Comparative analysis of HURWA versus MAKO robotic-assisted TKA: a prospective cohort study of surgical accuracy and clinical outcomes

Abstract

Background

Robotic-assisted total knee arthroplasty (RA-TKA) has emerged as an advanced surgical technique. However, direct comparisons between Chinese and American robotic systems are limited.

Objective

Exploratory comparison of surgical accuracy and short-term clinical outcomes between Chinese HURWA and American MAKO robotic systems in total knee arthroplasty.

Methods

A prospective exploratory cohort study was conducted from September 2023 to January 2024. Eighty-five patients were assigned to undergo RA-TKA with either the HURWA system (n = 43) or the MAKO system (n = 42) based on the sequential clinical introduction of the systems. The primary outcomes were radiographic accuracy, including the restoration of the mechanical axis within ± 3°. Secondary outcomes encompassed surgical efficiency metrics and patient-reported clinical scores (WOMAC, HSS, VAS) assessed preoperatively and at one-year postoperatively.

Results

All included patients have completed surgery and follow-up.The two systems achieved a comparable and high proportion of mechanical axis restoration within ± 3° (HURWA: 86.05% vs. MAKO: 90.48%). The HURWA system demonstrated a significant advantage in osteotomy time (6.03 ± 2.35 vs. 8.06 ± 2.73 min, P < 0.001). Both groups exhibited substantial and comparable improvements in all clinical outcome scores at the one-year follow-up. The HURWA system, as an open-platform design, offered flexibility in prosthesis selection, while the MAKO system demonstrated a highly standardized workflow.

Conclusion

This exploratory comparison indicates that the HURWA system achieves favorable and comparable one-year clinical outcomes and radiographic alignment to the MAKO system in TKA, with distinct advantages in osteotomy efficiency and implant flexibility. These preliminary findings support the clinical potential of HURWA as a viable option, though definitive conclusions regarding its equivalence require validation through larger, long-term studies.

Introduction

For end-stage knee osteoarthritis (KOA), total knee arthroplasty (TKA) has become the most common treatment, effectively reducing patients’ pain symptoms and improving knee joint function [1]. One of the key factors for success is the precise reconstruction of lower limb alignment, with the classic neutral mechanical axis being the most widely used alignment concept. A postoperative coronal mechanical axis deviation ≤ 3° optimizes load distribution across the medial and lateral compartments [2]. Exceeding this limit may lead to a series of complications such as joint stiffness and prosthesis loosening [3–5]. Currently, in traditional TKA, surgeons mainly use intramedullary guidance for the femur and extramedullary guidance for the tibia to achieve lower limb alignment during surgery. However, this method has inherent flaws: (1) there are inherent errors in surgical operation and significant anatomical variations among different patients, making it difficult to achieve ideal and precise lower limb alignment; (2) it heavily relies on the surgeon’s experience and surgical technique, and it is unable to achieve surgical reproducibility [6, 7].

To improve the precision of TKA implantation, computer navigation (N-TKA) and robotic-arm assisted (RA-TKA) technologies have been successfully applied to TKA [8–11]. In N-TKA, surgeons can complete the surgical planning in the system before the surgery, and receive real-time feedback on the positioning of key landmarks during the surgery to correct and improve the surgical precision. Although N-TKA can only offer real-time feedback during surgery and cannot correct or prevent intraoperative operator errors, RA-TKA has emerged as a new generation trend [9, 12, 13]. Robotic technology has made significant advancements since its first application in hip replacement in 1992, demonstrating effective results in the field of joint replacement [14–17].

In total knee arthroplasty, RA-TKA can be mainly divided into three types of robotic-assisted systems: passive, automated, and semi-automated [9]. Passive RA-TKA is primarily based on real-time monitoring of the surgical process through patient-specific 3D simulation information, providing the surgeon with a bone-cutting pathway. The surgeon uses conventional instruments during the procedure, such as the ROSA Knee Robot System (Zimmer Biomet, USA) and OMNIBotics (Corin Group, England). Automated RA-TKA involves the robotic arm independently performing bone cutting based on preoperative planning, patient-specific 3D information, and navigation guidance. Representations of this type include ROBODOC (Integrated Surgical Systems, Davis, CA, USA) and CASPAR (URS Ortho GmbH, Rastatt, Germany). Semi-automated RA-TKA is also based on patient-specific 3D information, preoperative planning, and provides a visual-tactile feedback interactive system. The surgeon operates the robotic arm to perform bone cutting and monitors the surgical process in real-time through intraoperative navigation. It helps in establishing restrictive virtual boundaries to prevent surgical errors and mechanical arm failures, making it one of the most widely used and mature systems in RA-TKA, such as RIO (Robotic Arm Interactive Orthopedic System, MAKO Surgical Corp., FL, USA) [18, 19]. Additionally, preoperative planning generated based on preoperative CT images can accurately predict component sizes, optimize the surgical workflow, and increase the accuracy of implant positioning [20].

The use of the MAKO system in the United States is expanding, and China’s artificial intelligence field is also following the global trend. In China, a series of joint robot operating systems with independent intellectual property, such as “HURWA”, “Tianji”, “Jianjia”, and “Honghu”, have emerged, achieving excellent accuracy and satisfactory clinical outcomes [21, 22]. However, direct comparative analyses of clinical outcomes between the American MAKO system and domestic robotic systems are lacking.This prospective cohort study analyzed 85 patients who underwent robot-assisted total knee arthroplasty (TKA) by the same chief surgeon at the Department of Orthopedics in Peking Union Medical College Hospital from September 2023 to January 2024. The patients were divided into two groups: MAKO group (n = 43) and HURWA group (n = 42). The study aimed to compare the advantages and disadvantages of the two approaches.

Materials and methods

Ethics statement

The surgical protocol of this study was approved by the Ethics Committee of Peking Union Medical College Hospital, and all patients provided informed consent for the surgery.

Acceptance and discharge standards

Inclusion criteria

The research subjects are patients with knee osteoarthritis who have undergone surgical treatment. When patients meet all of the following conditions, they can choose to enter the observational study:

1. Compliance with the OA diagnosis.

2. Age ≥ 18 years old, gender unlimited;

3. Perfect preoperative knee joint X-ray and bilateral lower limbs CT examinations.

4. Agreed to undergo robot-assisted TKA (HURWA or MAKO). All surgical staff involved completed appropriate training and certification for the specific robotic system used.

