STUDY DESIGN: Cross-sectional study among a population of scoliotic and nonscoliotic women aged 40 years or more with low back pain from a spine rehabilitation unit. OBJECTIVES: (1) To test the hypothesis that scoliotic (SW) women shrink faster than nonscoliotic women (NSW) in adulthood. (2) To investigate the effects of age and curve progression in the scoliotic group, and to develop a model to assess the natural history of scoliosis and shrinkage. SUMMARY OF BACKGROUND DATA: Little is known about the decrease in body height in adult scoliotic patients. A simple method to help predict the future course of the curvature in patients without radiograph follow-up could help clinicians make treatment decisions. METHODS: Sixty SW and 40 NSW women matched for age, with no history of vertebral fracture or osteoporosis, were questioned about their peak body height and measured. Total spine radiographs were performed, and compared with previous images if available. Correlations between self-reported peak body height and current height, shrinkage, age, time since menopause, and the Cobb angle were searched for. In women with documented radiograph follow-up, correlations between shrinkage and progression of the Cobb angle were sought to develop a predictive model of curve progression. RESULTS: Average shrinkage in the SW was twice that in the NSW (5.1 +/- 3.5 vs. 2.3 +/- 0.7 cm, P < 0.001), had begun early in adulthood, was due to the combined effect of age and scoliosis, and was strongly associated with rotatory olisthesis. In the 17 women with radiograph follow-up (19.7 +/- 7.3 years), curve progression was closely related to shrinkage (r = 0.74; P < 0.001; y = 0.7 + 2.7x; SE = 0.42). This allowed us to generate prediction limits of the scoliosis for a given individual. CONCLUSION: Determination of shrinkage could provide a simple, noninvasive, and cheap method to monitor the natural history of scoliosis in adults. It could therefore be used in routine clinical practice to help make treatment decisions for patients with no documented radiograph follow-up.