Intraforaminal location of thoracolumbar anterior medullary arteries.

Background: Transforaminal epidural steroid injection (TFESI) is a widely utilized interventional pain technique for radicular pain. Although the six o'clock position of the pedicle in the so-called "safe triangle" has been used as a target location, there have been a number of reported catastrophic complications of this procedure, including paraplegia. The mechanism of this has been attributed to the intravascular injection of steroids. The goal of this study was to examine the intraforaminal location of thoracolumbar medullary arteries which would help guide pain physicians in developing safer techniques and guidelines.

Methods: Twenty-four (24) embalmed cadavers were dissected and examined for the presence and distribution of thoracolumbar anterior medullary arteries. Access to the anterior surface of the spinal cord was made via anterior corpectomy from C2 to S5. Each medullary artery's course was determined by dissection from its origin, the anterior spinal artery, through the intervertebral foramen. The foramen was subsequently opened in the coronal plane, and the intraforaminal location of the artery, its diameter, and its relation to other foraminal structures were examined and measured.

Results: In the thoracolumbar foramina (T4-L2), 39 anterior medullary arteries were found, including 23 great medullary arteries (Adamkiewicz artery). One Adamkiewicz artery was found to be located in the left S2 foramen and was not included in the statistical analysis. Of the analyzed 39 anterior medullary arteries, 29 (74%) were located in the upper 1/3 of the foramen, 9 (23%) were located in the middle, and 1 (3%) artery was located in the lower 1/3. In relation to the dorsal root ganglion--ventral root complex, 21 (54%) arteries were located anterosuperiorly, 16 (41%) anteriorly, and 2 (5%) anteroinferiorly. The average intraforaminal artery diameter was 1.20 mm (0.84-1.91 mm). At thoracolumbar levels, the artery is almost always (92% ± 15%) located anterosuperior to the nerve. At typical thoracic levels, it is less often anterosuperior (38% ± 19%), but more often anterior to the nerve.

Conclusions: At thoracolumbar levels, if needles were to encounter an artery, they are most likely to do so if placed anterosuperior to the nerve. Encountering an artery anterosuperior to the nerve is less likely at typical thoracic levels, but the likelihood is far from negligible. Pain physicians should be cognizant of this when considering optimal needle placement during transforaminal epidural steroid injections.