Appraisal of cerebrospinal fluid alterations during aortic surgery with intrathecal papaverine administration and cerebrospinal fluid drainage Lars G. Svensson, MB, PhD, Daniel F. Grum, MD, Marilyn Bednarski, R N , Delos M. Cosgrove III, MD, and Floyd D. Loop, MD, Cleveland, Ohio We have previously described a technique for intrathecal administration of papaverine and cerebrospinal fluid drainage to prevent paraplegia after aortic surgery. Herein we report the cerebrospinal fluid and hemodynamic alterations that occurred in 11 patients who had 30 mg of a specially prepared papaverine hydrochloride 10% dextrose solution injected before aortic cross-clamping and also had cerebrospinal fluid drainage. A mean of 26.6 ml (SD - 7.1 ml) was drained before and 34,6 ml (SD -+ 24.1 ml) was drained during aortic cross-clamping. The cerebrospinal fluid pressure increased significantly with anesthetic induction (p < 0.03), during the period between anesthetic induction and cerebrospinal fluid drainage (p < 0.005), and with aortic cross-clamping (p < 0.05). These cerebrospinal fluid pressure alterations were similar to central venous pressure increases with a significant linear correlation between cerebral spinal fluid pressure and central venous pressure before anesthetic induction (r2 = 0.81, p < 0.005), and both before (r 2 = 0.94,p < 0.005) and after (~ = 0.74,p < 0.005) aortic cross-clamping. As expected, cerebrospinal fluid pressure was significantly reduced by cerebrospinal fluid drainage before aortic cross-clamping (p < 0.001). The admiuistration of intrathecal papaverine had no significant effect on mean arterial pressure, systemic vascular resistance, cerebrospinal fluid pressure, nor the pH of cerebrospinal fluid. Neither were there any complications noted related to the technique. All the patients survived, and no new immediate postoperative paraparesis or paraplegia occurred. The technique described may be used with other methods, such as distal aortic perfusion, in an attempt to protect the spinal cord against ischemie damage, and it may allow the surgeon a longer time to reanastomose critical segmental arteries to maintain spinal cord blood supply. (J VASC StatG 1990;11:423-9.) The protection of the spinal cord against ischemic injury when blood flow in the thoracic aorta is interrupted requires both, first, the provision of at least minimally adequate blood supply during the period of aortic interruption and, second, the restoration of normal blood flow to the spinal cord after the aortic circulation is restored by reattachment of critical segmental arteries.~,2 We have previously reported a new technique for protecting the spinal cord from prolonged ischemia during aortic surgery, which we began to use in humans in July 1985.1,3 The technique entailed the administration of intrathecal papaverine From the Departments of Thoracic and CardiovascularSurgery, and Anesthesia, ClevelandClinic Foundation. Reprint requests: Lars G. Svensson,MD, Departmentof Surgery, Baylor College of Medicine, One BaylorPlaza, Houston, TX 77030. 24/1/17246
combined with cerebrospinal fluid (CSF) drainage, ~ which we showed in nonhuman primates (Pupio ursinus) to cause both dilation of the anterior spinal artery and increased spinal cord blood flow thus preventing paraplegia? Similarly, in our recent report of studies in 11 humans) no immediate postoperative neurologic events occurred, although the sample size was small. In this study we report for the first time the hemodynamic and CSF alterations, including the changes in CSF pH and Po2 that occur with injection of intrathecal papaverine. The CSF pressure (CSFP) alterations that occur with CSF drainage during aortic cross-clamping in humans is also evaluated. Furthermore, we demonstrate and discuss the close relationship that exists between central venous pressure (CVP) and CSFP during the various periods of aortic surgery. Finally, the published reports of use of CSF 423
424
Journalof VASCULAR SURGERY
Svensson et al.
Table I. Alterations during aortic surgery using intrathecal papaverine and cerebrospinal fluid drainage (mean, _+SD). Catheter
MAP CVP CSFP
Drainage
B
A
B
A
B
105 22 7 3.3 --
106 23 8 2.6 12 3.6 2.8 161
107 20 8 1.9 13 d 2.3 2.8 159
93 28 9b 2.6 16 0.1) occurred with aortic cross-clamping. However, during the period between induction of anesthesia and aspiration of CSF, the cardiac index was increased (p < 0.05), and systemic vascular resistance was decreased (p < 0.01) by pharmacologic manipulation and administration of intravenous solutions, including mannitol. The significant increase (p < 0.0001) in the CVP during this period probably reflects the administration of fluid, including mannitol. As expected, MAP significantly fell with aortic unclamping (p < 0.01), as did systemic vascular resistance (p < 0.01), and cardiac index significantly increased (p < 0.01). The injection of intrathecal papaverine resulted in no significant measured effects on CSFP, CVP, MAP, cardiac index, systemic vascular resistance, CSF pH, nor CSF P Q (Table I). The CSFP correlated very significantly with the CVP before induction of general anesthesia (r 2 = 0.81, p < 0.005) and notably both before (r 2 = 0.94, p < 0.005) and after (r 2 = 0.74, p < 0.005) aortic cross-clamping. Immediately after unclamping of the aorta, no correlation between CVP and CSFP could be found, although, the CSFP tended to equilibrate over a period of time with the CVP. With the exception of a weak correlation between CSFP and MAP occurring sometime after CSF drainage, probably reflecting CSF secretion, first before injection of intrathecal ,papaverine (r 2 = 0.34, p = 0.05) and
