Between 1968 and 1977, 72 patients with idiopathic scoliosis underwent Harrington Instrumentation (HI). Between 1985 and 1988, 21 patients with idiopathic scoliosis had posterior spinal fusion with Cotrel-Dubousset instrumentation (CDI). All patients were operated by the same orthopedic surgeon. None of the CDI patients had postoperative brace or cast protection, the HI group had on average 6 months' postoperative brace treatment. The two groups of patients were comparable in age, sex, and type of curves. The HI group and CDI group were reexamined with clinical and radiological assessment after mean periods of 148 months and 60 months respectively. The average preoperative Cobb angle in the CDI group was 59.9 degrees (HI group 67.8 degrees), which improved to 20.8 degrees (HI group 33 degrees) postoperatively--a correction of 66.3% (HI group 51.3%). The loss of correction on reassessment amounted to 5% in the CDI group and 20.7% in the HI group. In both groups, the mean rib hump height was reduced to 2.2 cm. In 40% of the Harrington patients, a flat back was found, but this was not related to clinical back pain. The rate of complications and reintervention was 9.5% in the CDI group and 8.3% in the HI group. There were no neurological complications. Subjectively, 86% of the Harrington patients and 95.2% of the CDI patients rated the results of their operation as "good" or "very good." The CDI group showed better results in correction of the Cobb angle and loss of correction, while saving one mobile lumbar segment. The correction of the rib hump showed the same results for both techniques. Blood loss and operation time was much lower in the HI group. However, the rate of complications was similar in both groups.