7 preop, 4.8 postop, 9.9 FU) (Table 2). Table 2 Radiographic evaluation. In two patients, one selleck kinase inhibitor screw was found medial into the spine canal on the postoperative TC, without any clinical consequence. At the beginning of our experience, we planned to remove all implants including L2 or a lower vertebra, no implant above T10 and all the implants in the thoracolumbar junction showing clinical (local pain) or mechanical problems (hardware failure or screws mobilization). We planned hardware removal in the lumbar spine as we were afraid that posterior fixation without fusion in such a mobile part of the spine could lead to hardware failure and consequently to clinical problems. Overall, the instrumentation has been removed in 23 patients (19%), in 5 cases due to a local complication and in 17 cases, as scheduled, because of implantationin the lumbar spine (Figure 3).

The average delay from first surgery to implant removal was 9,5 months (range: 6�C36). In the 17 patients in which implant removal had been planned, only 3 showed screws mobilization, and only 2 had pain. None of them showed pain or loss of sagittal alignment at six-month followup. Figure 3 Percutaneous minimal invasive removal of the instrumentation. 4. Complications The complications were divided according to a temporal order of appearance in intraoperative and postoperative. The latter were divided into early if they appear within one month from the date of surgery and late when they occurred after that period [7]. Depending on the severity, we divided complications into major and minor [8].

Major complications were those involving an increased hospitalization, or a second operation not scheduled. We recorded 12 complications (9.8%) divided into 4 intraoperative (3.3%), 6 early postoperative (4.9%), 2 late postoperative (1.6%). Four complications were minor (3.3%) and 8 major (6.5%). Intraoperative complications were all minor, related to mechanical instruments, which lengthened the surgical time but without any consequence for the patients. Early postoperative complications were all major: 4 mechanical, 1 neurological and 1 infectious complication. In 2 patients the screw head disconnected from the stem in the first postoperative day. In one case, the patient was reoperated, while the other had to wear a brace for 3 months postoperatively.

In 2 patients we recorded a pullout of the pedicle screws, 15 days and 20 days after surgery respectively. The first case was a 63-year old patient with 2 noncontiguous type A1 fractures (T11 and L1) undergoing MIS from T10 �C L3 with bilateral pedicle screws in L1. The second case was a patient of 67 years fixed from T12 to L2 for a type A3 L1 fracture. In both cases, we performed the implant removal and a percutaneous augmentation of the vertebral bodies with cement. The neurologic complication was a cauda equina syndrome Carfilzomib which appeared in the second postoperative day in a patient treated for a type A L1 fracture by T12�CL2 MIS.