A modified minimally invasive subvastus approach shows superior early postoperative clinical outcomes compared to the medial parapatellar approach in total knee arthroplasty for osteoarthritis patients: a comparative study Authors: Wanyan Hu, Yi Nie, Yucong Li, Bingsheng Yang, Zexin Su, Zhaohua Zhu, Chao Xie, Lijun Lin Published: 2025 ABSTRACT ObjectiveThis study compared the modified minimally invasive subvastus approach (MMSA) versus the medial parapatellar approach (MPA) in total knee arthroplasty (TKA) for osteoarthritis, focusing on perioperative blood loss, postoperative pain, quadriceps strength, and knee mobility.MethodsIn this retrospective cohort study, we enrolled patients undergoing TKA from June 2020 to June 2022. Inclusion criteria were adults aged 40–80 with end-stage knee osteoarthritis undergoing primary TKA. Exclusion criteria included neuromuscular disorders, active infections, severe osteoporosis, prior knee surgery, and rheumatoid arthritis. We compared the MMSA group (n = 36) with the MPA group (n = 57) for primary outcomes including blood loss, postoperative pain, quadriceps strength, knee range of motion, and clinical scores including Oxford Knee Score (OKS) and Western Ontario and McMaster Universities (WOMAC) score.ResultsThe MMSA group showed reduced blood loss, less pain on day 3, and better quadriceps strength and knee mobility compared to the MPA group. Clinical scores (WOMAC and OKS) were significantly better in the MMSA group in the early postoperative period, indicating improved function and reduced pain. However, the MMSA group had longer operation times. Long-term recovery and prosthesis placement accuracy, including prosthesis angles, were similar in both groups.ConclusionThe MMSA offers superior short-term outcomes post-TKA without compromising long-term recovery, indicating its potential as a preferred approach for patients aiming to minimize postoperative complications and expedite recovery. This study provides evidence for the clinical superiority of MMSA in the early postoperative period.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12891-025-09428-8. Introduction Total knee arthroplasty (TKA) is an important surgical treatment for end-stage knee osteoarthritis [1]. The medial parapatellar approach (MPA) is a conventional and relatively safe technique for TKA that provides excellent surgical visualization and reduces the risk of complications such as prosthesis misalignment [2]. However, this approach may impact the knee extensor mechanism and blood supply to the peripatellar region, potentially leading to complications such as knee ischemic necrosis, patellar dislocation, and weakness of the knee extensors [3]. The subvastus approach is used as an alternative in TKA to reduce complications from the MPA [4]. This approach results in less trauma, preserves the extensor mechanism, and minimizes damage to the medial portion of the quadriceps muscle [5]. However, this technique still has several limitations, including challenging surgical field exposure and difficulty with patellar valgus [6]. Therefore, this surgical approach requires further refinement. Previously, several researchers [7] have investigated modifying the subvastus approach. However, a comprehensive comparative study between their modified approach and the MPA regarding clinical outcomes and radiographic results is still lacking in clinical practice. Moreover, their modified approaches also have several limitations: (1) restricted applicability for patients with patella baja, (2) high tension of the skin incision, (3) poor visualization of the posterolateral corner, and (4) limited surgical exposure [7]. Under the concepts of enhanced recovery after surgery (ERAS) [8] and minimally invasive surgery (MIS) [9], we present a modified minimally invasive subvastus (MMSA) and conduct a comparative study with the MPA in clinical practice. We hypothesize that in a well-designed retrospective, comparative study, our modified minimally invasive subvastus approach can improve early flexion function, reduce early postoperative pain, increase quadriceps strength, and achieve better clinical scores in the early postoperative period than the MPA. Methods Setting and patient selection The study was guided by the Declaration of Helsinki. The Medical Ethics Committee of Zhujiang Hospital of Southern Medical University approved this study (Approval ID: 2023-SYJS-156-01). This study was completed in the Department of Joint and Osteopathy, Zhujiang Hospital, Southern Medical University, from June 2020 to June 2022. The experimental group (n = 36) consisted of patients who underwent a minimally invasive subvastus approach for TKA between June 2020 and June 2022. In contrast, the control group (n = 57) underwent TKA via the MPA (Fig. 1). A skilled surgeon performed all surgeries for all patients via the same cemented posterior stabilized (PS) knee prosthesis. None of the patients underwent patellar resurfacing. The inclusion criteria were as follows: (i) aged between the ages of 40 and 80 years, regardless of sex, with no restrictions; (ii) skeletally mature; and (iii) end-stage osteoarthritis of the knees in patients who were undergoing primary TKA. The exclusion criteria were as follows: (i) patients presenting with neuromuscular insufficiency, such as paralysis, myolysis, or myasthenia. (ii) The presence of actively infectious lesions in the knee or other anatomical sites; (iii) severe osteoporosis, metabolic bone disease, radiation-induced bone disease, or tumor; (iv) history of previous knee surgery (e.g., high tibial osteotomy) or serious comorbidities (ASA ≥ GRADE III). (v) History of revision knee arthroplasty, posttraumatic knee pathology, and rheumatoid arthritis. The electronic medical records of hospitalized patients with knee osteoarthritis who underwent TKA were reviewed, and patient follow-up was performed. On the basis of previous experience, a power analysis was conducted to estimate the minimum sample size required for assessing the significant difference between the two groups. The significance level (α) was set at 0.05, and the power (1-β) was set at 0.8. The data were imported into GPower 3.1.9.7 software, which determined that a minimum sample size of 82 subjects would be needed (group 1 = 33, group 2 = 49). To reduce error, the experimental group consisted of 36 subjects, whereas the control group comprised 57 subjects (group 1 = 36, group 2 = 57). To avoid subjective bias, all patients were enrolled consecutively. Fig. 1Diagram of the study procedure Diagram of the study procedure Surgical procedures The surgical procedures were conducted with the patient in a supine position under spinal anaesthesia combined with continuous epidural anaesthesia, while the affected knee was flexed at 90 degrees. An inflatable tourniquet was applied to the affected limb at a pressure of 300 mmHg.The proximal incision was made slightly obliquely inwards, and the distal incision was made slightly obliquely outwards, based on the median skin incision of the knee joint, measuring approximately 12 cm in length (Fig. 2A). The skin, subcutaneous tissue, and deep fascia were subsequently incised to fully expose the vastus medialis muscle and medial retinaculum of the patella. Two cm below the vastus medialis muscle, an L-shaped incision was made on the inner side of the patella, with the upper end extending upwards at an angle of approximately 50° [10] relative to the long axis of the lower limb and the lower end extending parallel to the direction of the patellar tendon, ending at the tibial tuberosity (Fig. 2A, B). In addition, we started from the midpoint of the patella and added an auxiliary incision of approximately 2 cm along the inner upper edge of the patella (Fig. 2C, D).Fig. 2A Surgical incision length of the modified minimally invasive subvastus approach and intraoperative adductor block under direct vision. B Intraoperative adductor block under direct vision. C Anterior comparison of the original subvastu approach and modified subvastus approach. D Lateral comparison of the original subvastu approach and the modified subvastus approachThe infrapatellar fat pad was excised, and the release of the synovial sac band facilitated the outwards movement of the patella. The medial osteophytes were removed as part of a routine procedure along with the medial release. The anterior and