INTRODUCTION: A concern expressed about lumbar total disc replacement (TDR) has been safety. One measure of safety is the need for subsequent surgery for implant removal or revision. This may be of particular importance considering TDR removal/revision generally requires re-operation through the anterior approach with the corresponding increased risk of vascular injury. The purpose of this study was to analyze the incidence of, and reasons for, lumbar TDR removal/revision.

METHODS: Data were collected from a multi-site spine specialty practice. A consecutive series of 1,775 lumbar TDR patients, beginning with the first case experience in 2000, was reviewed to identify those undergoing re-operation for TDR removal or revision. Only patients who were at least 2 years postoperative were included. Contact by mail and/or telephone phone calls to collect current data was conducted for patients who did not have  a recent office visit. The mean follow-up was 85.9 months with a median of 73 months and maximum of 251 months. For each case of device removal/revision, the reason, duration from index surgery, and procedure performed were recorded.

RESULTS: In the series of 1,775 patients, 26 (1.46%) underwent TDR removal or revision. Removal was performed in 23 patients (1.29%) and 3 patients underwent TDR revision (0.17%). Based on the total number of 2,082 TDRs implanted in the 1,775 patients, rates of removals and revisions were 1.15% and 0.14%, respectively. Removals included: 10 for migration and/or loosening, 3 developed problems after a trauma, 3 had ongoing pain, 2 developed lymphocytic reaction to device materials, and there was one each of: TDR was too large and replaced with smaller device, vertebral body fractures (osteoporosis), developed a lytic lesion, device subsidence and facet arthrosis, and infection seeded from a chest infection at 146 month post-TDR. The 3 revisions were: Repositioning the core (technique error), repositioning device after displacement, and core replacement due to wear/failure. With respect to timing, 34.6% of removals/revisions occurred within one month after the index implantation. Of note, 42.3% of revisions/removals occurred in the first 25 TDR cases performed by individual surgeons. Vascular complication during TDR removal occurred in one patient with an iliac vein tear whose TDR was removed due to trauma including spinal fracture.  Removal/revision status was verified by office visit or mailing/telephone in 227 patients with a minimum of 15-year follow-up data. Only one patient underwent removal/revision surgery more than 15 years post-TDR. This was the patient who underwent TDR removal due to trauma-related spinal fracture.

DISCUSSION: In this large consecutive patient series, 1.5% of lumbar TDRs were removed/revised. Only one revision was related to device failure or wear. Many of the removals/revisions were performed within a month of implantation. Also of note, many occurred within the first 25 TDR cases for individual surgeons, suggesting a learning curve. The low rate of removal/revision in this large institutional experience over a 20 year period provides support for the safety of these devices.