426 Svensson et al.
second 10 minutes after aortic undamping (r 2 = 0.45,p < 0.03), no other correlation could be found with MAPs. No other correlations were found between CSFP and cardiac index nor systemic vascular resistance (p > 0. i). The mean CSF volume drained before aortic cross-clamping, including spillage during catheter insertion, was 26.6 ml _+ 7.1 ml. The mean CSF volume drained during aortic cross-clamping was 34.6 ml _+ 24.1 ml. The mean CSF p H values after aspiration of 20 ml of CSF, after papavcrine injection (aspirated during aortic cross-clamping), after unclamping, and at completion of the operation were respectively: 7.28 _+ 0.14, 7.29 + 0.08, 7.23 + 0.13, and 7.26 _+ 0.11. Similarly, the respective CSF partial gas pressures of oxygen (Po2) were: 125 _+ 26, 106 _+ 37, 128 _ 36, and 114 _+ 31. Clearly,, neither thc p H nor Po2 changes in CSF were significantly altered (p > 0.1) by the presence of the injected papaverine nor by potential ischcmia of the spinal cord during aortic cross-clamping. The mean aortic cross-clamp time was 43 minutes _+ 20 minutes. In 10 patients the intrathecal catheter was directed cranially and lay anterior to the spinal cord. In one patient the catheter had turned caudally. One of the 11 patients in the postoperative period had a recurrence of a unilateral temporary paraparesis that he had previously suffered after a coronary artery bypass operation. Another patient, who initially had no neurologic dysfunction, on the third postoperative day suffered a respiratory arrest requiring reintubation and readmission to intensive care unit. He was found to have a temporary paraparesis, from which he recovered by the time of his discharge from the hospital. There were no postoperative deaths. No patients suffered any complications related to the papaverine injection nor postoperative headaches nor other complications related to the spinal catheter. DISCUSSION
This study reports for the first time the CSF alterations that occur in humans with both intrathecal papaverine administration and also with CSF drainage during the course of aortic surgery. We found that neither hemodynamic nor CSFP alterations occurred with the injection of intrathecal papaverine, and therefore we believe that we can safely assume that the CSFP changes that occurred with aortic cross-clamping were not influenced by intrathecal papaverine. Thus it is of interest that CSFP increased progressively during the course of aortic surgery, and
Journal of VASCULAR SURGERY
that these increascs correlated best with incrcascs in the CVP. Before the author's prospective use of intrathecal papaverine and CSF drainage in humans, which began in July 1985 and was reported in 19864 and then described in detail, 1,3 CSF drainage had only been described in two patients having aortic cross-clamping in 1958 and 19595 without compiication. Since we have as yet seen no harmful effects from the use of intrathecal papaverine with CSF drainage in 34 patients and no immediate postoperative paraplegia, we believe that intrathecal papaverine can be safely used intrathecally for its vasodilatory effect and other possible protective effects. 1,4 However, strictly adhering to the precautions we have described in our methodology, induding attention to a totally sterile technique, adequate preinsertion intravenous fluid hydration, possible hemostasis abnormalities, and gentle insertion of the catheter we believe is crucial. We stress again that the papaverine and the catheter flushing solution must be at body temperature and be slowly injected as described. As with other spinal anesthetic techniques, the risk appears to be associated with the technique of catheter insertion and spinal anesthetics in general, such as severe hypotensive sympathetic responses, rather than the intrathecal administration of papaverine. 6"7 Our observations in humans suggest that the high CSFP values during aortic cross-clamping, also noted in animal studies, 8-~zare not solely due to aortic crossclamping, but that the induction of anesthesia and the subsequent period up to CSF drainage should be considered as additive factors that increase the CSFP. Of course, these alterations may have been different had we not drained CSF or pharmacologically manipulated the blood pressure during aortic cro~) clamping.13 Furthermore, we have noted that the use of distal aortic perfusion reduces the increase in CSFP that occurs with aortic cross-clamping? Despite drainage of 20 ml of CSF, the CSFP gradually returned to predrainage levels before cross-clamping, and this gradual rise in CSFP correlated with the mean arterial pressure (ta = 0.34, p < 0.05). We postulate that this was either because CSF was replenished (at the rate of approximately 25 ml an hour; a rate that is dependent on arterial pressure t4) or because the venous capacitance vessels surrounding the spinal cord had expanded to tal~e up the vacated space. We speculate that the possible reasons for the rapid rise in CSFP with aortic cross-clamping include the following: (1) The higher proximal MAP causing
Volume 11 Number 3 March 1990
either dilation of intracranial arteries or expansion of nervous tissue volume. However, cerebral autoregulation would probably prevent any expansion of neural tissue or arterial vessel volume from occurring because of increased systemic arterial pressure or cerebral blood flow) 3as (2) Increased CSF secretion, possibly related to increased blood flow to the brain. However, previous experiments purporting to support this, which mimicked either rapid tumor expansion or intracranial hemorrhage, cannot, in the true sense, be considered to be analogous to aortic cross-clamping, where there is no forceful injection ofintrathecal fluid. 16,17(3) The rise in CVP resulting in back pressure in the intracranial veins and venous sinuses. In our opinion, we tend to favor the third postulate. The classic Monro-Kellie hypothesis remains generally relevant, namely, that the total cranial and spinal columaa volumes remain constant within the closed space that the brain and spinal cord occupy. Any volume increase of the components (nervous tissue volume, blood volume, CSF volume) must therefore be at the expense of the other components. Normal CSFP will thus only be maintained until the compensatory mechanisms are exceeded. 