee flexion, so the score of knee flexion decreased. At the same time, the restorative training of muscle strength had just begun, and the muscle strength that had not fully recovered caused it to still have obstacles when going up and down the stairs. The setting of the scoring system is not designed for this objective change trend, resulting in a contradiction between the score and the actual postoperative change. Therefore, the evaluation system after TKA needs to be improved. The focus is to combine the rehabilitation condition of patients and set the scoring items with different key points at different time points.²⁰

Robot-assisted technology more accurate, and 3D gait analysis more efficacy

Achieving neutral alignment is important for the success of TKA, but it is still an unsolved issue with the current TKA technique.²¹ There are many factors that affect the efficacy of TKA. The iatrogenic factors recognized by most scholars include: lower limb alignment deviates from 180° for more than 3°, abnormal femoral prosthesis rotation alignment, unbalanced flexion and extension gap, unbalanced soft tissue tension, and over stuff caused by prosthesis that is too large. In conventional TKA, movement in instruments, especially inadvertent movement of fixation pins, can lead to realistic errors of 1°–4°, residual excessive varus and valgus deformity and malalignment of the lower limbs.²²

This study conducted research on patients receiving RATKA, aiming to improve the accuracy of TKA with the help of RA technology, and avoid the interference of the above-mentioned factors on the evaluation of its postoperative efficacy. We used a robotic system to achieve accurate angle of femoral and tibial cuts and counted the HKA angles pre- and post-TKA. It showed that the postoperative HKA angle was corrected to 180.71° ± 1.30°, which was within 2° of the preoperative target value of 180°; the postoperative LDFA and MPTA were both very close to 90; PCA was 0.64° ± 0.27°, achieving the expected goal pre-TKA; TCA was 3.49° ± 2.21°, and the average sTCA was about 3°. From the assessment of the prosthesis position in the sagittal plane, coronal plane, and transverse plane, it was confirmed that under RATKA, the angle of the prosthesis position was in the safe range, and the degree of variability was low. This result can minimize the impact of surgical accuracy factors on surgical efficacy. After removing the interfering factors of poor prosthesis position, our evaluation of the postoperative efficacy of TKA becomes more objective and closer to the real situation.

Hatfield et al.²³ reported that gait analysis assessments had shown improvements in walking velocity, peak magnitudes of joint loading, joint angles, and neuromuscular activation during gait moving toward more asymptomatic values. Moreover, most of the literatures analyzed gait and knee biomechanics at least 6 months post-TKA.¹² Bączkowicz et al.²⁴ observed that patients with knee OA present altered gait pattern, although with a progressive decrease of abnormalities after the TKA procedure.

Therefore, we adopted the method of 3D gait analysis to quantitatively analyze the patient's walking function from the spatiotemporal parameters (cadence, step length, single and double support time), kinematic parameters (knee flexion and extension at stance phase/swing phase, varus and valgus angle, internal and external rotation), and kinetic parameters (flexion and extension moment at stance phase/swing phase, rotation moment, varus and valgus moment). From the comparison between pre- and post-TKA, the postoperative double support time, maximum extension at stance phase, and maximum varus moment gradually decreased, while the maximum flexion at swing phase, knee joint range of motion, and external rotation gradually increased. In the same gait cycle, the reduction of double support time indicated the improvement of the walking function of the affected knee joint, and it was no longer dependent on the contralateral auxiliary support to maintain the balance required for walking. The maximum knee extension at stance phase and the maximum knee flexion at swing phase together constitute the knee joint ROM. During walking, the ROM of the knee joint increases, suggesting improvement of pain and stiffness, as well as the walking efficiency. The reduction of the maximum varus moment at stance phase suggests an improvement in the degree of knee varus, and the force on the medial of the knee joint is reduced. Bennell et al.²⁵ reported that higher peak and impulse of the knee adduction moment had been associated with higher odds for OA structural progression. Thus, as the TKA corrects the varus deformity of the patient, the postoperative adduction moment decreased during patient's walking. This is especially obvious in male patients.²⁶

In severe OA patients, the preoperative alignment is deviated, so the value of the adduction moment and the angle of adduction during walking are significantly increased. After the alignment is corrected by surgery, the patient's dynamic adduction angle should also be considered. Studies have confirmed that standing limb coronal alignment has limited effect in predicting the dynamic adduction angle of patients during postoperative walking. Due to the influence of soft tissue elasticity and walking habits, patients may have higher force on the medial compartment than the lateral one post-TKA. Therefore, from the perspective of kinetic analysis, greater force and greater coronal varus angle will result in greater varus moment, and the use of moment to reflect the postoperative dynamic varus angle is indirect evidence. In this study, we found that compared with before, the postoperative varus moment of RATKA patients was gradually reduced. Compared with the control group, although the varus moment was still greater than that of the healthy knee joint, the difference was constantly narrowed from 3 months to 6 months. On the premise that the lower limb alignment is corrected to 180°, the varus moment will continue to reduce through functional exercise and gait correction.

