To assess the effect of anxiety on response to the Dexamethasone Suppression Test (DST), cortisol concentrations were determined in patients who had various diagnoses, with anxiety as a secondary characteristic. Saliva was collected before and after venepuncture at 4:05 PM following completion of the Leeds Questionnaire at 3:00 PM. Matrix effects, which caused an initial artefactual decrease in cortisol levels in some saliva samples in patients with high anxiety, could be eliminated by repeated freezing/thawing. There was no significant difference between salivary cortisol concentrations before and after venepuncture, indicating that variability in response to the DST is not a correlate of anxiety and stressful venepuncture. There was no association between anxiety scores and plasma or salivary cortisol values: thus, anxiety is unlikely to be a major contributory cause of nonsuppression of hypercortisolemia in the DST.