The combined spinal epidural (CSE) block has received increasing attention in recent years. This is reflected in increasing numbers of published reports [1-4] and the availability of at least 10 "special" CSE needle sets on the market. The CSE block provides the reliability of a spinal block and the flexibility of the epidural catheter technique [1,4]. In most cases in which the CSE block is used as a regular spinal block, the epidural catheter is used only when needed. The presence of an epidural catheter permits the anesthesiologist to improve the quality, or extend the distribution, of spinal block during surgery and also to provide prolonged postoperative analgesia. Since the spinal needle makes a hole in the dura, the theoretical risk of epidural catheter penetration through this hole is a concern. Furthermore, the risk of migration may increase if several spinal puncture attempts are performed during a "difficult" CSE block. However there are no case reports of this complication.
In the present study, percutaneous epiduroscopy with video recording was performed in fresh cadavers with the following aims: 1) to visualize the anatomic structures of the epidural space and record the different stages during performance of epidural block with catheter and the CSE block; 2) to evaluate the risk of subarachnoid placement of the epidural catheter during CSE block after (a) a single dural puncture with the spinal needle through Tuohy needle, (b) multiple (five) punctures in a restricted area of dura with the spinal needle through Tuohy needle, and (c) dural puncture with Tuohy needle.
Fifteen fresh cadavers, aged 39-96 yr, were studied 18-36 h postmortem. The study was approved by the Orebro County Council Ethic Committee. With the subjects placed in the lateral position the epidural space was identified with a Tuohy needle in the lumbar region, using the midline or lateral approach and loss-of-resistance technique. A 4.0-mm diameter rigid endoscope (Storz Registered Trademark, Germany) with a 70 degrees angled lens was introduced into the epidural space percutanously with the aid of a sharp trochar and loss of resistance technique. A video camera (Concept Intravision Registered Trademark, USA) was attached to the endoscope and connected to a high mobility 9-in. video monitor with an 8-mm videocassette recorder. The endoscope was equipped with a 150-W halogen light source and a fiberoptic light guide (Universal Registered Trademark, USA). A syringe was attached to one of the lateral channels of the endoscope for the purpose of injecting air or water into the epidural space to facilitate visualization. The endoscope was inserted into the same or adjacent lumbar space as the Tuohy needle. By rotating the epiduroscope, the epidural space could be visualized in all directions. After visualization of the Tuohy needle in the epidural space, a combined spinal epidural (CSE) block was performed. The CSE block consists of a lumbar puncture with a long spinal needle through a Tuohy needle placed in the epidural space. After injecting the local anesthetic in the subarachnoid space, the spinal needle is withdrawn and an epidural catheter is inserted in the usual manner. The technique has been described elsewhere .
In our experimental sequence, the dura was punctured once by the spinal needle, followed by an attempt to insert the epidural catheter into the epidural space. With the aim to mimic a "difficult" CSE block, where multiple spinal puncture attempts are performed, four additional holes were made with the spinal needle followed by insertion of the epidural catheter. Finally, a dural puncture was made with the Tuohy needle and the insertion of epidural catheter was recorded. In every subject, the anatomic structures and experimental steps described above were video recorded at one or several lumbar levels Table 1. The needles used were 16 to 18-gauge Tuohy needles (Perican Registered Trademark; B. Braun Melsungen AG, Germany), 17 to 19-gauge soft-tip epidural catheters (Perifix Registered Trademark; B. Braun Melsungen AG or Portex, Great Britain), and long (120 mm) 25 to 26-gauge spinal needles (Spinocan Registered Trademark; B. Braun Melsungen AG).
In an attempt to mimic the circumstances in the epidural space in a clinical setting, an epidural catheter was inserted into the subarachnoid space at the thoracic level. Through the catheter, saline was injected in the subarachnoid space to create cerebrospinal fluid (CSF) pressure resembling that in a living patient. However, this was unsuccessful because the bulging of the dura and saline leakage down the epidural space quickly reduced the visual field, making further evaluation impossible.
In the two first subjects ligamentum flavum was identified by lateral laminectomy from L-2 to L-4 before inserting the endoscope. For confirmation of the various steps of the experiment, the spinal column was dissected and the meninges exposed in one cadaver.
