The diagnosis of adrenal haemorrhage complicating heparin therapy is often delayed, despite computed tomography (CT). Moreover, its pathogenesis is not clear. Adrenal haemorrhages are often seen in cases where there is no unduly excessive anticoagulation, and can be accompanied by a paradoxical thrombosis of the central adrenal vein. Symptoms usually occur within the first 8 to 12 days after starting heparin. The difficulty in establishing the diagnosis stems from the fact that symptoms are rather nonspecific: abdominal pain and backache, nausea, vomiting, lethargy, weakness, hypotension, hyperpyrexia. To confirm the diagnosis, both hormonal proof of adrenal failure and anatomic evidence of haemorrhage must be found. Early CT scans may show the haemorrhage. Several possible causes have been put forward to account for these adrenal haemorrhages. The degree of anticoagulation did not seem to be a prerequisite, 30 to 50% of patients showing no evidence of other bleeding or coagulation tests outside the therapeutic range. Capillary fragility of old age might be a factor. Stress would seem to be an important factor predisposing to adrenal haemorrhage. Many authors consider the paradoxical central vein thrombosis as a result of the haemorrhage rather than its cause, whereas other conclude the opposite. Unfortunately, to date coagulation studies are often incomplete; platelet counts were missing in most reports published before 1985. Since that date, a heparin induced thrombosis-thrombocytopaenia syndrome (HITTS), in which thrombosis may occur in any vascular bed, has been recognized with increasing frequency. Nine cases of adrenal haemorrhage associated with HITTS have been reported. It seems highly likely that a proportion of cases of heparin-related adrenal destruction are due to HITTS.(ABSTRACT TRUNCATED AT 250 WORDS)