Study design: To analyse the long term effect of Harrington Instrumentation and fusion to the lumbar spine in the treatment of idiopathic scoliosis.

Objectives: To demonstrate there is a relationship between the strategy used (determination of fusion area) and pain or degenerative changes.

Summary of background data: The literature has been fairly controversial in terms of pain and degenerative changes beyond a fusion for idiopathic scoliosis according as the lower level of fusion. This is the first study in which the results are analysed according as the “strategy used” and not the sole level of fusion.

Methods: 250 patients operated on by Harrington instrumentation were clinically and radiographically reviewed. Pain was classified (as Moskowitz and Moe). To be included they should have an idiopathic scoliosis, a minimum follow up of 20 years (mean 26, max 36), 37% over 30 years, had to have been under 20 years at the time of surgery, and should have a full set of radiographs.

Curves were classified according to our own classification (Salanova et al) 1973–2000 in single structural. Thoracic 114, thoraco-lumbar 21 and double structural thoracic and lumbar, true double major (52), false D.M. (45). The double thoracic was identified with permanent T1 tilt (18). On P.OP standing the lower level of fusion was identified: E.V. (Salanova et al 1973–2000) SV (King) other vertebra. On follow up radiographs standing coronal and sagittal, lumbar coronal and sagittal degenerative changes were evaluated, slipping lateral and sagittal, discopathy over 50% and classified as none, moderate, complete.

Results: Mean age at surgery 15 years + 6. Mean age at follow up 49 years. Ten patients were reoperated on for various reasons. Overall results: Pain none 70, episodic 82, frequent 42, permanent 46. Degenerative changes none 155, moderate 62, complete 23. These data were evaluated according to the strategy used; there is a strong statistical relationship between strategy and final results. Our study proves that King’s classification for so-called King II curves is misleading.

Conclusion: This study is the most important ever published in terms of patients, methodology, and follow up. It shows that if a clear analysis of curve(s) before surgery is effectuated for determination of fusion area, if for single curves the lower level of fusion is the good one and for double structural the choice between selective thoracic fusion and double fusion is correctly determinated the long term results are not so bad.