Cerebrospinal Fluid Drainage and Steroids Provide Better Spinal Cord Protection During Aortic CrossClamping Than Does Either Treatment Alone Thomas T. Woloszyn, MD, Corrado P. Marini, MD, Matthew S. Coons, MD, Ira M. Nathan, PhD, Samir Basu, MD, Anthony J. Acinapura, MD, and Joseph N. Cunningham, MD Department of Surgery, Maimonides Medical Center, and Division of Cardiothoracic Surgery, State University of New York, Health Science Center at Brooklyn, Brooklyn, New York
We investigated whether intravenous methylprednisolone (30 mg/kg) before 30 minutes of aortic crossclamping and after 4 hours could enhance the effects of cerebrospinal fluid drainage on spinal cord perfusion pressure and postoperative paraplegia when proximal blood pressure was controlled with sodium nitroprusside and partial exsanguination. Dogs were randomized into three groups: group 1 (n = 61, control; group 2 (n = 7), steroids; and group 3 (n = 61, steroids with cerebrospinal fluid drainage. During aortic cross-clamping, blood pressure proximal to the clamp decreased significantly in each group compared with baseline ( p < 0.051, but did not differ among groups (group 1 = 82.2, group 2 = 82.1, group 3 = 86.6 mm Hg, p > 0.05). Mean distal pressure decreased from systemic values to 8.4, 8.5, and 3.7 mm Hg, respectively, after aortic cross-clamping ( p < 0.05); these values did not differ from one another ( p > 0.05). During aortic cross-clamping, cerebrospinal fluid pressure in groups 1 and 2 did not differ significantly compared with baseline (12.2 versus 8.2,14.2 versus 10.7
mm Hg, p > 0.05), whereas in group 3 the baseline cerebral spinal fluid pressure of 10.7 mm Hg decreased to 0.4 mm Hg ( p < 0.05). Spinal cord perfusion pressure in group 3 was significantly higher than in groups 1 and 2 (3.3 versus -3.9 and -5.7 mm Hg, p < 0.05), but did not differ between groups 1 and 2 ( p > 0.05). Somatosensory evoked potential loss occurred significantly earlier in groups 1 and 2 than in group 3 (4 minutes 31 seconds and 4 minutes 18 seconds vs 11 minutes 16 seconds, p < 0.05). No significant difference in the time to somatosensory evoked potential loss was noted between groups 1 and 2 ( p > 0.05).Paraplegia rates differed significantly between groups 3 and 1 (1/6 versus 5/6 paralyzed, p < 0.05), but were not different when group 3 was compared with group 2 (1/6versus 3/7paralyzed, p > 0.05). This study shows that steroids with cerebrospinal fluid drainage provide spinal cord protection during aortic crossclamping, whereas steroids alone are ineffective.
C
stabilizing membranes [9], and reducing spinal cord edema [lo], but they have not been used either experimentally or clinically in association with CSFD. Because the postulated mechanism of CSFP increase during SNP infusion is a loss of compliance of the spinal canal secondary to edema of the spinal cord [3], and the ineffectiveness of CSFD may be related to the edema of the spinal cord, we designed a study to evaluate whether steroids could enhance the effects of CSFD on cerebrospinal fluid dynamics and neurological outcome in a canine model in which SNP and partial exsanguination were used to control proximal hypertension.
ross-clamping of the thoracic aorta during resection of thoracic or thoracoabdominal aneurysms causes proximal hypertension that must be controlled either pharmacologically or with use of shunts or bypasses. Sodium nitroprusside (SNP) remains the most widely used agent for control of blood pressure during aortic operations [l, 21. It has been experimentally shown to have detrimental effects on spinal cord perfusion, however, because it reduces distal aortic blood pressure (DBP) and causes an increase in cerebrospinal fluid pressure (CSFP) [3]. Cerebrospinal fluid drainage (CSFD) improves spinal cord perfusion pressure (SCPP) and decreases paraplegia rates after aortic operations [4, 51, but it is ineffective in an experimental model when SNP is used to control proximal hypertension [6]. Steroids can lengthen the warm ischemic time of the spinal cord [7] by acting as free radical scavengers [8], Presented at the Twenty-fifth Anniversary Meeting of The Society of Thoracic Surgeons, Baltimore, MD, Sep 11-13, 1989. Address reprint requests to Dr Cunningham, Department of Surgely, Maimonides Medical Center, 4802 Tenth Ave, Brooklyn, NY 11219.