18 However, with aortic cross-clamping, the one compensatory mechanism, namely compression of venous capacitance vessels (loss of blood volume together with the above), rather than compensating for any increases in the intracranial components, the venous capacitance vessels are the more likely cause of the increase in CSFP. Blaisddl and Cooley8 have noted in dogs that CSFP increased with both anesthetic induction and cross-clamping, and in our own (unpublished data) and other animal experiments s,19 it has been found .nat CVP increased with aortic cross-clamping. It is known that a close relationship normally exists under physiologic conditions between subarachnoid, intracranial sinus, CVP, and the C S F P , 2°'21 and it is likely that spinal CSFP is similarly related to spinal venous pressures 2t including in Batson's plexus, the azygous vein (which drains segmental intercostal and lumbar veins), and in the superior vena cava. Since CVP tends to increase with induction of anesthesia and cross-clamping, the mechanism of increased CSFP is probably similar to that of the Queckenstedt test = or heart failure, 2a in that decreased venous return as a result of raised intrathoracic pressure or, heart failure, results in backpressure in the intracranial and perichordal capacitance vessels because of increased impedence to venous outflow. In addition, there may also be an increased volume of blood return from the
Intrathecal papaverine and cerebrospinatfluid drainage 427 upper body via the superior vena cava24 because of increased shunting between the upper body arterial and venous systems as a result of pharmacologic vasodilators, further adding to the venous congestion. The significant linear correlation we noted between CSFP and CVP both before anesthetic induction, before cross-clamping, and with cross-clamping thus supports the concept that the increases in CSFP are probably due to increases in CVP. However, it should be noted that vasodilators themselves (nitroprusside, nitroglycerine, and trimethaphan) can cause an increased CSFPY ,26 The implied or measured lumbar spinal cord CSFP, the site at which ischemia is maximal during aortic cross-clamping, 4,9,24,27particularly without distal aortic perfusion, 27is controversial. Dunbar et al. 2° found in dogs that CSFP in the lumbar spinal cord sac fell during aortic occlusion, whereas CSFP measured in the cisterna magna or intraventricularty increased. These pressure changes were reversed by terminating aortic occlusion. 2° Others 81m6 have measured the cisterna magna pressure in dogs during aortic cross-clamping and have confirmed, as noted by Dunbar et al.,20 that CSFP in the cisterna magna increases. However, it has been implied from these pressure changes measured in the cisterna magna that lumbar spinal cord pressures also increase. Indeed, we did find that CSFP, measured alongside the lumbar spinal cord, increases with aortic cross-clamping from a mean of i5 mm H g to i9 mm Hg, but not as markedly when compared to the increase from 13 mm Hg before induction of anesthesia to 20 mm Hg before CSF drainage. It is clear from the above discussion that CSFP measured in the cisterna magna may not accurately reflect the CSFP alterations in the lumbar spinal cord vicinity. Furthermore, Killen et al.28 reported that the distal aortic pressure did not reflect the spinal cord arteriolar (nor capillary) blood pressure, but that the higher segmental intercostal or lumbar artery end pressures were a closer approximation. Therefore by implication, the double assumption by Miyamoto et al. s of "relative spinal cord perfusion pressure" being equal to distal aortic pressure minus the cistema magna pressure probably does not hold true. Moreover, Griffiths et al. 21have shown that perfusion pressure is a far greater determinant of blood flow than the possible inhibitory effect of a raised CSFP. They concluded that "spinal cord blood flow can accomodate any fluctuations in CSFP." There are even reports of increased cerebral blood flow with an increased intracranial pressure, la It should be
428
Journal of VASCULAR SURGERY
Svensson et al.
noted that the resistance to reabsorption of CSF in dogs is 17 times greater than that of man, 14and thus any increases in CSFP in man may be more quickly compensated for. Whether CSF drainage alone will reduce the incidence of paraplegia after aortic cross-clamping in humans is an unanswered question, although apparently good results have been reported by McCullough et al.10 in 24 patients, 10 of whom were low Crawford type III thoracoabdominal aortic aneurysms and six type IV (total or near total abdominal aneurysm) with an expected paraparesis or paraplegia rate of 3.2% o r 2 . 1 % 29, respectively. Experiments in dogs, draining CSF from the cisterna magna, have shown discrepant results. Blaisdell and Cooley 8 in 1962 and more recent studies by others 9"11,16 report the incidence of postocclusion neurologic dysfunction is reduced. Nevertheless, it is of note that Blaisdell, despite his earlier work, does not believe that "we can provide any protection by modifying spinal fluid pressure." (Commentary on Bower et al.9) Furthermore, Killen et al.28 in a similar dog model were tmable to show a protective effect by draining CSF from the cisterna magna. Similarly, Wadouh et al?9 used a porcine model, and in our own nonhuman primate model, CSF drainage by a limited lumbar laminectomy failed to reduce the incidence of paraplegia. ~ In conclusion, it would appear that the acute changes in CSFP are primarily a function of increased intracranial and intraspinal venous blood volume, which is determined by venous back pressure during anesthetic induction, intravenous fluid administration and aortic cross-clamping. Whether CSF drainage alone will reduce the incidence of postoperative paraplegia is not proved and can only be evaluated by a prospective randomized study. Thus it would appear to us that the best method of preventing paraplegia, with its multifactorial causes, is to use intrathecal papaverine with CSF drainage to prolong the safe period of aortic cross-clamping, reanastomose identified critical segemental arteries by a technique that has been presented by the author, s° and use distal aortic perfusion in selected complex situations, such as aortic dissection. 2However, distal aortic perfusion has not proved to be a panacea in preventing either paraplegia because of spinal cord anatomic reasons27 or acute renal failure.2,3t Robert W. Stewart, MD, and Bruce W. Lyric, MD, allowed their patients to be entered into the study and gave valuable advice. The preoperative and postoperative neurologic assessments were performed by Antonio Salgado,
MD, and Anthony Furlan, MD, from the Department ot Neurology, and Marion Robinson, PhD, performed the statistical analyses.