5. Subjects/guardians are able to understand the purpose of the trial, demonstrate sufficient compliance with the trial protocol, and sign the informed consent form (ICF).

Exclusion criteria

1. Patients with contraindications to traditional total knee arthroplasty (TKA) surgery.

2. Patients with severe deep vein thrombosis in the affected limb may require immediate medical intervention.

3. Subjects in an immunosuppressed state, suffering from autoimmune diseases, or immune suppression disorders;

4. Pregnant or lactating women;

5. The subject has participated in other drug, biological agent, or medical device clinical trials within the past three months.

6. Patients known for excessive alcohol consumption or drug abuse;

7. Participants who are deemed unsuitable to take part in this clinical trial will be excluded by the researchers.

Sample size

This study is a prospective cohort study with the Chinese HURWA robot group as the experimental group and the American MAKO robot group as the control group. Proportion of limbs restored to within 3° of the mechanical axis was calculated as the core indicator. According to the previous data of our research group, the lower limb correction rate of the experimental group was 81.2%²³, and the lower limb improvement rates of the control group was selected as 86.4% based on surgeries and studies of the same period and level [23–26]. The alpha value was set to 0.025, the beta value was set to 0.80.In the review of expert recommendations and further statistical research, we found that the initial non inferiority threshold value (δ = 0.15) and calculation could not meet the requirements of real-world clinical orthopedic research [27–30].In the The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) E9 and E10 guidelines and the US FDA [31], it is pointed out that the non inferiority threshold is the cornerstone of trial design, which should be based on clinical judgment and baseline data of standard therapy in historical data. For orthopedic surgical clinical outcomes (infection, thrombosis, false body survival rate, etc.), the threshold is usually limited to 0.05–0.1, and its value should be smaller than the advantage demonstrated by standard therapy compared to placebo [31–34].

Based on the final inclusion of 85 cases (HURWA: 43 cases, MAKO: 42 cases), this study was unable to meet the sample size requirements for formal non inferiority testing in ideal clinical studies. In view of this situation, we have re examined the positioning and analysis strategy of this study:

1. Research positioning: This study should be regarded as an exploratory comparative research aimed at providing preliminary and valuable real-world data and trend observations for the performance of two emerging robot systems, particularly in terms of surgical efficiency and workflow.

2. Statistical analysis: We compare the rates between two groups (chi square test or Fisher’s exact test, etc.) and their corresponding P-values, absolute risk differences, and 95% confidence intervals for the analysis of data results, which should be interpreted as descriptive and exploratory.

3. Post efficacy analysis: Conduct post efficacy analysis based on actual observed outcome indicators.

Despite the limitation of sample size, patients in this study were included based on their initial visit intention and naturally grouped in chronological order. All surgeries were performed by the same senior surgeon according to standard procedures, ensuring consistency of operations within the group. We will fully elaborate on the insufficient sample size as the main limitation of this study.

General information

The study included patients who underwent robot-assisted TKA at the Department of Orthopedics, Peking Union Medical College Hospital, from September 2023 to January 2024.A total of 95 cases were screened in this experiment, of which 3 were lost to follow-up and 7 were not suitable for this study due to complexity. Ultimately, 85 cases were included and were allocated to two groups: the HURWA group (the first 43 patients who underwent surgery using the HURWA surgical robot system) and the MAKO group (the next 42 patients who underwent surgery using the American MAKO surgical robot system).

The HURWA group consisted of 4 males and 39 females, with an average age of 69.53 ± 7.20 years (range 53–84 years) and BMI of 26.31 ± 4.64 kg/m2. The MAKO group consisted of 6 males and 36 females, with an average age of 69.17 ± 5.00 years (range 60–82 years) and BMI of 27.29 ± 3.46 kg/m2. There was no statistically significant difference in the overall preoperative basic situation, such as preoperative knee joint pain visual analogue scale (VAS), range of motion (ROM), Hospital for Special Surgery (HSS) score, Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score, and hip-knee-ankle angle (HKA) deviation between the two groups(Table 1).

Table 1.

General information before surgery

	Group (mean ± SD)		T/χ2	p
	HUWRA(n = 43)	MAKO(n = 42)
Age	69.53 ± 7.20	69.17 ± 5.00	0.275	0.784
Gender	M/F 4/39	M/F 6/36	0.508	0.476
BMI	26.31 ± 4.64	27.29 ± 3.46	−1.113	0.269
Extend	3.37 ± 4.04	4.83 ± 6.16	−1.289	0.202
Flexion	108.37 ± 9.62	105.60 ± 9.89	1.312	0.193
Pain (Rest)	5.00 ± 0.49	4.86 ± 0.87	0.929	0.356
Pain (activity)	6.05 ± 0.58	5.83 ± 0.93	1.263	0.211
HSS	55.56 ± 11.56	59.24 ± 7.48	−1.747	0.085
WOMAC	96.05 ± 26.78	101.64 ± 32.17	−0.872	0.385
HKA	172.51 ± 7.76	172.23 ± 8.04	0.167	0.868
Surgical side	L/R 23/20	L/R 19/23	0.579	0.447

* p < 0.05 ** p < 0.01

Patients were divided into two groups based on their admission time: from September 2023 to October 2023, the robot system used was HURWA (43 cases), designated as the HURWA group; and from November 2023 to January 2024, the robot system used was MAKO (42 cases), designated as the MAKO group. All patients are continuously tracked to complete a 1-year follow-up.The main surgeons in this study had all undergone training in robot-assisted total knee arthroplasty using HURWA/MAKO robots with 7 or more cases of experience prior to the start of the study.The main core personnel of the surgical team, anesthesia and nursing team have not changed, and some auxiliary personnel have been replaced by assistants of the same level.

Surgical methods

The principles of preoperative planning and the operating physician remained unchanged for the two groups of patients except for differences in the operating systems.

