J Thorac Cardiovasc Surg 1999;118:597-602
© 1999 Mosby, Inc.
SURGERY FOR ACQUIRED CARDIOVASCULAR DISEASE |
RETROGRADE PERFUSION OF THE SPINAL CORD DURING AORTIC CROSSCLAMPING: INITIAL OBSERVATIONS IN THE SWINE MODEL
F. Follis, MDa,
R. Dragan, BSa,
K. S. Blisard, MD, PhDb,
M. Hartshorne, MDc,
T. Temes, MDa,
S. B. Pett, Jr , MDa,
J. A. Wernly, MDa
From the University of New Mexico Health Sciences Center,a Department of Cardiothoracic Surgery, Albuquerque, Gila Regional Medical Center,b Silver City, and Veterans’ Administration Medical Center,c Albuquerque, NM.
Funding was provided by the Paralyzed Veterans of America/Spinal Cord Research Foundation. Work was performed at Veterans’ Administration Medical Center, Albuquerque, NM.
Address for reprints: Fabrizio Follis, MD, University of New Mexico Health Sciences Center, Department of Cardiothoracic Surgery, 2211 Lomas Blvd NE, Albuquerque, NM 87131 (E-mail: follis99{at}hotmail.com).
 |
Abstract
|
|---|
Background: Retrograde perfusion has emerged as a useful technique for the preservation of the heart and brain when arterial circulation is interrupted. Herein, this study was designed to test the hypothesis that retrograde perfusion of the azygos vein is sufficient to maintain viability of the spinal cord during aortic occlusion in the swine model.
Methods: Female swine, 17 to 22 kg, underwent left thoracotomy, creation of a shunt between the aortic arch and the azygos vein, and aortic crossclamping for 60 minutes: the shunt was open in the retrograde perfusion group (n = 5) and closed in the control group (n = 4). The animals were evaluated for neurologic function for 8 days and killed. Spinal cords were processed for histologic examination. Additional animals underwent left thoracotomy and injection of a casting solution in the azygos vein (n = 2), left thoracotomy and angiography of the azygos vein (n = 2), and a compartmentalization procedure to separate the azygos vein from the caval system followed by angiography (n = 2).
Results: Differences in the neurologic (2-sample t test, P = .11) and histologic (2-sample t test, P = .65) scores of retrograde perfusion and control groups were likely due to chance. Casting and angiography groups showed extensive collaterals between azygos and caval systems, only partially interrupted by compartmentalization.
Conclusions: Retrograde perfusion does not protect the spinal cord from ischemic injury. The collateral network between the azygos and caval systems prevents the oxygenated blood from reaching the cord. Surgical separation between the 2 systems was only partially successful in this study.
|
Introduction
|
|---|
Operations on the thoracic aorta demand 2 conflicting objectives, a bloodless field to perform the procedure and maintenance of blood flow to preserve cord viability. Conceivably, the paradox can be resolved by retrograde perfusion of the spinal cord through the azygos vein while the circulation in the aorta is interrupted to allow surgical repair.
Organ perfusion in a retrograde fashion, from the venous to the arterial compartment, is a new and revolutionary concept that has become the standard of care in the preservation of the heart during cardiac operations. Similarly, retrograde perfusion of the brain during circulatory arrest is being used with increasing frequency in the treatment of dissections and aneurysm of the arch.
1-5
|
Methods
|
|---|
Aortic crossclamping.
Female Duroc swine, weighing 17 to 22 kg, had anesthesia induced with a cocktail of tiletamine and zolazepam (Telazol, 11.0 mg/kg), xylazine (2.2 mg/kg), and atropine (0.04 mg/kg) given intramuscularly. They were intubated and their lungs were ventilated with supplemental oxygen with the use of a volume-cycled ventilator. Tidal volume and rate were adjusted to maintain normal blood gases. Anesthesia was continued with isoflurane 1% to 2%. The right femoral artery was exposed and a 20F Angiocath infusion catheter (Olympus America, Inc, Melville, NY) was inserted to monitor the distal aortic pressure. Likewise, the right brachial artery was cannulated for monitoring of proximal blood pressure.