REFERENCES
1. Svensson LG, Stewart RW, Cosgrove DM, et al. Intrathecal papaverine for the prevention of paraplegia after operation on the thoracic or thoracoabdominal aorta. J Thorac Cardiovasc Surg 1988;96:823-9. 2. Svensson LG, Loop FD. Prevention of spinal cord ischemia in aortic surgery. In: Bergan j'J, Yao JST, eds. Arterial surgery: new diagnostic and operative techniques. New York: Grune and Stratton, 1988:273-85. 3. Svensson LG, Stewart RW, Cosgrove DM, et al. Preliminary results and rationale for the use of intrathecal papaverine for the prevention of paraplegia after aortic surgery. S Aft J Surg 1988;26:153-9. 4. Svensson LG, Von Rirter CM, Groeneveld HT, et al. Crossclamping of the thoracic aorta. Influence of aortic shunts, laminectomy, papaverine, calcium channel blocker, allopurinol, and superoxide dismutase on spinal cord blood flow ~-,~t paraplegia in baboons. Ann Surg 1986;204:38-47. 5. Miyamota K, Keno A, Wada T, Kimoto S. A new and simple method of preventing spinal cord damage following temporary occlusion of the thoracic aorta by draining the cerebrospinal fluid. J Cardiovasc Surg 1960; 1:188-97. 6. Kane RE. Neurologic deficits following epidural or spinal anesthesia. Anesth Analg 1981;60:150-61. 7. Phillips OC, Ebner H, Nelson AT, Black MH. Neurologic complications following spinal anesthesia with lidocaine. Anesthesiology 1969;30:284-9. 8. Blaisdell FW, Cooley DA. The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid prssure. Surgery 1962;51:351-5. 9. Bower TC, Murray MJ, Glovicld P, Yaksh TL, Hollier LH, Pairolero PC. Effects of thoracic aortic occlusion and cerebrospinal fluid drainage on regional spinal cord blood flow in dogs: correlation with neurologic outcome. J VAsc SURG 1989;9:135-44. 10. McCullough JL, HoUier LH, Nugent M. Paraplegia aftc thoracic aortic occlusion: Influence of eerebrospinal fluid drainage. J VASCSuR~ 1988;7:153-60. 11. Grubbs PE, Matini C, Toporoff B, et al. Somatosensory evoked potentials and spinal cord perfusion pressure are significant predictors of postoperative neurologic dysfunction. Surgery 1988;104:216-23. 12. Molina JE, Cogordan J, Einzig S, et al. Adequacy of ascending aorta-descending aorta shunt during cross-clamping of the thoracic aorta for prevention of spinal cord injury. J Thorac Cardiovasc Surg 1985;90:126-36. 13. Lassen NA, Christensen MS. Physiology of cerebral blood flow. Br J Anaesth 1976;48:719-34. 14. Culfler RWP, Page L, Galicich J, Watters GV. Formation and absorption of cerebrospinal fluid in man. Brain 1968; 91:707-20. 15. Hickey R, Albin MS, Bunegin L, Gelineau J. Autoregulati~n of spinal cord blood flow: is the cord a microcosm of the brain. Stroke 1986;17:1183-9. 16. Oka Y, Miyamoto T. Prevention of spinal cord injury after cross-clamping of the thoracic aorta. Jpn J Surg 1984;14:15962.
Volume 11 Number 3 March 1990
17 Berendes JN, Bredee JJ, Schipperheyn JJ, Mashhour YAS. Mechanism of spinal cord injury after cross-clamping of the descending thoracic aorta. Circulation 1982;66(suppl I): 112-6. 18. Marmarou A, Shulman K, LaMorgese J. Compartmental analysis of compliance and outflow resistance of the cerebrospinal fluid system. J Neurosurg 1975;43:523-4. 19. Wadotth F, Lindemarm E-M, Amdt CF, Hetzer R, Borst HG. The arteria radicularis magna anterior as a decisive factor influencing spinal cord damage during aortic occlusion. J Thorac Cardiovasc Surg 1984;88:1-10. 20. Dunbar HS, Gutherie TC, Karpell B. A study of the cerebrospinal fluid pulse wave. Arch Neurol 1966;14:624-30. 21. Griffiths IR, Pitts LH, Crawford P_A,Trench JG. Spinal cord compression and blood flow. Neurology 1978;28:1145-51. 22. Bering EA. Choroid plexus and arterial pulsation ofcerebrospinal fluid. Arch Neurol Psychiat 1955;73:165-72. 23. Friedfeld L, Fishberg AM. The relation of the cerebrospinal and venous pressures in heart failure. J Clin Invest 1934; 13:495-501. 24. Geiman S, Rabbani S, Bradley EL. Inferior and superior vena caval blood flows during cross-clamping of the thoracic aorta in pigs, J Thorac Cardiovasc Surg 1988;96:387-92. 2f. Turner JM, Powell D, Gibson RM, McDowall DG. Intracranial pressure changes in neurosurgical patients during hypotension induced with sodium nitroprusside or trimetaphan. Br J Anaesth 1977;49:419-25.
