Longitudinal proneuroactive and neuroactive steroid profiles in medication-free women with, without and at-risk for perinatal depression: a liquid chromatography-tandem mass spectrometry analysis.

Abstract

Background Neuroactive steroids (NAS) are derivatives of cholesterol or steroidal precursors made in the gonads, adrenal gland, placenta and brain. We characterized longitudinal plasma proneuroactive and NAS in healthy perinatal comparison women (HPCW), women at-risk for perinatal depression (AR-PND), and women with PND with/without comorbid anxiety. We hypothesized that AR-PND women who either did or did not go on to develop PND would have elevated NAS concentrations as compared to HPCW and that NAS would be correlated to depressive and anxiety symptoms.Methods A prospective cohort study evaluated 75 medication-free perinatal women (HPCW, n = 30; AR-PND, n = 19; PND, n = 26). Standardized depression and anxiety assessments and blood samples were completed across 5 visits. Structured Clinical Interviews for DSM-IV TR Disorders were administered at study entry and exit. Plasma pregnenolone, progesterone, 5α- and 5β-dihydroprogesterone, pregnanolone, allopregnanolone, deoxycorticosterone and tetrahydrodeoxycorticosterone were quantified by liquid chromatography-tandem mass spectrometry. Longitudinal relationships between risk-group, depression and anxiety symptoms, and NAS concentrations were analyzed using generalized estimating equations to control for repeated measures correlations.Results Perinatal 5α-dihydroprogesterone, 5β-dihydroprogesterone, allopregnanolone, deoxycorticosterone, and tetrahydrodeoxycorticosterone concentrations were higher in AR-PND and PND women compared to HPCW (β = 3.57 ± 1.40 and β = 2.11 ± 1.12, p = 0.03; β = 0.18 ± 0.06 and β = 0.03 ± 0.05, p = 0.02; β = 1.06 ± 0.42 and β = 1.19 ± 0.47, p = 0.01; β = 0.17 ± 0.07 and β = 0.11 ± 0.06, p = 0.05; β = 0.03 ± 0.01 and β = 0.03 ± 0.01, p = 0.05, respectively). Perinatal allopregnanolone, 5α-dihydroprogesterone and tetrahydrodeoxycorticosterone were positively associated with HAM-D 17 (all p < 0.02). HAM-A was positively associated with 5α- and 5β-dihydroprogesterone, pregnanolone, allopregnanolone, deoxycorticosterone and tetrahydrodeoxycorticosterone (all p < 0.05). A history of depression was associated with increased 5α-dihydroprogesterone (2.20 ± 1.09, p = 0.05), deoxycorticosterone (0.13 ± 0.06, p = 0.03) and tetrahydrodeoxycorticosterone (0.03 ± 0.01, p = 0.02).Conclusion To our knowledge, this study represents the largest prospective study of 5-α and 5-β reductase products of progesterone and deoxycorticosterone in HPCW and women AR-PND. Data suggest that PND is associated with both a reduction of progesterone to 5β-dihydroprogesterone, 5α-dihydroprogesterone, and allopregnanolone, and the 21-hydroxylation to deoxycorticosterone and tetrahydrodeoxycorticosterone. The shift towards 5α-dihydroprogesterone, deoxycorticosterone and tetrahydrodeoxycorticosterone was associated with a history of depression, a significant risk factor for PND.