Bilateral sampling of the inferior petrosal sinuses (IPSS) to distinguish Cushing's disease from the ectopic ACTH syndrome is accurate but risky and technically difficult. Bilateral sampling of the internal jugular vein (JVS) is simpler and presumably safer. To compare jugular and petrosal sinus venous sampling for distinguishing Cushing's disease from ectopic ACTH syndrome, we studied 74 patients with surgically proven Cushing's disease, 11 with surgically confirmed, and three with occult ectopic ACTH secretion. Patients underwent JVS and IPSS with administration of CRH on separate days. Ratios of central-to-peripheral ACTH in venous samples were calculated. At 100% specificity, IPSS correctly identified 61 of 65 patients with Cushing's disease [sensitivity, 94%; confidence interval (CI), 84-98%]. When patients with abnormal venous drainage were excluded, sensitivity was 98% (CI, 90-100%). JVS had a sensitivity of 83% (CI, 71-91%) at 100% specificity. Receiver operated characteristics plot areas under the curve were similar (0.968 +/- 0.020 and 0.974 +/- 0.016, area under the curve +/- se, JVS vs. IPSS). Although petrosal sampling had better diagnostic accuracy, CIs overlapped (95% CI, 90-100% vs. 86% CI, 78-94%). Centers with limited sampling experience may choose to use the simpler JVS and refer patients for IPSS when the results are negative.