Comparison of computer-navigated and conventional total knee arthroplasty in patients with Ranawat type-II valgus deformity: medium-term clinical and radiological results

Between January 2002 and January 2009, patients who had arthritic knees with Ranawat type-II valgus deformity (a preoperative valgus angulation of anatomical axis ≥10° based on hip-to-ankle standing radiography) and who underwent primary TKA were enrolled in this retrospective study. The cost of CN was not reimbursed in our national health insurance, the type of surgery was chosen by the patients themselves after well explaining the merits of both techniques to patients.

The clinical and radiographic data and functional outcomes were reviewed. The clinical data included age, gender, diagnosis, intraoperative procedures, tourniquet time (the tourniquet was inflated before skin incision and was deflated after the setting of cement), difference in perioperative hemoglobin level (which was checked before the operation and in the morning of the second postoperative day), and antibiotic prophylaxis. The radiographic and functional assessments before surgery and at the time of each follow-up were analyzed and recorded by a research associate (Yu-Shuan, Lin) who was not part of the operative team and was blinded to technique allocation. This study was approved by the Ethics Committee and Institutional Review Board of Chang Gung Memorial Hospital (99-2386B). Informed consent for participation in the study was obtained from all participants.

A tourniquet was used for all patients. The procedure was carried out through a midline skin incision with the use of a medial parapatellar arthrotomy. All cases received the same cruciate-retaining type of cemented total knee prosthesis (DePuy PFC Knee Systems, DePuy Orthopaedics, Warsaw, Indiana).

In all of the conventional TKA procedures, the prosthesis was implanted using an intramedullary guidance system for the femoral component and an extramedullary guide for the tibial side. After removal of osteophytes from the femur, the soft tissues of the lateral compartment were then released from the proximal tibia. The distal femoral cut was based on the valgus correction angle of the distal femur (measured on hip-to-ankle standing radiography [23]) (Figure 1). Instead of 3 degrees of external rotation relative to the posterior aspect of the condyles, femoral component rotational alignment was determined according to the transepicondylar axis. The Whiteside’s line, posterior condylar line and the tibial cutting plane were supplementary to judge the femoral rotation. A spacer block was then used to check the rectangular gap in flexion and extension. The soft tissue balance was assessed at trial reduction and achieved by sequential release of the tight structures (iliotibial band, popliteus, lateral collateral ligament, and the lateral head of the gastrocnemius) as recommended by Whiteside [24]. The posterior cruciate ligament (PCL) was assessed using the pull-out lift-off test reported by Scott and Chmell [25], and the PCL was then recessed as needed from its insertion site in the tibia to obtain the desired tension. The assessment of bony defect are concerning because of the potential for later mechanical failure. It is important to determine the extent of the bony defect, the integrity of the cortex, and the stability of the component fixation. If the bone loss is minimal and the cortical rim is intact, we addressed the small cavitary defects with use of bone cement or cancellous bone chips. For patients with greater bone loss with poor integrity of cortical rim, structural allografts were applied to for restoration of the proximal part of the tibia or the distal part of the femur. After the cement had fully set, the tourniquet was then deflated and hemostasis. Assessment of patellar tracking were performed. The indication for release the lateral retinaculum is according to “The no thumb test”.

The CN-TKA was implanted with the use of a computed tomography (CT)-free navigation system (BrainLAB, Munich, Germany). After exposing the knee joint by using an anterior midline longitudinal skin incision and a medial parapatellar arthrotomy, the synovium was adequately removed to make registration precise. The femoral reference arrays were fixed to the distal femur using two bi-cortical half-pins, and the tibial reference arrays were fixed to the proximal tibia. An infrared camera was used to track the femoral and tibial reference arrays. The implant size and orientation were identified by dragging the pointer along the bone surface to reconstruct the three-dimensional bone model. The rotational alignment of the femur was checked by the transepicondylar axis under navigation. The contracted soft tissue was sequentially released using Whiteside’s method [24], and the tension of the PCL was assessed using the method of Scott and Chmell [25] with real-time and quantitative feedback of the navigation. The assessment and management of bony defect were performed. The femoral and tibial reference arrays were retained until the cement had fully set, and were removed after the alignment was verified under navigation. The tourniquet was then deflated and hemostasis. Assessment of patellar tracking were performed by using “The no thumb test”. All procedures were performed by a senior surgeon (Hsu RWW) who has extensive experience in the use of both conventional mechanical guides and computer-assisted navigation.

A continuous passive motion machine was used for passive range of motion exercises four times daily (30 minutes per exercise period) from the day of surgery until the day of discharge from the hospital. Under the supervision of a physical therapist, the patients started active knee-motion exercises and began standing at the bedside or walking with crutches or a walker twice daily for 30-minute periods. All patients used crutches or a walker with full weight-bearing for six weeks, and a cane when needed thereafter.

All patients received hip-to-ankle standing radiography along with anteroposterior and lateral radiography of the knees and the skyline view of the patellofemoral joint before and after surgery, following a protocol described in detail in a previous publication by the senior author [26]. The lateral patellar tilt and displacement were measured according to the method of Laurin et al. [27, 28]. The postoperative patellar tilting angle was measured by the method described by Kawahara S et al. [29]. Radiographic parameters including the mechanical axis and four component alignment angles (the femoral valgus angle, tibial valgus angle, femoral flexion angle, and tibial flexion angle [30]) were assessed. The position of the prosthetic joint line was measured on radiographs taken before surgery and at the last follow-up using the method proposed by Figgie et al. [31]. In brief, the length of the joint line measured in the lateral view was the perpendicular distance from a weight bearing surface of the tibial plateau to the tibial tubercle of the natural tibia. The relative change in joint line distance was recorded. All measurements were done by a blinded observer using digital radiographs.