Next, a left thoracotomy was performed. After systemic heparinization with a 250 mg/kg dose of heparin, the distal aortic arch was cannulated with an 8F Bio-Medicus cannula (Medtronic Bio-Medicus, Eden Prairie, Minn) and the vein azygos sinistra with a 13F DLP retrograde cardioplegia catheter (DLP, Inc, Walker, Mich). Azygos vein pressure was monitored through a separate port built into the catheter. The 2 cannulas were connected with 
-inch bypass 70 tubing that included a probe (Bio-Medicus flow probe, inch) to measure flow in the shunt (Fig 1).
After these procedures, the descending thoracic aorta at the level of the ligamentum arteriosum was crossclamped for 60 minutes. Complete occlusion was confirmed by a drop in the blood pressure in the distal aorta measured by the femoral arterial line and a simultaneous rise in the proximal aortic pressure (Fig 2). In the control group the shunt was kept closed. In the retrograde perfusion group the shunt was open and flow was measured. At the end of aortic occlusion, the shunt was clamped and the aortic crossclamp released. Rectal temperature was monitored and kept constant with a warming blanket. Heparin was reversed with protamine. All the animals recovered and were observed for 7 days. The pigs were evaluated daily by means of a quantitative neurologic score devised to appraise neurologic deficits of the hind limbs. At the end of the observation period the animals were anesthetized, perfused with a fixative, and the spinal cords were explanted and submitted to histologic examination.
View larger version (34K):
[in this window]
[in a new window]
|
Fig. 2. Simultaneous recording of proximal, distal aortic, and azygos vein pressures before and after aortic crossclamping with the shunt open.
|
|
Casting procedure.
Each animal was anesthetized and a left thoracotomy was performed. After systemic heparinization with a 250 mg/kg dose of heparin, the descending thoracic aorta was cannulated with a DLP right-angled aortic cannula and the vein azygos sinistra with a 13F DLP retrograde cardioplegia catheter. The cannulas were connected with a roller pump. A No. 28 Pacifico right-angled cannula was inserted in the left atrial appendage and connected to a drainage bag. After these procedures, the animal’s blood was exsanguinated into the drainage bag while the aorta and azygos vein were perfused with normal saline solution until all the blood was cleared from the animal. Then the femoral, axillary, and jugular veins were ligated. Afterward, a mixture (250 mL) composed of base solution, catalyst, promoter, and the pigment was made and injected only through the cannula in the azygos vein with a roller pump. The solution was left to "cure" for 12 hours. The carcasses were reduced to a torso and submerged in 32% potassium hydroxide for 1 week and then in 6N hydrochloric acid for 3 weeks to corrode the tissues and bones and obtain a colored cast of the vessels in the study.
Angiography.
Each animal was anesthetized, a left thoracotomy was performed, and the azygos vein was cannulated with a 13F catheter after systemic heparinization. Then the animal was moved to the angiography suite where a dose (50 mL) of high-osmolality contrast material was injected into the azygos vein cannula while digital subtraction angiograms were taken. Afterward, the animal was killed.
Compartmentalization procedure.
Each animal was anesthetized and placed in the supine position; next, both internal jugular veins were skeletonized from the base of the skull to the thoracic inlet. The external jugular veins were ligated. The animal was then turned to a left lateral position and a right retroperitoneal approach was used to dissect and ligate the communicating veins between the azygos system and the inferior vena cava. Attention was then directed to the right side of the chest, which was entered through the 4th intercostal space. Again, the communicating veins between the superior vena cava and the azygos system were divided. On completion, the animal awakened.
Neurologic assessment.
The assessment included clinical criteria of hind limb neurologic function according to a modified Tarlov scale:
Grade 1: No voluntary function (complete randitemlysis)
Grade 2: Movements of joints perceptible
Grade 3: Active movement of joints (inability to stand)
Grade 4: Able to stand (unable to walk)
Grade 5: Complete recovery (able to stand, walk, run)
Histologic examination.
