In July 1998, Cortef oral suspension (Pharmacia & Upjohn) was reformulated changing the suspending agent tragacanth to xanthan gum. We subsequently observed suboptimal control of hormone levels in a group of children with classic congenital adrenal hyperplasia, despite increasing doses of Cortef suspension and stringent instructions to parents regarding shaking of the bottles of medication. Nineteen children receiving Cortef and fludrocortisone therapy were changed to hydrocortisone tablets and fludrocortisone, with a 10 percent reduction in hydrocortisone dose. A significant decrease in 17-hydroxyprogesterone (235 +/- 120 vs. 27 +/- 7 nmol/L; p</=0.001) and androstenedione (18.9 +/- 18.0 vs. 3.5 +/- 3.5 nmol/L; p=0.002) was observed 4-6 weeks later. Twenty-one percent (4/19) had 17-hydroxyprogesterone and androstenedione levels at or below the detection limit of the assay. Despite a significant reduction in glucocorticoid dose (19.6 +/- 4.7 vs. 17.6 +/- 3.9 mg/M(2)/day; p<0.001), eight children experienced significant weight gain and appetite increase, three experienced trouble sleeping, four experienced moodiness, and three developed hypertension requiring a decrease in fludrocortisone therapy. Hydrocortisone dose was further decreased to 15.2 +/- 2.6 mg/M(2)/day with resolution of symptoms. We conclude that Cortef suspension and hydrocortisone tablets are not bioequivalent and the reformulated form of hydrocortisone oral suspension was inadequate in the control of children with congenital adrenal hyperplasia. Cortef suspension has been recalled as a result of these data.