posterior cruciate ligaments as well as the medial and lateral menisci were subsequently resected while ensuring balance between the medial and lateral soft tissues.The osteotomy sequence followed a movable window technique: starting from the distal femur - proximal tibia - the remaining portion of the femur - and finally addressing the patellar surface.The alignment of the lower limb appeared to be within normal limits, and the knee joint exhibited a range of motion from 0° to 150°. The bone cement prosthesis was successfully implanted.Because the position of the adductor canal can be visualized during the operation (Fig. 2B), the adductor canal block (Figure S1) can be placed on the operating table by the operator during the operation. A drainage tube was then placed. According to the anatomical position, the wound was sutured from the tip of the L-shaped incision to avoid excessive contraction of the medial soft tissue, and the wound was sutured layer by layer.After the procedure was complete, Busch’s cocktail formula [11] was administered via periprosthetic intra-articular injection. All tourniquets were released after wound closure, and a pressure dressing was applied. For both the traditional MPA and the modified approach, continuous passive motion (CPM) [12] exercise was performed on the second day after the operation. The strains were removed after 24 h. The surgical procedures were conducted with the patient in a supine position under spinal anaesthesia combined with continuous epidural anaesthesia, while the affected knee was flexed at 90 degrees. An inflatable tourniquet was applied to the affected limb at a pressure of 300 mmHg. The surgical procedures were conducted with the patient in a supine position under spinal anaesthesia combined with continuous epidural anaesthesia, while the affected knee was flexed at 90 degrees. An inflatable tourniquet was applied to the affected limb at a pressure of 300 mmHg. The proximal incision was made slightly obliquely inwards, and the distal incision was made slightly obliquely outwards, based on the median skin incision of the knee joint, measuring approximately 12 cm in length (Fig. 2A). The skin, subcutaneous tissue, and deep fascia were subsequently incised to fully expose the vastus medialis muscle and medial retinaculum of the patella. Two cm below the vastus medialis muscle, an L-shaped incision was made on the inner side of the patella, with the upper end extending upwards at an angle of approximately 50° [10] relative to the long axis of the lower limb and the lower end extending parallel to the direction of the patellar tendon, ending at the tibial tuberosity (Fig. 2A, B). In addition, we started from the midpoint of the patella and added an auxiliary incision of approximately 2 cm along the inner upper edge of the patella (Fig. 2C, D).Fig. 2A Surgical incision length of the modified minimally invasive subvastus approach and intraoperative adductor block under direct vision. B Intraoperative adductor block under direct vision. C Anterior comparison of the original subvastu approach and modified subvastus approach. D Lateral comparison of the original subvastu approach and the modified subvastus approach The proximal incision was made slightly obliquely inwards, and the distal incision was made slightly obliquely outwards, based on the median skin incision of the knee joint, measuring approximately 12 cm in length (Fig. 2A). The skin, subcutaneous tissue, and deep fascia were subsequently incised to fully expose the vastus medialis muscle and medial retinaculum of the patella. Two cm below the vastus medialis muscle, an L-shaped incision was made on the inner side of the patella, with the upper end extending upwards at an angle of approximately 50° [10] relative to the long axis of the lower limb and the lower end extending parallel to the direction of the patellar tendon, ending at the tibial tuberosity (Fig. 2A, B). In addition, we started from the midpoint of the patella and added an auxiliary incision of approximately 2 cm along the inner upper edge of the patella (Fig. 2C, D).Fig. 2A Surgical incision length of the modified minimally invasive subvastus approach and intraoperative adductor block under direct vision. B Intraoperative adductor block under direct vision. C Anterior comparison of the original subvastu approach and modified subvastus approach. D Lateral comparison of the original subvastu approach and the modified subvastus approach A Surgical incision length of the modified minimally invasive subvastus approach and intraoperative adductor block under direct vision. B Intraoperative adductor block under direct vision. C Anterior comparison of the original subvastu approach and modified subvastus approach. D Lateral comparison of the original subvastu approach and the modified subvastus approach The infrapatellar fat pad was excised, and the release of the synovial sac band facilitated the outwards movement of the patella. The medial osteophytes were removed as part of a routine procedure along with the medial release. The anterior and posterior cruciate ligaments as well as the medial and lateral menisci were subsequently resected while ensuring balance between the medial and lateral soft tissues. The infrapatellar fat pad was excised, and the release of the synovial sac band facilitated the outwards movement of the patella. The medial osteophytes were removed as part of a routine procedure along with the medial release. The anterior and posterior cruciate ligaments as well as the medial and lateral menisci were subsequently resected while ensuring balance between the medial and lateral soft tissues. The osteotomy sequence followed a movable window technique: starting from the distal femur - proximal tibia - the remaining portion of the femur - and finally addressing the patellar surface. The osteotomy sequence followed a movable window technique: starting from the distal femur - proximal tibia - the remaining portion of the femur - and finally addressing the patellar surface. The alignment of the lower limb appeared to be within normal limits, and the knee joint exhibited a range of motion from 0° to 150°. The bone cement prosthesis was successfully implanted. The alignment of the lower limb appeared to be within normal limits, and the knee joint exhibited a range of motion from 0° to 150°. The bone cement prosthesis was successfully implanted. Because the position of the adductor canal can be visualized during the operation (Fig. 2B), the adductor canal block (Figure S1) can be placed on the operating table by the operator during the operation. A drainage tube was then placed. According to the anatomical position, the wound was sutured from the tip of the L-shaped incision to avoid excessive contraction of the medial soft tissue, and the wound was sutured layer by layer. Because the position of the adductor canal can be visualized during the operation (Fig. 2B), the adductor canal block (Figure S1) can be placed on the operating table by the operator during the operation. A drainage tube was then placed. According to the anatomical position, the wound was sutured from the tip of the L-shaped incision to avoid excessive contraction of the medial soft tissue, and the wound was sutured layer by layer. After the procedure was complete, Busch’s cocktail formula [11] was administered via periprosthetic intra-articular injection. All tourniquets were released after wound closure, and a pressure dressing was applied. For both the traditional MPA and the modified approach, continuous passive motion (CPM) [12] exercise was performed on the second day after the operation. The strains were removed after 24 h. After the procedure was complete, Busch’s cocktail formula [11] was administered via periprosthetic intra-articular injection. All tourniquets were released after wound closure, and a pressure dressing was applied. For both the traditional MPA and the modified approach, continuous passive motion (CPM) [12] exercise was performed on the second day after the operation. The strains were removed after 24 h. An example of a minimally invasive subvastus approach is shown in Figure S2. Outcome collection The main outcomes of this study were the clinical score, range of motion (ROM), and straight leg raise angle (SLR) of the patients. Additionally, we measured the radiographic data. Pain assessment utilized the 10 cm