BackgroundThe mobile Oxford unicompartmental knee arthroplasty (UKA) implant has been widely used with an intramedullary guide for femoral preparation. We modified the femoral guide technique based on the tibial cut first and spacer block technique. This study was performed to determine the radiographic accuracy and early clinical outcomes of the extramedullary method.MethodsWe retrospectively evaluated 50 consecutive patients who underwent UKA using the extramedullary technique. An equal number of patients who underwent UKA with the conventional technique were matched as the control group. Clinical outcomes were evaluated in terms of the operating time, blood loss, range of motion, and Hospital for Special Surgery score. Radiographic accuracy was evaluated by the implant position and alignment in the coronal and sagittal planes.ResultsThe mean follow-up period was 39.76 5.77months. There were no differences in the postoperative Hospital for Special Surgery score, range of motion, or hip-knee-ankle angle between the two groups. The operating time in the extramedullary group was shorter than that in the conventional group (54.78 7.95 vs. 59.14 10.91min, respectively; p = 0.025). The drop in hemoglobin after 3days was only 12.34 +/- 4.98g/L in the extramedullary group which was less than that in the conventional group (p = 0.001). No significant differences were found in the postoperative coronal and sagittal angles between the two groups. Acceptable radiographic accuracy of the implant alignment and position was achieved in 92% of patients in the extramedullary group and 96% of patients in the conventional group.ConclusionsThe radiographic and clinical results of the extramedullary technique were comparable with those of the conventional technique with the advantage of no intramedullary interruption, less blood loss, a shorter operating time, and more rapid recovery. As the technique depends on the accurate tibial cut and overall alignment, we do not recommend it to surgeons without high volume experiences.Trial registration Retrospectively registeredLevel of evidenceIV, retrospective study