This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Murata, H. Articles by Sumikawa, K. Search for Related Content PubMed PubMed Citation Articles by Murata, H. Articles by Sumikawa, K. Related Collections Anesthetic Techniques Regional Anesthesia Technology

---

REGIONAL ANESTHESIA

Three-Dimensional Computed Tomography for Difficult Thoracic Epidural Needle Placement

Hiroaki Murata, MD, PhD^*, Tetsuya Sakai, MD, PhD^*, Shinichi Goto, MD, PhD, and Koji Sumikawa, MD, PhD^*

From the *Department of Anesthesiology, Nagasaki University School of Medicine, Nagasaki, Japan; and Department of Anesthesia, the Japanese Red Cross Nagasaki Atomic Bomb Hospital, Nagasaki, Japan.

Address correspondence and reprint requests to Hiroaki Murata, Department of Anesthesiology, Nagasaki University School of Medicine, 1-7-1 Sakamoto, Nagasaki-City, Nagasaki 852-8501, Japan. Address e-mail to hiroaki_muramura{at}yahoo.co.jp.

	Abstract

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
BACKGROUND: Thoracic epidural anesthesia is often used as a postoperative analgesic technique in thoracic surgery. However, the structure of the overlapping spinous processes, resulting in interlaminar space occlusion, often makes thoracic epidural needle placement difficult. With the development of multi-detector row spiral computed tomography (CT), three-dimensional (3D) thoracic images can be readily obtained, providing potentially useful clinical information. Therefore, we conducted this study to evaluate the correlation between difficult thoracic epidural needle placement and anatomical findings obtained by 3DCT image processing techniques.

METHODS: Seventy-eight patients were studied. The number of new skin puncture attempts required for successful catheter insertion into the epidural space and the time spent during the procedure were recorded for each patient. The patients were defined as a first-level success when the needle placement was successful at the spinal level initially attempted. The others were defined as a first-level failure. The number of occluded mid-thoracic interlaminar spaces and the existence of mid-thoracic supraspinous and interspinous ligament ossification on the 3DCT images were also evaluated.

RESULTS: The percentage of first-level success was 84.6%. The number of occluded mid-thoracic interlaminar spaces was significantly greater in the first-level failure than in the first-level success (P < 0.001). The incidence of ossification of the mid-thoracic supraspinous ligament was significantly more frequent in first-level failure than in the first-level success (P = 0.001). The number of attempts and the time spent during the procedure significantly correlated to the number of occluded mid-thoracic interlaminar spaces (P < 0.001).

CONCLUSION: Preoperative 3DCT imaging may be useful in predicting difficult thoracic epidural needle placement.

	Introduction

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
Thoracic epidural anesthesia is often used as an analgesic technique in thoracic surgery.1–3 However, the structure of the overlapping spinous vertebral processes often makes thoracic epidural needle placement difficult.4–6 Few reports have evaluated the potential technical problems in the performance of epidural needle placement. Sprung et al.7 reported that examination of the patient's back for specific landmarks and spinal deformity predicts difficult epidural needle placement. Axial computed tomography (CT) imaging was reported to be useful in predicting the depth of thoracic epidural insertion through the median8 or paramedian5 approach by calculating the distance from the skin to the epidural space using basic trigonometric principles. However, these investigations did not describe specific structural features associated with difficult thoracic epidural needle placement. With the development of the multi-detector row spiral CT (MDCT), three-dimensional CT (3DCT) images can be readily obtained to provide very useful thoracic information.9–11 We conducted this study to evaluate the correlation between difficult thoracic epidural needle placement and anatomical findings observed by using currently available 3DCT image processing techniques.

	METHODS

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
After obtaining institutional approval and written informed consent from each patient, 81 consecutive patients scheduled for pulmonary resection for lung cancer under general anesthesia in combination with epidural anesthesia between January 2005 and March 2006 were included in this study. Patients with obvious abnormal spinal anatomy were excluded. Because in our institution the surgical incision was usually made at the fourth or fifth intercostal space, all patients underwent an initial epidural needle placement attempt at the T4/5 or T5/6 interlaminar space. The line drawn between the lower borders of the scapula that approximately crosses the seventh vertebral body was used to identify the spinal level.12 All punctures were performed by the same experienced anesthesiologist with >8 yr of professional experience with the patent in the flexed left lateral decubitus position through a left paramedian approach, before the induction of general anesthesia. The left lateral position was fixed to facilitate the epidural needle placement in order to replicate the same technical conditions in all patients.