Each animal was anesthetized and a left thoracotomy was performed. The descending thoracic aorta was cannulated with a 12F Bio-Medicus cannula and the left atrial appendage with a No. 28 Pacifico cannula connected to a drainage bag. Heparin was given and the animal’s blood was exsanguinated into the drainage bag while 4 L of normal saline solution was infused into the thoracic aorta with a roller pump at a pressure of 30 mm Hg. Once the blood was completely removed, 1 L of formalin was infused into the aorta. Then the spinal cord was explanted, fixed in formalin, and processed for histologic examination. Spinal cord sections were stained with hematoxylin and eosin and luxol fast blue with a periodic acid–Schiff (PAS) counterstain. The histologic changes were scored according to the following criteria:
Gray matter gliosis
Gray matter PAS-positive material
Number of motor neurons
White matter myelination, white matter gliosis
White matter PAS-positive material
Each element was scored on a scale of negative (–), sl +, +, ++, and +++. The total scores were then combined to yield an overall histologic score as follows:
0: all – or 1 sl +
1: both gray and white sl + or 1 +
2: both gray and white +
3: ++ or +++
Experimental protocol.
The animals received humane care in compliance with the "Guide for the Care and Use of Laboratory Animals" published by the National Academy Press, 1996. Fifteen animals were divided into the following groups: retrograde perfusion (n = 5) and control (n = 4), casting (n = 2), angiography (n = 2), and compartmentalization plus angiography (n = 2) groups.
|
Results
|
|---|
Retrograde perfusion and control groups.
A population mean was calculated from individual lesion scores on postoperative day 7 for the retrograde perfusion (2.0 ± 0.4) and control (1.0 ± 0.0) groups. The difference (P = .11) observed by Wilcoxon 2-sample t test could be due to chance. Eight spinal cords, 5 in the retrograde perfusion and 3 in the control, were available for pathologic examination: complete loss of motor neurons, pronounced gliosis, and accumulation of macrophages in both anterior and posterior horns throughout the length of the lumbar cord were seen. The surrounding white matter displayed loss of myelin and gliosis. PAS-positive material (breakdown products) was abundant in both gray and white matter. The mean histologic score of the retrograde perfusion group was 14.6 ± 0.8 versus 14.0 ± 0.7 of the control. Wilcoxon 2-sample test yielded a P value of .65.
With the exception of the azygos vein pressure and shunt flow during aortic occlusion, there were no significant differences between the physiologic variables (weight, rectal temperature, proximal and distal aortic pressure) of the 2 groups by 2-sample t test (Table I).
Casting group.
Two animals underwent the casting procedure. As expected, injection of the azygos vein resulted in cast formation of the inferior and superior venae cavae, liver, and kidneys, but it also resulted in cast formation of an extensive perimedullar and perivertebral venous network. Several of the veins connecting the azygos to the cava systems were visible (Fig 3).
View larger version (62K):
[in this window]
[in a new window]
|
Fig. 3. A cast of the inferior vena cava, liver, superior vena cava (left), and the azygos system with perimedullary and perivertebral venous network (right). Because of postmortem distortion, the 2 systems are closer than in vivo. Note the collaterals joining the 2 systems.
|
|
Angiography group.
Two animals were subjected to angiographic study of the azygos system. Contrast media injected into the azygos vein quickly opacified the inferior vena cava through a collateral bed (Fig 4).
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 4. Before compartmentalization, rapid filling of the inferior vena cava (left) through collaterals on injection of the azygos vein (right).
|
|
Compartmentalization plus angiography.
Two animals underwent the compartmentalization procedure and then were subjected to angiographic study of the azygos system. Despite separation of the majority of the communicating veins, few collaterals could be identified (Fig 5).
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 5. After compartmentalization, successful interruption of the collaterals to the inferior vena cava, but rapid opacification of the superior vena cava through the anastomotic bed in the cervical region.
|
|
|
Discussion
|
|---|
Despite the indisputable appeal of this approach, we were not able to prove the hypothesis that retrograde perfusion of the spinal cord reduces ischemic injury produced by 1 hour of crossclamping of the aorta in the pig model. The major obstacle to the achievement of this goal was the extensive collateral circulation between the azygos system and the superior and inferior venae cavae, which shunted the oxygenated blood to the cavae through a path of lesser resistance. Further modifications of the animal preparation designed to separate the 2 systems and "compartmentalize" the azygos bed were unsuccessful owing to the complexity in identifying and dividing all the collateral vessels connecting the 2 systems.
