To present the results of surgical correction in patients with double or triple thoracic/lumbar AIS (Lenke types 2,3,4) with the use of a novel convex/convex unilateral segmental screw correction technique in a single surgeon's prospective series.

We reviewed the medical records and spinal radiographs of 92 consecutive patients (72 female-20 male). We measured scoliosis, thoracic kyphosis, lumbar lordosis, scoliosis flexibility and correction index, coronal and sagittal balance before and after surgery, as well as at minimum 2-year follow-up. SRS-22 data was available preoperatively, 6-month, 12-month and 2-year postoperatively for all patients. Surgical technique. All patients underwent posterior spinal fusion using pedicle screw constructs. Unilateral screws were placed across the convexity of each individual thoracic or lumbar curve to allow for segmental correction. ‘Corrective rod’ was the one attached to the convexity of each curve with the correction performed across the main thoracic scoliosis always before the lumbar. Maximum correction of main thoracic curves was always performed, whereas the lumbar scoliosis was corrected to the degree required to achieve a balanced effect across the thoracic and lumbar segments and adequate global coronal spinal balance. Concave screws were not placed across any deformity levels. Bilateral screws across 2 levels caudally and 1–2 levels cephalad provided proximal/distal stability of the construct.

Mean age at surgery was 14.9 years with mean Risser grade 2.8. The distribution of scoliosis was: Lenke type 2–26 patients; type 3–43 patients; type 4–23 patients. Mean preoperative Cobb angle for upper thoracic curves was 45°. This was corrected by 62% to mean 17° (p<0.001). Mean preoperative Cobb angle for main thoracic curves was 70°. This was corrected by 69% to mean 22° (p<0.001). Mean preoperative Cobb angle for lumbar curves was 56°. This was corrected by 68% to mean 18° (p<0.001). No patient lost >2° correction at follow-up. Mean preoperative thoracic kyphosis was 34° and lumbar lordosis 46°. Mean postoperative thoracic kyphosis was 45° (p<0.001) and lumbar lordosis 46.5° (p=0.69). Mean preoperative coronal imbalance was 1.2 cm. This corrected to mean 0.02 cm at follow-up (p<0.001). Mean preoperative sagittal imbalance was −2 cm. This corrected to mean −0.1 cm at follow-up (p<0.001). Mean theatre time was 187 minutes, hospital stay 6.8 days and intraoperative blood loss 0.29 blood volumes (1100 ml). Intraoperative spinal cord monitoring was performed recording cortical and cervical SSEPs and transcranial upper/lower limb MEPs and there were no problems. None of the patients developed neurological complications, infection or detected non-union and none required revision surgery to address residual or recurrent deformity. Mean preoperative SRS-22 score was 3.6; this improved to 4.6 at follow-up (p<0.001). All individual parameters also demonstrated significant improvement (p<0.001) with mean satisfaction rate at 2-year follow-up 4.9.

The convex-convex unilateral pedicle screw technique can reduce the risk of neurological injury during major deformity surgery as it does not require placement of screws across the deformed apical concave pedicles which are in close proximity to the spinal cord. Despite the use of a lesser number of pedicle fixation points compared to the bilateral segmental screw techniques, in our series it has achieved satisfactory scoliosis correction and restoration of global coronal and sagittal balance with improved thoracic kyphosis and preserved lumbar lordosis. These results have been associated with excellent patient satisfaction and functional outcomes as demonstrated through the SRS-22 scores.