Preoperative planning

Preoperatively, the patient’s lower extremity CT is imported into the surgical planning software compatible with the surgical robot, and a three-dimensional model of the lower extremities is reconstructed based on the CT data. The surgeon manually identified key points on the 3D model, including the center point of the femoral head rotation, the center point of the knee joint, and the center point of the ankle joint. By connecting the femoral head rotation center, the knee joint center, and the ankle joint center, the coronal lower limb force line is determined. The selection of key points is based on the respective standards of the two companies. After the key points are selected, preoperative prosthesis positioning is performed on the segmented reconstructed 3D model, aiming to reconstruct the knee joint movement function. A suitable preoperative osteotomy plan is designed based on the patient’s knee joint anatomical morphology and the experience of the chief surgeon, including the prosthesis model and appropriate varus/valgus, flexion/extension, and internal/external rotation angles.

Surgical procedure

Robot registration and calibration

The patient is under general anesthesia and positioned supine. The robotic surgery table and navigation console are placed on both sides of the patient. The MAKO system required additional spatial positioning steps and registration for the robotic arm, femur, and tibia targets. The femur and tibia targets are then inserted into the patient’s femur and tibia. The femur is rotated around the femoral head to determine the center point of the femoral head. A blunt-tipped target pen is used to mark “medial malleolus” and “lateral malleolus” to determine the center point of the ankle joint. It is important to select different characteristic points for registration based on the system requirements.Implant bone positioning nails to facilitate real-time checking of whether there is any deviation in the femur and tibia targets during the operation.

Surgical plan confirmation

Prior to the osteotomy procedure, the lead surgeon modifies the preoperative plan taking into consideration their expertise, established osteotomy protocols, and the individual circumstances of the patient observed during the surgery.

Osteotomy

According to the surgical plan, the lead surgeon held the foot pedal in the down position, moved the robotic arm to the specified position, and adjusts the end cutting guide to the cutting state. During the movement of the robotic arm, if the lead surgeon recognizes any errors in the movement position or other situations that require stopping, they should release the foot pedal, and the robotic arm will cease movement.In case of an emergency, the emergency brake button can be pressed. Once the robotic arm is in position, the lead surgeon can use a spacer to confirm if the cutting guide is in the correct position, and after confirmation, the surgeon operates the oscillating saw for bone cutting.During operation, both systems can display the position of the saw blade and the cutting plan in real time(Figs. 1 and 2). It is noteworthy that when using the MAKO system for bone cutting, the blade needs to be replaced once, while with HURWA, no replacement is required.

Fig. 1.

HURWA Robot Osteotomy’s Data Interface

Fig. 2.

MAKO Robot Osteotomy’s Data Interface

The balance of soft tissues

Soft tissue balance was assessed and adjusted as necessary by carefully analyzing the relationship between them and making any necessary adjustments to achieve optimal harmony.

The HURWA system provides flexibility by allowing a selection of implant manufacturers, whereas the MAKO system restricts choices to fixed implant models (specifically, Stryker: Triathlon in this study) [35]. To ensure consistent surgical soft tissue balance after installing the femoral implant trial model, we utilized the knee joint soft tissue tension tester from AiQiao Company for measurement. Based on the data and the lead surgeon’s experience, adjustments were made to the soft tissue balance or secondary osteotomy was performed [7, 36–38].

Implant installation and incision closure

After the prosthesis trial fitting, irrigate the wound with a pulse gun, and then install the total knee joint prosthesis. The installation method is consistent with traditional surgery. After the bone cement solidifies, reassess the soft tissue balance in the varus stress state of the knee joint, then remove the femoral and tibial targets and marking nails, irrigate, and close the incision.

Postoperative management

Postoperative routine treatment includes anti-thrombotic, anti-infective, and pain management. Activities of daily living with the assistance of a walker are initiated 1–2 days postoperatively based on the patient’s recovery status. Non-weight-bearing knee joint flexion and extension exercises are performed to prevent joint stiffness, and postoperative lower limb weight-bearing full-length images are taken. Monitor for complications such as surgical site infection and deep vein thrombosis.

Evaluation indicators

To ensure measurement consistency, the pre- and post-operative imaging index measurements in this study were performed by two observers. The final result is the average value of the two observers.

Imaging metrics

All radiographic indices were measured using full-length X-ray of the lower limbs. Standard bilateral weight-bearing full-length X-rays of the lower limbs were taken before surgery, and on postoperative day 3, standard weight-bearing full-length X-rays and anteroposterior view were obtained. Full-length X-rays of the lower limbs were taken from all patients using a standardized protocol. Patients were positioned standing upright on the imaging stand with their back against the stand, arms naturally hanging down, knees extended, feet al.igned with shoulders, and lower limbs internally rotated by approximately 15°. The fibular head overlapped the proximal one-third of the tibia, and the patella pointed vertically forward. Postoperative indices were measured and recorded to assess the accuracy of the robotic surgical system. Radiographic indices included the following: FFC (Frontal Femoral Component angle/femoral mechanical axis deviation angle, preoperative standard 87°, postoperative 90°), FTC (Frontal Tibia Component angle/tibial mechanical axis deviation angle, preoperative standard 87°, postoperative 90°), LFC (Lateral Femoral Component angle/anatomical axis of the femur in reference to the knee joint, preoperative approximately 83°, postoperative standard 90°, tends to anterior tilt, may be slightly less such as 88–89°), LTC (Lateral Tibia Component angle/anatomical axis of the tibia in reference to the knee joint, preoperative approximately 81°, postoperative standard 87°), etc [39]. The absolute difference between postoperative radiographic measurement indices and intraoperative design indices was defined as the deviation for that parameter, with a smaller offset indicating higher accuracy.

Clinical efficacy indicators

Primary outcomes were the proportion of limbs restored to within 3° of the mechanical axis and the use of the thinnest insert; secondary indicators consist of assessment of postoperative length of hospital stay, drainage volume, and patient’s WOMAC score, HSS, VAS, and satisfaction at 6 months postoperatively.

The side of femoral prostheses

To assess the accuracy of preoperative planning and execution in robotic surgery, collect the final implanted prosthetic model of the patient, including femoral prostheses, tibial prostheses, and liner models. This assessment will be based on whether the thinnest liner of that model is used.

Statistical analysis

Using SPSS 26.0 statistical software (SPSS Inc., USA) for statistical analysis of data, we tested the normality of metric data. Variables that follow a normal distribution are expressed as x ± s. The comparison of clinical scores before and after surgery is conducted using paired sample t-tests.The significance level α is set at 0.05 for the tests.

