Objective To evaluate the effectiveness of pedicle subtraction osteotomy (PSO) and non-osteotomy techniques in treatment of medium-to-severe kyphoscoliosis by retrospective studies. Methods Between January 2005 and January 2009, 99 patients with medium-to-severe kyphoscoliosis were treated by PSO (PSO group, n=46) and non-osteotomytechnique (non-osteotomy group, n=53) separately. There was no significant difference in sex, age, Cobb angle of scol iosis on coronal plane, and Cobb angle of kyphosis on saggital plane between 2 groups (P gt; 0.05). The operation time and blood loss were recorded; the Cobb angle of scol iosis on coronal plane and kyphosis on sagittal plane were measured at pre- and postoperation to caculate the rates of correction on both planes. Results The operation was successfully completed in all the patients. The operation time and blood loss of the patients in PSO group were significantly greater than those of the patients in non-osteotomy group (P lt; 0.05). All patients were followed up 12-56 months (mean, 22.4 months); no spinal cord injury occurred, and bone fusion was achieved at last follow-up. The Cobb angles of scol iosis and kyphosis at 2 weeks and last follow-up were significantly improved when compared with the preoperative angles in the patients of 2 groups (P lt; 0.05). There was no significant difference in Cobb angle of scol iosis and the rate of correction between 2 groups (P gt; 0.05), but the correction loss of PSO group was significantly smaller than that of non-osteotomy group (P lt; 0.05) at last follow-up. At 2 weeks and last follow-up, the Cobb angle of kyphosis, the rate of correction, and correction loss were significantly better in PSO group than in non-osteotomy group (P lt; 0.05). Conclusion There is no signifcant difference in scol iosis correction between PSO and non-osteotomy techniques.PSO can get better corrective effect in kyphosis correction than non-osteotomy technique, but the operation time and blood losswould increase greatly.