OBJECTIVE We have previously reported on 30 patients with Cushing's syndrome suggesting an aetiological role for stressful life events. The investigation about life events in the year before the first signs of disease onset was extended to a larger population of patients with Cushing's syndrome, allowing us to differentiate patients with pituitary-dependent and pituitary-independent forms. METHODS Case-control study. METHODS Sixty-six consecutive patients with Cushing's syndrome of various aetiologies (46 with pituitary-dependent forms and 20 with primary adrenal hyperfunction or ectopic ACTH production) and a control group of 66 healthy subjects, matched for sociodemographic variables, were studied. METHODS Paykel's Interview for Recent Life Events (a semistructured research interview covering 64 life events) was administered after the acute phase of illness while in remission. RESULTS Patients with Cushing's syndrome reported significantly more stressful life events (P < 0.001), both events that had an objective negative impact (P < 0.001) and independent events (P < 0.001), than controls, confirming previous findings. Patients with pituitary-dependent Cushing's disease were compared with their matched controls and reported significantly more total events, events with an objective negative impact and independent events (all at P < 0.001). There were no significant differences between patients with pituitary-independent forms and their matched controls. CONCLUSIONS These findings indicate a causal role for stressful life events exclusively in pituitary-dependent Cushing's disease, and suggest a limbic-hypothalamic involvement in the pathogenesis of this condition. The results are similar to those obtained in major depression, and add to other analogies between the two disorders.