The purpose of this study was to compare ACTH levels in unstimulated samples obtained from the cavernous sinuses (CS) to unstimulated and CRH-stimulated samples obtained from the inferior petrosal sinuses (IPS) in 15 patients with surgically proven Cushing's disease. After unstimulated samples were obtained through 5-French catheters placed in both IPS, tracker catheters were introduced into both cavernous sinuses, and unstimulated samples were obtained within 5 min of the initial set. The Tracker catheters were removed, CRH was administered, and CRH-stimulated samples were obtained from the IPS. We compared the central to peripheral ACTH ratios in unstimulated samples from the cavernous sinuses to unstimulated and CRH-stimulated samples from the IPS as a basis for distinguishing pituitary from ectopic ACTH production. In addition, we compared the ability of the intercavernous and interpetrosal ACTH ratios to correctly predict the site of the microadenoma. Unstimulated levels of ACTH in the cavernous sinuses were generally higher than unstimulated levels of ACTH in the petrosal sinuses. However, 3 of 15 patients failed to show central to peripheral ACTH ratios greater than 2 in unstimulated samples from the cavernous sinuses and were thus falsely negative for the diagnosis of Cushing's disease (test sensitivity, 80%). By comparison, the test sensitivity for the petrosal sinus samples was 87% for the unstimulated samples and 100% for the CRH-stimulated samples. Lateralization was correct in 6 of 15 patients based on CS samples and in 9 of 15 patients based on IPS samples. Because of the 20% false negative rate, CS sampling without CRH stimulation is not recommended for the differential diagnosis of ACTH-dependent hypercortisolism.