The accuracy of postoperative imaging indicators (HKA, FFC, and LTC) is not related to preoperative differences and is not subjected to statistical processing. Instead, only the commonly used ratio in current research on knee arthroplasty robots is used to describe the accuracy of lower limb alignment reconstruction.

In this study, the inter-observer agreement of repeated measurements is evaluated using the intraclass correlation coefficient (ICC) value. ICC > 0.75 indicates good agreement; 0.4 < ICC < 0.75 suggests moderate agreement; ICC < 0.4 indicates poor agreement.

Results

Consistency check

In this study, the intra-class correlation coefficients (ICC) of preoperative and postoperative values of HKA, FFC, LFC, FTC, and LTC measured between two observers were all > 0.75(Table 2), indicating good inter-observer reliability.

Table 2.

Result of ICC intra group correlation coefficient

category	Bidirectional mixing/random consistency	ICC intra group correlation coefficient	95% CI
Preoperative coronal deformation	ICC(C,1)	0.985	0.978 ~ 0.990
	ICC(C, K)	0.993	0.989 ~ 0.995
Postoperative coronal deformation	ICC(C,1)	0.870	0.806 ~ 0.913
	ICC(C, K)	0.930	0.893 ~ 0.955
FFC	ICC(C,1)	0.894	0.842 ~ 0.930
	ICC(C, K)	0.944	0.914 ~ 0.964
FTC	ICC(C,1)	0.927	0.889 ~ 0.952
	ICC(C, K)	0.962	0.941 ~ 0.975
LFC	ICC(C,1)	0.955	0.932 ~ 0.971
	ICC(C, K)	0.977	0.965 ~ 0.985
LTC	ICC(C,1)	0.912	0.868 ~ 0.942
	ICC(C, K)	0.954	0.929 ~ 0.970

General results

All patients successfully underwent surgery, with the incision healing in the initial stage. Mean total surgical time for the 85 procedures was 86.82 ± 12.97 min(Fig. 3). The osteotomy time was 7.03 ± 2.72 min, and the average length of hospital stay was 6.45 ± 2.1 days.No postoperative infections, severe deep vein thrombosis in the lower limbs, or serious complications such as prosthetic loosening occurred in any of the patients(Table 3).

Fig. 3.

The total surgical time for the HURWA and MAKO groups

Table 3.

Surgical parameters and outcomes

	Group	Sample size	Mean value	Standard deviation	Mean value difference	95% CI	t	df	p
Bone 3D registration time(min)	HURWA	43	4.46	0.76	−0.59	−0.935 ~−0.236	−3.329	83	0.001**
	MAKO	42	5.05	0.85
	TOTAL	85	4.75	0.86
Osteotomy time(min)	HURWA	43	6.03	2.35	−2.03	−3.123 ~−0.930	−3.674	83	0.000**
	MAKO	42	8.06	2.73
	TOTAL	85	7.03	2.72
Sewing time(min)	HURWA	43	21.95	4.69	−0.21	−2.072 ~ 1.645	−0.228	83	0.82
	MAKO	42	22.17	3.88
	TOTAL	85	22.06	4.28
Surgery time(min)	HURWA	43	90.3	9.55	7.04	1.579 ~ 12.502	2.571	69.222	0.012*
	MAKO	42	83.26	15.03
	TOTAL	85	86.82	12.97
Intraoperative blood loss(ml)	HURWA	43	52.33	10.65	−27.44	−39.660 ~−15.213	−4.515	47.282	0.000**
	MAKO	42	79.76	37.95
	TOTAL	85	65.88	30.82
HB difference(g/l)	HURWA	43	28.74	15.85	3.55	−2.873 ~ 9.981	1.1	83	0.275
	MAKO	42	25.19	13.85
	TOTAL	85	26.99	14.91
Volume of drainage(ml)	HURWA	43	155.48	88.22	−19.12	−61.490 ~ 23.247	−0.897	89	0.372
	MAKO	43	174.6	113.85
	TOTAL	42	164.93	101.59
Postoperative hospitalization time(days)	HURWA	43	5.85	2.65	−0.6	−1.637 ~ 0.429	−1.162	83	0.249
	MAKO	43	5.85	2.65
	TOTAL	42	6.45	2.1

* p < 0.05 ** p < 0.01

Primary outcomes

The accuracy of lower limb force line reconstruction (Table 4, 5).

Table 4.

Cross (chi square) analysis results about ‘HKA improvement rate’ and ‘amount of use the thinnest insert’

	classification	Group(%)		Total(%)	χ2▢	p▢
		HURWA	MAKO
HKA Improvement Rate	1°	23(53.49)	29(69.05)	52(61.18)	5.251	0.154
	2°	10(23.26)	3(7.14)	13(15.29)
	3°	4(9.30)	6(14.29)	10(11.76)
	4°	6(13.95)	4(9.52)	10(11.76)
Use The Thinnest Insert	NO	3(6.98)	6(14.29)	9(10.59)	1.199	0.274
	YES	40(93.02)	36(85.71)	76(89.41)
Both core results are met	NO	9(20.93)	9(21.43)	18(21.18)	0.003	0.955
	YES	34(79.07)	33(78.57)	67(78.82)

* p < 0.05 ** p < 0.01

Table 5.

Imaging information of HKA and HKA deviation

	Group	Sample size	Mean value	Standard deviation	Mean value difference	95% CI	t	df	p
HKA	HURWA	43	180.10	2.03	−0.15	−0.982 ~ 0.680	−0.362	83	0.718
	MAKO	42	180.25	1.82
	TOTAL	85	180.17	1.92
HKA Deviation	HURWA	43	1.31	1.54	0.13	−0.504 ~ 0.763	0.407	83	0.685
	MAKO	42	1.18	1.40
	TOTAL	85	1.24	1.46

* p < 0.05 ** p < 0.01

Statistical analysis was performed on postoperative HKA and preoperative planning schemes for all 85 cases.

Core outcome measures for this study include the HKA improvement rate and the decision on whether to use the thinnest insert.

