Thirty-six knee arthrodeses performed using an external fixator with an average followup of 48 months were reviewed retrospectively. A single plane fixator was used in 19 cases and a biplane fixator in 17 cases. The reasons for fusion included an infected total knee arthroplasty (21 cases), aseptic loosening of a total knee arthroplasty (9 cases), posttraumatic osteoarthritis (3 cases), and a neuropathic joint, an infected unicondylar knee arthroplasty, and a tuberculous joint (1 case each). A fusion was obtained after the initial procedure in 22 patients (61%). With additional procedures, a fusion was obtained eventually in 27 patients (75%). The fusion rate decreased with an increasing number of prior knee procedures. Single and biplane external fixator designs had similar initial fusion rates (single 58%, biplane 65%). Complications included 14 nonunions (5 fused with additional procedures), 6 pin tract infections, 5 delayed unions, 1 stress fracture through a pin site, and 1 persistent infection resulting in an above-knee amputation. Despite biomechanical advances in external fixator design, knee arthrodesis remains difficult to achieve in patients who have had multiple previous procedures, a failed total knee arthroplasty, or an infected total knee arthroplasty with significant bone loss.