DHEAS and cortisol/DHEAS-ratio in recurrent depression: State, or trait predicting 10-year recurrence?

BACKGROUND Major depressive disorder (MDD) has been associated with low dehydroepiandrosterone-sulphate (DHEAS), - particularly relative to high cortisol - although conflicting findings exist. Moreover, it is unclear whether low DHEAS is only present during the depressive state, or manifests as a trait that may reflect vulnerability for recurrence. Therefore, we longitudinally tested whether low DHEAS and high cortisol/DHEAS-ratio in recurrent MDD (I) reflects a trait, and/or (II) varies with depressive state. In addition, we tested associations with (III) previous MDD-episodes, (IV) prospective recurrence, and (V) effects of cognitive therapy. METHODS At study-entry, we cross-sectionally compared morning and evening salivary DHEAS and molar cortisol/DHEAS-ratio of 187 remitted recurrent MDD-patients with 72 matched controls. Subsequently, patients participated in an 8-week randomized controlled cognitive therapy trial. We repeated salivary measures after 3 months and 2 years. We measured clinical symptoms during a 10-year follow-up. RESULTS Remitted patients showed steeper diurnal DHEAS-decline (p<.005) and a flatter diurnal profile of cortisol/DHEAS-ratio (p<.001) than controls. We found no state-effect in DHEAS or cortisol/DHEAS-ratio throughout follow-up and no association with number of previous episodes. Higher morning cortisol/DHEAS-ratio predicted shorter time till recurrence over the 10-year follow-up in interaction with the effects of cognitive therapy (p<.05). Finally, cognitive therapy did not influence DHEAS or cortisol/DHEAS-ratio. CONCLUSIONS Diurnal profiles of DHEAS and cortisol/DHEAS-ratio remain equally altered in between depressive episodes, and may predict future recurrence. This suggests they represent an endophenotypic vulnerability trait rather than a state-effect, which provides a new road to understand recurrent depression and its prevention. TRIAL REGISTRATION www.isrctn.com/ISRCTN68246470.