1. HKA improvement rate: All patients in this study were treated with surgical osteotomy plans based on the classical mechanical axis principle. Therefore, the postoperative HKA deviation is improved within 3°, the improvement rates are as follows: Both systems demonstrated excellent accuracy in achieving the primary endpoint of mechanical axis restoration within 3° (HURWA: 86.05% [37/43] vs. MAKO: 90.48% [38/42]). The mean postoperative HKA deviation was 1.24°±1.46° overall, with no significant difference between groups (HURWA: 1.31°±1.54° vs. MAKO: 1.18°±1.40°, 95% CI=−0.504 ~ 0.763,P = 0.685).The data from both the HURWA group and MAKO group were superior to the HKA deviation of the chief surgeon in convention TKA(C-TKA) surgery in our group (N = 74 cases, 3.02 ± 2.74)²³.

2. When the thinnest insert of knee prosthesis is implanted in patients, it is demonstrated that this surgery achieves the appropriate minimum bone resection and causes the least bone loss to the patients. The use of the thinnest polyethylene insert showed no significant difference between groups (HURWA: 93.02% [40/43] vs. MAKO: 85.71% [36/42]; absolute difference: 7.31%, 95% CI: −5.2% to 19.8%; χ²=1.197, P = 0.274; Cramer’s V = 0.12).

3. Post efficacy analysis: Meet core indicator one: improve the lower limb force line to within 3 °, meet core indicator two: use the thinnest gasket, and both meet the post work efficiency analysis separately. The results are shown in Table 6.

Table 6.

Post effectiveness analysis

	HURWA	MAKO	δ	power
HKA Improvement Rate	86.05(37/43)	90.48(38/42)	0.05	0.06
			0.08	0.13
			0.1	0.2
Use The Thinnest Insert	93.02(40/43)	85.71(36/42)	0.05	0.58
			0.08	0.74
			0.1	0.83
Both core results are met	79.07(34/43)	78.59(33/42)	0.05	0.15
			0.08	0.25
			0.1	0.32

Note: δ is the threshold value, power is the efficiency

The two core outcome measures preset in this study. The statistical power analysis of these two indicators reveals different testing powers:

1. The primary outcome measure (HKA Improvement rate): Post hoc efficacy analysis showed that the statistical efficacy of this study was extremely low (6% −20%) for detecting clinically relevant differences (δ = 0.05–0.10). Therefore, in terms of force line alignment accuracy, this study lacks sufficient statistical power to draw any conclusions about the non inferiority or equivalence of the HURWA system to the MAKO system. The observed small differences (86.05% vs. 90.48%) are more likely to be due to opportunities rather than real effects, and the current data cannot rule out the possibility of a clinically significant negative difference.

2. Key secondary outcome measure (using the thinnest Insert): For this measure, this study demonstrated moderate to good statistical efficacy. At δ = 0.10, the efficiency reaches 83%. This indicates that this study has a reasonable ability to detect inter group differences at this threshold. The observed data (HURWA: 93.02% vs. MAKO: 85.71%) suggests a potential advantage of the HURWA system in achieving better joint space balance. However, it should be emphasized that at a stricter threshold (δ = 0.05), the efficacy decreased to 58%. Therefore, conclusions about its clear clinical advantages still need to be treated with caution and confirmed through larger scale studies.

3. For composite indicators, their effectiveness is similar to that of the primary endpoint indicator, at a lower level.

Overall conclusion: The sample size of this study severely limited the statistical testing power, resulting in a significant lack of efficacy in the primary outcome measure. Therefore, this study should be regarded as an exploratory preliminary comparative study, and its findings (especially those related to alignment of force lines) need to be validated by future larger sample studies. In addition, the positive signal observed on a key indicator of soft tissue balance (using the thinnest Insert) is worth noting, as it points to a potential area of advantage for the HURWA system, providing a clear and worthy hypothesis for future research.

Secondary outcomes

Surgical efficiency metrics

1. Osteotomy Time: HURWA demonstrated significantly shorter osteotomy time compared to MAKO (6.03 ± 2.35 vs. 8.06 ± 2.73 min; mean difference: −2.03 min, 95% CI: −3.13 to −0.93; t = 3.614, P < 0.001; Cohen’s d=−0.80 [large effect size]). This represents a 25.2% reduction in bone cutting time.

2. Total Surgical Time: Conversely, total surgical time was significantly longer with HURWA (90.30 ± 9.55 vs. 83.26 ± 15.03 min; mean difference: +7.04 min, 95% CI: 1.59 to 12.49; t = 2.576, P = 0.012; Cohen’s d = 0.56 [medium effect size]).

3. Registration Time: HURWA required less registration time (4.46 ± 0.76 vs. 5.05 ± 0.85 min; mean difference: −0.59 min, 95% CI: −0.97 to −0.21; t = 3.125, P = 0.001; Cohen’s d=−0.72 [medium-to-large effect size]).

Imaging remaining indicators: FFC, LFC, FTC, LTC (Table 7; Fig. 4)

Table 7.

Deviation of FFC, LFC, FTC, LTC

	Group	Sample size	Mean value	Standard deviation	Mean value difference	95% CI	t	df	p
FFC Deviation	HURWA	43	1.06	1.2	−0.02	−0.545 ~ 0.507	−0.073	83	0.942
	MAKO	42	1.08	1.24
	TOTAL	85	1.07	1.21
FTC Deviation	HURWA	43	0.82	1.17	−0.23	−0.729 ~ 0.274	−0.904	83	0.369
	MAKO	42	1.05	1.16
	TOTAL	85	0.93	1.16
LFC Deviation	HURWA	43	2.7	2.01	−0.72	−1.683 ~ 0.252	−1.471	83	0.145
	MAKO	42	3.41	2.46
	TOTAL	85	3.05	2.26
LTC Deviation	HURWA	43	1.26	1.2	−1.07	−1.773 ~ −0.370	−3.049	68.048	0.003**
	MAKO	43	2.33	1.94
	TOTAL	42	1.79	1.69

* p < 0.05 ** p < 0.01

Fig. 4.

Comparison of postoperative imaging angle deviation

Femoral component coronal (FFC) deviation was 1.07°±1.21°, with the FFC deviation in the HURWA group being 1.06°±1.2° and in the MAKO group being 1.08°±1.24° (P > 0.05), showing no statistical difference.

Tibial component coronal (FTC) deviation was was 0.93°±1.16°, with the HURWA group having an angle of 0.82°±1.17° and the MAKO group having an angle of 1.05°±1.16°(P > 0.05), showing no statistical difference.

