CSF samples from ten healthy volunteers and 22 patients with major depression were collected by lumbar puncture at 9 a.m. and the content of monoamine metabolites, corticotropin releasing factor (CRF) and somatostatin (SRIF) was analyzed. Plasma concentrations of TSH following a TRH challenge test (200 micrograms) and plasma cortisol following dexamethasone (1 mg; DST) were also analyzed. No relationships were observed between the CRF or SRIF concentrations and either basal or post-dexamethasone cortisol concentrations. Fourteen of 21 depressed patients were DST nonsuppressors using a plasma cortisol concentration cut off point greater than or equal to 138 nmol/l. If a more conservative cut off point was used (greater than 290 nmol/l) seven out of 21 patients revealed a severity-related cortisol nonsuppression. No significant difference was observed between healthy volunteers and depressed patients with regard to TSH response to TRH. The CSF content of CRF was elevated and the content of SRIF reduced in the depressed patients. In the healthy volunteers an inverse relationship was observed between CSF concentrations of CRF and MHPG (r = -0.72; P = 0.019); no relationship was observed between the concentrations of CRF and 5-HIAA or HVA. In the depressed patients positive correlations were found between CSF concentrations of CRF and 5-HIAA (r = 0.59; P = 0.004) and between CRF and HVA (r = 0.44; P = 0.042). These data are concordant with the view that norepinephrine and serotonin may be involved in the regulation of CRF secretion.(ABSTRACT TRUNCATED AT 250 WORDS)