| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
,
,
Departments of
*Anesthesiology,
Neurosurgery, and
||Pathology, and the
§Research Center, Hospital General Universitario de Valencia; and the
Department of Surgery, University of Valencia, Valencia, Spain
Address correspondence and reprint requests to Roberto García-Aguado, MD, PhD, Servicio de Anestesiología, Hospital General Universitario, c/Tres Cruces, s/n, 46014 Valencia, Spain. Address e-mail to aguador{at}arrakis.es
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
Implications: In this swine study, we repaired a cerebrospinal fluid leak after a dural puncture by percutaneously injecting tissue adhesive. The technique of percutaneous injection of fibrin glue seems promising for the prophylaxis of headache associated with cerebrospinal fluid leakage, and may be an alternative to an epidural blood patch.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Several percutaneous treatments for persistent dural leakage or PDPH have been recommended, such as injection of dextran (4), saline (2), or blood (5) in the epidural space. The epidural blood patch (EBP) is most commonly used because of its effectiveness, safety, and duration (5). Fibrin glue has been used for the percutaneous repair of chronic CSF fistulae after intrathecal catheterization (6) or secondary to spine surgery (7).
Prophylactic percutaneous sealing of a dural leak might theoretically decrease complications. Our group has studied the feasibility of this repair with fibrin glue in an in vitro model of a continuing CSF fistula (8). The purpose of this paper was to study in vivo the effect of percutaneously applying fibrin glue to stop CSF leakage after dural puncture with a 17-gauge Tuohy needle in swine.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Fasting animals were premedicated with intramuscular midazolam (0.5 mg/kg), ketamine (10 mg/kg), and glycopyrrolate (0.2 mg). After 15 min, general anesthesia was induced with IV propofol (1 mg/kg), the tracheas were intubated, and the lungs were mechanically ventilated with a mixture of oxygen (35%) in air, with a minute ventilation to achieve an end-tidal CO2 between 30 and 40 mm Hg.
Anesthesia was maintained with a continuous IV infusion of propofol (10 mg · kg-1 · h-1), pancuronium bromide (0.3 mg · kg-1 · h-1), and fentanyl (0.0075 mg · kg-1 · h-1). Electrocardiogram, oxygen saturation, end-tidal CO2, and nasopharyngeal temperature were continuously monitored. With the animal in the right lateral decubitus position, a suboccipital cisterna magna puncture was performed with a 18-gauge Tuohy needle, and a 22-gauge catheter was inserted 2 cm past the needle tip and connected to a pressure measuring device. The animal was then secured in the sitting position.
Fibrin glue was prepared from a kit (Tissucol® Fibrin Sealant, Inmuno AG, Vienna, Austria) containing: 1) freeze-dried protein concentrate of human fibrinogen, Factor XIII, fibronectin, and plasminogen; 2) freeze-dried bovine thrombin (4 IU/mL or 500 IU/mL); 3) aprotinin solution (3000 IU/mL); and 4) calcium chloride solution (40 mmol/mL). The mixture was prepared according to the manufacturer’s instructions. Thrombin and aprotinin were used in large concentration (500 IU/mL and 3000 IU/mL, respectively). A double-barreled syringe with a common piston (Duploject®) was used to enable simultaneous mixture of the two end-stage components.
The animals were randomly assigned to one of three groups: the control group (9 animals), the experimental group (9 animals), and the sham-operated group (4 animals).
Animals in the control and experimental groups were subjected to lumbar puncture through either the L4-L5 or the L5-L6 interspace in the sitting position. The epidural space was identified with a 17-gauge needle by the loss of resistance technique with air, and the needle advanced until the dura mater was pierced and CSF obtained. The needle was immediately withdrawn until no CSF outflow was observed. Through this epidural needle, animals belonging to the control group received 1.4 mL saline, and those in the experimental group received 1.4 mL of previously heated and mixed fibrin glue. In both groups, the needle was withdrawn 15 min later. Then, 15 mL of 0.3% Evans blue in saline was infused through the cisterna magna catheter with the aid of an infusion pump at 1.5 mL/min over 10 min. Animals belonging to the sham-operated group were infused with the same Evans blue solution, but did not receive a lumbar puncture.
CSF pressures were determined at the following stages: 1) in the lateral decubitus position after occipital puncture, 2) sitting position (basal reading), 3) immediately after lumbar dural puncture, 4) 15 min after injection of either fibrin glue or saline, 5) 10 min after Evans blue infusion, and 6) 1 h after this last measure. In the sham-operated group, the third measure was not done. CSF pressure was measured through the occipital catheter, except the third measure, which was performed through the lumbar needle.
