BACKGROUND: Prosthesis restricted selection, soft tissue release, patellar trajectory recovery, and bone defect reconstruction were need to be optimized in total knee arthroplasty for moderate and severe valgus knee. The medial parapatellar approach has disadvantages in the treatment of valgus knee, such as aggravating the medial soft tissue relaxation. In recent years, it has been found that the lateral parapatellar approach has advantages in the treatment of valgus knee, such as exposure and release. OBJECTIVE: To observe the efficacy of the lateral and medial parapatellar approach in total knee arthroplasty for moderate and severe valgus knee, and to explore a more suitable surgical approach for moderate and severe valgus knee. METHODS: Totally 56 patients with moderate and severe valgus knee underwent total knee arthroplasty and would take turns performing surgery through the medial and lateral parapatellar approach according to the order of admission. The lateral group (n=28) underwent total knee arthroplasty through lateral parapatellar approach, and the medial group (n=28) through medial parapatellar approach. Posterior stablized knee prosthesis was used in all patients. The restricted types of prosthesis, thickness of polyethylene, operation time, amount of blood loss, femoro-tibia angle, patellar tilt angle, range of motion, Hospital for Special Surgery score, Western Ontario and McMaster Universities Arthritis Index (WOMAC) score, and complications were collected as observation indexes for statistical analysis. RESULTS AND CONCLUSION: (1) The utilization rate of condyle-restricted prosthesis in the lateral group was significantly lower than that in the medial group. The average thickness of polyethylene liner was lower, and the operation time was shorter in the lateral group compared with the medial group. There were significant differences between the two groups (P < 0.05), and there was no significant difference in the amount of surgical bleeding. (2) 56 patients were followed up for an average of 2-5 years. There were no signs of prosthesis loosening or bone resorption in all patients. The average patellar tilt angle and femoral tibial angle of the lateral group were lower than those of the medial group, and the average Hospital for Special Surgery score and WOMAC score of the lateral group were higher than those of the medial group, with significant differences (P < 0.05). There was no significant difference in the range of motion of the knees between the two groups. (3) Incision fat liquefaction and calf intermuscular venous thrombosis occurred in one case in each group. No infection, poor incision healing, incisional hematoma, prosthesis dislocation, iatrogenic nerve injury, ectopic ossification, or periprosthesis fracture occurred during follow-up. (4) In conclusion, the treatment of moderate and severe valgus knee by lateral parapatellar approach can better protect the tension of the medial soft tissue of the knee, use less condylar restrictive prostheses, and have a more friendly patellar trajectory and higher postoperative function score. [ABSTRACT FROM AUTHOR]