Background: Epidural needle insertion is usually a blind technique where the rate of adverse events depends on the experience of the operator. A novel ultrasound method to guide epidural catheter insertion is described. Methods: An ultrasound transducer (40 MHz, a -6 dB fractional bandwidth of 50%) was placed into the hollow chamber of an 18-gauge Tuohy needle. The single crystal was polished to a thickness of 50 μm, with a width of 0.5 mm. Tissue planes were identified from the reflected signals in an A-mode display. The device was inserted three times into both the lumbar and thoracic regions of five pigs (average weight, 20 kg) using a paramedian approach at an angle of 35-40°. The epidural space was identified using signals from the ligamentum flavum and dura mater. Epidural catheters were placed with each attempt and placement confirmed by contrast injection. Results: The ligamentum flavum was identified in 83.3% of insertions and the dura mater in all insertions. The dura mater signal was stronger than that of the ligamentum flavum and served as a landmark in all epidural catheter insertions. Contrast studies confirmed correct placement of the catheter in the epidural space of all study animals. Conclusions: This is the first study to introduce a new ultrasound probe embedded in a standard epidural needle. It is anticipated that this technique could reduce failed epidural blocks and complications caused by dural puncture.