A prospective randomized trial on 128 patients with end-stage osteoarthritis was conducted to assess the accuracy of patient-specific guides. In cohort A (n = 64), patient- specific guides from four different manufacturers (Subgroup A1 Signature ®, A2 Trumatch ®, A3 Visionaire ® and A4 PSI ®) were used to guide the bone cuts. Surgical navigation was used as an intraoperative control for outliers. In cohort B (n = 64), conventional instrumentation was used. All patients of cohorts A and B underwent a postoperative full-leg standing X-ray and CT scan for measuring overall coronal alignment of the limb and three-planar alignment of the femoral and the tibial component. Three-planar alignment was the primary endpoint. Deviation of more than three degrees from the target in any plane, as measured with surgical navigation or radiologic imaging, was defined as an outlier.

In 14 patients (22%) of cohort A, the use of the patient-specific guide was abandoned because of outliers in more than one plane. In 18 patients (28%), a correction of the position indicated by the guide, was made in at least one plane. A change in cranial-caudal position was most common. Cohort A and B showed a similar percentage of outliers in long-leg coronal alignment (24.6%, 28.1%, p = 0.69), femoral coronal alignment (6.6%, 14.1%, p = 0.24) and femoral axial alignment (23%, 17.2%, p = 0.50). Cohort A had more outliers in coronal tibial alignment (14.6%) and sagittal tibial alignment (21.3%) than cohort B (3.1%, p = 0.03 and 3.1%, p = 0.002, respectively). These data indicate that patient specific guides do not improve accuracy in total knee arthroplasty.