VAS. Patients marked the scale according to their knee pain, with scores calculated as rounded distances from the “0 (no pain)” endpoint. Assessments were performed by 2 resident physicians. Evaluations were conducted at three time points: baseline (24 h preoperatively), early postoperative period (postoperative days 3 and 7 on-site in the ward following a 30-minute rest), and long-term follow-up (2 years postoperatively, either on-site during outpatient visits or via telephone if patients were unable to attend). The Oxford Knee Score (OKS) and the Western Ontario and McMaster Universities (WOMAC) score were used to assess knee function preoperatively, 1 month postoperatively, 3 months postoperatively, 6 months postoperatively, 1 year postoperatively, and at the last follow-up at 2 years postoperatively. The ROM and SLR were chosen as the evaluation indicators for the rehabilitation of the extensor mechanism. ROM assessment employed a standardized method with a medical metal goniometer. Patients were placed in a supine position with the knee joint in a natural, relaxed state; the goniometer’s axis was fixed at the knee’s lateral condyle, while its fixed arm was aligned along the femur and movable arm along the tibia. Patients flexed the knee to the maximum pain-limited angle; measurements were repeated 3 times, and the mean value was used as the final ROM (expressed as the maximum flexion angle). Consistent with pain assessment, 2 resident physicians conducted ROM evaluations. Assessment timing was synchronized with VAS evaluations: baseline (24 h preoperatively, concurrent with VAS assessments), early postoperative period (immediately after VAS scoring on postoperative days 3 and 7), and long-term follow-up (2 years postoperatively, with on-site measurements at outpatient visits). The SLR was measured preoperatively and on the first day after surgery, as well as at 3-day, 7-day, and 2-year after surgery. The ROM of the knee was evaluated at 3 days, 7 days, and 2 years after surgery. We also recorded the basic demographic characteristics of the patients, including age, height, weight, BMI, sex, and affected knee side, and recorded perioperative-related outcome indicators, such as surgical duration, skin incision length, and preoperative and postoperative hemoglobin levels. The radiographic examination included an anteroposterior (AP) view of the entire leg to assess the leg axis, as well as AP and lateral views of the knee to assess the alignment. The alignment of the femoral and tibial components was assessed via the femoral coronal alignment angle (FCA), femoral sagittal alignment angle (FSA), tibial coronal alignment angle (TCA), and tibial sagittal alignment angle (TSA), as measured via AP and lateral radiographs of the knee. Data processing and statistical analysis Data were gathered by retrieving medical records, conducting outpatient visits, and making telephone follow-ups. To ensure the reliability of the assessment results, all evaluators underwent standardized pre-employment training to achieve consistency in their understanding and application of the assessment protocol. Additionally, a formal data verification procedure was implemented in this study, involving multiple rounds of cross-validation to perform rigorous and systematic quality control on the collected data, thereby ensuring its accuracy and consistency. At the same time, to ensure the validity of the assessment results, a single-blind assessment protocol was implemented, in which assessors were unaware of participants’ group allocation (with treatment assignment concealed), ensuring that evaluations were based exclusively on original data and effectively minimizing the potential for expectation bias. Continuous variables are summarized as the means ± standard deviations (SDs), whereas categorical variables are expressed as frequencies (%). We used SPSS 29.0 software to compare frequencies via the chi-square test. The data were subjected to statistical analysis via the Kolmogorov‒Smirnov test (p < 0.05) to assess normality. To compare the significance between the two groups for normally distributed data, we used Student’s t-test, and for skewed data, we used the Mann‒Whitney test. Additionally, we used one-way ANOVA to assess the significance of differences between multiple groups. Results Demographic characteristics and blood loss The demographic details of the patients in the two groups at baseline are presented in Tables 1 and 2. Height, weight, hemoglobin, age, and BMI did not significantly differ between the two groups (p > 0.05).Table 1Comparison of preoperative general data between the two groupsMini-modified subvastusMedial parapatellarVariablesn%Mean ± SDn%Mean ± SDpAge (year)68.89 ± 6.76968.35 ± 6.5700.7055(t = 0.380)Height (meter)1.6269 ± 0.045601.6239 ± 0.071230.613(z=−0.506)Weight (kilogram)62.664 ± 11.401360.123 ± 5.33540.063(z=−1.862)BMI (kg/m2)23.16 ± 3.6222.85 ± 2.130.114(z=−1.582)Sex(male/female) Female877.81119.3 Male2822.24680.7Operated knee Right2466.72543.9 Left1233.33256.1Table 2Comparison of postoperative hematological indexes and operation times between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpHemoglobin (g/L) Preoperative129.67 ± 11.278131.07 ± 14.0560.896(z=−0.130) Postoperative113.31 ± 12.217110.72 ± 15.3270.395(t = 0.855) Amount of change16.36 ± 6.77020.35 ± 11.8720.042* (t = 2.061)Hematocrit Preoperative 0.39417 ± 0.039234 0.40361 ± 0.0374290.248(t=−1.164) Postoperative 0.34944 ± 0.042624 0.34126 ± 0.0436240.377(t = 0.889) Theoretical red blood cell loss (TRBCL) 0.0447 ± 0.03525 0.0624 ± 0.034160.005*z=−2.808) Length of incision11.044 ± 1.041817.767 ± 1.9046< 0.001*(z=−8.096) Operative time (minute)69.89 ± 5.59557.19 ± 4.373< 0.001*(t = 12.222) Table 1Comparison of preoperative general data between the two groupsMini-modified subvastusMedial parapatellarVariablesn%Mean ± SDn%Mean ± SDpAge (year)68.89 ± 6.76968.35 ± 6.5700.7055(t = 0.380)Height (meter)1.6269 ± 0.045601.6239 ± 0.071230.613(z=−0.506)Weight (kilogram)62.664 ± 11.401360.123 ± 5.33540.063(z=−1.862)BMI (kg/m2)23.16 ± 3.6222.85 ± 2.130.114(z=−1.582)Sex(male/female) Female877.81119.3 Male2822.24680.7Operated knee Right2466.72543.9 Left1233.33256.1 Comparison of preoperative general data between the two groups Table 2Comparison of postoperative hematological indexes and operation times between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpHemoglobin (g/L) Preoperative129.67 ± 11.278131.07 ± 14.0560.896(z=−0.130) Postoperative113.31 ± 12.217110.72 ± 15.3270.395(t = 0.855) Amount of change16.36 ± 6.77020.35 ± 11.8720.042* (t = 2.061)Hematocrit Preoperative 0.39417 ± 0.039234 0.40361 ± 0.0374290.248(t=−1.164) Postoperative 0.34944 ± 0.042624 0.34126 ± 0.0436240.377(t = 0.889) Theoretical red blood cell loss (TRBCL) 0.0447 ± 0.03525 0.0624 ± 0.034160.005*z=−2.808) Length of incision11.044 ± 1.041817.767 ± 1.9046< 0.001*(z=−8.096) Operative time (minute)69.89 ± 5.59557.19 ± 4.373< 0.001*(t = 12.222) Comparison of postoperative hematological indexes and operation times between the two groups of patients There was no significant difference in postoperative hemoglobin levels between the experimental and control groups. However, the experimental group presented significantly lower changes in hemoglobin levels, theoretical loss of red blood cells (Fig. 3A), and incision length than did the control group. Notably, the operation time in the experimental group was significantly longer than that in the control group.Fig. 3A The theoretical loss of red blood cells (TRBCL) in different groups. B VAS pain scores of the different groups. C The straight leg raising angle in the different groups. D) The traditional parapatellar medial approach requires patella rollover, whereas the modified subvastus does not A The theoretical loss of red blood cells (TRBCL) in different groups. B VAS pain scores of the different groups. C The straight leg raising angle in the different groups. D) The traditional parapatellar medial approach requires patella rollover, whereas the modified subvastus does not Follow-up VAS score There was no significant difference in the preoperative VAS pain score between the experimental and control groups (p > 0.05). The VAS pain score of the experimental group was significantly lower than that of the control group on postoperative day 3 in this trial, whereas no significant difference was observed between