Background: Treatment of children with classic congenital adrenal hyperplasia (CAH) with glucocorticoids is a difficult balance between hypercortisolism and hyperandrogenism. Biochemical monitoring of treatment is not well defined. Achievement of a normal growth rate is the most important therapeutic goal.

Methods: We retrospectively evaluated 123 24-hour gas chromatography-mass spectrometry urinary steroid metabolome analyses together with their corresponding 1-year height velocity (HV) z scores in 63 prepubertal children aged 7.2 +/- 1.6 years with classic CAH due to 21-hydroxylase deficiency treated with hydrocortisone and fludrocortisone.

Results: Multivariate linear mixed effects model analysis revealed a positive influence of CAH-specific z scores of summed urinary androgen metabolites (B = 0.97 +/- 0.20, t = 4.87, P< 0.0001) and a negative influence of the cortisol metabolite tetrahydrocortisol (B = -1.75 +/- 0.79, t = -2.20, P = 0.03) on HV z scores. Receiver operating characteristic analysis demonstrated that adrenal androgen excess, defined as HV >1.5 z, was best determined by a z score of all urinary androgen metabolites of >0.512 [accuracy, 66.2%; sensitivity, 57.1%; specificity, 74.4%; positive prediction value (PPV), 66.7%; negative prediction value (NPV), 65.9%]. Tetrahydrocortisol excretion >1480 mu g/m(2) BSA/d in conjunction with suppressed urinary androgen metabolites <0.163 z indicated overtreatment, defined as HV < -1.5 z (accuracy, 79.6%; sensitivity, 40.0%; specificity, 94.9%; PPV, 75.0%; NPV, 80.4%).

Conclusion: We established target values for urinary steroid metabolite excretions in children with CAH based on their growth rate. Urinary steroid metabolome analysis represents a highly suitable method for monitoring metabolic control in children with CAH.