The gradual increase in the external rotation of the knee joint during the gait cycle indicates improvement of the knee joint rotation alignment. The rotation angle of the knee joint at stance phase is larger than that at stance phase, which indicates that the knee rotation is greater under weight-bearing conditions, which helps to improve the trajectory of the patella movement and reduce the incidence of knee pain. When the WOMAC score cannot accurately reflect the change trend of the patient's curative effect post-TKA, the use of 3D gait analysis can accurately quantify the patient's walking function. Through controlled studies, parameters with clinical curative effect evaluation values can be screened.

Compared with the control group, although the maximum rotation at swing phase of the experimental group continued to increase after the operation, it was still smaller than that of the control group, suggesting remnant habits of aid-supported gait were exhibited despite the observed significant improvement in overall gait function. This indicator may reflect the specificity of the postoperative efficacy of TKA. In the early and middle postoperative period, it may become an important indicator of efficacy evaluation. The maximum flexion and extension moment and rotation moment of the experimental group were greater than those of the control group at the 3 time points of the gait analysis. However, the results of the comparative analysis pre- and post-TKA showed that there was no statistical difference between the two indicators. The flexion and extension moment and rotation moment of the experimental group pre-TKA were higher than those of the control group, and this difference still exists between the two groups post-TKA. Another reason is that some patients in the control group also have OA lesions. This may be explained as an effect of a deliberate strategy of antalgic gait to reduce knee moments and joint load in the painful knee(s) typical for patients with OA.²⁷

Yeo et al.²⁸ noted that with 8.7 years of the median follow-up duration, the mechanical alignment after robotic TKA had no significant difference in kinematic or kinetic parameters. This suggested that even with the assistance of robot precision technology might be difficult to find differences in kinematic or kinetic parameters in the long-term follow-up results. However, the sample size of gait analysis in their study was only 10 cases, and it is necessary to conduct more in-depth research on this issue.

Significance of 3D gait analysis and WOMAC score correlation

In this study, the 3D gait analysis after TKA showed that some parameters of the knee joint on the operated side had positive improvement changes compared with the preoperative knee, but there were still some differences compared with the knee joint without surgery. The data of a single gait cycle would have a certain degree of deviation. In order to reduce this deviation, we took the average value of 10 cycles with the most natural walking and the most complete gait measurement data, which can represent the real knee joint motion state of the patient. Whether such differences will bring about differences in clinical efficacy remains to be studied. Zanasi²⁹ reported specific gait impairments following surgery may help to determine adequate patient care including rehabilitation strategies focused on improving knee function and gait retraining to optimize recovery following TKA and to prevent the development of compensatory mechanisms. Some indirect correlation between gait velocity and function in the WOMAC sub-scale was found.

This study found that at 6 months postoperatively, the cadence and the WOMAC score were negatively correlated. The increase of cadence reflected the improvement of the overall knee joint function; the reduction of pain during walking and the increase in knee joint support and muscle strength during the gait cycle were reflected in the WOMAC scores, and there would be an improvement in the comprehensive score of lower limb function, and the WOMAC score would be lower.³⁰

The longer the single support time of the patient pre-TKA, the better the strength and balance of the affected knee joint preoperatively, which will have a positive impact on the early recovery postoperatively, and the WOMAC score at 3 months would be lower. At the same time, probably due to the improvement in patient-reported functional status (including pain relief), these relationships are weaker. Although the single support time at 3 months is negatively correlated with the WOMAC score at 6 months, the correlation gradually decreases. It is suggested that before TKA, intensive training of unilateral knee joint strength may improve the function of the knee joint in the early postoperative period.³¹

