Stressor-cortisol concordance among individuals at clinical high-risk for psychosis: novel findings from the NAPLS cohort

Abstract

Whilst elevations in basal cortisol levels have been reported among individuals at-risk for psychosis, the extent to which this represents hyperresponsivity of the hypothalamic-pituitary-adrenal (HPA) axis to psychosocial stressors encountered in the natural environment is currently unclear. We aimed to examine stressor-cortisol concordance among youth at clinical high-risk (CHR) for psychosis in the North American Prodrome Longitudinal Study 2 (NAPLS 2) and the relationship with clinical outcome. At baseline, CHR (N = 457) and healthy (N = 205) individuals provided salivary cortisol samples and completed daily stressor, life event, and childhood trauma measures. CHR youth were categorised as remitted, symptomatic, progression of positive symptoms, or psychosis conversion at the two-year follow-up. Within-group regression models tested associations between psychosocial stressors and cortisol; standardised beta coefficients (Stβ) were subsequently derived to enable within-group pooling of effect sizes across stressor types. After adjustment for potential confounders, all CHR subgroups reported greater exposure to life events and daily stressors, and more distress in relation to these events, relative to controls. All CHR groups were also more likely to experience childhood trauma; only CHR converters, however, were characterised by elevated basal cortisol. Daily stressor distress was significantly associated with cortisol in controls (β = 0.60, 95% CI: 0.12-1.08) and CHR youth who converted to psychosis (β = 0.91, 95% CI: 0.05-1.78). In controls only, life event exposure was associated with cortisol (β = 0.45, 95% CI: 0.08-0.83). When pooled across stressors, stressor-cortisol concordance was substantially higher among CHR converters (Stβ = 0.26, 95% CI: 0.07 to 0.44) relative to CHR progressed (Stβ = 0.02, 95% CI: -0.11 to 0.15), symptomatic (Stβ = 0.01, 95% CI: -0.11 to 0.12), and remitted groups (Stβ = 0.00, 95% CI: -0.13 to 0.13); however, unexpectedly, healthy controls showed intermediate levels of concordance (Stβ = 0.15, 95% CI: 0.05 to 0.26). In conclusion, whilst all CHR subgroups showed increased psychosocial stress exposure and distress relative to controls, only those who later converted to psychosis were characterised by significantly elevated basal cortisol levels. Moreover, only CHR converters showed a higher magnitude of stressor-cortisol concordance compared to controls, although confidence intervals overlapped considerably between these two groups. These findings do not support the notion that all individuals at CHR for psychosis show HPA hyperresponsiveness to psychosocial stressors. Instead, CHR individuals vary in their response to stressor exposure/distress, perhaps driven by genetic or other vulnerability factors.