Spacer-based gap balancing is useful in total knee arthroplasty: a 3-year follow-up of a retrospective study

This study was approved by the institutional ethics committee, and each participant signed an informed consent form.

Between May 2016 and August 2018, 124 consecutive patients underwent primary TKA with one of two surgical techniques: the GB technique assisted by a modified spacer block (n = 61) or the traditional MR technique (n = 63). The inclusion criteria were patients who were 50–75 years old with primary knee osteoarthritis with knee varus, who had a poor response to conservative treatment and a severely impaired ability to perform daily activities, and who had participated in a systematic follow-up for at least 3 years. All patients underwent a preoperative physical examination (to determine the collateral ligament integrity through varus and valgus stress testing with the knee at 20° of flexion) and radiographic evaluation (standing full-length anteroposterior and knee lateral X-rays). Patients who did not want to participate in the study or who had collateral ligament dysfunction and knee varus > 20° were excluded. There were no significant differences in baseline patient demographics between the two groups in terms of age, sex, side of surgery or preoperative BMI (Table 1).

All surgeries were performed by the same senior knee arthroplasty surgeon using a posterior-stabilized TKA prosthesis (XN, Chunlizhengda Medical Instruments Co. Ltd., Beijing, China). Both groups used a median anterior knee incision with a medial parapatella approach.

After the knee was exposed, the osteophytes were removed, and the same intramedullary and extramedullary guidance systems were used to perform distal femoral and proximal tibial resections. Then, the GB and MR techniques were used for the femoral rotation resection.

After completing the distal femur and proximal tibia resection, sequential medial releases were performed as required to create a rectangular extension gap, which was verified by inserting a 19-mm traditional spacer block (Fig. 2A). Then, the knee was flexed 90°. The measuring units of the modified spacer blocks were sequentially inserted into the medial and lateral joint space in order from small to large. As the medial and lateral joint spaces were stretched, the tension on the medial and lateral ligaments gradually increased. When the tension of the medial and lateral ligaments were balanced (under valgus stress, the medial compartment was stretched within 1 mm, and under varus stress, the lateral compartment was allowed to be stretched by 1–2 mm; Fig. 2B), the thickness of the two spacers determined the external rotation resection angle of the femur (the angle between the PCA and the cut tibial surface, Fig. 3B). Then, a specially designed condylar measuring device was placed on the handle of the two spacers. Two temporary nails were driven into the nail holes of the device (refer to the thickness of the expansion gap that has been obtained, such as mark “19”) to determine the AP position of the 4-in-1 resection block, and the size of the femoral component was obtained by measuring the AP diameter of the condyle (Fig. 2C). We also used the AP axis as an additional visual reference to confirm the component rotation (Fig. 2D). Next, an appropriately sized 4-in-1 resection block was installed in the nail hole (Fig. 2E), and the block was utilized to perform anterior, posterior, and chamfer bone cuts. Finally, a rectangular flexion gap equivalent to the extension gap was obtained (Fig. 2F). No soft tissue was released after this step. The thickness of the medial and lateral posterior condylar bone resection was measured with callipers in both groups.

The traditional posterior condylar referencing jig was conventionally set to 3° of external rotation to the PCA to determine the rotation of the femur [23]. In this process, we paid special attention to whether the posterior condyle had hypoplasia or erosion and rechecked that rotation with the AP or TEA. After cutting the anterior, posterior, and chamfered bones with the appropriate 4-in-1 resection block, the ligaments were released as needed to balance the knee.

The following procedures for the two groups were the same, mainly including the processing of the tibial keel, patella and patella trajectory.

No patients had drainage tubes placed. Prophylactic antibiotics were intravenously administered within 24 h after surgery. Rivaroxaban (10 mg/day) was orally administered for 21 days to prevent deep venous thrombosis in the lower extremities. Two days after the operation, the patient began to take the initiative to perform quadriceps muscle contractions, CPM training, and ambulation with a walking aid.

Intraoperative data collection included the operative time, blood loss, thickness of the cut posterior condyle, and the angle of femoral rotation resection relative to the PCA. (This value is the degree of rotation set on the reference jig in the MR group but in the GB group it is calculated by the trigonometric function formula based on the thickness of the medial and lateral modified spacers; Fig. 3B.)

Patients were followed at 6 weeks, 6 months, 1 year, and then annually. Any complication of the treatment was recorded. Each clinical follow-up examination included the range of motion (ROM), knee stability tests (subjective varus–valgus stress testing can be performed to assess the stability of the knee), and full-length AP and lateral X-rays of the knee for radiological evaluation. The hip–knee–ankle angle (HKA, Fig. 3C), medial distal femoral angle (MDFA, the angle between the distal articular surface and the mechanical axis of the femur), medial proximal tibial angle (MPTA, the angle between the proximal articular surface and the mechanical axis of the tibia) and joint line displacement [24] (Fig. 3D) were measured by one reviewer blinded to the surgical technique using a picture archiving and communication system (PACS) presenting the preoperative and postoperative radiographs. Patients whose mechanical axial alignment was not in the range of 180.0° ± 3.0° were considered as HKA outliers [25]. In addition, the Knee Society Score (KSS) and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) scores of patients preoperatively and postoperatively at 6 weeks, 6 months and 2 years were recorded.

All patients were followed up. The follow-up time ranged from 36 to 53 months, with an average of 39 months. There were no significant differences in the ROM, KSS scores or WOMAC scores between the groups at 6 weeks, 6 months, or 2 years post-surgery (Table 2).

No patients in either group had undergone revision surgery by the end of the last follow-up. After the surgery, complications of knee anterior pain (2 knees) occurred in the GB group (2/61), and slight postoperative knee laxity (1 knee), knee anterior pain (1 knee), and periprosthetic joint infection (1 knee) occurred in the MR group (3/63). There was no difference in the number of complications between the two groups (Table 2). The postoperative knee laxity patient in the MR group showed slight valgus laxity in flexion, but he did not complain of knee instability, and his clinical score was similar to that of the other patients. One patient developed an acute infection 28 days after implantation and was treated with debridement and implant retention (DAIR), and he had recovered at the 6-month follow-up. Three patients with anterior knee pain caused by patellar arthritis were relieved after taking Celebrex capsules for 3 months.

The mean value of the postoperative mechanical axis (HKA°) was 178.65° ± 1.30° in the GB group and 178.35° ± 1.71° in the MR group (P = 0.275). There was no significant difference in the number of HKA outliers between the two groups (P = 0.795). The postoperative MDFA (P = 0.495) and MPTA (P = 0.253) were similar between the groups. There was also no significant difference in the joint line displacement between the groups (P = 0.270, Table 2).

There were no significant differences between the GB and MR groups in terms of the operative time (P = 0.075) or intraoperative blood loss (P = 0.251, Table 3). The angle of femoral rotation resection relative to PCA in the GB group was statistically greater than that in the MR group (4.06 ± 1.10° vs. 3.19 ± 0.59°, P < 0.001). The thickness of the cut posterior medial condyle was larger than that in the MR group (9.72 ± 0.84 mm vs. 9.25 ± 0.62 mm, P = 0.001). In contrast, the thickness of the cut posterior lateral condyle was similar in the two groups, measuring 6.91 ± 0.71 mm in the GB group and 7.08 ± 0.53 mm in the MR group (P = 0.137, Table 3).