the two groups on day 7 or at later time points (Fig. 3B; Table 3).Table 3Comparison of postoperative VAS pain scores between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpVisual analog scale (VAS) Preoperative6.94 ± 1.1207.21 ± 0.8610.231(z=−1.198) Postoperative3d5.72 ± 1.1117.11 ± 1.030< 0.001*(z=−5.093) Postoperative7d1.69 ± 1.1421.91 ± 1.2430.372(z=−0.892) Postoperative2y0.36 ± 0.6390.56 ± 0.9070.464(z=−0.732) Table 3Comparison of postoperative VAS pain scores between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpVisual analog scale (VAS) Preoperative6.94 ± 1.1207.21 ± 0.8610.231(z=−1.198) Postoperative3d5.72 ± 1.1117.11 ± 1.030< 0.001*(z=−5.093) Postoperative7d1.69 ± 1.1421.91 ± 1.2430.372(z=−0.892) Postoperative2y0.36 ± 0.6390.56 ± 0.9070.464(z=−0.732) Comparison of postoperative VAS pain scores between the two groups of patients Postoperative 3d Postoperative 7d Postoperative 2y Functional evaluation Straight leg Raising angle (SLR) In this study, the straight leg raising angle was significantly greater in the experimental group than in the control group on the 1 st and 3rd days after the operation, but there was no significant difference between the two groups on the 7th day (Fig. 3C; Table 4).Table 4Comparison of the postoperative straight leg Raising angle between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpAngle of postoperative straight leg elevation Preoperative88.67 ± 1.09588.61 ± 1.1460.860(z=−0.176) Postoperative1d46.67 ± 18.82219.30 ± 15.908< 0.001*(z=−5.707) Postoperative3d41.94 ± 13.69531.40 ± 12.739< 0.001*(z=−3.571) Postoperative7d56.67 ± 10.14255.44 ± 10.7020.412(z=−0.820) Table 4Comparison of the postoperative straight leg Raising angle between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpAngle of postoperative straight leg elevation Preoperative88.67 ± 1.09588.61 ± 1.1460.860(z=−0.176) Postoperative1d46.67 ± 18.82219.30 ± 15.908< 0.001*(z=−5.707) Postoperative3d41.94 ± 13.69531.40 ± 12.739< 0.001*(z=−3.571) Postoperative7d56.67 ± 10.14255.44 ± 10.7020.412(z=−0.820) Comparison of the postoperative straight leg Raising angle between the two groups of patients Postoperative 1d Postoperative 3d Postoperative 7d The knee range of motion (ROM) There was no significant difference in the range of motion of the current knee between the experimental group and the control group (P > 0.05), whereas the knee range of motion of the experimental group was significantly greater than that of the control group on the 3rd and 7th days after the operation (Fig. 4A; Table 5).Table 5Comparison of postoperative knee joint function between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpROM Preoperative141.58 ± 4.948140.75 ± 4.6030.615 (z=−0.502) Postoperative3d73.47 ± 7.72969.25 ± 10.1120.037*(z=−2.085) Postoperative7d107.64 ± 8.236101.32 ± 6.018< 0.001*(z=−3.650) Postoperative2y131.83 ± 14.262130.65 ± 12.7570. 678(t = 0.416)Fig. 4A Knee range of motion in different groups. B OKS scores of the different groups. C, D The WOMAC scores of the different groups Table 5Comparison of postoperative knee joint function between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SDpROM Preoperative141.58 ± 4.948140.75 ± 4.6030.615 (z=−0.502) Postoperative3d73.47 ± 7.72969.25 ± 10.1120.037*(z=−2.085) Postoperative7d107.64 ± 8.236101.32 ± 6.018< 0.001*(z=−3.650) Postoperative2y131.83 ± 14.262130.65 ± 12.7570. 678(t = 0.416) Comparison of postoperative knee joint function between the two groups of patients Postoperative 3d Postoperative 7d Postoperative 2y A Knee range of motion in different groups. B OKS scores of the different groups. C, D The WOMAC scores of the different groups Clinical score outcomes Oxford knee score (OKS) There was no significant difference in the 6th month, the 1 st year, the 2nd year, or the status between the two groups (P > 0.05). The OKS of the experimental group was significantly lower than that of the control group at the 1 st and 3rd months (P < 0.05, Fig. 4B; Table 6). Table 6Comparison of postoperative knee joint function between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SD p OKS Preoperative9.89 ± 3.0879.70 ± 2.9270 0.742 (z= −2.850) Postoperative1 m19.22 ± 3.95817.00 ± 3.2950.004 *(t = 2.928) Postoperative3 m32.64 ± 6.61230.00 ± 5.9220.048*(t = 2.000) Postoperative6 m39.25 ± 5.33139.77 ± 5.2240.610(z=−0.510) Postoperative1y41.67 ± 1.39441.74 ± 1.4080.955(z=−0.057) Postoperative2y43.14 ± 1.55242.53 ± 1.9000.121(z= −1.552) Table 6Comparison of postoperative knee joint function between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SD p OKS Preoperative9.89 ± 3.0879.70 ± 2.9270 0.742 (z= −2.850) Postoperative1 m19.22 ± 3.95817.00 ± 3.2950.004 *(t = 2.928) Postoperative3 m32.64 ± 6.61230.00 ± 5.9220.048*(t = 2.000) Postoperative6 m39.25 ± 5.33139.77 ± 5.2240.610(z=−0.510) Postoperative1y41.67 ± 1.39441.74 ± 1.4080.955(z=−0.057) Postoperative2y43.14 ± 1.55242.53 ± 1.9000.121(z= −1.552) Comparison of postoperative knee joint function between the two groups of patients Postoperative 1 m Postoperative 3 m Postoperative 6 m Postoperative 1y Postoperative 2y The Western Ontario and McMaster universities (WOMAC) In this trial, the total WOMAC score of the experimental group was significantly better than that of the control group in the first month after surgery (Fig. 4C; Table 7). There was no significant difference in the total WOMAC score between the experimental group and the control group at the third month, the sixth month, the first year, or the second year after surgery (Fig. 4C). The pain and stiffness in the first month after the operation were significantly better than those in the control group (Fig. 4D), the stiffness in the third month after the operation was significantly better than that in the control group, and the function in the first year after the operation was significantly better than that in the control group. Table 7Comparison of postoperative knee joint function between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SD p WOMAC Preoperative51.86 ± 13.26050.61 ± 12.9130.655 (t = 0.449) pain12.81 ± 2.70313.79 ± 2.5620. 081 (t=−1.766) stiffness2.97 ± 1.4643.16 ± 1.5670.615(z=−0.503) function36.08 ± 13.52533.67 ± 12.7470.387 (t = 0.870) Postoperative1 m27.28 ± 6.10231.61 ± 7.5990.005* (t=−2.885) pain5.97 ± 2.7107.09 ± 2.4300.026* (z=−2.223) stiffness2.430 ± 1.1953.44 ± 1.7730.028* (z=−2.195) function18.64 ± 5.33021.09 ± 6.7090.122 (z=−1.546) Postoperative3 m18.19 ± 5.61520.39 ± 5.9090.062(z=−1.866) pain3.11 ± 1.9093.39 ± 1.7500.301 (z=−1.035) stiffness1.64 ± 1.0182.93 ± 1.163< 0.001* (−4.768) function13.44 ± 5.42714.07 ± 5.8790.489 (z=−0.692) Postoperative6 m15.50 ± 2.83316.98 ± 4.8460.066(t=−1.860) pain2.44 ± 0.7352.70 ± 1.2390.256(z=−1.136) stiffness1.56 ± 0.9091.86 ± 0.9720.108 (z=−1.606) function11.50 ± 2.81312.42 ± 4.3260.260(t=−1.134) Postoperative1y16.78 ± 3.68117.98 ± 3.8380.138 (t= −1.498) pain2.11 ± 1.2372.04 ± 1.3880.809 (z=−0.242) stiffness1.56 ± 1.1321.42 ± 1.1790.593 (z=−0.535) function13.11 ± 3.34514.53 ± 3.3280.045* (z=−2.008) Postoperative2y15.08 ± 5.69416.56 ± 5.4870.216(t= −1.247) pain1.44 ± 0.9981.67 ± 1.0580.324 (z=−0.987) stiffness1.50 ± 1.2071.46 ± 1.1190.861 (z=−0.175) function12.139 ± 5.81213.44 ± 5.5070.260 (z=−1.126) Table 7Comparison of postoperative knee joint function between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SD p WOMAC Preoperative51.86 ± 13.26050.61 ± 12.9130.655 (t = 0.449) pain12.81 ± 2.70313.79 ± 2.5620. 