We retrospectively reviewed 150 patients with osteoarthritis who had a primary TKA performed using the same cruciate-retaining cemented total knee system (VanguardTM; Biomet Inc, Warsaw, IN, USA) between January 2009 and September 2010. These 150 patients were selected from those who had a total of 285 primary TKAs performed during the study period meeting the inclusion and exclusion criteria and who had a full series of adequate radiographs. We included only patients with primary osteoarthritis with or without prior open or arthroscopic meniscectomy. We excluded patients with previous fractures of the femoral condyles, tibia plateau, or femoral/tibial shaft and patients with previous osteotomy around the knee. Patients with a pacemaker (owing to their inability to have MRI) and patients with metal around the knee (owing to potential metallic artifact) also were excluded. We established three cohorts of patients for comparison by choosing an unselected group of consecutive patients who had primary TKA and a scout CT scan on the day of discharge or within 6 weeks of surgery. Scans met criteria for being of sufficient quality for an accurate reading (had to include entire femoral head, entire ankle, and the femoral component flange had to be centered on the ovals of the posterior condyle).
In Group 1 (n = 50), we used conventional instrumentation and the distal femur was prepared with an intramedullary rod using a 5°-valgus resection cut. The proximal tibia was prepared with an extramedullary cutting jig perpendicular to the mechanical axis of the tibia. In Group 2 (n = 50) we used the patient-specific SignatureTM system (Biomet Inc) based on restoring the mechanical axis. In Group 3 (n = 50) we used the patient-specific OtisMedTM system (OtisMed Corp, Alameda, CA, USA) based on shape matching to restore the kinematic axis.
In Group 2, each patient had preoperative MRI using the SignatureTM system manufacturing protocol, which uses 1-mm high-resolution slices at the knee and selected 5-mm spot images at the hip and ankle to allow the engineers to derive the mechanical axis and correct for rotation of the extremity. In all of these cases, the surgeon’s predetermined default settings were used to construct the disposable patient-specific blocks. The default settings for the femoral preparation incorporated a 9-mm posterior resection measured from the cartilage, a 9-mm medial distal resection measured from the bone, and the femoral rotation was set at 0° parallel to the transepicondylar axis. The default settings for the tibial preparation were perpendicular to the tibial mechanical axis (0° varus/valgus), 12 mm below the lateral plateau high point, and 2° posterior slope.
In Group 3 each patient had preoperative MRI using the OtisMedTM system manufacturing protocol, which used 2-mm high resolution slices localized to the knee only and included a 16-cm field of view centered on the joint line, which is similar to using a short knee radiograph for preoperative surgical templating. The alignment of the coronal and axial plane was perpendicular to the corticocancellous junction of the distal femur and posterior femur, respectively. Using proprietary software, the arthritic knee was reconstructed by filling articular defects, removing osteophytes, and approximating the joint surface to its prearthritic natural alignment, and not the mechanical alignment. The proprietary software then shape matched the femoral and tibial components to the reconstructed natural knee model to align the flexion-extension rotation axis and then the corresponding patient-specific disposable cutting guides are manufactured to fit the arthritic knee.
One surgeon (RLB) performed all of the TKAs for Groups 1 and 2, whereas a different surgeon (SMH) performed all of the TKAs in Group 3. Both surgeons had experience with patient-specific instrumentation before the start of this study and we did not include any of their early cases during the learning curve for this new technology. In accordance with standard protocol for both surgeons, all patients had a postoperative coronal CT scanogram with a field of view from the hip to the ankle (Fig. ). To minimize projection errors and standardize radiographic measurements, the leg was internally or externally rotated until the two augment holes on the posterior condyles of the femoral component were partially visible on either side of the anterior flange of the femoral component to control for rotation at the time of the scout CT scan (Fig. ).
A CT scout image with a field of view from the hip to the ankle was used to make measurements for this study and was adjusted to correct for rotation.
Fig. 2 A CT scout image shows the two augment holes (arrows) on the posterior condyles of the femoral component, which are visible on either side of the anterior flange of the femoral component. These two augment holes are used to help the radiology technicians (more ...)
Two reviewers (JZ and BMW) blinded to the surgical method performed the following three radiographic measurements on the AP projection of each CT scanogram [29
]: (1) femorotibial angle (FTA), which was the angle formed by a line that bisected the distal fourth of the femur and a line that bisected the proximal fourth of the tibia (Fig. ); (2) hip-knee-ankle angle (HKA), which was the angle formed by the mechanical axis of the femur (line between the center of the femoral head and the center of the knee) and the mechanical axis of the tibia (line between the center of the ankle and the center of the knee) (Fig. ); and (3) zone of mechanical axis (ZMA), which was the zone of the tibial base plate (divided into five equal regions) where the mechanical axis of the limb (straight line from the center of the femoral head to the center of the ankle) intersected the tibial base plate (Fig. ). Valgus angular values were ascribed a negative value (−), whereas varus angular values were ascribed a positive value (+). Measurements denoting lateral displacement from the center point of the tibial plateau were ascribed a positive value (+), whereas measurements indicating a medial distance from the center of the tibial plateau were considered negative (−). In this study, the accepted values for normal or in-range alignment were: (1) 2° to 8° valgus for FTA [1
]; (2) 0° ± 3° varus/valgus for HKA [2
]; and (3) the central zone for ZMA [25
]. We determined interrater reliability by calculating the intraclass correlation coefficient (ICC) using a two-way mixed effects model. The highest level of agreement among raters was seen for FTA (ICC = 0.87; 95% CI, 0.81–0.91; p < 0.001). HKA assessment had a good level of agreement among raters (ICC = 0.86; 95% CI, 0.81–0.89; p < 0.001), as did ZMA (ICC = 0.83; 95% CI, 0.79–0.87; p < 0.001).
The femorotibial angle (FTA) angle measurement between the anatomic axis of the femur and the anatomic axis of the tibia was used to measure coronal alignment on a short image.
Fig. 4 The hip-knee-ankle (HKA) angle measurement between the mechanical axis of the femur (line between the center of the femoral head and the center of knee) and mechanical axis of the tibia (line between the center of the ankle and the center of the knee) (more ...)
Fig. 5 The zone of the mechanical axis (ZMA) was determined by dividing the width of the tibial base plate into five equal regions and then drawing a line through the mechanical axis of the lower extremity to determine through which zone it passes. In this image (more ...)
We used a single-factor ANOVA and Tukey’s post hoc test to determine whether the FTA and HKA differed between the three surgical techniques. Fisher’s exact test was used to determine whether the percentage of in-range, varus, and valgus outliers of the FTA, HKA, and ZMA differed between the three surgical techniques. Statistical analyses were performed using SPSS for Windows, version 20 (SPSS Inc, Chicago, IL, USA).