Intrathecal papaverine and cerebrospinatfluid drainage
429
26. Dohi S, Matsumoto M, Takahashi T. The effects of nitroglycerin on cerebrospinal fluid pressure in awake and anesthetized humans. Anesthesiology 1981;54:511-4. 27. Svensson LG, Rickards E, Coull A, Rogers G, Fimmel CJ, Hinder RA. Relationship of spinal cord blood flow to vascular anatomy during thoracic aortic cross-clamping and shunting. J Thorac Cardiovasc Surg 1986;91:71-8. 28. Killen DA, Edwards RH, Tinsley EA, Boehm FH. Effect of low molecular weight dextran, heparin, urea, cerebrospinal fluid drainage, and hypothermia on ischemic injury of the spinal cord secondary to m o b ~ a t i o n of the thoracic aorta from the posterior parietes. J Thorac Cardiovasc Surg 1965;50:882-7. 29. Crawford ES, Crawford JL, Sail HJ, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J VASC SURe 1986;3:389-404. 30. Svensson LG, Patel V, Coselli JS, Crawford ES. Preliminary report of localization of spinal cord blood supply by hydrogen during aortic surgery. Ann Thorac Surg (In press)° 31. Svensson LG, Coselli JS, Sail HJ, Hess KR, Crawford ES. Appraisal of adjuncts to prevent acute renal failure after surgery on the thoracic or thoracoabdominal aorta. J VAsc SuRe 1989;10:230-9.
combined with cerebrospinal fluid (CSF) drainage, ~ which we showed in nonhuman primates (Pupio ursinus) to cause both dilation of the anterior spinal artery and increased spinal cord blood flow thus preventing paraplegia? Similarly, in our recent report of studies in 11 humans) no immediate postoperative neurologic events occurred, although the sample size was small. In this study we report for the first time the hemodynamic and CSF alterations, including the changes in CSF pH and Po2 that occur with injection of intrathecal papaverine. The CSF pressure (CSFP) alterations that occur with CSF drainage during aortic cross-clamping in humans is also evaluated. Furthermore, we demonstrate and discuss the close relationship that exists between central venous pressure (CVP) and CSFP during the various periods of aortic surgery. Finally, the published reports of use of CSF 423
424
Journalof VASCULAR SURGERY
Svensson et al.
Table I. Alterations during aortic surgery using intrathecal papaverine and cerebrospinal fluid drainage (mean, _+SD). Catheter
MAP CVP CSFP
Drainage
B
A
B
A
B
105 22 7 3.3 --
106 23 8 2.6 12 3.6 2.8 161
107 20 8 1.9 13 d 2.3 2.8 159
93 28 9b 2.6 16 0.1) occurred with aortic cross-clamping. However, during the period between induction of anesthesia and aspiration of CSF, the cardiac index was increased (p < 0.05), and systemic vascular resistance was decreased (p < 0.01) by pharmacologic manipulation and administration of intravenous solutions, including mannitol. The significant increase (p < 0.0001) in the CVP during this period probably reflects the administration of fluid, including mannitol. As expected, MAP significantly fell with aortic unclamping (p < 0.01), as did systemic vascular resistance (p < 0.01), and cardiac index significantly increased (p < 0.01). The injection of intrathecal papaverine resulted in no significant measured effects on CSFP, CVP, MAP, cardiac index, systemic vascular resistance, CSF pH, nor CSF P Q (Table I). The CSFP correlated very significantly with the CVP before induction of general anesthesia (r 2 = 0.81, p < 0.005) and notably both before (r 2 = 0.94, p < 0.005) and after (r 2 = 0.74, p < 0.005) aortic cross-clamping. Immediately after unclamping of the aorta, no correlation between CVP and CSFP could be found, although, the CSFP tended to equilibrate over a period of time with the CVP. With the exception of a weak correlation between CSFP and MAP occurring sometime after CSF drainage, probably reflecting CSF secretion, first before injection of intrathecal ,papaverine (r 2 = 0.34, p = 0.05) and