Femoral component sagittal (LFC) deviation was was 3.05°±2.26°, with the HURWA group having an angle of 2.7°±2.01° and the MAKO group having an angle of 3.41°±2.46°(P > 0.05), showing no statistical difference.

Tibial component sagittal (LTC) deviation was was 1.79°±1.69°, with the LTC deviation in the HURWA group being 1.26°±1.2° compared to the MAKO group which was 2.33°±1.94°. The two groups exhibited a significant difference in LTC offset angle at the 0.01 level (t=−3.049, p = 0.003). Specifically, the average value of HURWA (1.26) was significantly lower than the average value of MAKO (2.23).

Accuracy of prosthesis model implantation

When comparing the final implant model of the 85 patients who received femoral and tibial prostheses with the preoperative planned prosthesis model, it was found that all patients used the actual implant model during surgery consistent with the preoperative planned size. This greatly proves the accuracy and reliability of the surgical robot system in estimating the surgery.This is consistent with the research results of the same period in China [40].

The result of clinical efficacy

The comparison of preoperative and postoperative clinical efficacy among all enrolled patients is shown in Table 8.There was no significant difference in follow-up time between the two groups of patients(HUWRA 365.65 ± 3.70 vs. MAKO 367.05 ± 4.61 days; mean difference: −1.4 days, 95% CI:−3.205 ~ 0.412; t=−1.537, P = 0.128;).The WOMAC pain score decreased significantly from 98.81 ± 29.53 preoperatively to 13.18 ± 12.70 at final follow-up (t = 23.307, P < 0.001). HSS joint function score increased from preoperative (57.38 ± 9.88) points to the last follow-up of (96.15 ± 5.08) points, with a statistically significant difference (t=−31.225, P < 0.001). Activity and pain scores also improved.It can be seen from the chart that there is a statistically significant difference in clinical scores before and after surgery in both groups.

Table 8.

Comparison of clinical efficacy (Preoperative vs. Postoperative)

Item	Mean value ± Standard deviation		D-value	t	p
	Preoperative	Postoperative
HKA	172.37 ± 7.86	180.17 ± 1.92	−7.8	−9.897	0.000**
Range of motion(Extend)	4.09 ± 5.22	0.15 ± 0.73	3.94	6.943	0.000**
Range of motion(Flexion)	107.00 ± 9.80	130.21 ± 8.59	−23.21	−16.293	0.000**
Pain(Rest)	4.93 ± 0.70	0.35 ± 1.19	4.58	31.289	0.000**
Pain(Activity)	5.94 ± 0.78	0.79 ± 1.44	5.15	29.157	0.000**
HSS	57.38 ± 9.88	96.15 ± 5.08	−38.78	−31.225	0.000**
WOMAC	98.81 ± 29.53	13.18 ± 12.70	85.64	23.307	0.000**

* p < 0.05 ** p < 0.01

However, there is no significant difference in clinical scores between the two groups(Table 9; Fig. 5).We rate patients based on postoperative pain, activity, endurance, and other aspects, with a total score of 100 points [41]. A score of ≥ 80 is considered very satisfied;≥60 points is considered satisfactory;< 60 points is considered common;< 40 points is dissatisfaction;< 20 points is very dissatisfied.The group of patients with a score ≥ 80 points is (HURWA group: 40/43 vs. MAKO group: 38/42). The overall satisfaction rate is 100%, with a Satisfaction Score of is 94.96 ± 7.81 (HUWRA 94.14 ± 8.07 vs. MAKO 95.81 ± 7.54, P = 0.327,no significant difference), From the current average follow-up days, there is no significant difference in satisfaction between the two groups, and both groups have achieved high satisfaction.

Table 9.

Comparison of postoperative clinical outcome scores (HURWA vs. MAKO)

	Group	Sample size	Mean value	Standard deviation	Mean value difference	95% CI	t▢	df	p▢
Range of motion(Extend)	HUWRA	43	0.14	0.68	−0.03	−0.345 ~ 0.291	−0.17	83	0.866
	MAKO	42	0.17	0.79
	TOTAL	85	0.15	0.73
Range of motion(Flexion)	HUWRA	43	129.84	9.85	−0.76	−4.484 ~ 2.968	−0.405	83	0.687
	MAKO	42	130.6	7.17
	TOTAL	85	130.21	8.59
Pain(Rest)	HUWRA	43	0.35	1.41	−0.01	−0.526 ~ 0.509	−0.032	83	0.975
	MAKO	42	0.36	0.93
	TOTAL	85	0.35	1.19
Pain(Activity)	HUWRA	43	0.77	1.63	−0.04	−0.667 ~ 0.583	−0.134	83	0.894
	MAKO	42	0.81	1.23
	TOTAL	85	0.79	1.44
HSS	HUWRA	43	96.3	5.24	0.3	−1.900 ~ 2.505	0.273	83	0.786
	MAKO	42	96	4.97
	TOTAL	85	96.15	5.08
WOMAC	HUWRA	43	11.95	11.81	−2.48	−7.963 ~ 3.013	−0.897	83	0.372
	MAKO	42	14.43	13.59
	TOTAL	85	13.18	12.7
FJS	HUWRA	43	14.95	6.16	−0.81	−3.192 ~ 1.576	−0.674	83	0.502
	MAKO	42	15.76	4.78
	TOTAL	85	15.35	5.51
Satisfaction Score	HUWRA	43	94.14	8.07	−1.67	−5.042 ~ 1.702	−0.985	83	0.327
	MAKO	42	95.81	7.54
	TOTAL	85	94.96	7.81
Follow-up Time	HUWRA	43	365.65	3.7	−1.4	−3.205 ~ 0.412	−1.537	78.503	0.128
	MAKO	43	367.05	4.61
	TOTAL	42	366.34	4.21

* p < 0.05 ** p < 0.01

Fig. 5.

Postoperative clinical efficacy evaluation

Brief summary

Clinical Significance Analysis: Despite statistical significance in lateral tibial component (LTC) deviation (P = 0.003), the mean difference of 1.07° falls below the established minimum clinically important difference of 2° for angular measurements, suggesting limited clinical relevance. Similarly, the observed differences in femoral component sagittal positioning (0.89°, P = 0.006) remain within clinically acceptable ranges.

