Diurnal cortisol variation and cortisol response to an MRI stressor in schizophrenia and bipolar disorder

Markers of HPA axis function, including diurnal cortisol rhythm and cortisol responses to stress or pharmacological manipulation, are increasingly reported as disrupted in schizophrenia (SZ) and bipolar disorder (BD). However, there has been no direct comparison of cortisol responses to stress in SZ and BD in the same study, and associations between cortisol dysfunction and illness characteristics remain unclear. In this study we used spline embedded linear mixed models to examine cortisol levels of SZ and BD participants at waking, during the first 45min after waking (representing the cortisol awakening response; CAR), during the period of rapid cortisol decline post the awakening response, and in reaction to a stressor (MRI scan), relative to healthy controls (HC). Contrary to expectations, neither SZ nor BD showed differences in waking cortisol levels, CAR, or immediate post-CAR decline compared to HC; however, waking cortisol levels were greater in BD relative to SZ. In response to the MRI stressor, the SZ group showed a significant absence of the expected increase in cortisol responsivity to stress, which was seen in both the BD and HC groups. Clinical factors affecting the CAR differed between SZ and BD. In SZ, higher antipsychotic medication dosage was associated with a steeper incline of the CAR, while greater positive symptom severity was associated with a more blunted CAR, and greater levels of anxiety were associated with the blunted cortisol response to stress. In BD, longer illness duration was associated with a steeper incline in CAR and lower levels of waking cortisol. These results suggest that cortisol responses may normalize with medication (in SZ) and longer illness duration (in BD), in line with findings of aberrant cortisol levels in the early stages of psychotic disorders.