After skin infiltration with 1% lidocaine, a 17-gauge Tuohy needle was introduced 1 to 2 cm lateral to the superior margin of the spinous process perpendicular to the skin in all planes until the lamina was encountered.13 The needle was "walked off" the superior edge of the lamina and into the epidural space.14 A loss-of-resistance technique was used to confirm epidural needle placement position using a saline-filled glass syringe. An epidural catheter was advanced threaded 5 cm into the epidural space.

If epidural needle placement could not be completed at the initial level, the next level was attempted. The decision to change the insertion site at the same level or to change the spinal level was made by the anesthesiologist. Neither time limits nor limitation to the number of attempts at each spinal level were preset. After a test aspiration was performed with the syringe to confirm negative blood or cerebrospinal fluid, an epidural test dose of 3 mL of 2% lidocaine with 1:80,000 epinephrine was injected through the catheter. Thermal hypesthesia was confirmed with the cold test 10 min after the test dose injection in all patients.

We assessed difficult epidural needle placement as described by Sprung et al.7 First, a record was kept as to whether needle placement was successful at the initial spinal level. We termed this either a first-level success or a first-level failure. Second, we counted the number of attempts required for successful catheter insertion into the epidural space. Each new skin puncture was considered another separate and distinct attempt, even if at the same spinal level. Needle redirection without a new skin puncture was not considered an additional attempt. Additionally, the total procedural time spent was measured beginning with the first skin puncture by the Tuohy needle to the initiation of the test dose injection through the epidural catheter.

The 3DCT images originally generated for the evaluation of pulmonary vessels branching patterns to ensure a safe surgical procedure by the surgeons were used. An Aquilion 16 MDCT scanner (Toshiba Medical Systems Corporation, Tokyo, Japan) imaged the patients. With the patient in the supine position, scan volumes extending from the thoracic inlet to the lung base were obtained during full inspiration. The following CT imaging variables were used: 120 kVp, 1.0-mm reconstructed slice width, a helical pitch of 15, a beam pitch of 0.93, and a rotation time of 0.5 s.

The raw data were reconstructed by using true cone beam tomography image reconstruction algorithm represented by a 512 x 512 matrix at 1 mm intervals and subsequently transmitted to a workstation (ZIOSOFT M900 QUADRA; ZIOSOFT, Tokyo, Japan). Volume rendering was used for 3D reconstruction. Three-dimensional images were reconstructed from the scans by using all isotropic voxels more than the selected minimum threshold of 116 Hounsfield units. The minimum opacity threshold value that we used in the present study was preset in the workstation memory and was used to assess the contrast-enhanced vessels, bones and ossified structures. Then, the isotropic voxels from the costal ribs, sternum, vertebral bodies and contrast-enhanced vessels were manually removed from the 3DCT images to extract the vertebral arches and spinous processes (Fig. 1).

View larger version (80K): [in this window] [in a new window]   Figure 1. Three-dimensional images were constructed from the scans by using all isotropic voxels higher than the selected minimum threshold of 116 Hounsfield units, at which the enhanced blood vessels and the bone were sufficiently segmented (A). Then, the isotropic voxels from the scapula and costal ribs (B), sternum, enhanced blood vessels and vertebral bodies (C) were manually removed from the three-dimensional computed tomography images. Finally, we extracted vertebral arches and spinous processes (D).

Interlaminar space occlusion was defined as complete closure on the 3DCT image. We did not define the minimum voxel number that allowed us to state that the interlaminar space was occluded. Occlusion of the mid-thoracic interlaminar spaces (T3/4-T8/9, 6 at the maximum) was evaluated by viewing the 3DCT images from the dorsal left side of the body rotated inward at 25° to the midline of the sagittal plane and 25° to cephalad. This viewing protocol considered physiological thoracic kyphosis and the position of the patient during supine CT, avoiding inappropriate evaluation of interlaminar space occlusion (Fig. 2).

