Lumbar-level axial section of the combined spinal–epidural technique in situ. It is of utmost clinical importance that the spinal needle not be advanced too far into the subdural space so that mechanical damage is done to the neural tissue by the needle point. At levels caudal to the cauda equina, the neural tissue may move upon contact with the needle, thus avoiding damage; however, this is not the case above the termination of the spinal cord. Clearly, the Tuohy needle (16-gauge) is much larger than the spinal needle (27-gauge). Hence, the punctures made in the dura and arachnoid mater by the respective needles will differ substantially, which may result in post–dural puncture headache; large-gauge epidural needles should not puncture the dura mater if possible

Spinal needle (top), epidural needle (center), and spinal needle inserted through the cannula of the epidural needle (bottom). The needles include centimeter distance markings, as well as a locking hub design, to ensure that the clinician is aware of the distance the spinal needle advances past the end of the epidural needle and that this may be fixed once the dura is breached. CSEcure Combined Spinal Epidural System, Portex (Reproduced courtesy of Smiths Medical, Ashford, Kent, United Kingdom)

Volunteer positioned ((a) sitting, (b) supine, and (c) laterally recumbent) in a Fonar Upright Multi-Position MRI scanner (Fonar Corporation, Melville, NY) with an antenna attached around the thorax. The volunteers were not limited by curvature of their backs in the sitting position, and they had more curved-back and flexed-head (head-down) positions during the measurements than the straight-backed position illustrated in this figure. fMRI scans are effective for imaging volunteers in postures resembling those in clinical practice during insertion of epidural and spinal needles (From Lee et al. [1]; with permission)

(a) Segmentation of the axial MRI images. This MRI image is from the actual dataset and shows the spinal cord and dural sac modeled using an ellipsoid approximation, whereas the measurements were taken automatically. The anterior dura mater is bordered by the posterior longitudinal ligament. 1 Spinal cord; 2 subarachnoid space with cerebrospinal fluid; 3 posterior longitudinal ligament/anterior dura mater; 4 posterior dura mater; 5 posterior epidural space; 6 flavum ligament; 7 neuroforamen, lateral dural space, sleeve region; 8 costa. (b) Magnified median MRI of the thoracic spine of a volunteer in the supine position. The arrows indicate the pertinent distances for neuraxial anesthesia. A skin to dura, B epidural space, C posterior dura to cord, D cord diameter, E anterior dura to cord, and F posterior dura to cord in needle path. Note that the craniocaudal slope of the spinal processes (analogous to roofing tiles) at the thoracic level increases the separation of the dura from the spinal cord relative to a horizontal approach. 1 Spinal cord, 2 anterior arachnoid space, 3 posterior arachnoid space, 4 anterior dura mater, 5 posterior dura mater, 6 posterior dura mater next to vertebral arch, 7, posterior longitudinal ligament, 8 posterior epidural space (fat, connective tissue, veins), 9 basivertebral veins(s), 10 flavum ligament, 11 interspinous ligament, 12 supraspinous ligament, 13 intervertebral disk (Panel a from Lee et al. [1]; with permission)