Steroid Metabolomic Disease Signature of Nonsyndromic Childhood Obesity.

CONTEXT: The profile of urinary steroids as measured by gas chromatography-mass spectrometry defines a subject's "steroidal fingerprint."

OBJECTIVE

Here, we clustered steroidal fingerprints to characterize patients with nonsyndromic childhood obesity by "steroid metabolomic signatures."

HYPOTHESIS: Nonsyndromic obesity is a symptom of different diseases and conditions, some of them will have their own signature.

DESIGN

A total of 31 steroid metabolites were quantified by gas chromatography-mass spectrometry, and their excretion rates were z-transformed. Using MetaboAnalyst 3.0, we divided the subjects into 5 distinctive groups by k-means clustering. Steroidal fingerprints and clinical/biochemical data of patients in each cluster were analyzed.

PATIENTS: A total of 87 obese children (44 females), aged 8.5-17.9 years, were clinically characterized, and their 24-hour urine was collected.

RESULTS

Cluster 1 (n = 39, 21 females) had normal steroid profile. Cluster 2 (n = 20, 11 females) showed mild, nonspecific elevation of C19 and C21 steroids, females' resistance to polycystic ovary morphology, and hirsutism. Cluster 3 (n = 7 female), with relative 21-hydroxylase insufficiency, was characterized by partial or full polycystic ovary syndrome. Cluster 4 (n = 4 males), showed markedly elevated C21 steroids and imbalance in the 11β-hydroxysteroid dehydrogenase system, higher insulin, increased frequency of glucose/insulin index more than 0.3, γ-glutamyl transpeptidase activity, systolic blood pressure, and tendency to liver steatosis. Cluster 5 (n = 17, 5 females) had elevated dehydroepiandrosterone and 17-OH-pregnenolone metabolites, suggesting 3β-hydroxysteroid dehydrogenase insufficiency but no clinically unique phenotype. Z-score body mass index values were not significantly different between the clusters.

CONCLUSIONS

We defined a novel concept of disease-specific steroid metabolomic signature based on urinary steroidal gas chromatography-mass spectrometry. Clustering by software designed for metabolic data analysis reclassified childhood obesity into 5 groups with distinctive signatures; groups require further definition and may require cluster-specific therapeutic strategies.