Component malalignment remains a major concern in total knee arthroplasty (TKA). Patient-specific guides (PSG) were developed to increase accuracy of bone resections and component placement, but available evidence is contradictory. We assessed the accuracy of 3D component placement in TKA with PSG compared to conventional surgery using virtual 3D bone models.
Fifty patients were randomly assigned to the PSG or conventional instrumentation group, 44 were finally analyzed. Preoperatively, MRI and CT scans were converted into virtual 3D models and a surgical plan was developed. Surgery was performed and changes in component sizing were recorded. Postoperative CT images were converted to 3D models and aligned to the planned, preoperative models and implant orientation. Differences between planned and postoperative implant orientations were calculated in 3D.
PSG allowed significantly more accurate varus/valgus placement for the femoral component (PSG: 0.14 ± 1.47; control: 1.40 ± 1.99; p < 0.05), but more slope was introduced (PSG: 2.82 ± 2.42; control: 0.90 ± 2.28; p < 0.05). Less variability in positioning accuracy for femoral flexion angle and tibial rotation was found with PSG, indicating a result closer to the planned position, but no significant differences in positioning accuracy were found. PSG allowed more accurate prediction of the femoral (PSG: 100%; control: 64%) and tibial (PSG: 79%; control 56%) component size.
PSG led to adequate component positioning accuracy compared to the pre-operative plan. For the femoral component, the positioning was significantly closer to the planned position in the coronal plane, a similar trend was observed for the sagittal plane. But, for the tibial component, significantly more slope was introduced. A better prediction of component sizing was found with PSG compared to conventional surgery.