Evaluation of the predictors of postoperative aggravation of shoulder imbalance in severe and rigid thoracic or thoracolumbar scoliosis

Objective To study the predictors of postoperative aggravation of shoulder imbalance in severe and rigid thoracic or thoracolumbar scoliosis. Methods In this study, 49 patients with severe and rigid thoracic or thoracolumbar scoliosis were analyzed retrospectively. The patients underwent whole-spine anteroposterior and lateral radiography preoperatively and postoperatively. On the radiographs, we measured parameters, including T1 tilt, radiographic shoulder height (RSH), proximal curve, middle curve, distal curve, apical vertebral translation (AVT) of the middle curve, thoracic trunk shift (TTS), coronal balance, and sagittal balance. We regarded RSH and T1 tilt as postoperative shoulder balance parameters and divided the patients into improved and aggravated groups of shoulder imbalance. Univariate analysis, receiver operating characteristic (ROC) curve, and multivariate logistic regression analyses were used in the statistical analyses. Results The RSH was −17.01 ± 21.85 mm before surgery and 4.76 ± 18.11 mm at follow-up. The T1 tilt angle was −10.20° ± 19.53° before surgery and −2.72° ± 13.48° at follow-up. The results of the univariate analysis suggest that preoperative RSH and proximal to middle curve change ratio were significantly higher in the patients in the improved RSH group ( p < 0.01). In addition, preoperative RSH, preoperative T1 tilt, and apical vertebral translation of the middle curve were significantly higher, and preoperative proximal curve, postoperative proximal curve, and preoperative distal curve were significantly lower in the patients with improved T1 tilt group ( p < 0.01). In a binary logistic regression analysis, preoperative RSH [ B = −0.120, odds ratio (OR) = 0.887, p = 0.006] was found to be an independent predictor of postoperative aggravation of RSH. Similarly, preoperative T1 tilt ( B = −0.488, OR = 0.614, p = 0.001) was found to be an independent predictor of postoperative aggravation of T1 tilt. Moreover, the relationship between changes in RSH and T1 tilt was either concordant or discordant. Conclusion Several radiographic parameters were found to affect postoperative aggravation of RSH and T1 tilt. In particular, preoperative RSH and T1 tilt were found to be independent predictive factors of postoperative aggravation of RSH and T1 tilt, respectively.