426 Svensson et al.
second 10 minutes after aortic undamping (r 2 = 0.45,p < 0.03), no other correlation could be found with MAPs. No other correlations were found between CSFP and cardiac index nor systemic vascular resistance (p > 0. i). The mean CSF volume drained before aortic cross-clamping, including spillage during catheter insertion, was 26.6 ml _+ 7.1 ml. The mean CSF volume drained during aortic cross-clamping was 34.6 ml _+ 24.1 ml. The mean CSF p H values after aspiration of 20 ml of CSF, after papavcrine injection (aspirated during aortic cross-clamping), after unclamping, and at completion of the operation were respectively: 7.28 _+ 0.14, 7.29 + 0.08, 7.23 + 0.13, and 7.26 _+ 0.11. Similarly, the respective CSF partial gas pressures of oxygen (Po2) were: 125 _+ 26, 106 _+ 37, 128 _ 36, and 114 _+ 31. Clearly,, neither thc p H nor Po2 changes in CSF were significantly altered (p > 0.1) by the presence of the injected papaverine nor by potential ischcmia of the spinal cord during aortic cross-clamping. The mean aortic cross-clamp time was 43 minutes _+ 20 minutes. In 10 patients the intrathecal catheter was directed cranially and lay anterior to the spinal cord. In one patient the catheter had turned caudally. One of the 11 patients in the postoperative period had a recurrence of a unilateral temporary paraparesis that he had previously suffered after a coronary artery bypass operation. Another patient, who initially had no neurologic dysfunction, on the third postoperative day suffered a respiratory arrest requiring reintubation and readmission to intensive care unit. He was found to have a temporary paraparesis, from which he recovered by the time of his discharge from the hospital. There were no postoperative deaths. No patients suffered any complications related to the papaverine injection nor postoperative headaches nor other complications related to the spinal catheter. DISCUSSION
This study reports for the first time the CSF alterations that occur in humans with both intrathecal papaverine administration and also with CSF drainage during the course of aortic surgery. We found that neither hemodynamic nor CSFP alterations occurred with the injection of intrathecal papaverine, and therefore we believe that we can safely assume that the CSFP changes that occurred with aortic cross-clamping were not influenced by intrathecal papaverine. Thus it is of interest that CSFP increased progressively during the course of aortic surgery, and
Journal of VASCULAR SURGERY
that these increascs correlated best with incrcascs in the CVP. Before the author's prospective use of intrathecal papaverine and CSF drainage in humans, which began in July 1985 and was reported in 19864 and then described in detail, 1,3 CSF drainage had only been described in two patients having aortic cross-clamping in 1958 and 19595 without compiication. Since we have as yet seen no harmful effects from the use of intrathecal papaverine with CSF drainage in 34 patients and no immediate postoperative paraplegia, we believe that intrathecal papaverine can be safely used intrathecally for its vasodilatory effect and other possible protective effects. 1,4 However, strictly adhering to the precautions we have described in our methodology, induding attention to a totally sterile technique, adequate preinsertion intravenous fluid hydration, possible hemostasis abnormalities, and gentle insertion of the catheter we believe is crucial. We stress again that the papaverine and the catheter flushing solution must be at body temperature and be slowly injected as described. As with other spinal anesthetic techniques, the risk appears to be associated with the technique of catheter insertion and spinal anesthetics in general, such as severe hypotensive sympathetic responses, rather than the intrathecal administration of papaverine. 6"7 Our observations in humans suggest that the high CSFP values during aortic cross-clamping, also noted in animal studies, 8-~zare not solely due to aortic crossclamping, but that the induction of anesthesia and the subsequent period up to CSF drainage should be considered as additive factors that increase the CSFP. Of course, these alterations may have been different had we not drained CSF or pharmacologically manipulated the blood pressure during aortic cro~) clamping.13 Furthermore, we have noted that the use of distal aortic perfusion reduces the increase in CSFP that occurs with aortic cross-clamping? Despite drainage of 20 ml of CSF, the CSFP gradually returned to predrainage levels before cross-clamping, and this gradual rise in CSFP correlated with the mean arterial pressure (ta = 0.34, p < 0.05). We postulate that this was either because CSF was replenished (at the rate of approximately 25 ml an hour; a rate that is dependent on arterial pressure t4) or because the venous capacitance vessels surrounding the spinal cord had expanded to tal~e up the vacated space. We speculate that the possible reasons for the rapid rise in CSFP with aortic cross-clamping include the following: (1) The higher proximal MAP causing
Volume 11 Number 3 March 1990
either dilation of intracranial arteries or expansion of nervous tissue volume. However, cerebral autoregulation would probably prevent any expansion of neural tissue or arterial vessel volume from occurring because of increased systemic arterial pressure or cerebral blood flow) 3as (2) Increased CSF secretion, possibly related to increased blood flow to the brain. However, previous experiments purporting to support this, which mimicked either rapid tumor expansion or intracranial hemorrhage, cannot, in the true sense, be considered to be analogous to aortic cross-clamping, where there is no forceful injection ofintrathecal fluid. 16,17(3) The rise in CVP resulting in back pressure in the intracranial veins and venous sinuses. In our opinion, we tend to favor the third postulate. The classic Monro-Kellie hypothesis remains generally relevant, namely, that the total cranial and spinal columaa volumes remain constant within the closed space that the brain and spinal cord occupy. Any volume increase of the components (nervous tissue volume, blood volume, CSF volume) must therefore be at the expense of the other components. Normal CSFP will thus only be maintained until the compensatory mechanisms are exceeded. 