081 (t=−1.766) stiffness2.97 ± 1.4643.16 ± 1.5670.615(z=−0.503) function36.08 ± 13.52533.67 ± 12.7470.387 (t = 0.870) Postoperative1 m27.28 ± 6.10231.61 ± 7.5990.005* (t=−2.885) pain5.97 ± 2.7107.09 ± 2.4300.026* (z=−2.223) stiffness2.430 ± 1.1953.44 ± 1.7730.028* (z=−2.195) function18.64 ± 5.33021.09 ± 6.7090.122 (z=−1.546) Postoperative3 m18.19 ± 5.61520.39 ± 5.9090.062(z=−1.866) pain3.11 ± 1.9093.39 ± 1.7500.301 (z=−1.035) stiffness1.64 ± 1.0182.93 ± 1.163< 0.001* (−4.768) function13.44 ± 5.42714.07 ± 5.8790.489 (z=−0.692) Postoperative6 m15.50 ± 2.83316.98 ± 4.8460.066(t=−1.860) pain2.44 ± 0.7352.70 ± 1.2390.256(z=−1.136) stiffness1.56 ± 0.9091.86 ± 0.9720.108 (z=−1.606) function11.50 ± 2.81312.42 ± 4.3260.260(t=−1.134) Postoperative1y16.78 ± 3.68117.98 ± 3.8380.138 (t= −1.498) pain2.11 ± 1.2372.04 ± 1.3880.809 (z=−0.242) stiffness1.56 ± 1.1321.42 ± 1.1790.593 (z=−0.535) function13.11 ± 3.34514.53 ± 3.3280.045* (z=−2.008) Postoperative2y15.08 ± 5.69416.56 ± 5.4870.216(t= −1.247) pain1.44 ± 0.9981.67 ± 1.0580.324 (z=−0.987) stiffness1.50 ± 1.2071.46 ± 1.1190.861 (z=−0.175) function12.139 ± 5.81213.44 ± 5.5070.260 (z=−1.126) Comparison of postoperative knee joint function between the two groups of patients Postoperative 1 m Postoperative 3 m Postoperative 6 m Postoperative 1y Postoperative 2y Radiographic findings The statistical results revealed that measurements were taken for preoperative patellar tendon length, patella length, the Insall–Salvati index, the posterior tibial slope (PTS), and the hip–knee–ankle angle (HKA) (Fig. 5A; Table 8). Additionally, the lateral distal femoral angle (LDFA), medial proximal tibial angle (MPTA), posterior condylar offset (PCO) (Fig. 5B), posterior condylar offset ratio (PCOR), postoperative PCO, PCOR, femoral coronal alignment angle (FCA), femoral sagittal alignment angle (FSA), tibial coronal alignment angle (TCA), and tibial sagittal alignment angle (TSA) exhibited no statistically significant differences (Fig. 5C, D; Table 9). Table 8Comparison of radiographic outcomes between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SD p Preoperative patellar tendon length54.8144 ± 9.5061055.1025 ± 10.978200.810(z=−0.241) length of the patella43.5164 ± 5.8965844.2114 ± 4.728370.381(z=−0.876) Insall-Salvati index1.28 ± 0.281.26 ± 0.280.850(z=−0.189) posterior tibial slope (PTS)8.04 ± 3.559.31 ± 2.980.067(t=−1.856) hip-knee-ankle angle (HKA)173.5342 ± 7.66303171.7886 ± 7.651830.287 (t = 1.071) the lateral distal femoral angle (LDFA)88.6719 ± 3.2015888.9309 ± 3.522190.722(t=−0.357) the medial proximal tibial angle (MPTA)86.4150 ± 3.7480186.5758 ± 3.141440.824(t=−0.223)preoperative posterior condylar offset, (PCO)32.5731 ± 4.8991033.4914 ± 5.573050.420(t=−0.810) posterior condylar offset ratio (PCOR)0.53 ± 0.0590.53 ± 0.0650.801(z=−0.252)Postoperative posterior condylar offset, (PCO)32.6597 ± 4.7783934.8119 ± 5.344500.052(t=−1.969) posterior condylar offset ratio (PCOR)0.52 ± 0.0480.54 ± 0.0680.158(z=−1.412)Postoperative femoral coronal alignment angle (FCA)95.42 ± 1.07995.65 ± 1.0770.245(z=−1.163) Femoral sagittal alignment angle (FSA)1.14 ± 1.6760.77 ± 0.7800.308(z=−1.019) tibial coronal alignment angle (TCA)88.67 ± 1.37388.86 ± 1.2020.499(z=−0.676) tibial sagittal alignment angle (TSA)88.11 ± 1.43088.04 ± 1.3750.759(z=−0.307) Table 8Comparison of radiographic outcomes between the two groups of patientsMini-modified subvastusMedial parapatellarVariablesMean ± SDMean ± SD p Preoperative patellar tendon length54.8144 ± 9.5061055.1025 ± 10.978200.810(z=−0.241) length of the patella43.5164 ± 5.8965844.2114 ± 4.728370.381(z=−0.876) Insall-Salvati index1.28 ± 0.281.26 ± 0.280.850(z=−0.189) posterior tibial slope (PTS)8.04 ± 3.559.31 ± 2.980.067(t=−1.856) hip-knee-ankle angle (HKA)173.5342 ± 7.66303171.7886 ± 7.651830.287 (t = 1.071) the lateral distal femoral angle (LDFA)88.6719 ± 3.2015888.9309 ± 3.522190.722(t=−0.357) the medial proximal tibial angle (MPTA)86.4150 ± 3.7480186.5758 ± 3.141440.824(t=−0.223)preoperative posterior condylar offset, (PCO)32.5731 ± 4.8991033.4914 ± 5.573050.420(t=−0.810) posterior condylar offset ratio (PCOR)0.53 ± 0.0590.53 ± 0.0650.801(z=−0.252)Postoperative posterior condylar offset, (PCO)32.6597 ± 4.7783934.8119 ± 5.344500.052(t=−1.969) posterior condylar offset ratio (PCOR)0.52 ± 0.0480.54 ± 0.0680.158(z=−1.412)Postoperative femoral coronal alignment angle (FCA)95.42 ± 1.07995.65 ± 1.0770.245(z=−1.163) Femoral sagittal alignment angle (FSA)1.14 ± 1.6760.77 ± 0.7800.308(z=−1.019) tibial coronal alignment angle (TCA)88.67 ± 1.37388.86 ± 1.2020.499(z=−0.676) tibial sagittal alignment angle (TSA)88.11 ± 1.43088.04 ± 1.3750.759(z=−0.307) Comparison of radiographic outcomes between the two groups of patients Subgroup analysis was conducted for the modified minimally invasive subvastus approach. The results of the subgroup analysis revealed a statistically significant difference in the FSA between patients with patella alta and patella baja, although the mean value fell within the normal range. Table 9Comparison of postoperative prosthesis-related angles among patients with patella alta, normal patella, and patella Baja in the modified subvastus approach groupMini-modified subvastusVariablesPatella Alta(IS > 1.2) (n = 26)Normal patella (1.2 > IS > 0.8) (n = 5)Patella Baja (IS < 0.8)(n = 5) p PostoperativeANOVA femoral coronal alignment angle (FCA)95.46 ± 1.17495.40 ± 0.89495.20 ± 0.8370.889 femoral sagittal alignment angle (FSA)0.81 ± 0.7491.20 ± 0.4472.80 ± 4.0870.047* tibial coronal alignment angle (TCA)88.69 ± 1.49088.60 ± 1.14088.60 ± 1.1400.985 tibial sagittal alignment angle (TSA)87.92 ± 1.41288.80 ± 1.30488.40 ± 1.670.415 Table 9Comparison of postoperative prosthesis-related angles among patients with patella alta, normal patella, and patella Baja in the modified subvastus approach groupMini-modified subvastusVariablesPatella Alta(IS > 1.2) (n = 26)Normal patella (1.2 > IS > 0.8) (n = 5)Patella Baja (IS < 0.8)(n = 5) p PostoperativeANOVA femoral coronal alignment angle (FCA)95.46 ± 1.17495.40 ± 0.89495.20 ± 0.8370.889 femoral sagittal alignment angle (FSA)0.81 ± 0.7491.20 ± 0.4472.80 ± 4.0870.047* tibial coronal alignment angle (TCA)88.69 ± 1.49088.60 ± 1.14088.60 ± 1.1400.985 tibial sagittal alignment angle (TSA)87.92 ± 1.41288.80 ± 1.30488.40 ± 1.670.415 Comparison of postoperative prosthesis-related angles among patients with patella alta, normal patella, and patella Baja in the modified subvastus approach group Subgroup analysis was conducted for the modified minimally invasive subvastus approach. The results revealed no significant differences in the mean values of FCA, FSA, TCA, and TSA. Table 10Comparison of postoperative prosthesis-related angles in patients with varus, valgus, and normal knees in the modified subvastus approach groupMini-modified subvastusVariablesgenu varum(n = 27)Normal patella(n = 6) p PostoperativeANOVA femoral coronal alignment angle (FCA)95.41 ± 1.11895.67 ± 1.2110.692 femoral sagittal alignment angle (FSA)1.26 ± 1.8730.67 ± 0.8160.739 tibial coronal alignment angle (TCA)88.70 ± 1.38288.67 ± 1.6330.911 tibial sagittal alignment angle (TSA)88.41 ± 0.63287.00 ± 0.6320.076 Table 10Comparison of postoperative prosthesis-related angles in patients with varus, valgus, and normal knees in the modified subvastus approach groupMini-modified subvastusVariablesgenu varum(n = 27)Normal patella(n = 6) p PostoperativeANOVA femoral coronal alignment angle (FCA)95.41 ± 1.11895.67 ± 1.2110.692 femoral sagittal alignment angle (FSA)1.26 ± 1.8730.67 ± 0.8160.739 tibial coronal alignment angle (TCA)88.70 ± 1.38288.67 ± 1.6330.911 tibial sagittal alignment angle (TSA)88.41 ± 0.63287.00 ± 0.6320.076 Comparison of postoperative prosthesis-related angles in patients with varus, valgus, and normal knees in the modified subvastus approach group Fig. 5A Schematic diagram of the hip–knee–ankle angle. B Schematic diagram of patellar tendon length, length of the patella, and posterior condylar offset. C Schematic diagram of the femoral coronal alignment angle and tibial coronal alignment angle. D Schematic diagram of the femoral sagittal alignment angle and tibial sagittal alignment angle A Schematic diagram of the hip–knee–ankle angle. B Schematic diagram of patellar tendon length, length of the patella, and posterior condylar offset. C Schematic diagram of the femoral coronal alignment angle and tibial coronal alignment angle. D Schematic diagram of the femoral sagittal alignment angle and tibial sagittal alignment angle Discussion The modified minimally invasive subvastus approach plays an important role in blood conservation during the perioperative period The theoretical red blood cell loss (TRBCL) is considered a crucial indicator for measuring surgical bleeding, and its calculation formula is as follows: TRBCL = PBV (preoperative total blood volume) × (preoperative haematocrit - postoperative haematocrit). The PBV calculation method is as follows: PBV = K1 × height (m)^3 + K2 × body mass (kg) + K3. For males, K1 = 0.3669, K2 = 0.03219, and K3 = 0.6041; for females, K1 = 0.3561, K2 = 0.03308, and K3 = 0.1833 [13]. Compared with control patients, patients who underwent the modified minimally invasive subvastus approach presented significantly lower TRBCL, decreased hemoglobin changes, and smaller incisions. These findings indicate a