A randomized prospective study. To compare safety, efficacy, and cost between 1-stage and 2-stage anterior and posterior spinal fusion in treating rigid thoracic adolescent idiopathic scoliosis (AIS). Retrospective study issued different outcomes comparing 1-stage with 2-stage surgical procedures in treating neuromuscular and congenital scoliosis. A randomized prospective study in AIS has rarely been reported. METHODS.: Twenty-four patients were randomly divided into 2 groups. Group A underwent a 2-stage procedure, group B a 1-stage procedure. Inclusion criteria were: (1) AIS; (2) rigid single thoracic curve, > or =90 degrees on anteroposterior and > or =60 degrees on side bending view; and (3) pulmonary function tests, vital capacity > or =500 mL or FEV1 > or =50%. Exclusion criteria were: (1) curves either <90 degrees on anteroposterior or <60 degrees on bending view; and (2) pulmonary function tests, both vital capacity <500 mL and FEV1 <50%. The mean follow up was 3.3 years (range 2-5.1). Clinical outcomes between the 2 groups were analyzed. A statistical t test demonstrated no significant difference in age (P = 0.299), curves on anteroposterior (P = 0.908), on bending (P = 0.905), after surgery (P = 0.699), and at last follow-up (P = 0.946). No significant difference in blood loss (P = 0.089), transfusion (P = 0.334), operation time (P = 0.172), fusion segments anterior (P = 0.161), posterior (P = 0.546), wound drainage (P = 0.557), thoraco drainage (P = 0.948), and loss of correction at follow-up (P = 0.596). However, significant differences were found in length of stay (P < 0.001), patient cost of hospitalization (P < 0.001) between 2 groups, and curve correction (P < 0.001) in each group. Perioperative complications included 1 patient each with pneumonia, pneumothorax (group A), and ileus (group B). Screw nut loosing was found in 1 patient in group B, but no curve deteriorated. No neurologic deficit occurred in either group. A comparison of 1-stage versus 2-stage fusion in treating rigid thoracic curves in AIS did not show significant differences in safety or efficacy but did demonstrate significant difference in length of stay and patient cost.