The best test for the assessment of the hypothalamic-pituitary-adrenal (HPA) axis remains a matter of controversy. We compared the performance of the short synacthen test (SST, 250 mg) with the insulin stress test (IST) to assess the reliability of the former as a first line test. Patients with pituitary disease underwent both the SST and the IST. The results in patients who had both tests within 4 weeks of each other, and where these were not separated by a therapeutic intervention, were compared. Basal, 30 and 60 min cortisol levels were obtained from the SST. Basal and maximal cortisol level after adequate hypoglycaemia (glucose < 2.2 mmol/l) were recorded for the IST.Sixty-nine paired test results were available for analysis. With a 30 min 'pass' plasma cortisol value of 500 nmol/l on the SST, 7/69 (10%) patients who passed the SST failed the IST set at a 'pass' maximum value of 500 nmol/l. At a 'pass' cortisol value of 600 nmol/l on the SST, 3/69 (4%) who passed the SST failed the IST. Assuming the IST as the gold standard, the sensitivity of an SST 'pass' of 600 nmol/l is 85% with a specificity of 96%.During the conventional dose SST (250 mg) a 30 min plasma cortisol value of 600 nmol/l is more reliable than a value of 500 nmol/l, and using the former criterion the SST can safely be used as a first line test for the evaluation of the HPA axis in patients with pituitary disease. However, if the result is borderline or there is clinical suspicion of mild hypocorticotrophism an IST or other test of the HPA axis may be warranted.