OBJECTIVE: To report the 5-year experience of a diagnostic programme for adrenal incidentaloma with special emphasis to diagnose hormonally active and malignant lesions. DESIGN: A prospective study in which new cases of adrenal incidentalomas in Sweden have been evaluated by a standardized diagnostic protocol between January 1996 and July 2001. SETTING: Thirty-three different Swedish hospitals have contributed with cases. SUBJECTS: A total of 381 patients (217 females, 164 males) with adrenal incidentalomas were studied. INTERVENTIONS: Diagnostic procedures were undertaken according to a standardized programme. Operation was recommended if the incidentaloma had a size of more than 3-4 cm or if there was a suspicion of a hypersecreting tumour. MAIN OUTCOME MEASURES: The size of the incidentaloma, clinical characteristics of the patients and results of biochemical diagnostic tests were registered. RESULTS: The median age of the patients was 64 years (18-84 years), and the median size of the incidentalomas was 3 cm (1-20 cm). A total of 85(22%) patients were operated. Twenty of these patients were diagnosed with a benign hypersecreting tumour and 14 with a malignant tumour. Fourteen of 15 operated patients with diagnosed pheochromocytoma had elevated 24-h urinary noradrenaline and all of the patients operated because of a biochemical suspicion of aldosterone or cortisol hypersecretion (n = 6) were found to have adrenal adenomas. Of the 14 operated patients with malignant diseases, 10 were adrenal carcinomas (median size 10 cm; range 4-16 cm). In a multiple logistic regression model, incidentaloma size was significantly associated with the risk of a malignant tumour (P = 0.009), and there was a tendency of an association between age/male sex and the risk of a malignancy (both, P = 0.07). CONCLUSION: In this Swedish multicentre study of 381 cases with adrenal incidentalomas, 5% had benign hypersecreting tumours and nearly 4% had malignant tumours. The results of the biochemical diagnostic tests used had a high compatibility with the histological diagnosis found at operation in the patients with hypersecreting tumours. Tumour size, male gender and high age were predictive for the risk of a malignant tumour. A follow-up of the patients is warranted in order to establish whether there are undiscovered cases of malignant or hypersecreting tumours amongst the nonoperated patients.