significant reduction in perioperative blood loss and adherence to the international standard for minimally invasive incision length (< 14 cm) [12]. The reduction in blood loss during the modified minimally invasive subvastus approach compared with the traditional MPA is attributed mainly to the specific anatomical approach employed in the procedure (Fig. 6A). The descending geniculate artery (DGA) typically has three branches that serve to nourish the knee joint. In a cadaver dissection study, DGA was found to be present in 89% of the samples. In the remaining 11%, the DGA was absent, and the main trunk of the artery originated separately from the femoral artery, forming the joint branch, the saphenous branch, and the muscular branch [14] (classified as Type 3 according to Dubois et al. [15]). Compared with the traditional approach, the following are the reasons why our modified approach reduces bleeding. First, it reduces interference with the quadriceps extensor apparatus, avoiding dissection and disruption of the rectus femoris and the vastus intermedius. Second, it uses small incisions, which minimize interference with soft tissues and reduce the risk of damaging the skin, fat layer vessels, and nerves [16]. Third, it involves anatomical structures; the traditional approach cuts the distal end of the rectus femoris and the superior end of the patellar tendon, whereas our modified approach, which is based on the medial muscle-splitting technique, starts from the patellar midpole and adds a small oblique incision (1–2 cm) without cutting the distal end of the rectus femoris or the superior end of the patellar tendon. In general, the fewer disturbances to the muscle, the less blood loss. Typically, the MPA divides the DGA, the superomedial genicular artery (SMG), the inferomedial genicular artery (IMG), and the inferolateral genicular artery (ILG) while leaving the superolateral genicular artery (SLG) patent (Fig. 6) [16]. In contrast, the subvastus approach retains the patency of both the DGA and SMG [17]. Finally, this approach does not require dislocation of the patella, thus reducing unnecessary dissection of the lateral retinaculum and other structures and minimizing interference with the retinaculum and its insertion points [18]. Fewer blood vessels are severed, resulting in a corresponding reduction in patient bleeding (Fig. 6B). In terms of operation time, the operation time of the experimental group was significantly longer than that of the control group. Previous studies have shown that the subvastus approach may affect surgical observation and operation because it maintains the integrity of the vastus medialis tendon. For example, the patella cannot be inverted when the patella is cleaned [6], the operation should be careful not to damage the adjacent muscles and affect the measurement of the osteotomy position when performing femoral osteotomy, and tibial osteotomy should be more careful to prevent damage to the knee collateral ligament. This results in prolonged operation time. Although the surgical method of the experimental group in our study was the modified minimally invasive subvastus approach, which was more fully exposed than the unmodified subvastus approach was, the exposure to the field of view was inferior to that of the traditional MPA [19] because of the relatively minimally invasive incision, which could be solved by the symbiotic use of retractors [20]. However, it still has an impact on the extension of the operation time, thereby causing the extension of the operation time to a certain extent. All surgeries were performed by a skilled surgeon for all patients. The surgeon at our center has used the MPA for a long time, and the operation time is relatively fast. The modified minimally invasive subvastus approach, in general, may result in a prolonged operation time and requires technical mastery that adheres to a specific learning curve; however, the experience gained can further reduce the duration of the procedure. Fig. 6A Injury to the vessels caused by the medial parapatellar approach. B Comparison of vascular injuries caused by the medial parapatellar approach and the modified subvastus approach A Injury to the vessels caused by the medial parapatellar approach. B Comparison of vascular injuries caused by the medial parapatellar approach and the modified subvastus approach The modified minimally invasive subvastus approach can more effectively alleviate patient pain during the early postoperative period Postoperative pain after TKA is an important factor affecting early postoperative functional exercise. Many factors, such as the surgical approach and incision size, affect postoperative pain. According to previous studies, the subvastus approach is better than the traditional approach during the early postoperative period [21]. Our study revealed that our modified minimally invasive subvastus approach can effectively reduce short-term postoperative pain after total knee replacement but has no significant effect on long-term pain relief. Specifically, for the VAS score, the improved group performed better on the third postoperative day; for the pain score in the WOMAC score, the improved group performed better in the first postoperative month. However, neither of the two pain assessment indicators showed significant differences in long-term follow-up. From an anatomical perspective, the main difference between the modified minimally invasive approach and the traditional medical approach is the reduction of the incision on the quadriceps tendon and the rectus femoris muscle. Additionally, the incision is smaller in line with the principles of minimally invasive surgery [22]. Compared with traditional approaches, our modified minimally invasive approach primarily enhances the protection of the quadriceps tendon superior to the patella while reducing injury to the corresponding tendons of the rectus femoris and vastus intermedius. Our improved minimally invasive approach includes a curved incision along the lateral aspect of the knee, which provides better exposure to the surgical field and reduces the risk of scar tissue formation and pain associated with extensive soft tissue release around the knee. As one of the most commonly used nerve block methods, continuous femoral nerve block (CFNB) is considered the preferred analgesic after TKA. However, it may weaken the strength of the quadriceps muscles, thereby increasing the risk of falls [23]. The adductor canal block (ACB) is a new sensory block technique that can effectively relieve postoperative pain while maintaining quadriceps strength [24]. Ultrasound-guided adductor canal block (High-ACB) is an effective option for pain control in TKA, but it is typically performed as a standalone perioperative procedure under ultrasound guidance, which can increase the cost of TKA by up to $2,000 [25, 26]. Therefore, an intraoperative adductor canal block (Low-ACB) performed by the operating surgeon has been described as an alternative method [27]. In terms of time efficiency, performing the block intraoperatively has the potential to substantially reduce the operating room turnaround time in institutions where the block was previously administered by anesthesia in the operating room [26]. In addition, compared with low-ACB, which is performed by the operating surgeon, high-ACB requires more operating time and has difficulty controlling accuracy [28]. Additionally, in terms of safety, there may be more complications, such as vascular injury, nerve injury, and catheter site infections [29]. The elimination of the ultrasound requirement and the reduced surgical costs make ACB under direct vision a more cost-effective option than ACB with high ultrasound guidance. This approach requires fewer specialized skills and equipment, enabling it to be performed by the operator alone without involving an anesthesiologist [26]. We found that the saphenous nerve can be easily exposed beyond the adductor canal via our modified minimally invasive subvastus approach. The nerve can be located nearly 4 cm above the condyle line in the deep fascia of the muscle, where it can be directly visualized and blocked to achieve complete anesthesia [30]. This method is simple, intuitive, safe, and effective and does not require extending the incision. The drug can reach its target directly, ensuring 100% accuracy and effectiveness in