18 However, with aortic cross-clamping, the one compensatory mechanism, namely compression of venous capacitance vessels (loss of blood volume together with the above), rather than compensating for any increases in the intracranial components, the venous capacitance vessels are the more likely cause of the increase in CSFP. Blaisddl and Cooley8 have noted in dogs that CSFP increased with both anesthetic induction and cross-clamping, and in our own (unpublished data) and other animal experiments s,19 it has been found .nat CVP increased with aortic cross-clamping. It is known that a close relationship normally exists under physiologic conditions between subarachnoid, intracranial sinus, CVP, and the C S F P , 2°'21 and it is likely that spinal CSFP is similarly related to spinal venous pressures 2t including in Batson's plexus, the azygous vein (which drains segmental intercostal and lumbar veins), and in the superior vena cava. Since CVP tends to increase with induction of anesthesia and cross-clamping, the mechanism of increased CSFP is probably similar to that of the Queckenstedt test = or heart failure, 2a in that decreased venous return as a result of raised intrathoracic pressure or, heart failure, results in backpressure in the intracranial and perichordal capacitance vessels because of increased impedence to venous outflow. In addition, there may also be an increased volume of blood return from the
Intrathecal papaverine and cerebrospinatfluid drainage 427 upper body via the superior vena cava24 because of increased shunting between the upper body arterial and venous systems as a result of pharmacologic vasodilators, further adding to the venous congestion. The significant linear correlation we noted between CSFP and CVP both before anesthetic induction, before cross-clamping, and with cross-clamping thus supports the concept that the increases in CSFP are probably due to increases in CVP. However, it should be noted that vasodilators themselves (nitroprusside, nitroglycerine, and trimethaphan) can cause an increased CSFPY ,26 The implied or measured lumbar spinal cord CSFP, the site at which ischemia is maximal during aortic cross-clamping, 4,9,24,27particularly without distal aortic perfusion, 27is controversial. Dunbar et al. 2° found in dogs that CSFP in the lumbar spinal cord sac fell during aortic occlusion, whereas CSFP measured in the cisterna magna or intraventricularty increased. These pressure changes were reversed by terminating aortic occlusion. 2° Others 81m6 have measured the cisterna magna pressure in dogs during aortic cross-clamping and have confirmed, as noted by Dunbar et al.,20 that CSFP in the cisterna magna increases. However, it has been implied from these pressure changes measured in the cisterna magna that lumbar spinal cord pressures also increase. Indeed, we did find that CSFP, measured alongside the lumbar spinal cord, increases with aortic cross-clamping from a mean of i5 mm H g to i9 mm Hg, but not as markedly when compared to the increase from 13 mm Hg before induction of anesthesia to 20 mm Hg before CSF drainage. It is clear from the above discussion that CSFP measured in the cisterna magna may not accurately reflect the CSFP alterations in the lumbar spinal cord vicinity. Furthermore, Killen et al.28 reported that the distal aortic pressure did not reflect the spinal cord arteriolar (nor capillary) blood pressure, but that the higher segmental intercostal or lumbar artery end pressures were a closer approximation. Therefore by implication, the double assumption by Miyamoto et al. s of "relative spinal cord perfusion pressure" being equal to distal aortic pressure minus the cistema magna pressure probably does not hold true. Moreover, Griffiths et al. 21have shown that perfusion pressure is a far greater determinant of blood flow than the possible inhibitory effect of a raised CSFP. They concluded that "spinal cord blood flow can accomodate any fluctuations in CSFP." There are even reports of increased cerebral blood flow with an increased intracranial pressure, la It should be
428
Journal of VASCULAR SURGERY
Svensson et al.
noted that the resistance to reabsorption of CSF in dogs is 17 times greater than that of man, 14and thus any increases in CSFP in man may be more quickly compensated for. Whether CSF drainage alone will reduce the incidence of paraplegia after aortic cross-clamping in humans is an unanswered question, although apparently good results have been reported by McCullough et al.10 in 24 patients, 10 of whom were low Crawford type III thoracoabdominal aortic aneurysms and six type IV (total or near total abdominal aneurysm) with an expected paraparesis or paraplegia rate of 3.2% o r 2 . 1 % 29, respectively. Experiments in dogs, draining CSF from the cisterna magna, have shown discrepant results. Blaisdell and Cooley 8 in 1962 and more recent studies by others 9"11,16 report the incidence of postocclusion neurologic dysfunction is reduced. Nevertheless, it is of note that Blaisdell, despite his earlier work, does not believe that "we can provide any protection by modifying spinal fluid pressure." (Commentary on Bower et al.9) Furthermore, Killen et al.28 in a similar dog model were tmable to show a protective effect by draining CSF from the cisterna magna. Similarly, Wadouh et al?9 used a porcine model, and in our own nonhuman primate model, CSF drainage by a limited lumbar laminectomy failed to reduce the incidence of paraplegia. ~ In conclusion, it would appear that the acute changes in CSFP are primarily a function of increased intracranial and intraspinal venous blood volume, which is determined by venous back pressure during anesthetic induction, intravenous fluid administration and aortic cross-clamping. Whether CSF drainage alone will reduce the incidence of postoperative paraplegia is not proved and can only be evaluated by a prospective randomized study. Thus it would appear to us that the best method of preventing paraplegia, with its multifactorial causes, is to use intrathecal papaverine with CSF drainage to prolong the safe period of aortic cross-clamping, reanastomose identified critical segemental arteries by a technique that has been presented by the author, s° and use distal aortic perfusion in selected complex situations, such as aortic dissection. 2However, distal aortic perfusion has not proved to be a panacea in preventing either paraplegia because of spinal cord anatomic reasons27 or acute renal failure.2,3t Robert W. Stewart, MD, and Bruce W. Lyric, MD, allowed their patients to be entered into the study and gave valuable advice. The preoperative and postoperative neurologic assessments were performed by Antonio Salgado,
MD, and Anthony Furlan, MD, from the Department ot Neurology, and Marion Robinson, PhD, performed the statistical analyses.
