The longitudinal growth pattern during the first 36 months of life was studied in 24 patients (17 females) with congenital adrenal hyperplasia (CAH) due to 21-hydroxylase deficiency by analyzing the mean required daily dose of cortisone with respect to steroid suppression, height and weight growth velocities and bone age maturation. All patients were treated with cortisone acetate and 9-fluorohydrocortisone. The standard deviation score for length (SDS-L), the percentage of ideal body weight (% IBW) and biochemical parameters, 17-hydroxy-progesterone (17-OHP) and androstenedione (A) were evaluated every 3 months; bone age (BA) was evaluated annually. At diagnosis, the female population of patients with respect to the males were younger (chronological age (CA): 15 +/- 14 vs. 45 +/- 16 days, p < 0.005) and had a higher % IBW (91.7 +/- 8.0 vs. 76.3 +/- 16.7%, p < 0.05). At 3 months of age (45 days after initiating treatment) the % IBW in males normalized (97 +/- 19%) and was similar to that found in females (101 +/- 12.8%). No differences were noted in SDS-L at the moment of diagnosis (females-1.1 +/- 1.1 vs. males -0.5 +/- 0.7); however, at 3 months of age the SDS-L in females increased (0.41 +/- 0.88, p < 0.005 vs. diagnosis SDS-L) whereas that of males progressively decreased to reach the nadir at 6 months (-1.41 +/- 0.96). No differences between males and females were noted throughout this time with regard to: (a) A or 17-OHP levels (neither of which were suppressed to 'control values'); (b) the dosage of cortisone received (13.5-17.8 mg/m2/day), and (c) change in BA/CA ratio. In all patients the SDS-target height (TH) correlated with the SDS-L at 2 years (r = 0.74, p < 0.0005) and at 3 years (r = 0.60, p < 0.02) of age. In 12 patients who reached 7 years of age the SDS-L correlated with both SDS-predicted adult height (PAH) (r = 0.75, p < 0.001) and SDS-TH (r = 0.80, p < 0.005). Although the commonly accepted definition of "good control' for patients with CAH has generally included, in addition to adequate suppression of hormone markers, normal growth and skeletal maturation, the present data suggest that normal growth and BA maturation are the most useful parameters to follow and not necessarily strive for hormone suppression. Early diagnosis and replacement therapy using cortisone doses less than those currently recommended allow normal growth within the genetic potential at least for the first 7 years of life.