One hour after the infusion had ended, animals were killed with an IV injection of 40 mEq KCl. A squared block of tissue measuring 3 cm3 was cut from the puncture site after dissecting the spinous ligament and the attachments of the paraspinal muscles from the spinous processes. These latter laminae, together with the vertebral laminae, were removed with a rongueur, and the inner fascia of the periostium was exposed. This was cut longitudinally, and the epidural fat and the dura were inspected. A 4-cm-long block of the spinal cord, together with the dura and the fibrin glue adhered to it, was cut and removed.
The specimen was inspected with the aid of a loupe, and macro photographs were taken during all the previously described processes, searching for the appearance of blue pigmentation. These observations were categorized and rated as follows:
The observations were done independently by two blinded observers.
The skin interspinous tissue and the dura spinal cord blocks were fixed in 10% formaldehyde after the animal’s death, set in paraffin blocks, cut in 4 µm sections transversally to the spinal cord axis, and stained with hematoxylin and eosin for histologic analysis.
Continuous data from physiological measures were compared with the Mann-Whitney U-test (Kruskal-Wallis test for the three-group comparisons).
Discrete variables were compared by means of the 
2 test (Fisher’s two-tailed exact test). Statistical significance was set to the 95% confidence interval. Data are expressed as the median ± range unless otherwise stated.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Five of six animals of the control group showed dye outflow through the cutaneous puncture site. In the experimental group, only one of eight and none of the sham-operated animals showed such an outflow. All animals in the control group, one of eight animals from the experimental group, and no sham-operated animals showed coloration of the subcutaneous tissue and the interspinous ligament.
The epidural fat and periostium were intensely dyed in all control animals, and the dural orifice, as well as the underlying blue-dyed CSF were evident. Macro photographs showed oval holes (1.5 x 0.7 mm).
Only one animal from the experimental group showed an intense coloration of the periostium and the epidural fat. The rest of the animals of this group showed a scarcely dyed periostium and epidural fat. Dye was more intense at the points where the spinal roots exit the dura. This last dye pattern was identical to all animals from the sham-operated group.
In six of eight animals in the experimental group, there was no dyed CSF in the epidural space, and in seven the puncture hole could not be identified, because it was covered by a mucoid white material. This material was located under the periostium, and extended over both neighboring vertebral spaces and laterally into the lateral recesses reaching the foramina. In one case, the fibrin glue material had migrated into the dural sac, and the dura was bulging at the puncture site. In this case, the dye pattern was identical to that in the control group.
Table 1 shows the comparison among the categorized variables in the control, experimental, and sham-operated groups. No significant differences were found between the experimental and the sham-operated groups for any variables studied; both showed significant differences when compared with the control group.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Evans blue dye was continuously infused at 1.5 mL/min over 10 min to maintain an elevated CSF pressure. This infusion rate was within the range reported by Czernicki et al. (11), who used this technique to offset CSF resorption. The CSF infusion possibly outweighed the CSF outflow through the fistula. This may explain the lack of difference in CSF pressure measured among the groups in this acute experiment, although a drop in the intradural pressure would be expected in the fistula group should the experiment be prolonged.
The histologic study showed the fibrin glue forming a sheath localized around the puncture hole, extending over the outer dural surface, and sometimes over the inner dural surface also. Only in one case was the fibrin glue found inside the dural sac, probably due to inadvertent intradural injection. In this animal, sealing did not occur; we hypothesize that the needle prevented fibrin glue adherence to the dural hole.
The presence of intensely blue-dyed periostium and fat was a clear marker of the outflow of CSF through the dural puncture site and coincided with a visible puncture hole at the dura. It was clearly different from the slightly bluish pigmentation of the epidural fat observed in all animals from the sham-operated group and in most of the animals from the experimental group. This pigmentation was more marked at the dural root outlets, possibly due to draining of CSF through the perineurium at this site (12).