In 15 subjects, 24 recordings at different lumbar levels were performed Table 1. In four cases it was not possible to visualize and record the complete experimental sequence due to technical difficulties, such as poor visibility (two cases) and dural puncture with the endoscope (two cases). In the remaining 20 recordings, the visibility was good or excellent when the epiduroscope was kept close to the needle. The contrast between the yellow fat deposits, the whitish connective tissue, and the darker blood vessels was distinct. Little or no problem was encountered in identifying these structures. The holes made by spinal and epidural needles (single and multiple attempts) were clearly identifiable. The ease or difficulty of catheter penetration through these holes was also clearly visible. Spinal column dissection showed correct placement of epiduroscope, Tuohy needle, and epidural catheter in the epidural space.
The epidural space appeared to be only a potential space which was opened by the Tuohy needle or epiduroscope. In approximately 30% of the cases it had to be kept open by repeated injections of air or water. Epidural fat distribution varied greatly among the subjects, and did not seem to be related to general fat distribution (obesity) of the subject. Connective tissue strands between the dura and ligamentum flavum were richly distributed, these strands being generally thin and fragile. However, in two recordings a somewhat denser connective tissue strand was noted in the midline, suggesting a dorsomedian connective tissue band . Blood vessels were also richly distributed, although the distribution varied greatly among the subjects.
Due to the narrowness of the lumbar epidural space, extensive tenting of the dura, both with Tuohy needle and with the epidural catheter, was noted in all 20 recordings Figure 1 and Figure 2. The dura appeared to be highly resilient but of uneven thickness. In three recordings (15%) the spinal nerves were discernible through parts of the dura and moved when the dura was touched by the catheter tip. During introduction of the catheter small particles of fat and "debris" emerged from the bevel of the Tuohy needle in eight cases (40%) Figure 1. In two recordings, the catheter tip was caught in areas of densely distributed epidural fat tissues. Approximately 15 cm of the epidural catheter could be introduced easily and seen to coil in the epidural space while the tip of the catheter was stationary Figure 3.
In 3 of 20 recordings (15%) difficulties in introducing the catheter into the epidural space were encountered. When the catheter tip was "bent" prior to introduction, the insertion was smooth. Under these circumstances, the catheter was caught in the needle tip and began to shear on withdrawal of the catheter through the Tuohy needle. It was not possible to predict the direction the catheter would take in the epidural space. When the needle was turned with the bevel in directions other than cephalad (in order to make catheter introduction easier), the catheter still turned cephalad due to the narrowness of the epidural space. No unintentional perforation of the dura by the Tuohy needle occurred.
After intentional dura perforation with the Tuohy needle the epidural catheter was seen to pass through this hole and enter the subarachnoid space in 9 of 20 recordings (45%). However in several cases the epidural catheter had to be directed toward the hole and considerable force had to be applied for the catheter to pass through the dural hole Figure 4 A, B.
When a single dural puncture hole was made with a 25- or 26-gauge spinal needle through the epidural needle Figure 5, it was impossible to force a 16- or 18-gauge epidural catheter through the hole. Although the catheter was guided toward the dural hole and considerable pressure applied, the catheter remained in the epidural space in each case Figure 6. When five holes were made close to each other in the same area of the dura by 25- or 26-gauge spinal needles Figure 7 the catheter was seen to get caught in the perforated area, causing it to coil when pressure was applied. This occurred in five recordings (25%). In 1 of 20 (5%) recordings, the epidural catheter penetrated the dura in the area of multiple holes when the catheter was guided toward the holes.
When turning the Tuohy needle 180 degrees after making the dural holes, the catheter tip seemed to "hit" the dura further away from the holes, but this distance could not be measured. It was not possible to establish a reliable method for marking the dural holes and the place where the tip of catheter "hit" the dura and then measure the distance between these marks on the video screen.
Epiduroscopy has been performed in cadavers as well as in living human patients with the aid of rigid lenses or flexible scopes [5-9]. Blomberg  suggested that postmortem changes in the cadaver might influence the anatomic structures found in the epidural space; Blomberg and Olsson  pointed out that good visualization of the epidural space in the living specimen is more complicated to achieve than in the cadaver. In earlier epiduroscopy studies a thinner lens (1.5-2.0 mm) has been used. For our particular purpose; these lenses did not provide adequate light conditions and field of vision. To be able to interpret the dynamic changes correctly during a CSE block recorded with video technique, we decided to use a larger epiduroscope. This modification provided a greater field of vision and improved light conditions. Although fiberoptic techniques are being increasingly tested for epiduroscopy it is still not possible to study the CSE technique or the risk of catheter migration into the subarachnoid space in patients [8,9]. The large amount of injected air or saline in the epidural space necessary to achieve acceptable visual conditions in living humans may make the technique unsafe.