A discussion of the anatomic premises, as well as the clinical and experimental experience with retrograde perfusion of the brain, is useful to further analyze the problem. The surface veins of the spinal cord, which include the venous plexus of the pia mater and 2 large venous trunks (the ventral and the dorsal median spinal veins), drain into the radicular veins; they, in turn, empty into the large epidural venous plexus. The latter communicates with the vertebral vein in the cervical region, with the azygos and hemiazygos veins in the thoracic region, and with the inferior vena cava in the lumbosacral region. 6,7 In the swine, both azygos veins drain into the vein azygos sinistra, which empties into the coronary sinus. 8 Both the azygos and the spinal veins are valveless.
In the published literature, several experimental studies of retrograde perfusion of the brain have been carried out in the dog model, 9-14 whereas only 2 investigations have reported data on retrograde perfusion of the spinal cord. In a protocol of selective perfusion of the superior vena cava in dogs, regional tissue blood flow was measured with the colored microsphere method 13: at a pressure of 30 mm Hg, retrograde perfusion provided one third of spinal cord flow supplied by antegrade flow of cardiopulmonary bypass. In an anatomic study on cadavers, 600 mL of latex was injected into the superior vena cava at 100 mL/min with and without inferior vena cava ligation (groups I and II). In both groups massive injection of latex was observed in the intrarachidian, extrarachidian, and perimedullar venous plexuses of the cervical and thoracic cords. 15
On the basis of these anatomic and experimental observations, it can be speculated that the spinal and azygos veins represent an open valveless system through which the spinal cord could be perfused during aortic crossclamping. Conversely, some would dispute that the experimental data of Usui and associates 10-12 are unfounded and their experimental protocol was likely to count recirculation volume of the blood flow. Others could argue that all the above premises are false because retrograde perfusion can never reach the spinal cord tissue owing to the characteristics of the microcirculation or because, even in retrograde perfusion of the brain, there is no proof that the protective effect is based on the supply of nutritive blood flow rather than more effective regional cooling.
In fact, effective retrograde perfusion of the brain during circulatory arrest has been more difficult to demonstrate than in other organs like the heart, and the evidence rests mainly on the observation of blood return from the head vessels (which could be the result of arteriovenous shunting outside the brain) and a better neurologic outcome in cases of prolonged arrest. Here the issue of wasted perfusion through the collateral venous circulation becomes crucial and presents aspects similar to those we encountered during retrograde perfusion of the spinal cord. There is experimental and clinical evidence to expect that most of the retrograde cerebral perfusion flow is directed in a descending manner through the azygos-caval connection into the large-capacitance low-resistance systemic venous bed. 15 In a setting of total circulatory arrest, this drawback can be partially overcome by total body pressurization of the entire venous system with a roller pump. Under these circumstances, retrograde perfusion of the azygos system could reach the spinal cord as the experiment by Oohara and colleagues 13 in dogs and the findings in our casting group seem to indicate.
In presence of the beating heart and normal circulation, however, the problem is compounded by pressure differences in the azygos-caval venous systems, which leads to obligatory stealing from the azygos vein to the systemic venous bed through a path of lesser resistance.
|
Conclusion
|
|---|
In conclusion, although this investigation reports negative results, it has helped to define the problems associated with such an approach and perhaps the insurmountable limits of this idea but could, however, serve as a springboard for incoming research in this rather novel field. Thus, if the ischemic insult at normothermia in this study was too severe to allow differentiation between groups, a future study should compare the 2 groups under hypothermia. Furthermore, retrograde perfusion of the azygos vein as a way of cooling the spinal cord may prove to be a worthwhile line of investigation and provide an efficient method, superior to those already proposed. In this respect, a recent report by Ross and coworkers 16 supports the validity of this approach. Finally, the relationship between azygos vein, distal aortic and central venous pressures, and their variation with respect to neurologic outcome during perfusion in the swine model merit a more in-depth understanding and may lead to interesting findings.
