Background. Established hip and knee arthroplasty registers exist in many countries but this is not the case with spinal implants. Moreover, in the case of a rod intended to guide spinal growth in a child and then be removed, the definition of ‘failure’ (revision) used for hip or knee arthroplasty is inappropriate. How can the performance of such spinal implants be judged?. Methods. Ninety-six MAGnetic Expansion Control (MAGEC) spinal rods were obtained from multiple centres after removal from the spines of 52 children with scoliosis. Clinical details were assessed and divided between unplanned revision operations (‘failures’) and those which were planned. Of the explanted rods, 49 were tested for the amount of force they could output, using the manufacturer's supplied test jig. Sixty-five rods were cut apart so that the internal components (bearings, O-ring seals, drive pins) could be assessed, alongside if there was evidence of internal wear. Results. Seventy-four per cent of revision operations were unplanned. Eighty per cent of explanted rods were unable to produce the force expected from a new rod. All rods (100%) that were successfully cut open showed signs of internal wear. Non- functional bearings were seen in 74% of cases, obvious seal damage in 57% of cases and broken drive pins in 47% of cases. Conclusion. Despite potential clinical benefits, explanted
