Secondary Logo

Journal Logo

Regional Anesthesia: Brief Report

The Anatomy of the Thoracic Spinal Canal Investigated with Magnetic Resonance Imaging

Imbelloni, Luiz Eduardo MD*,†; Quirici, Marcelo Bianco MD; Ferraz Filho, Jose Roberto MD§; Cordeiro, José Antonio PhD; Ganem, Eliana Marisa PhD

Author Information
doi: 10.1213/ANE.0b013e3181d5aca6

The anatomical characteristics between the thoracic and lumbar spine are significantly different. The anatomy of the thoracic spinal canal was recently investigated with magnetic resonance imaging (MRI) in 19 patients.1 In 113 thoracic epidural blocks performed in anesthetized patients, there was a 4.4% occurrence of accidental puncture of the dura mater without neurological sequelae.2 In our study, we retrospectively investigated the distance of the dura mater to the spinal cord at the 3 thoracic segments, analyzing the magnetic resonance images of patients without spinal or medullar disease.


The hospital's ethical committee approved our retrospective study. Between January 2001 and July 2008, 1232 MRI examinations were performed at the Hospital de Base de São José do Rio Preto. A group of 50 (Table 1) healthy patients, in supine position, were imaged with the MRI 1.5-T superconducting system scanner (Gyroscan Intera, Philips Medical Systems, Best, the Netherlands). Thoracic spine images were acquired through sagittal spin echo at the 2nd, 5th, and 10th thoracic segments (Fig. 1), and the perpendicular distance from the posterior dura to cord was measured. In 10 patients, the angle from the skin to the subarachnoid space was determined using a compass on the printed magnetic resonance image.

Table 1
Table 1:
Patients Characteristics
Figure 1
Figure 1:
Magnetic resonance imaging of the spinal column.

Comparisons of the mean dura–spinal cord distance were performed by paired t test, and between genders by 2-sample t test. For multiple comparisons of segments, i.e., all comparisons of 2 segments, the Bonferroni correction was applied through multiplication of the actual P value by 3 (Bonferroni P value). The adopted significance level was α = 0.05, which means that multiple comparisons were significant when the actual P value ≤0.05/3 = 0.0167.


The posterior dural–spinal cord distance is significantly greater at the middle thoracic region (5th thoracic segment = 5.8 ± 0.8 mm) than at the upper (2nd thoracic segment = 3.9 ± 0.80 mm) and lower thoracic levels (10th thoracic segment = 4.1 ± 0.1 mm) (Table 2) (P < 0.015). The T2 and T10 distances were not significantly different. There is evidence of correlations between the measured distance at T2 and both T5 and T10 (same r = 0.5; P < 0.001), and between T5 and T10 (r = 0.4; P = 0.004). There is no correlation between patient age or gender and the measured distance between the dura mater and the spinal cord (Table 3).

Table 2
Table 2:
Minimum, Mean (SD), and Maximum (in mL) of Sagital Distance to the Cord (SDC) and Needle Path
Table 3
Table 3:
Minimum Distance (mm) of the Dura Mater to the Spinal Cord According to Gender

Based on subsamples of 6 women and 4 men, the angle relative to the tangent at the insertion point on the skin showed significant differences between T5 (45.0° ± 7.4°) and T2 (9.0° ± 2.5°) and T10 (9.5° ± 4.2°) (Bonferroni P < 0.0015). The T2 and T10 angles were not significantly different (Table 2). Dividing the cord distance by the cosine of the angle, the length of the needle trajectory was calculated. At T5 (8.2 ± 2.1 mm), the length of the trajectory was more than at T2 (3.8 ± 0.9 mm) and T10 (4.2 ± 0.6 mm) (Bonferroni P < 0.0015). The T2 and T10 needle trajectory lengths were not significantly different. There was no correlation between gender and the angle of insertion or length of needle trajectory.


This study confirmed that there was a greater depth of the posterior subarachnoid space at midthoracic levels (T5 = 5.8 mm) than upper (T2 = 3.9 mm) and lower (T10 = 4.1 mm) thoracic levels. The distance at midthoracic levels (T5) with a 45° angle of insertion also increased this distance. Our results confirm those of a previous study1 that found a greater depth of the posterior subarachnoid space at midthoracic levels than at lumbar and upper thoracic levels.

In this study, we investigated measurements with the patient positioned supine. However, most neuraxial blockades are performed with the patient in a lateral decubitus or a sitting position. Previous studies have demonstrated that the spinal cord and cauda equina move with gravity and also ventrally with leg flexion.35 It is noteworthy that the spinal cord lies approximately most anterior at the apex of the thoracic curve. Considering the geometry, it is expected that with the patient in a lateral or sitting position with exaggerated curvature of the back, the cord would tend to lie even further anteriorly.

Because of the risk of a medullary lesion caused by needle-tip injury, the distance between the dura mater and the cord is of great importance. This space may prevent lesion formation during accidental dural puncture in thoracic and cervical epidural anesthesia or thoracic combined spinal epidural anesthesia.68

This study demonstrated that there is a great distance from the dura mater to the spinal cord at the T2, T5, and T10 levels, with the greatest distance at T5. The main finding of our study is that using a 45° angle when accessing the midthoracic epidural space will help protect the spinal cord against accidental dural perforation.


This article is a portion of a thesis submitted by the first author to the Faculdade de Medicina, Universidade Estadual Paulista as partial fulfillment of the requirements for a PhD degree.


1. Lee RA, van Zundert AAJ, Breedveld P, Wondergem JHM, Peek D, Wieringa PA. The anatomy of the thoracic spinal canal investigated with magnetic resonance imaging (MRI). Acta Anaesth Belg 2007;58:163–7
2. Bessa PRN, Costa VV, Arci ECP, Fernandes MCBC, Saraiva RA. Thoracic epidural block performed safely in anesthetized patients.A study of a series of cases. Rev Bras Anestesiol 2008;58:354–62
3. Takiguchi T, Yamaguchi S, Hashizume Y, Kitajima T. Movement of the cauda equine during the lateral decubitus position with fully flexed leg. Anesthesiology 2004;101:1250
4. Takiguchi T, Yamaguchi S, Okuda Y, Kitajima T. Deviation of the cauda equine by changing position. Anesthesiology 2004;100:754–5
5. Takiguchi T, Yamaguchi S, Tezuka M, Kitajima T. Measurement of shift of the cauda equina in the subarachnoid space by changing position. Reg Anesth Pain Med 2009;34:326–9
6. van Zundert AAJ, Stultiens G, Jakimowicz JJ, van den Borne BEEM, van der Hamk WGJM, Wildsmith JAW. Segmental spinal anaesthesia for cholecystectomy in a patient with severe lung disease. Br J Anaesth 2006;96:464–6
7. van Zundert AAJ, Stultiens G, Jakimowicz JJ, Peek D, van der Ham WGJM, Korsten HHM, Wildsmith JAW. Laparoscopic cholecystectomy under segmental thoracic spinal anaesthesia: a feasibility study. Br J Anaesth 2007;98:682–6
8. Imbelloni LE, Fornasari M, Fialho JC. Combined spinal epidural anesthesia during colon surgery in a high-risk patient. Case report. Rev Bras Anestesiol 2009;59:741–5
© 2010 International Anesthesia Research Society