administration. Moreover, research indicates that the anatomical course of the saphenous nerve after exiting the adductor canal is relatively constant, consistently running between the sartorius muscle and the femoral artery [31]. This consistent anatomical position allows us to perform an adductor canal block during surgery, ensuring that our operations are repeatable. Our modification has significantly enhanced our daily clinical practice and provided clear benefits to patients. Compared with the conventional method, the modified minimally invasive subvastus approach results in superior preservation of quadriceps strength during the early postoperative phase Given that direct measurement of quadriceps strength via a dynamometer, while highly accurate, may cause significant discomfort to the patient [32], we indirectly assessed the recovery of quadriceps strength by measuring the angle of the straight leg raise. In this study, the modified approach group exhibited superior recovery of quadriceps muscle strength compared with the traditional approach group, particularly on the first and third postoperative days, with a significantly greater straight leg raise angle than the traditional approach group. On the one hand, previous studies comparing patients treated with the minimally invasive subvastus approach with those treated with the traditional parapatellar approach have also suggested that the subvastus technique results in faster recovery of quadriceps function than does the MPA [33]. However, Young Bok Jung et al. [34] modified the subvastus approach and compared it with the traditional approach to evaluate the time of postoperative active SLR and reported that patients with their modified approach were able to perform active SLR earlier than those with the medial patellofemoral approach (an average of 2.2 days vs. an average of 0.5 days). Although we modified the minimally invasive subvastus approach by making a curved auxiliary incision at the medial-superior border of the patella, we still observed statistically faster muscle strength recovery with this approach than with the traditional parapatellar approach, indicating that our modified minimally invasive subvastus approach can better maintain quadriceps muscle strength in the short term after TKA and has no significant effect on long-term results. Regarding the surgical approach, the MPA separates the medial 1/3 of the quadriceps tendon from the medial upper edge of the patella and requires the patella to be turned over (Fig. 3D) during the operation [35]. Compared with the MPA, the modified minimally invasive approach primarily increases protection of the quadriceps tendon at the superior part of the patella. It reduces injury to the corresponding tendons of the rectus femoris and vastus intermedius. Reducing muscle interference enhances the protection of nerves running within the muscle (Fig. 7), thereby indirectly protecting the nerve supply structures [36]. Consequently, this represents an additional factor contributing to enhanced muscular strength. Our modified approach is more anatomy-focused than other modified approaches are, as it cuts the vastus medialis muscle at a 50-degree [10] angle to the femoral axis, which better aligns with the course of the quadriceps fibers and reduces interference with the quadriceps tendon and muscle. We did not use lateral cuts blindly, as this would sever a large number of muscle fibers, and we did not need to cause excessive interference at the superior pole of the patella. Specifically, we inserted the knife along the direction of the patellar insertion into the quadriceps, cut at a 50-degree angle down to the middle of the patella, and then added a 2-cm curved incision along the inner side of the patella to the superior pole, thus increasing surgical exposure and reducing interference with the vastus medialis obliquus and quadriceps tendon. Like traditional subvastus procedures, we do not need to turn the patella over during surgery [4]. Studies have shown that continuous turnover of the patella for more than 1 h can easily cause permanent damage to the quadriceps femoris [37]. In addition, Yoshioka et al. [38] reported a 62% reduction in quadriceps muscle strength 1 month after TKA via a standard MPA. Fig. 7A Peripheral nerves around the knee. B Comparison of nerve injuries caused by the medial parapatellar approach and the modified subvastus approach A Peripheral nerves around the knee. B Comparison of nerve injuries caused by the medial parapatellar approach and the modified subvastus approach Compared with the MPA, our modified approach allows for direct visualization of the ACB. The target is the sensory branch of the saphenous nerve, while the motor branch is avoided as much as possible [39]. This further reduces soft tissue interference during postoperative ultrasound-guided blocks and achieves precise anesthesia and analgesia with minimal impact on quadriceps muscle strength. This significantly enhances daily clinical practice and provides direct benefits to patients. A nerve block is generally retained for 2‒3 days after the operation [40]. During this period, the adductor canal block has little effect on quadriceps strength or the large straight leg raising angle. After the nerve block was discontinued, there was no significant difference in the angle of straight leg raising between the two groups in terms of the effects of analgesic drugs or the reduction in the impact of the femoral nerve block on quadriceps strength. In summary, our modified minimally invasive subvastus approach can maintain better quadriceps strength in the early postoperative period and does not affect the long-term results. Better muscle strength helps patients perform limb activities earlier, leading to higher levels of satisfaction. This finding is in line with the concept of enhanced recovery after surgery [8]. The modified minimally invasive subvastus approach can accelerate recovery of the knee joint in the early postoperative period Our study also revealed that patients who underwent TKA through the modified approach were able to walk and climb stairs earlier, with or without assistance, than those who received TKA through the MPA. Previous studies have shown that the ROM of the subvastus approach is superior to that of the MPA in the short term after TKA [41]. At the same time. While some studies have shown that there is no difference in short-term or midterm functional outcomes, the subvastus approach seems to increase the ROM at 6 months after surgery [42]. To distinguish the long-term and short-term clinical effects of ROM, we measured ROM at 3 days, 7 days, and 2 years after surgery. We found that there was no significant difference in knee ROM between the experimental group and the control group at present (P > 0.05). However, the knee ROM in the experimental group was significantly greater than that in the control group on days 3 and 7 postoperatively. These findings suggest that our modified minimally invasive approach provides better ROM in the early postoperative period without compromising long-term clinical outcomes. Compared with the MPA, our modified minimally invasive subvastus approach results in better flexion and extension angles in the short term, which is in line with the concept of ERAS [8]. Moreover, early mobilization may reduce the risk of long-term immobilization-related complications and improve cardiopulmonary function. Notably, regardless of the surgical approach, CPM was administered on the second postoperative day in both groups. CPM has positive biological effects on tissue healing, edema, and hematoma resolution [43]; although most studies indicate that CPM may enhance the slight flexion ability of the knee joint and improve patients’ quality of life, its therapeutic effect is not substantial [44]. Nevertheless, it remains widely used because of patient subjective factors and rehabilitation expectations [12]. According to the Cochrane review by Harvey et al. [44], CPM has short-term minor effects on the knee joint active and passive flexion range of motion, specifically 2 degrees and 3 degrees, respectively. A good ROM may be partially attributed to the prevention of joint adhesion to a very slight degree. However, in this trial, CPM was used for both the modified approach group and the traditional MPA group to control for these variables. Compared with