[No Title]

The study of Strickland et al (2002) underpins the considerable inconsistency in the literature that addresses the area of peripheral markers in depression.

It has been argued that neuroendocrine challenge tests (such as the prolactin response to dexfenfluramine used by the authors) is not a valid probe of central neuronal function (for discussion see Weiss & Coccaro, 1997) and this may account for the negative findings of the study. More perplexing is the lack of association between depression and a reliable index of hypothalamic—pituitary—adrenal (HPA) axis activation, late-night salivary cortisol (reviewed by Kirschbaum & Hellhammer, 1994). In both humans and animals various models of acute and chronic stress (e.g. physical trauma, public speaking, caregiver stress in carers of people with Alzheimer's disease) are reliably associated with hyper-cortisolaemia (Kirschbaum & Hellhammer, 1994). If depression is considered to be an extreme form of chronic stress, why is there so little consistency between studies examining cortisol in populations with depression (Haskett, 1993)?

Strickland et al may have serendipitously discovered a crucial, if seemingly trivial, psycho-biological ‘co-factor’ in depression that dramatically distinguishes between cases with or without HPA axis activation: perceived stress (as measured by the Life Events and Difficulties Schedule (LEDS), see Strickland et al; p. 170, Fig. 1c). Equally depressed patients may not be equally ‘stressed’ and this may have biological as well as clinical consequences. The increased cardiac (Carney et al, 1997) and oncological (Persky et al, 1987) morbidity and mortality associated with depression may particularly apply to the depressed—stressed—hypercortisolaemic sub-group. Clearly more research is needed to explore this possibility.