Abstract
Glucocorticoid replacement therapy is the mainstay of treatment for congenital adrenal hyperplasia (CAH) but has a narrow therapeutic index and dose optimisation is challenging. Metabolomic profiling was carried out on plasma samples from 117 adults with 21-hydroxylase deficiency receiving their usual glucocorticoid replacement therapy who were part of the CaHASE study. Samples were profiled by using hydrophilic interaction chromatography with high resolution mass spectrometry. The patients were also profiled using nine routine clinical measures. The data were modelled by using both multivariate and univariate statistics by using the clinical metadata to inform the choice of patient groupings. Comparison of 382 metabolites amongst groups receiving different glucocorticoid doses revealed a clear distinction between patients receiving ≤5 mg (n = 64) and >5 mg (n = 53) daily prednisolone-equivalent doses. The 24 metabolites which were statistically significantly different between groups included free fatty acids, bile acids, and amino acid metabolites. Using 7 metabolites improved the receiver operating characteristic with area under the curve for predicting glucocorticoid dose of >0.9 with FDR adjusted P values in the range 3.3 E-04 -1.9 E-10. A combination of seven plasma metabolite biomarkers readily discriminates supraphysiological glucocorticoid replacement doses in patients with CAH.