Fifty patients (14-55 years of age) with unstable thoracolumbar fractures were studied: 24 patients treated conservatively 1971-1977 and 26 patients treated surgically with Harrington instrumentation 1977-1981. The treatment groups were comparable in all respects. Radiologic evaluation showed that Harrington distraction rods restored the fractured vertebra almost to its original shape, and the gibbus and scoliosis were significantly reduced. However, at the follow-up examination at least 2 years after the injury, the gibbus angle had recurred almost to the value at admission in patients with the rods removed. The conservatively treated patients showed a continuous increase of the gibbus angle and of the anterior and central vertebral compression. At the follow-up evaluation, all fractures in both treatment groups were healed. There was no difference between the treatment groups regarding neurologic improvement. Thirteen of 14 patients with severe or moderate paraparesis considerably improved their neurological status. A rehabilitation index with special reference to paraparetic patients showed no difference between the treatment groups already three months after the injury. Thoracolumbar fatigue, thoracolumbar pain and stiffness, skin problems, and pain at direct pressure at the fracture site occurred equally in the conservative and Harrington groups. The overall complications were few. The aseptic intermittent catheterization method introduced in 1977 considerably diminished the frequency of upper urinary tract infections. The treatment with open reduction, fusion, and stabilization with Harrington rods considerably reduced the immobilization and hospitalization times. The average immobilization time was reduced from 67 to 18 days. The hospitalization time in neurologically intact patients was reduced from 80 to 30 days.