From March 2015 to September 2019, 527 consecutive patients were selected and examined in our department were screened for research. The institutional review boards of the participating hospitals reviewed the study and approved the protocol. Informed consent was obtained from each patient. Our criteria for eligibility included 1) adult patients; 2) severe (stage 4 osteoarthritis ) knee joint osteoarthritis confirmed on X-ray, CT, or MRI; 3) needed a primary knee replacement surgery or knee arthroplasty; 4) a fixed valgus knee > 10°; 5) and unilateral or bilateral involvement. Exclusion criteria were 1) stage 0 to 3 osteoarthritis; 2) revision knee arthroplasty; 3) active local or widespread infection; 4) medical conditions that put the patient at high risks of complications; 5) discontinued intervention; 6) severe disease affecting peripheral blood vessels or nerves; and 7) lost to follow-up.
A total of 123 patients (139 knees) were randomly and blindly allocated into experimental group (60 patients; 68 knees) and control group (57 patients; 65 knees) using a computational pseudorandom number generator (Fig. 1). One surgeon who did not attended the assessments and treatments generated the random allocation. All operations were performed by the same surgical team. Implants were selected preoperatively based on radiological and clinical evaluations, but the final decision was made after osteotomy and soft tissue balancing.
Step 1 in both groups
As described by Sheth et al , operations were performed through the anteriomedian incision of the knee and the lateral para-condylar approach. We exposed the medial aspect of the knee by subperiosteally elevating the anteromedial capsule. The anterior and posterior cruciate ligaments, residual meniscus and osteophytes were removed, and the medial collateral ligament of deep layer was released. Using the measured resection technique, we performed the osteotomy followed by the sequence of tibia osteotomy and femur osteotomy. We performed the tibial osteotomy with extramedullary alignment technique, and the osteotomy plane was perpendicular to the axis of tibia. The posterior corner osteotomy was set at 7°, and the thickness was 10 mm. Distal femoral valgus osteotomy for valgus deformity of the knee was set at 6°. The anterior and posterior femoral osteotomies were positioned using the anterior reference method. The size of the prosthesis was determined, and the reference was 3° of external rotation to determine the external femoral osteotomy using the posterior condylar line. After osteotomy, we used a gap module to test the tension and balance of the medial and lateral gaps during extension and flexion of the knee.
Step 2 in experimental group (figure-of-four position)
We took the ankle and rest it across the contralateral leg with the knee bent to 90° in the figure-of- four position (Fig. 2). In order to visualize the posteromedial corner structures, we dislocated the knee anteriorly using a Hoffman hook placed behind the tibial plateau. We subperiosteally released the structures (posterior half of the superficial medial collateral ligament, semi-membrane tendon insertion, and posteromedial joint capsule) of the posteromedial corner.
Step 2 in control group (conventional full flexion)
As described by Tang et al , we fully flexed the knee to subluxate the joint forward and outward using a Hoffman hook. The tibial plateau and posteromedial corner structures were completely exposed, and subperiosteal release was performed as that in experimental group.
Step 3 in both groups
If soft tissue balance was not completely achieved or the medial gap was still tight, additional loosening techniques, such as pie-crusting technique, tibial stripping of the superficial medial collateral ligament, and medial femoral epicondyle up-sliding osteotomy, were used to achieve symmetric medial and lateral spaces . The femoral component, tibial component, patellar component, and plastic spacer (German Link Inc) were implanted. The wound was closed in layers. After surgery, the routine therapies and rehabilitation program were used in both groups. The patients were followed up 3 months, 6 months, 1 year, and 5 years after surgery .
Assessments were performed by a senior orthopedic surgeon who did not attend the treatments. Varus knee deformity was measured on the frontal X-ray (varus/valgus position) as the angle formed from the intersecting femoral and tibial mechanical axes . The femoral axis was deﬁned as a line drawn from the center of the femoral head to the center of the femoral intercondylar notch. Time for balancing the medial soft tissue was defined as the time from the start of spacer test to the end of balance test. Length of release was defined as the distance from the osteotomy surface of the tibial plateau to the farthest structures released. The rating system of Hospital for Special Surgery (HSS) knee score was used to evaluate the clinical results of TKA .
Quantitative variables were described as mean and standard deviation for symmetric distribution or median and interquartile range for asymmetric distribution. We used the one-way analysis of variance to determine whether there were any significant differences between the groups. Differences were considered statistically significant at p< 0.05. The collected data were analyzed with SPSS Version 11.0 (SPSS, Inc., Chicago, Ill.).