Although the absence of circulating blood in veins and arteries and the very low CSF pressure in cadavers are drawbacks , we think several conclusions can be made. Epiduroscopy combined with video recording provided good quality visualization of anatomic structures in the epidural space, and gave a clearer understanding of the various stages during the performance of epidural and CSE blocks. Epiduroscopy videos may be a good educational tool and also provide a useful method for evaluating newer developments in spinal and epidural needles and catheters. The risk of catheter shearing as it is being withdrawn through the Tuohy needle was clearly demonstrated with the video technique.
It is conceivable that the hanging drop technique for localization of the epidural space  is somewhat unreliable because the bevel of the Tuohy needle is sometimes obstructed by tissue fragment. The amount and distribution of epidural fat seems to be an important factor that influences the success of epidural catheter insertion. Our study showed that epidural fat distribution may not be related to the general fat distribution (obesity) of the patient, which is consistent with the opinion of Blomberg and Olsson . Our study also demonstrated that excessive epidural fat may cause the epidural catheter to coil upon itself in the epidural space. Thus, it is not possible to predict catheter tip placement based on the length of epidural catheter introduced or on the direction of needle bevel. Coiling of the epidural catheter has been demonstrated by radiographic techniques in patients [11,12]. We agree with Bridenbaugh et al.  and Gielen et al.  that only a limited length of catheter should be inserted into the epidural space. Other factors that may influence the position of the catheter tip are depth of epidural space, distribution of blood vessels and connective tissues structures in the epidural space, and paramedian or median approach.
There is some controversy about the presence of a connective tissue band that divides the epidural space in the posterior midline. The presence of this band has been claimed [5,7,13-16] and denied [17-19]. This band is believed to increase the risk of dural puncture and of unilateral analgesia. It may also be associated with increased difficulty in introducing the epidural catheter . The dorso median band has been demonstrated by resin casts , epiduroscopy [5,7], epidurography [14,15], and computed tomography scan . Contradicting these reports are those that show total absence of such a membrane after resin casts , cryomicrotome sections , and clinical studies . Our video recordings demonstrated that in the majority of cases thin fibrous strands are richly distributed in the epidural space. In only 10% of our cases could a somewhat dense dorso median tissue band be visualized Figure 2.
Our results indicates that the distribution of epidural fat is a more important factor influencing the position an epidural catheter than any connective tissue band. These findings are consistent with those of Rosenberg et al.  who used fiberoptic epiduroscopes in dogs and humans.
Although this study was performed in cadavers, the dura appeared to be surprisingly resilient. In every recording, varying degrees of tenting of the dura were noted. This was consistently noted every time a spinal or epidural needle or epidural catheter was introduced in the epidural space. Since the epiduroscope is already in the epidural space, pushing the dura away from the ligamentum flavum, and there is low CSF pressure in the cadaver, one would theoretically suspect an even greater tenting effect in living patients. It seems obvious that even slight movements with the needle, such as rotating it in the epidural space, might cause unintentional damage or perforation of the dura, as has been suggested by Meiklejohn .
To achieve acceptable visual conditions during epiduroscopy, we kept the epidural space open by repeated air or saline injections. In a living patient, the dura probably lies closer to ligamentum flavum. Hence the risk of tenting or shearing the dura during both epidural and CSE block is always present.
It was impossible to force the epidural catheter into the subarachnoid space after a single perforation of the dura with the spinal needle. This is consistent with earlier reports on tests with isolated dural tissue . Although all of these results are based on in vitro studies, it seems that the risk of subarachnoid placement of the epidural catheter during the performance of an uncomplicated CSE block appears to be very small. The video recordings in this study suggest that the risk of subarachnoid migration of the epidural catheter increases after multiple dural punctures with the spinal needle.
The risk of dura penetration with the epidural catheter after a dural puncture with the Tuohy needle was clearly demonstrated in the study. Although the risk of catheter penetration was similar, whether single or multiple dural punctures were performed by the spinal needle, our study showed that the risk of the epidural catheter tip getting caught in the dura is much higher after multiple holes. This can be expected to increase the risk of epidural drugs being deposited in the subarachnoid space. Careful aspiration and small incremental doses of drugs injected through the epidural catheter are therefore important, as suggested by van Zundert et al.  and others.
In conclusion, percutaneous epiduroscopy with video recording is a useful teaching tool. Our results indicate that the risk of epidural catheter migration through the dural hole during uncomplicated combined spinal epidural block is very small.