that of the traditional MPA, the postoperative ROM of the modified approach group was significantly greater than that of the control group. Therefore, we can still draw this conclusion. The OKS and WOMAC are widely used in TKR research, supported by evidence of their reliability, structural validity, and lack of bias [45]. Our study compared the clinical efficacy of the modified minimally invasive subvastus approach with that of the MPA for TKA in knee osteoarthritis patients via the validated OKS and WOMAC scores. Previous studies have indicated that the mean OKS was significantly greater in patients receiving the subvastus approach than in patients receiving MPA at the 3- and 6-month follow-ups (P < 0.05), but there was no significant difference at the 12-month follow-up [46]. To provide a more comprehensive evaluation of the clinical outcome scores, we measured the OKS and WOMAC scores at 1 month, 3 months, 6 months, 12 months, and 24 months postoperatively. In the modified approach group, there was a statistically significant difference in the OKS scores at 1 and 3 months, but there was no statistically significant difference at 6, 12, and 24 months. These results suggest that our modified minimally invasive subvastus approach is superior to the traditional MPA in terms of short-term outcomes. The WOMAC score is primarily used to evaluate postoperative knee rehabilitation, particularly after TKA, including pain, stiffness, and function [47]. Studies have reported that quadriceps tendon preservation reduces postoperative pain for up to 1 month, which is the time required for tendon healing; however, there is no difference between the two approaches at 3 months after surgery [48]. In our study, the pain and stiffness in the first month after the operation were significantly greater in the experimental group than in the control group. There was no significant difference in the total WOMAC score between the experimental group and the control group at the third month; however, there was a significant difference in the stiffness subscale score of the WOMAC at the third month, which may be related to the formation of a tendon scar following tendon injury caused by the MPA. In addition, our modified curved incision can accelerate the early rehabilitation of knee patients compared with the MPA and has no significant effect on long-term results. Combined with the enhanced recovery after the surgery program of TKA [8], this study can help patients exercise for a short period in the early perioperative period to obtain better rehabilitation results and further improve daily clinical practices. The modified minimally invasive subvastus approach accelerates the recovery of knee joint activity and function in the early stages after TKA, aligns with the international concept of enhanced recovery after surgery [14], and does not affect the long-term outcomes of TKA. The modified minimally invasive subvastus approach does not affect the accuracy of prosthesis placement One of the most significant findings of the study is that patients who undergo TKA through modified approaches do not experience a greater risk of implant malposition or limb misalignment. With the maturation of traditional TKA techniques, orthopedic surgeons have increasingly shifted their focus to minimally invasive TKA techniques, which provide reduced postoperative pain and faster recovery. However, the smaller incision size poses a challenge in achieving adequate exposure to the surgical field [49]. In general, the FCA, FSA, TCA, and TSA [50] reflect the evaluation of the placement of the knee prosthesis in patients, except for the rotation angle. In this trial, there were no significant differences in any of these angles, indicating that the modified minimally invasive subvastus approach had no significant effect on the position or angle of the prosthesis (Tables 9 and 10). Previously, Pan et al. conducted a prospective randomized controlled study involving 35 patients who underwent minimally invasive subvastus approach TKA with a minimum follow-up of two years [51]. Their study indicated that the minimally invasive subvastus approach may not ensure optimal prosthetic positioning, potentially increasing the risk of associated complications. We modified the minimally invasive subvastus approach by creating an auxiliary incision at the medial-superior margin of the patella, thereby enhancing the surgical field of view. Therefore, precise prosthesis positioning and optimal prosthesis placement are needed. Some modified approaches have difficulty exposing the lateral support ligament and fail to fully expose the posterior lateral capsule, leading to incomplete removal of cement residue [52]. Our modified approach expanded the exposure area, allowing the insertion of a nerve retractor after prosthesis implantation and effectively removing the cement residue. There was no residual bone cement detected on the postoperative DR images. The patellar position of the knee joint was classified according to the Insall-Salvati index [53] as patellar alta, patella baja, or normal patella [54]. Subgroup analysis was conducted for the modified minimally invasive subvastus approach. The results of the subgroup analysis revealed a statistically significant difference in the FSA between patients with patella alta and patella baja, although the mean value fell within the normal range. Conversely, no statistically significant difference was observed among the other groups. Aglietti P et al. believe that the mini-subvastus approach is not suitable for patients with low-positioned patella [52]. Jung YB et al. modified the traditional subvastus approach, but after discussion, it was concluded that the approach should be avoided for patients with low-positioned patella [34]. Our findings suggest that our modified approach is suitable for patients whose patella baja leads to a relatively greater FSA. Overall, our modification has significantly enhanced daily clinical practice. The hip-knee-ankle angle at the patellar position of the knee joint was categorized into varus (HKA < 177), valgus (> 183), and normal knees [55]. Subgroup analysis was conducted for the minimally invasive approach. The results of the subgroup analysis revealed no significant differences among the groups, with all values falling within the normal range. These findings suggest that our modified minimally invasive subvastus approach, which involves making an additional incision from the midpoint of the patella along its medial side up to the superior margin, can enhance the surgical exposure area, simplify the osteotomy procedure, and reduce the risk of implant placement. Limitations Several limitations of this study may impact the generalizability of the findings. One limitation pertains to the retrospective design, which could have introduced bias and constrained the ability to establish causality. As outlined previously, the reliability and validity of the assessment indicators have been guaranteed through rigorous methodological measures. Another limitation is associated with the relatively small sample size, potentially affecting the statistical power and limiting the generalizability of the findings to other populations. Additionally, as this study was conducted in a single center, its applicability to other settings may be restricted. Therefore, an ongoing related multicenter prospective clinical study aims to further promote the adoption of the modified approach. Looking ahead, we plan to include cruciate retaining (CR) prostheses in our future research to expand the scope of our study and to contribute to the growing body of evidence comparing CR and PS prostheses in total knee arthroplasty. Conclusion In this trial, the implementation of the modified minimally invasive subvastus approach exhibited a notable perioperative blood-conserving effect for patients undergoing TKA. This technique can effectively relieve initial pain, maintain optimal quadriceps femoris muscle strength, and accelerate knee joint mobility and functional rehabilitation without compromising long-term recovery or causing patient discomfort. Despite potentially prolonged operative time due to increased technical complexity, it does not negatively affect prosthesis placement while offering significant advantages to patients. Supplementary Information Supplementary Material 1 Supplementary Material 1