Power Analysis Results: Post-hoc power analysis revealed that the study had sufficient power (> 80%) to detect large effect sizes in surgical timing outcomes but was underpowered (β < 20%) to detect clinically meaningful differences in radiological parameters and functional scores. To achieve 80% power for detecting a 10% difference in HKA restoration rates, approximately 200 patients per group would be required.

Discussion

Comparison of the domestic surgical robot system “HURWA” and the surgical robot system “MAKO “

The HURWA and MAKO systems share fundamental principless, component structure, and usage process. Both are semi-automatic robotic assistance systems that require preoperative CT imaging of the lower limbs to build a three-dimensional model for preoperative design of the osteotomy plan. Additionally, both systems require positioning reference frame and system registration before osteotomy. Intraoperatively, the surgical osteotomy plan can be adjusted at any time based on soft tissue gaps and actual situations. Both systems also have safety boundary protection to ensure surgical safety and reduce the operating burden on the primary surgeon, and may reduce surgeon fatigue while ensuring surgical accuracy [42].

During the surgical operation, Registration and osteotomy times were significantly shorter with HURWA (HURWA: MAKO registration time, 4.46 ± 0.76 min VS 5.05 ± 0.85 min, p = 0.014; HURWA: MAKO osteotomy time, 6.03 ± 2.35 min VS 8.06 ± 2.73 min, p < 0.01). Potential reasons for these time differences include: 1) The MAKO system increased the operation time for secondary positioning verification and changing surgical saw blades to achieve higher accuracy and safety, while the HURWA system significantly shortened the time with one-time verification and no blade change;2)The MAKO system provided more detailed intraoperative gap balance data, frequently necessitating adjustments to the surgical plan, which indirectly increased osteotomy time. However, total surgical time was significantly shorter with MAKO (HURWA: MAKO total operation time, 90.3 ± 9.55 min VS 83.26 ± 15.03 min, p < 0.01).Factors potentially contributing to the shorter total time with MAKO include: 1)The HURWA system has a variety of adaptable prostheses, which is a major advantage. However, different changes of prostheses will bring corresponding changes in the surgical trial model operation, altering the consistency of the surgical process and increasing the corresponding time, indirectly prolongs the learning curve time [43]. In contrast, all prostheses of the MAKO system are Stryker: Triathlon prostheses, which have higher homogeneity and consequently reduce the subsequent operation time [44].2)We observed that the MAKO system provided more precise gap balance data. This allows for more precise adjustments to the surgical plan, resulting in a more balanced knee joint gap after osteotomy. As a result, the time needed for trial balancing adjustments is reduced, ultimately leading to a shorter overall operation time [45];3)Due to selection bias in the order of enrollment and learning curve effects, the surgical operations in the MAKO group were all performed after those in the HURWA group, which may increase the proficiency of the lead surgeon, reduce surgical time, and achieve better results.

In terms of bone-cutting operations performed by surgical robots, the HURWA system robot stands out for its efficiency and ease of learning, leading to a more effective reduction in operation time. Additionally, its ability to accommodate a wider range of prosthesis models is a significant advantage, this feature can provide surgical teams with more freedom and flexibility in prosthesis selection and surgical planning, and to some extent increase the competitiveness of prosthesis manufacturers, thereby reducing medical costs and promoting the internationalization of China’s medical equipment industry.On the other hand, the MAKO system offers a higher level of security with its dual verification feature and boasts an excellent gap balancing system, resulting in reduced trial balance time and ultimately shorter surgical procedures.

Comparison of accuracy between the “HURWA” surgical robot system and the “MAKO” surgical robot system

The surgical robot system can ensure higher accuracy of implantation through preoperative three-dimensional planning and intraoperative precise bone cutting navigation. The improvement rate of HKA and the use of the thinnest spacer are defined as core outcome indicators. The restoration of the HKA angle, designed through mechanical alignment, is considered a classic standard for measuring surgical outcomes. Additionally, we believe that using the thinnest spacer indicates minimal bone loss resulting from this bone cutting procedure, effectively minimizing damage to patients and facilitating quicker postoperative recovery.

In this study, there were two core outcome measures: HKA improvement rate [HURWA: 86.05% (37/43) vs. MAKO: 90.48% (38/42)] and thinnest pad implantation rate [HURWA: 93.02% (40/43) vs. MAKO: 85.71% (36/42)]. Postoperative HKA deviations observed with both robotic systems were lower than those reported for the lead surgeon’s C-TKA procedures 63.5% (47/74), which proves that the surgical robot system brings more accurate and effective recovery of lower limb force lines [46].

In other radiological measurements, the The LTC deviation was 1.79°±1.69°, with 1.26°±1.2° in the HURWA group and 2.33°±1.94° in the MAKO group (t=−3.049, p = 0.003). The HURWA system demonstrated greater surgical accuracy in this aspect. Consultation with engineers from both companies suggested the primary difference in LTC measurement likely stems from variations in coordinate system definitions between the platforms. MAKO engineers utilized the “center of femoral condyle - center of femoral head” as the axis to define the sagittal plane rotation of the femoral component, with the ankle joint center point distributed as 44% medial malleolus and 56% lateral malleolus.In contrast, the HURWA system defined the sagittal plane rotation of the femoral component using the “anatomical axis of femur” and the ankle joint center point as the distance from the proximal cortical center of the tibia (50% medial malleolus: 50% lateral malleolus) [47]. In postoperative sagittal plane measurements of the LFC, we used the “anatomical axis of femur” and found it challenging to precisely define the proportion points for measurement when using the MAKO system. Consequently, measurements based on the anatomical axis may show greater apparent deviation for MAKO-placed components under this methodology, potentially favoring HURWA in this specific comparison.

Comparison of clinical efficacy between the “HURWA” surgical robot system and the “MAKO” surgical robot system

Significant improvements in WOMAC, HSS, and VAS scores were observed postoperatively. The WOAMC pain score decreased from (98.81 ± 29.53) points before surgery to (13.18 ± 12.70) points at the last follow-up, with a significant difference (t = 23.307, P < 0.001). The HSS joint function score increased from (57.38 ± 9.88) points before surgery to (96.15 ± 5.08) points at the last follow-up, with a significant difference (t=−31.225, P < 0.001). Activity and pain scores also improved. The clinical scores before and after surgery in both groups showed statistically significant differences, while there were no significant differences in clinical scores and satisfaction between the two groups.