At 6 months postoperatively, the knee extension will have an impact on the functional status of the knee joint. The higher the extension degree of the knee joint, the better the improvement of the flexion contracture. At stance phase, the increase in extension degree will be beneficial to reduce the fatigue of the quadriceps muscle, thereby improving the functional performance of the lower limb, including walking, flexion and extension and weight-bearing. At 6 months, the increase in knee external rotation at stance phase will affect the WOMAC score at 6 months. The knee external rotation at stance phase can truly reflect the knee joint rotation and alignment state under load and stress. The increase in external rotation is more conducive to the improvement of the patella tracking and the improvement of the knee flexion and extension function during walking. Therefore, the increase in the knee external rotation is negatively correlated with the WOMAC score. Compared with the contralateral knee joint without surgery, the external rotation of the operated knee joint at stance phase recovered better, and there was no significant difference. The external rotation of the knee joint at swing phase was still lower than the control group. The possible reason is that in the non-weight-bearing state, the external rotation control force of the knee joint has not been fully restored. However, this difference is not correlated with the WOMAC score, indicating that the external rotation at swing phase does not have a substantial impact on the various indicators of the WOMAC score. The 3D gait analysis parameters show that with the decrease in the external rotation at swing phase, the moment during the flexion and extension of the knee joint increases, indicating that the knee joint bears more force during the flexion and extension process, so the external rotation at swing phase may reduce the moment in flexion and extension of the knee joint. In the case that the patient's knee extension muscle strength has not fully recovered after the operation, reducing the flexion and extension moment will be more conducive to the patient's functional recovery of the knee flexion and extension, which is closer to the knee joint function of the unoperated side. Winby et al.³² reported that external rotation moments were often interpreted as a “quadriceps avoidance” gait pattern, so that the patients may be avoiding creating a large external rotation moment when they have insufficient knee extensor strength or activation to balance the moment internally.

There are some limitations of the study. Firstly, all the surgeries were performed by one surgeon, using a single implant, of which the conclusion cannot be generalized to all TKAs. Increased simple size and more types of prothesis should be involved into studies in the future. Secondly, this study adopted robotic arm-assisted method to ensure the accuracy of the lower limb alignment and rotation alignment. This is a new technology that causes the sample size included in the study to be very limited. Thirdly, gait analysis has a high degree of patient specificity. In order to achieve error control, the best way is to select patients with one knee joint receiving TKA and the other healthy knee joint for a comparative study. In the Kellgren-Lawrence (KL) classification of the control group, 18 of the 31 patients were classified as KL 1–2, 9 at KL 3 and 4 at KL 4, which has caused a certain impact on the research results to a certain extent. Finally, this is not a randomized trial and the case-control design of the trial may introduce potential bias.

In a conclusion, there is a paucity of data on 3D joint dynamic differences between preoperative and postoperative at early stages of RATKA. Our results indicate that the effect of RATKA on knee joint kinematics, kinetics patterns during gait is valuable.

Patients after RATKA were characterized by significant improvements in reduction of gait abnormalities, including double support time, maximum knee extension and flexion, knee ROM, maximum knee rotation, and maximum knee varus and valgus moment, and there was no statistically significant difference even with self-reported scale. Knee rotation may be the most valuable evaluation indicator of clinical efficacy. Correlation analysis results show that increasing knee strength and knee extension training in postoperative rehabilitation will help improve postoperative knee joint function. Nevertheless, this finding needs to be confirmed by studies with larger sample size and longer-term follow-up results.

Funding

This study was funded by the National Key R&D Program of China (2017YFC0110705).

Ethical statement

All procedures in this study involving human participants were performed in accordance with the ethical standards of the institutional and/or national research committee standards and the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. This study was approved by the Ethics Committee of the First Hospital Affiliated to Army Military Medical University (No. KY2019163). All patients gave informed consent and signed an informed consent form. The study was registered on http://www.chictr.org.cn/(ChiCTR2100054391).

Declaration of competing interest

No benefits in any form had been received from a commercial party related to this article. The authors declared that there were no conflicts of interest.

Author contributions

Liu Yang conceived of the idea and developed the rationale. Rui He completed the surgical procedures and wrote the manuscript. Ran Xiong collected gait data. Mao-lin Sun performed computations. Jun-jun Yang and Hao Chen collected radiological data and conducted measurements. Peng-fei Yang collected the clinical follow-up data. The authors would like to acknowledge the contributions of Xin Chen from Center for Joint Surgery, Southwest Hospital, Third Military Medical University (Army Medical University) for the language support.