We considered a tibial component positioned at a valgus angle of 90° in the coronal plane and at a flexion angle of 87° in the sagittal plane (3° of posterior slope) to be ideal. A femoral component positioned at a flexion angle of 0° in the sagittal plane was considered to be ideal, and the desired valgus angle in the coronal plane was determined according to the valgus correction angle of the distal femur which was measured by hip-to-ankle standing radiography. The goal of alignment was to correct the postoperative mechanical axis to within 3° of neutral, and all component angles to within 3 degrees of the ideal.

Preoperative and postoperative functional scores were calculated using the International Knee Society (IKS) scoring system [32] and patellar score [33]. The patellar score was designed by Feller et al. to assess the management of the patellofemoral joint. Active maximum range of motion of the knee was measured using a goniometer. The patients who had extra-articular deformity of the femur or tibia related to trauma or previous surgery and those with incomplete medical records, radiographic analysis and functional evaluation were excluded.

A total of 62 patients (70 knees) were enrolled in this study, with a mean age of 71.3 years (range, 50 to 86 years) at the time of surgery. The diagnosis was degenerative osteoarthritis in 68 knees and rheumatoid arthritis in the other two. The mean body height was 156.2 cm (range, 133.0 to 174.0 cm), mean body weight 64.8 kg (range, 35.0 to 95.0 kg), and mean body mass index 27.5 kg/m² (range, 19.8 to 38.1 kg/m²). The mean follow-up time was 6.2 years (range, 5–11 years).

Thirty-six knees underwent conventional TKA and thirty-four knees underwent CN-TKA. There were no statistically significant differences between the two groups with regards to age at the time of surgery, body height, body weight, body mass index, duration of hospital stay, and length of follow-up (Table 1).

The difference in perioperative hemoglobin level was less in the CN-TKA group than in the conventional TKA group, however the difference did not reach statistical significance. No significant difference in tourniquet time was noted. Theoretically, the reference base attachments and registration should have led to a longer tourniquet time in the CN-TKA cohort. However, intraoperatively repeated assessments of the position of the femoral component according to multiple reference lines in the conventional cohort may have also resulted in a longer tourniquet time. Thirteen knees (36.1%) in the conventional TKA group required release of the lateral retinaculum to obtain adequate patellar tracking, while only 2 knees (5.9%) required release in the CN-TKA group (p = 0.001). Bone grafting was performed in 7 knees (19.4%) in the conventional TKA group and 11 knees (32.4%) in the CN-TKA group (Table 2).

In the radiographic evaluation, we found no differences between the two groups in preoperative anatomical axis and postoperative mechanical axis. The mean preoperative anatomical axis was 195.4° (range, 191°-204°) in the conventional TKA group and 194.3° (range, 192°-202°) in the CN-TKA group. The postoperative mechanical axis was corrected to 179.9° (range, 176°-184°) in the conventional TKA group and 180.4° (range, 178°-181°) in the CN-TKA group. The mean preoperative patellofemoral congruence angles were 12.6° (range, 1°-39°) in group A and 11.5° (range, 2°-38°) in group B. There was no statistical significance (p = 0.716). The postoperative patellar component-tilting angle was 2.7 (range, 2°-5°) in group and 2.4° (range, 1°-4°) in group B. There was still no statistical significance (p = 0.855)”. There was less joint line elevation in the CN-TKA group than the conventional TKA group (p = 0.020). No postoperative differences were noted between the two groups with regards to the four component alignment angles including femoral valgus angle (p = 0.448), femoral flexion angle (p = 0.165), tibial valgus angle (p = 0.960) and tibial flexion angle (p = 0.573) (Table 2). The percentage of procedures achieving the ideal reconstructed mechanical axis and component alignments were similar between the two groups (Table 3).

Clinically, the active range of motion improved from 97.6° to 113.3° in the conventional TKA group, and from 96.9° to 114.3° in the CN-TKA group. The mean patellar score improved markedly after surgery in both groups (from 17.3 to 26.2 in the conventional TKA group, and from 16.4 to 26.7 in the CN-TKA group). On the IKS scoring system, the mean pain score improved from 16.4 to 47.4 in the conventional TKA group, and from 14.3 to 48.1 in the CN-TKA group, the mean clinical knee score improved from 40.5 to 95.2, and from 39.3 to 96.7, respectively, and the mean functional knee score improved from 35.3 to 94.6, and from 33.4 to 96.3, respectively (Table 4).

No patient required advancement of MCL or was converted to a posterior-stabilized prosthesis or varus-valgus constrained components because of excessive bony resection or inadequate soft tissue release during the operation. There were no cases of wound infection, peroneal nerve neurapraxia, pulmonary emboli, or deep vein thrombosis. No periprosthetic fractures, joint instability, or patellar tracking problems were encountered. No patients showed loosening or osteolysis on radiography at the time of the last follow-up, and no patients received revision surgery for any reason.