Beards et al. (13) noted a higher incidence of PDPH when the EBP was applied prophylactically than when it was applied 24–48 h later. This difference was attributed to the higher transdural pressure in the earlier stages due to the CSF outflow. In our study, fibrin glue was injected against a fistula driving pressure of 24.5 [17–31] cm H2O, which was not an impediment for sealing. Moreover, the final pressure supported by the fibrin glue patch might be in the order of 60 cm H2O, since the median final pressure measured at the cisterna magna 10 min after Evans blue infusion was 46.5 [34–68] cm H2O in the sitting position, and, accordingly, the pressure at the lumbar puncture site must be higher (in fact, we measured a CSF pressure difference of 14 [12–19] cm H2O between the cisterna magna and the lumbar puncture site in the sitting position before Evans blue infusion). This would imply maintained pressure values that are higher than those measured in humans in the sitting position, even after transient Valsalva maneuvers: 8.5–17.8 cm H2O after cough and 18–20 cm H2O after uterine contractions (14).
A possible advantage of the prophylactic injection of fibrin glue through the same needle provoking the tear would be the immediacy of the needle tip to the dural tear site. This may limit the volume of glue necessary to obtain an appropriate seal. Clinical reports of the use of a delayed fibrin glue injection to attempt to seal fistulae secondary to surgery or to prolonged use of the indwelling intrathecal catheter (6,7) suggest that higher volumes have a lower sealing rate under these conditions. The reason may be that the interspace injected is a different one from that where the fistula is located. In our study, the volume injected did not span more than two intervertebral spaces. This may suggest that percutaneous patching with fibrin glue, at least with the volumes we used, may only seal tears located at the same vertebral level.
Also, different volumes of blood have been used for EPB, and higher volumes were more effective in limitating the PDPH (15,16). The injection of large volumes of blood may be detected by magnetic resonance imaging (17), and injection of 18–20 mL is associated with epidural compression from 3 min to 3 h after injection (18). Thus, stopping the CSF leak could be due to an increase in epidural pressure when such large volumes are injected. In our study, CSF pressure recordings were not significantly changed after epidural injection of 1.4 mL of fibrin glue. The lack of increase in CSF pressure we found suggests fibrin glue seals the leak by a mechanism other than pressure effect. However, this does not imply a higher efficacy of fibrin glue over blood to obtain a good patch, an issue that remains to be clarified in comparative studies between the volumes of fibrin glue or blood needed to seal the same CSF leaks with a similar success rate using either technique. Furthermore, the ability of a fibrin glue patch effectively stopping a CSF leak to clinically improve PDPH remains to be demonstrated.
Several reservations before introducing fibrin glue patch in the clinical practice should be considered. The need to freshly prepare the mixture may be cumbersome in an acute situation, such as an incidental dural puncture. The possibility of inducing arachnoiditis, fibrous adhesions, infection, or a mass effect intradurally may be of concern, although no documented case has been reported (6,7). Fibrin glue components, fibrinogen, and thrombin are extracted from human pooled plasma, which constitutes the main problem with its clinical application. This may lead to infectious diseases transmission (19) or the induction of immune reactions. In fact, Mitsuhata et al. (20) reported an anaphylactic reaction after topical fibrin glue application. Also, immunization against bovine components of fibrin glue may occur (21). However, Tissucol Immuno® fibrin glue, both in the selection of the plasma samples and in its storage and transport, fulfills the World Health Organization specifications, including deactivation of several viruses by steam treatment. Also, the polymerase chain reaction is used to check a very low level of detectable viral load (22). Until now, no documented case of viral transmission using Tissucol Immuno® fibrin glue has been reported (6).
In conclusion, an epidural patch of 1.4 mL of fibrin glue applied directly through the same 17-gauge needle used to provoke a dural tear was able to seal the tear and prevent CSF leakage against pressures of 24.5 [17–31] cm H2O in 87.5% of cases in a swine model. Further studies may show if this can be applied to the prophylaxis of PDPH associated with accidental dural tears during epidural anesthesia in humans.
|
Acknowledgments |
|---|
Tissucol® was donated by Laboratorios Immuno, Barcelona, Spain. The authors wish to thank Professor E. Guijarro and Ms. Alicia Esparza for their help during the experimental procedure.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. Worner, K. Kothbauer, and H. Gerber Intrathecal Morphine Pump Malfunction Due to Leakage at the Catheter Connection Site: A Rare Problem and Its Prevention Anesth. Analg., June 1, 2009; 108(6): 1994 - 1995. [Full Text] [PDF]
|
|
|
|
|
|
J. S. Kroin, S. K.S. Nagalla, A. Buvanendran, R. J. McCarthy, K. J. Tuman, and A. D. Ivankovich The Mechanisms of Intracranial Pressure Modulation by Epidural Blood and Other Injectates in a Postdural Puncture Rat Model Anesth. Analg., August 1, 2002; 95(2): 423 - 429. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
