Ten-year survival and clinical outcome of the AcroFlex lumbar disc replacement for the treatment of symptomatic disc degeneration

Abstract

BACKGROUND CONTEXT We have previously reported on the osseointegration, stability, and preserved motion of the AcroFlex lumbar disc replacement (LDR) in a nonhuman primate model. Detailed biomechanical testing of the device predicted implant survival for at least 10 years of in vivo use. Significant improvements in the clinical outcome were reported at 2 years. However, mechanical failure of the polyolefin rubber was detected by fine-cut computed tomography (CT) in a number of subjects within 2 years. As a result, no further devices were implanted. PURPOSE To report on the 10-year survival and clinical outcome of the AcroFlex elastomeric LDR when used for the treatment of one- or two-level symptomatic disc degeneration between L4 and S1. STUDY DESIGN Prospective nonrandomized clinical trial with a mean 10-year follow-up. PATIENT SAMPLE Twenty-eight patients with symptomatic disc degeneration who underwent AcroFlex LDR at one or two levels. OUTCOME MEASURES Clinical: Visual Analog Score for back pain, Oswestry Disability Index (ODI), Low Back Outcome Score (LBOS), and Short Form-36 (SF-36). Survival: Kaplan-Meier analysis over 10 years with first revision surgery as the end point. Radiographic: Dynamic flexion/extension radiographs at 2 years. Magnetic resonance imaging (MRI) and CT scans at 10 years. METHODS Twenty-eight subjects (14 male, mean age 41 years) with symptomatic disc degeneration unresponsive to nonsurgical treatment were enrolled into a prospective nonrandomized trial of the AcroFlex LDR. Visual analog score for back pain, ODI, LBOS, and SF-36 questionnaires were administered preoperatively at 6 months, 1, 2, and 10 years after the index procedure. All subjects were invited to undergo an MRI and for those with the device remaining in situ, a lumbar CT scan. Kaplan-Meier survival analysis was performed with first revision surgery as the end point. RESULTS At a mean of 9 years, 8 months (range, 8 years, 8 months–11 years, 3 months) after surgery, 17 of 28 patients did not require a revision surgery, representing a cumulative survival of 60.7%. In contrast, 11 of 28 patients (39.3%) underwent a total of 14 revision procedures; 9 of 11 patients underwent a conversion to anterior lumbar interbody fusion supplemented with pedicle screw fixation. Indications for a revision included device failure in seven and disabling pain in four patients. Mean time to revision was 3 years, 10 months (range, 23 months–8 years, 4 months). Mean ODI at 10 years for nonrevision cases was 27.5 (±17.6) compared with 41.8 (±26) for revision cases. Mean improvement over 10 years in the ODI for nonrevision cases was 17.9 (±16.9) compared with 12 (±16.1) for revision cases. Similar trends were observed in LBOS and SF-36 scores. Radiographic findings in the revision group included midsubstance tears in the rubber, osteolysis, and implant displacement. CT findings in 11 of 17 survivors included heterotopic bone formation (85%), osteolysis (50%), and subsidence (14%). Magnetic resonance imaging in 14 of 23 subjects at the final follow-up demonstrated an adjacent-level disc degeneration in 68% of those with the AcroFlex LDR in situ and in 40% of those who had been converted to fusion. Skip-level disc degeneration was present in 44% of those with AcroFlex device in situ and in 20% of those who had been converted to fusion. CONCLUSIONS The cumulative survival was 60.7% at 10 years when the first revision surgery was taken as the end point. The etiology of the implant failure prompting the revision included failure of osseointegration, midsubstance elastomeric tears, and osteolysis. Further use of this implant is not justified. The incidence of adjacent-level disc degeneration for the AcroFlex was comparable with that observed adjacent to the spinal fusio n. Salvage procedures involving conversion to spinal fusion are technically demanding, but appear to improve outcomes modestly.