REFERENCES
1. Svensson LG, Stewart RW, Cosgrove DM, et al. Intrathecal papaverine for the prevention of paraplegia after operation on the thoracic or thoracoabdominal aorta. J Thorac Cardiovasc Surg 1988;96:823-9. 2. Svensson LG, Loop FD. Prevention of spinal cord ischemia in aortic surgery. In: Bergan j'J, Yao JST, eds. Arterial surgery: new diagnostic and operative techniques. New York: Grune and Stratton, 1988:273-85. 3. Svensson LG, Stewart RW, Cosgrove DM, et al. Preliminary results and rationale for the use of intrathecal papaverine for the prevention of paraplegia after aortic surgery. S Aft J Surg 1988;26:153-9. 4. Svensson LG, Von Rirter CM, Groeneveld HT, et al. Crossclamping of the thoracic aorta. Influence of aortic shunts, laminectomy, papaverine, calcium channel blocker, allopurinol, and superoxide dismutase on spinal cord blood flow ~-,~t paraplegia in baboons. Ann Surg 1986;204:38-47. 5. Miyamota K, Keno A, Wada T, Kimoto S. A new and simple method of preventing spinal cord damage following temporary occlusion of the thoracic aorta by draining the cerebrospinal fluid. J Cardiovasc Surg 1960; 1:188-97. 6. Kane RE. Neurologic deficits following epidural or spinal anesthesia. Anesth Analg 1981;60:150-61. 7. Phillips OC, Ebner H, Nelson AT, Black MH. Neurologic complications following spinal anesthesia with lidocaine. Anesthesiology 1969;30:284-9. 8. Blaisdell FW, Cooley DA. The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid prssure. Surgery 1962;51:351-5. 9. Bower TC, Murray MJ, Glovicld P, Yaksh TL, Hollier LH, Pairolero PC. Effects of thoracic aortic occlusion and cerebrospinal fluid drainage on regional spinal cord blood flow in dogs: correlation with neurologic outcome. J VAsc SURG 1989;9:135-44. 10. McCullough JL, HoUier LH, Nugent M. Paraplegia aftc thoracic aortic occlusion: Influence of eerebrospinal fluid drainage. J VASCSuR~ 1988;7:153-60. 11. Grubbs PE, Matini C, Toporoff B, et al. Somatosensory evoked potentials and spinal cord perfusion pressure are significant predictors of postoperative neurologic dysfunction. Surgery 1988;104:216-23. 12. Molina JE, Cogordan J, Einzig S, et al. Adequacy of ascending aorta-descending aorta shunt during cross-clamping of the thoracic aorta for prevention of spinal cord injury. J Thorac Cardiovasc Surg 1985;90:126-36. 13. Lassen NA, Christensen MS. Physiology of cerebral blood flow. Br J Anaesth 1976;48:719-34. 14. Culfler RWP, Page L, Galicich J, Watters GV. Formation and absorption of cerebrospinal fluid in man. Brain 1968; 91:707-20. 15. Hickey R, Albin MS, Bunegin L, Gelineau J. Autoregulati~n of spinal cord blood flow: is the cord a microcosm of the brain. Stroke 1986;17:1183-9. 16. Oka Y, Miyamoto T. Prevention of spinal cord injury after cross-clamping of the thoracic aorta. Jpn J Surg 1984;14:15962.
Volume 11 Number 3 March 1990
17 Berendes JN, Bredee JJ, Schipperheyn JJ, Mashhour YAS. Mechanism of spinal cord injury after cross-clamping of the descending thoracic aorta. Circulation 1982;66(suppl I): 112-6. 18. Marmarou A, Shulman K, LaMorgese J. Compartmental analysis of compliance and outflow resistance of the cerebrospinal fluid system. J Neurosurg 1975;43:523-4. 19. Wadotth F, Lindemarm E-M, Amdt CF, Hetzer R, Borst HG. The arteria radicularis magna anterior as a decisive factor influencing spinal cord damage during aortic occlusion. J Thorac Cardiovasc Surg 1984;88:1-10. 20. Dunbar HS, Gutherie TC, Karpell B. A study of the cerebrospinal fluid pulse wave. Arch Neurol 1966;14:624-30. 21. Griffiths IR, Pitts LH, Crawford P_A,Trench JG. Spinal cord compression and blood flow. Neurology 1978;28:1145-51. 22. Bering EA. Choroid plexus and arterial pulsation ofcerebrospinal fluid. Arch Neurol Psychiat 1955;73:165-72. 23. Friedfeld L, Fishberg AM. The relation of the cerebrospinal and venous pressures in heart failure. J Clin Invest 1934; 13:495-501. 24. Geiman S, Rabbani S, Bradley EL. Inferior and superior vena caval blood flows during cross-clamping of the thoracic aorta in pigs, J Thorac Cardiovasc Surg 1988;96:387-92. 2f. Turner JM, Powell D, Gibson RM, McDowall DG. Intracranial pressure changes in neurosurgical patients during hypotension induced with sodium nitroprusside or trimetaphan. Br J Anaesth 1977;49:419-25.
Intrathecal papaverine and cerebrospinatfluid drainage
429
26. Dohi S, Matsumoto M, Takahashi T. The effects of nitroglycerin on cerebrospinal fluid pressure in awake and anesthetized humans. Anesthesiology 1981;54:511-4. 27. Svensson LG, Rickards E, Coull A, Rogers G, Fimmel CJ, Hinder RA. Relationship of spinal cord blood flow to vascular anatomy during thoracic aortic cross-clamping and shunting. J Thorac Cardiovasc Surg 1986;91:71-8. 28. Killen DA, Edwards RH, Tinsley EA, Boehm FH. Effect of low molecular weight dextran, heparin, urea, cerebrospinal fluid drainage, and hypothermia on ischemic injury of the spinal cord secondary to m o b ~ a t i o n of the thoracic aorta from the posterior parietes. J Thorac Cardiovasc Surg 1965;50:882-7. 29. Crawford ES, Crawford JL, Sail HJ, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J VASC SURe 1986;3:389-404. 30. Svensson LG, Patel V, Coselli JS, Crawford ES. Preliminary report of localization of spinal cord blood supply by hydrogen during aortic surgery. Ann Thorac Surg (In press)° 31. Svensson LG, Coselli JS, Sail HJ, Hess KR, Crawford ES. Appraisal of adjuncts to prevent acute renal failure after surgery on the thoracic or thoracoabdominal aorta. J VAsc SuRe 1989;10:230-9.
