A screw position change at an early postoperative stage preceding the subsequent occurrence of screw loosening.

PURPOSE The authors recently proposed the novel radiologic assessment method to measure chronological screw position changes precisely. The aim of this study was to predict the late occurrence of screw loosening, which was diagnosed by the radiographic lucent zone, by evaluating screw position changes at an early postoperative stage using the novel method. METHODS Forty-three patients who underwent thoracolumbar screw fixation and follow-up computed tomography (CT) scans on the day, between 1 and 5 weeks, and at more than 6 months after surgery were retrospectively evaluated. Screw images were generated from CT data. Screw position changes were evaluated by superposing screw images on the day and between 1 and 5 weeks after surgery. Screw loosening was diagnosed by the radiographic lucent zone on CT images at 6 months or later post-surgery, and patients were classified into screw loosening and non-loosening groups. The early screw position changes were compared between the two groups. RESULTS Significant differences in early screw position changes were found between the screw loosening and non-loosening groups in Mann-Whitney U test (p = 0.001). On the receiver operating characteristic (ROC) curve analysis, the area under the ROC curve was 0.791, and the best cutoff value of early screw position change for the prediction of screw loosening was 0.83 mm with a sensitivity of 64.0% and a specificity of 88.9%. CONCLUSION We calculated a cutoff value of the screw position changes at an early postoperative stage for the prediction of subsequent development of screw loosening with the radiographic lucent zone.