The authors wish to express their gratitude to A. Amilon, MD, Department of Hand Surgery, T. Dolk, MD, Department of Orthopedic Surgery, and B. Risberg, MD, Department of Pathology, Orebro Medical Center Hospital, for their support. We also wish to thank A. Spanholz and H. Otto of B. Braun Melsungen AG, M. Bain, Germany, and U. Larsson, RN, for making this study possible. We are grateful to the operating room staffs at the Department of Hand Surgery, Orebro Medical Centre Hospital and the Department of Surgery, Lindesbergs Hospital. BBS Medical Electronics, Sweden, are acknowledged for the use of their video printer equipment. The assistance of L. G. Jogsten and L. Lofstedt at the Department of Pathology is also appreciated.
1. Rawal N, Schollin J, Westrom G. Epidural versus combined spinal epidural block for caesarean section. Acta Anaesthesiol Scand 1988;32:61-6.
2. Carrie LES. Extradural, spinal or combined block for obstetric surgical anaesthesia. Br J Anaesth 1990;65:225-33.
3. Vandermeersch E, Kick O, de Gouw N, Van Aken H. Combined spinal and epidural anaesthesia. Reg Anaesth 1991;14:108-12.
4. Holmstrom, Laugaland K, Rawal N, Hallberg S. Combined spinal epidural block versus spinal and epidural block for orthopaedic surgery. Can J Anaesth 1993;40:601-6.
5. Blomberg R. The dorsomedian connective tissues band in the lumbar epidural space of humans: an anatomical study using epiduroscopy in autopsy cases. Anesth Analg 1986;65:747-52.
6. Pick C-F, Bockhorn J, Schumacher M. Endoskopische und computear-tomographische Untersuchungen zur Lagerungsbehandlung akuter Halswirbelsaulenverletzungen. Arch Orthop Trauma Surg 1980;97:43-9.
7. Blomberg RG, Olsson SS. The lumbar epidural space in patients examined with epiduroscopy. Anesth Analg 1989;68:157-60.
8. Shimoji K, Fujioka H, Onodera M, et al. Observation of spinal canal and cisternae with the newly developed small-diameter, flexible fiberscopes. Anesthesiology 1991;75:341-4.
9. Rosenberg PH, Heavner JE, Chokhavatia S, et al. Epiduroscopy with a thin flexible and deflectable fiberscope. Br J Anaesth 1994;72(1 Suppl):121.
10. Bromage PR. Identification of the epidural space. In: Bromage PR, ed. Epidural Analgesia. Philadelphia: WB Saunders, 1978:176-214.
11. Bridenbaugh LD, Moore DC, Bagdi P, Bridenbaugh PO. The position of plastic tubing in continuous-block techniques: an x-ray study of 552 patients. Anesthesiology 1968;29:1047-9.
12. Gielen MGM, Slappendel R, Merx JL. Asymmetric onset of sympathetic blockade in epidural anaesthesia shows no relation to epidural catheter position. Acta Anaesthesiol Scand 1991;35:81-4.
13. Husemeyer RP, White DC. Topography of the lumbar epidural space. Anaesthesia 1980;35:7-11.
14. Roberson GH, Hatten HP, Hesselink JH. Epidurography: selective catheter technique and review of 53 cases. Am J Radiol 1979;132:787-93.
15. Gallart L, Blanco D, Samoso E, Vidal F. Clinical and radiologic evidence of the epidural plica mediana dorsalis. Anesth Analg 1990;71:698-70.
16. Savoliaine ER, Pandya JB, Greenblatt SH, Conover SR. Anatomy of human lumbar epidural space: new insights using CT-epidurography. Anesthesiology 1988;68:217-20.
17. Harrison GR, Parkin IG, Shah JL. Resin injection studies of the lumbar extradural space. Br J Anaesth 1985;57:333-6.
18. Hogan QH. Lumbar epidural anatomy. A new look by cryomicrotome section. Anesthesiology 1991;75:767-75.
19. Asato F, Hirakawa N, Oda M, et al. A median epidural septum is not a common cause of unilateral epidural blockade. Anesth Analg 1990;71:427-9.
20. Meiklejohn BH. The effect of rotating of an epidural needle. Anaesthesia 1987;42:1180-2.
21. van Zundert A, Vaes L, Soetens M, et al. Every dose given in epidural analgesia for vaginal delivery can be a test dose. Anesthesiology 1987;67:436-40.