|
Acknowledgments
|
|---|
We thank Clifford Qualls, PhD (Professor of Statistics, Emeritus, and Biostatistician, Research Office, Veteran’s Affairs Medical Center, Albuquerque, NM) for assistance in the statistical evaluation of data and Deborah L. Heuser for technical assistance.
|
References
|
|---|
-
Takamoto S, Matsuda T, Harada M, Shimamura Y, Miyata S. Simple hypothermic retrograde cerebral perfusion during aortic arch surgery. J Cardiovasc Surg 1992;33:560-7. [Medline]
-
Yasuura K, Ogawa Y, Okamoto H, et al. Clinical perfusion of total body retrograde perfusion to operation for aortic dissection. Ann Thorac Surg 1992;53:655-8. [Abstract]
-
Lin PJ, Chang C-H, Tan PPC, et al. Protection of the brain by retrograde cerebral perfusion during circulatory arrest. J Thorac Cardiovasc Surg 1994;108:969-74. [Abstract/Free Full Text]
-
Safi HJ, Heather HW, Winter JN, et al. Brain protection via cerebral retrograde perfusion during aortic arch aneurysm repair. Ann Thorac Surg 1993;56:270-6. [Abstract]
-
Murase M, Maeda M, Koyama T, et al. Continuous retrograde cerebral perfusion for protection of the brain during aortic arch surgery. Eur J Cardiothorac Surg 1993;7:597-600. [Abstract]
-
Bowsher D. A comparative study of the azygos venous system in man, monkey, dog, cat, rat and rabbit. J Anat 1954;88:400-6. [Medline]
-
Batson OV. The vertebral veins system. AJR 1957;78:195-212.
-
Nickel R. Lehrbuch der Anatomie der Haustiere, Band III: Kreislaufsystem, Haut und Hautorgane. Berlin: Verlag Paul Parey; 1984. p. 191-3.
-
Nojima T, Mori A, Watarida S, Onoe M. Cerebral metabolism and effects of pulsatile flow during retrograde cerebral perfusion. J Cardiovasc Surg 1993;34:483-92. [Medline]
-
Usui A, Hotta T, Hiroura M, et al. Retrograde cerebral perfusion through a superior vena caval cannula protects the brain. Ann Thorac Surg 1992;52:47-53.
-
Usui A, Oohara K, Liu T, et al. Determination of optimum retrograde cerebral perfusion conditions. J Thorac Cardiovasc Surg 1994;107:300-8. [Abstract/Free Full Text]
-
Usui A, Oohara K, Liu T, et al. Comparative experimental study between retrograde cerebral perfusion and circulatory arrest. J Thorac Cardiovasc Surg 1994;107:1228-36. [Abstract/Free Full Text]
-
Oohara K, Usui A, Tanaka M, Abe T, Murase M. Determination of organ blood flows during retrograde inferior vena caval perfusion. Ann Thorac Surg 1994;58:139-45. [Abstract]
-
Safi HJ, Iliopoulos DC, Gopinath SP, et al. Retrograde cerebral perfusion during profound hypothermia and circulatory arrest in pigs. Ann Thorac Surg 1995;59:1107-12. [Abstract/Free Full Text]
-
De Brux J-L, Subayi J-B, Pegis J-D, Pillet J. Retrograde cerebral perfusion: anatomic study of the distribution of blood to the brain. Ann Thorac Surg 1995;60:1294-8. [Abstract/Free Full Text]
-
Ross SD, Kern JA, Gangemi JJ, et al. Hypothermic retrograde venous perfusion cools the spinal cord and reduces paraplegia during thoracic aortic surgery (abstract). American Association for Thoracic Surgery, 1999; 79th Annual Meeting. p. 202.
Received for publication Feb 22, 1999. Revisions requested May 10, 1999; revisions received May 26, 1999. Accepted for publication June 16, 1999.
Top
Abstract
Introduction