View larger version (32K): [in this window] [in a new window]   Figure 2. (A) Three-dimensional computed tomography image of the thoracic vertebral arches and spinous processes of a patient defined as first-level success. Occlusion of the mid-thoracic interlaminar spaces (T3/4-T8/9) was evaluated by viewing the image from the dorsal left side of the body rotated inward at 25° to the midline of the sagittal plane and 25° to cephalad. All mid-thoracic interlaminar spaces are unobstructed and no ligament ossification was observed. (B) Three-dimensional computed tomography image of the thoracic vertebral arches and spinous processes of a patient defined as first-level failure. Occlusion of the mid-thoracic interlaminar spaces was evaluated by the same method as Figure 2A. All mid-thoracic interlaminar spaces are occluded. Ossification of the supraspinous ligament is observed between T8 and T9 spinous processes (arrow).

When there were ossified structures between the tips of spinous processes, they were regarded as ossified supraspinous ligament. In a similar way, when the ossified structures existed on the line drawn between the bodies of neighboring spinous processes, they were regarded as ossified interspinous ligament. The acquired 3DCT images were freely rotated on the display monitor to identify the ossification of the supraspinous and interspinous ligaments.

Statistical analysis was performed using SPSS 14.0 for Windows (SPSS Japan, Tokyo, Japan). Data were analyzed using the unpaired Student's t-test or Fisher's exact test when appropriate. The Spearman rank correlation was calculated to assess the correlation of the data. Results were expressed as means ± sd, where values of P < 0.05 were considered statistically significant.

	RESULTS

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
Eighty-one patients were included in this evaluation. No patient had abnormal spinal anatomy, such as scoliosis. Three of 81 patients required immediate replacement of the epidural catheter because of IV catheter placement (two detected by continuous aspiration of blood through the catheter and one suspected by an increase in heart rate and arterial blood pressure by about 20% after test dose injection). As a result, 78 patients completed the study. In all patients, thermal hypesthesia between T3–7 dermatome was confirmed with the cold test 10 min after test dose injection. Adequate postoperative epidural analgesia was obtained for at least 24 h with 0.15% bupivacaine at 2 or 3 mL/h and morphine at 2 or 3 mg daily (pain at rest was <4/10 using the numerical rating scale from 0 to 10, in which 0 = no pain and 10 = excruciating pain.). There were no complications from needle or catheter placement.

The percentage of first-level success was 84.6% (66 of 78 patients). Only one attempt at the same spinal level was required for 95.5% (63 of 66) of the patients for first-level success. The remaining three patients required two attempts at the same spinal level. In 12 patients defined as first-level failure, 5 required 2 spinal levels, another 5 required 3 spinal levels, 1 required 5 spinal levels, and 1 patient required 6 spinal levels to complete the procedure. Patient characteristics did not differ between first-level success and first-level failure (Table 1). Although we did not evaluate the quality of spinal landmarks, spinous processes were palpable in all patients studied.

The number of occluded mid-thoracic interlaminar spaces observed in the 3DCT images was significantly greater in the first-level failure than in first-level success (P < 0.001) (Table 2). The incidence of mid-thoracic supraspinous ligament ossification was significantly more frequent in first-level failure than in first-level success (P = 0.001). The number of attempts was significantly greater in first-level failure than in first-level success (P < 0.001). The time spent during the procedure was significantly longer in first-level failure than in first-level success (P < 0.001) (Table 2, Fig. 3). No ossification of the interspinous ligament was observed.

View larger version (7K): [in this window] [in a new window]   Figure 3. (A) Distribution of the number of attempts required for a successful epidural needle placement as a function of the number of occluded interlaminar spaces in the three-dimensional computed tomography images. (B) Distribution of the time spent during the procedure required for a successful epidural needle placement as a function of the number of occluded interlaminar spaces in the three-dimensional computed tomography images. (C) Relationship between the number of attempts and the time spent during the procedure required for a successful epidural needle placement.

