In a 5-year prospective study on idiopathic scoliosis, an attempt was made to elucidate the natural history of the disease and to determine which factors contribute to curve progression. A total of 85,622 children were examined for scoliosis in a prospective school screening study carried out in northwestern and central Greece. Curve progression was studied in 839 of the 1,436 children with idiopathic scoliosis of at least 10 degrees detected from the school screening program. Each child was followed clinically and roentgenographically for one to four follow-up visits for a mean of 3.2 years. Progression of the scoliotic curve was recorded in 14.7% of the children. Spontaneous improvement of at least 5 degrees was observed in 27.4% of them, with 80 children (9.5%) demonstrating complete spontaneous resolution. Eighteen percent of the patients remained stable, while the remaining patients demonstrated nonsignificant changes of less than 5 degrees in curve magnitude. A strong association was observed between the incidence of progression and the sex of the child, curve pattern, maturity, and to a lesser extent age and curve magnitude. More specifically, the following were associated with a high risk of curve progression: sex (girls); curve pattern (right thoracic and double curves in girls, and right lumbar curves in boys); maturity (girls before the onset of menses); age (time of pubertal growth spurt); and curve magnitude (> or = 30 degrees). On the other hand, left thoracic curves showed a weak tendency for progression. In conclusion, the findings of the present study strongly suggest that only a small percentage of scoliotic curves will undergo progression. The pattern of the curve according to curve direction and sex of the child was found to be a key indicator of which curves will progress.