Bilateral inferior petrosal sinus sampling (BIPSS) is the most reliable procedure for distinguishing Cushing's disease from ectopic ACTH secretion. However, it is less reliable at predicting the lateralization of the pituitary corticotroph microadenoma. We sought to determine whether this could be improved by taking into account the pattern of venous drainage and the precise location of the catheters. We retrospectively studied data from 86 patients who underwent BIPSS. Cushing's disease was predicted in 74 patients, of whom 69 underwent transsphenoidal surgery. Surgical cure was obtained in 65 patients, with identification of a corticotroph microadenoma in 58 cases. In 49 patients the location of the microadenoma predicted by the intersinus ACTH gradient could be compared with the pathologist's data. BIPSS accurately predicted the lateralization of the microadenoma in only 57% of these patients. Prediction was improved to 71% when both venograms and catheters were symmetric (35 patients). In this subgroup accuracy was 86% in patients with both catheters in the inferior petrosal sinuses compared with 50% in patients with both catheters in the cavernous sinuses (CS). Two transient sixth nerve palsies occurred during CS catheterization. Our data suggest that BIPSS results are much improved when venous drainage is symmetric. Catheterization of CS did not improve the results and was less safe.