The number of attempts (r_s = 0.50, P < 0.001) and the time spent during the procedure (r_s = 0.48, P < 0.001) were significantly correlated with the number of occluded mid-thoracic interlaminar spaces, though these results do not support a strong correlation between the mentioned factors (Figs. 3A and B). When all mid-thoracic interlaminar spaces were unobstructed, epidural needle placement at the initial attempt were successful in 93.3% of the patients (42 of 45, Fig. 3A). As shown in Figure 3C, the number of attempts significantly correlated with the time spent during the procedure (r_s = 0.70, P < 0.001). Age, height, weight, and Body Mass Index did not correlate with the number of attempts nor with the number of occluded mid-thoracic interlaminar spaces.

	DISCUSSION

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 
We demonstrated a relationship between the anatomical findings observed by using currently available 3DCT image processing techniques with the difficulty in thoracic epidural needle placement. The number of attempts and the time spent during the procedure were significantly correlated with the number of occluded mid-thoracic interlaminar spaces. The incidence of supraspinous ligament ossification visible on the 3DCT image is also correlated with difficult thoracic epidural needle placement.

MDCT has become popular and is changing the performance and interpretation9 of routine thoracic imaging. Volume rendering is one of the most useful functions of MDCT for clarifying complex 3D anatomic relationships.9,15 Because each data component of a volume-rendered image is assigned an opacity value, a threshold should be selected to determine the minimum opacity displayed.15 The minimum opacity threshold value used in this study was preset in the workstation's internal memory to assess the contrast-enhanced organs and vessels, which sufficiently displayed bone and ossified structures.15

Segmentation is the process of selecting data to be included in a 3DCT image. Opacity threshold is also available as a method of segmentation to remove the background structures in 3DCT images. A region of interest can be removed by manually drawing a rectangular, elliptical, or other shape from within the data set using a type of virtual scalpel to cut the defined region. In the present study, segmentation techniques allowed us to evaluate the interlaminar space without interference from vertebral bodies and ribs (Fig. 2).

Although 3DCT imaging techniques provide stereoscopic information, one of the problems encountered in this study was the positioning of the patient during supine CT which differs from that in spine flexion during an epidural needle placement. An unobstructed interlaminar space in the flex lateral decubitus position could be inappropriately evaluated as occluded in the supine position. Thus, the number of occluded interlaminar spaces may be over-estimated in the supine position. However, the number of occluded interlaminar spaces was very low even in the supine position for first-level successes, and significantly more interlaminar spaces were occluded for first-level failures. Indeed, successful needle placement was achieved at the initial skin puncture in 93.3% of patients who had no occluded mid-thoracic interlaminar spaces visible on the 3DCT images. Therefore, 3DCT imaging techniques may be useful in evaluating difficult thoracic epidural needle placement.

Because the same anesthesiologist performed all epidural needle placements, it is possible that technical bias might have been introduced. However, technical variables, such as time spent during the procedure and number of attempts, were significantly correlated with the number of occluded mid-thoracic interlaminar spaces. Therefore, we consider that occlusion of the mid-thoracic interlaminar space was positively correlated to difficult thoracic epidural needle placement.

Examination of 3DCT images only for epidural needle placement may be neither reasonable nor practical. Indeed, all patients had successful epidural placement without 3DCT guidance in this study. However, 3DCT images were reconstructed to evaluate pulmonary vessel branching patterns to ensure a safe surgical procedure; the imaging was not performed specifically to evaluate difficulty of epidural needle placement. Processing of existing reconstructed 3DCT images is not technically demanding, and once instructed by a radiologist, an anesthesiologist can perform this simple task alone. Thus, radiologists performed little additional work and the patients did not receive additional radiation exposure, nor were they billed for this study. These factors increase the feasibility of 3DCT images for the evaluation of spinal anatomy.

In conclusion, the number of occluded mid-thoracic interlaminar spaces and the incidence of the supraspinous ligament ossification observed in the supine 3DCT images were correlated with difficult thoracic epidural needle placement, suggesting that 3D anatomical information obtained by 3DCT imaging techniques may be useful for predicting difficult thoracic epidural needle placement.