In the postoperative clinical follow-up of this study, we found that some patients make choices that are beneficial to the overall development of the event when scoring, such as choosing acceptable or satisfactory surgical outcomes even if there is still some pain present. This may be an interesting point, and the impact of different people’s personalities on prognosis may also be worth exploring [48].

Finally, Statistical considerations reinforce these findings: while some individual parameters showed statistical differences, effect size analysis suggests most differences lack clinical significance. These results should be interpreted within the context of study limitations, including insufficient power for definitive equivalence testing and short-term follow-up. Nevertheless, the comparable accuracy outcomes and complementary advantages of each system support the clinical viability of both robotic platforms in contemporary total knee arthroplasty.Robot assisted total knee arthroplasty has brought better accuracy in prosthesis implantation and certain short-term clinical benefits. It has also been optimized and alleviated in the surgical process to alleviate the fatigue of the surgical team. However, its extra cost is also one of the problems that need to be solved in the future [49–51].

It must be clearly stated that this study is an exploratory comparison rather than a confirmatory non inferiority or equivalence test. Although we initially considered non inferiority test design, after combining expert review opinions with orthopedic clinical research standards, we realized that using a rigorous non inferiority threshold with clinical significance (such as Δ = 5%) would require a sample size far beyond the inclusion size of this study. Therefore, we explicitly reposition this study as an exploratory work that provides preliminary real-world data for emerging robot systems.Post effectiveness analysis further confirms the limitations brought by sample size: the statistical power of this study in detecting clinically relevant differences in force line accuracy is insufficient, therefore all comparative results should be considered as exploratory interpretations. However, in terms of the key indicator reflecting joint balance - “using the thinnest pad”, the HURWA group showed a potential advantage (93.02% vs. 85.71%), which provides preliminary evidence for further mechanism exploration and confirmatory research. Overall, the final conclusions regarding the surgical accuracy and balance performance of the two systems still require further validation through large-scale research in the future.

Limitations

This study is subject to several limitations that should be considered when interpreting the results:

Sample size

The sample size of this study is its main limitation. As discussed in the discussion section, the sample size does not meet the requirements of formal non inferiority testing, which directly prompts us to position the study as an exploratory comparison. Therefore, all inter group comparisons in the article should be considered as preliminary trends indicating future research directions, rather than final conclusions.

Study design

As a single-center, non-randomized cohort study with sequential patient allocation (rather than randomization), our findings may be influenced by selection and temporal biases. While this design was pragmatic for the initial comparative evaluation of the two platforms, it precludes definitive causal inferences. Future multi-center randomized controlled trials are needed to validate our conclusions.

Follow-up duration

While the short-term follow-up (mean 1 year) is sufficient to assess radiographic alignment and early complications, it is inadequate for evaluating long-term critical outcomes such as implant survivorship, functional longevity, and patient satisfaction. A longer-term follow-up study is currently underway to address these concerns.

Learning curve effect

The sequential implementation of the systems (with MAKO surgeries following HURWA procedures) introduces a potential confounding “learning curve” effect. Although all participating surgeons underwent standardized certification and procedural training on both systems, and the study included a run-in period, the inherent advantage of increased surgical experience with the latter cases cannot be fully discounted.

Lack of economic analysis

The absence of a cost-effectiveness comparison limits the broader healthcare policy implications of our findings. However, the open-platform nature of the HURWA system, which allows for the use of various implant brands, presents a compelling hypothesis for significant cost-saving potential that warrants dedicated future investigation.

Generalizability

The study was conducted at a high-volume tertiary referral center with experienced surgeons, which may limit the generalizability of the results to community hospitals or lower-volume settings with less robotic surgery experience.

Absence of conventional TKA control

The study design does not include a direct comparison with conventional TKA. Therefore, while both robotic systems demonstrated improved alignment accuracy, the magnitude of their advantage over conventional methods cannot be quantified within our cohort and is inferred from the existing literature and the previous experience of our surgical team.

Despite the aforementioned limitations, this study provides the first detailed head to head comparison between the HURWA and MAKO robot systems. This study provides valuable quantitative data on efficiency indicators such as operating room time and osteotomy time for two systems. Research has confirmed that imaging evaluation based on postoperative CT has extremely high inter observer reliability, providing a reliable evaluation method for similar studies in the future.

Conclusion

This prospective exploratory comparison indicates that both the Chinese HURWA and American MAKO robotic systems can achieve favorable and comparable early performance in lower limb alignment restoration, with excellent short-term clinical outcomes. Key observations include:

1. Comparable Early Alignment Accuracy: Both systems achieved a high proportion (> 85%) of mechanical axis restoration within 3°.

2. Distinct System Characteristics: The HURWA system demonstrated advantages in osteotomy efficiency and, as an open-platform system, offers flexibility in prosthesis selection. The MAKO system showed a highly standardized workflow.

3. Significant Clinical Improvement: Patients in both cohorts experienced substantial short-term improvements in pain relief, functional recovery, and satisfaction scores.

These preliminary findings support the clinical potential of the HURWA system as a viable option in robot-assisted TKA. However, definitive conclusions regarding its equivalence to established platforms require validation through larger, long-term studies.

Author contributions

Z.XL. wrote this manuscript and conducted data collection and analysis; M.QD assist in writing and data measurement; L.J. &Q.WW.designed this experiment and made revisions to the manuscript; W.W. &Z. XL. assisted in revising the manuscript and conducting data analysis.

Funding

This project is funded by Peking Union Medical College Hospital, with project number 2022-PUMCH-C-016.

Data availability

This study has disclosed some relevant data that does not involve patient privacy. If you need all the data, please contact the corresponding author and the ethics committee of the supporting unit upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.

Ethics statement

The study protocol was approved by the Ethics Committee of Peking Union Medical College Hospital in accordance with the Chinese “Ethical Review Measures for Research Involving Human Life Sciences and Medical Sciences“(Number: I-23PJ975), and all patients have provided written informed consent.

Consent for publication

Not applicable.

Clinical trials

This project has been registered on the International Clinical Trials Registry Platform (ICTRP) in China, with registration number: ChiCTR2300075227(Date:20230830).