	ACKNOWLEDGMENTS

 
We gratefully acknowledge the technical assistance of the following radiologists for their time in evaluating the three-dimensional computed tomography images; Mayumi Ohtsubo, MD, Director, Masakazu Mori, MD, Staff Radiologist, and Atsushi Miyazaki, MD, Staff Radiologist, Department of Radiology, The Japanese Red Cross Nagasaki Atomic Bomb Hospital, Nagasaki, Japan.

	Footnotes

 
This work was supported in part by a Grant-in-Aid for Scientific Research from Nagasaki University, Japan.

Accepted for publication October 3, 2007.

	REFERENCES

Top Abstract Introduction METHODS RESULTS DISCUSSION REFERENCES

 

1. Kavanagh BP, Katz J, Sandler AN. Pain control after thoracic surgery. A review of current techniques. Anesthesiology 1994;81:737–59[Web of Science][Medline]
2. Azad SC. Perioperative pain management in patients undergoing thoracic surgery. Curr Opin Anaesthesiol 2001;14:87–91[Medline]
3. Senturk M. Acute and chronic pain after thoracotomies. Curr Opin Anaesthesiol 2005;18:1–4[Medline]
4. Grau T, Leipold RW, Delorme S, Martin E, Motsch J. Ultrasound imaging of the thoracic epidural space. Reg Anesth Pain Med 2002;27:200–6[Web of Science][Medline]
5. Kao MC, Tsai SK, Chang WK, Liu HT, Hsieh YC, Hu JS, Mok MS. Prediction of the distance from skin to epidural space for low-thoracic epidural catheter insertion by computed tomography. Br J Anaesth 2004;92:271–3[Abstract/Free Full Text]
6. Nagaro T, Yorozuya T, Kamei M, Kii N, Arai T, Abe S. Fluoroscopically guided epidural block in the thoracic and lumbar regions. Reg Anesth Pain Med 2006;31:409–16[Web of Science][Medline]
7. Sprung J, Bourke DL, Grass J, Hammel J, Mascha E, Thomas P, Tubin I. Predicting the difficult neuraxial block: a prospective study. Anesth Analg 1999;89:384–9[Abstract/Free Full Text]
8. Carnie J, Boden J, Gao Smith F. Prediction by computerised tomography of distance from skin to epidural space during thoracic epidural insertion. Anaesthesia 2002;57:701–4[Web of Science][Medline]
9. Gruden JF. Thoracic CT performance and interpretation in the multi-detector era. J Thorac Imaging 2005;20:253–64[Web of Science][Medline]
10. Watanabe S, Arai K, Watanabe T, Koda W, Urayama H. Use of three-dimensional computed tomographic angiography of pulmonary vessels for lung resections. Ann Thorac Surg 2003;75:388–92[Abstract/Free Full Text]
11. Remy-Jardin M, Remy J. Spiral CT angiography of the pulmonary circulation. Radiology 1999;212:615–36[Abstract/Free Full Text]
12. Brown DL. Atlas of Regional Anesthesia. Philadelphia: WB Saunders; 1997:283–92
13. Ramamurthy S. Thoracic epidural nerve block. In: Waldman SD, ed. Interventional Pain Management. 2nd ed. Philadelphia: WB Saunders; 2001:390–5
14. Bernards CM. Epidural and spinal anesthesia. In: Barash PG, Cullen BF, Stoelting RK, eds. Clinical Anesthesia. 4th ed. Philadelphia: Lippincott Williams & Wilkins; 2001:689–713
15. Dalrymple NC, Prasad SR, Freckleton MW, Chintapalli KN. Informatics in radiology (infoRAD): introduction to the language of three-dimensional imaging with multidetector CT. Radiographics 2005;25:1409–28[Abstract/Free Full Text]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Murata, H. Articles by Sumikawa, K. Search for Related Content PubMed PubMed Citation Articles by Murata, H. Articles by Sumikawa, K. Related Collections Anesthetic Techniques Regional Anesthesia Technology