0 1990 by The Society of Thoracic Surgeons
(Ann Thorac Surg 1990;49:78-83)
Material and Methods Nineteen mongrel dogs weighing 25 to 35 kg were anesthetized with intravenous pentobarbital sodium (20 mg/ kg) and fentanyl (4 pg/mL). Animals were intubated and placed on a Harvard ventilator using room air. Respirator settings included a tidal volume of 12 mL/kg, respiratory rate to maintain the partial pressure of carbon dioxide between 25 and 35 mm Hg, and 5 cm H,O of positive end-expiratory pressure. 0003-4975/90/$3.50
79
WOLOSZYN ET AL STEROIDS WITH CSFD
Ann Thorac Surg 1990:49:78-83
Pressure-monitoring catheters were inserted into the left carotid and femoral arteries and left jugular vein to monitor aortic pressures proximal and distal to the crossclamp and central venous pressure. A 20-gauge, 1.5-inch (37.5 mm) spinal needle was inserted percutaneously into the cisterna cerebellomedullaris to monitor CSFP. All pressures were monitored with Bentley Trantec physiological pressure transducers (model 60-800; Santa Ana, CA). Intravenous fluids (lactated Ringer's solution) were administered to maintain a central venous pressure of 3 to 5 mm Hg. Somatosensory evoked potentials (SEPs) were monitored using previously described techniques [111; SEP waveforms were recorded every minute during the crossclamp interval. The time required to identify a morphologically flat waveform (on two consecutive measurements) was designated as SEP loss. Through a left thoracotomy in the fourth intercostal space, the aorta was isolated 1 cm distal to the left subclavian artery. Baseline SEPs, proximal aortic blood pressure (P-BP)and D-BP, and CSFP were measured after a ten-minute stabilization period. Proximal blood pressure was controlled to a mean of 85 mm Hg by removing 40% of the estimated circulating blood volume through the left carotid arterial catheter before application of the aortic cross-clamp. In addition, SNP (50 mg in 250 mL of 5% dextrose in water) was infused at a rate of 15 to 45 pglkglmin throughout the experimental interval. Methylprednisolone sodium succinate (30 mg/kg) was administered intravenously ten minutes before crossclamping and repeated after four hours in groups 2 (n = 7) and 3 (n = 6). In addition, in group 3 cerebrospinal fluid was drained, using previously described techniques [ 121. Group 1 (n = 6) controls received no treatment. The aorta was cross-clamped at the previously isolated site for 30 minutes. All variables were monitored every minute until termination of the experiment. One minute before release of the aortic cross-clamp, 44.6 mEq of sodium bicarbonate was administered intravenously, and the shed blood was slowly retransfused. The animals were allowed to recover, and at 48 hours postoperatively their neurological status was evaluated by an observer unaware of the experimental protocol using Tarlov's criteria [13]: grade 0, spastic paraplegia and no movement of the lower limbs; grade 1, spastic paraplegia and slight movement of the lower limb joint; grade 2, good movement of the lower limbs, but unable to stand; grade 3, able to stand but not able to walk normally; grade 4, complete recovery. Animals were then killed in compliance with the guidelines of the 1986 report of the American Veterinary Medical Association Panel on Euthanasia [14]. All animals received humane care and treatment in compliance with the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH Publication No. 85-23, revised 1985). All data are reported as the mean the standard deviation, and were analyzed using analysis of variance, Student-Newman-Keuls test, and Fisher's exact test. Sta-
*
Table 1. Hemodynamics" Baseline Group
1 2 3
P-BP
Aortic Cross-Clamping D-BP
P-BP
D-BP
104 f 15.2 105 2 15.2 82.2 2 5.4b 8.4 f 6.0b 104.6 f 15.7 105.7 2 15.8 82.1 2 5.0b 8.5 f 4.1b 107.2 2 11.8 107.3 2 14.6 86.6 2 5.0b 3.7 rt lBb
a Data are shown as mean values plus or minus standard deviation in Significance: p < 0.05 versus baseline. millimeters of mercury. P-BP = proximal blood pressure. D-BP = distal blood pressure;
tistical significance was accepted at a p value of less than 0.05.
Results Hemodynamics and Cerebrospinal Fluid Dynamics There was no significant difference in mean P-BP, D-BP and CSFP among the three groups at baseline. After aortic cross-clamping, while SNP was being infused, P-BP decreased significantly from systemic baseline values to 82.2 f 5.4 mm Hg in group 1, 82.1 & 5 mm Hg in group 2, and 86.6 ? 5 mm Hg in group 3 (p < 0.05 versus baseline). These values were not significantly different from each other (p > 0.05). Mean D-BP decreased from systemic values to 8.4 ? 6 mm Hg, 8.5 & 4.1 mm Hg, and 3.7 1.8 mm Hg, respectively, in the three groups ( p < 0.05 versus baseline, p > 0.05 versus each other) (Table 1). Cerebrospinal fluid pressure did not change significantly from its baseline value while the aorta was crossclamped in groups 1 and 2; CSFP was significantly reduced to 0.4 ? 0.4 mm Hg by CSFD in group 3 ( p < 0.05 versus groups 1and 2) (Table 2). This difference remained statistically significant throughout the aortic cross-clamp interval ( p < 0.05). The CSFPs at each five-minute interval were not significantly different within or between groups 1 and 2 (Fig 1).No significant change at similar intervals was noted in group 3 after the initial statistically significant decrease (p < 0.05) caused by CSFD. Spinal cord perfusion pressure (defined as D-BP minus 5.8 and -5.7 5.5 mm Hg during the CSFP) was -3.9 aortic cross-clamp interval in groups 1and 2, respectively, whereas it remained significantly higher at 3.3 & 1.9 mm
*
*
*
Table 2. Cerebrospinal Fluid Dynamics" Baseline Group
1 2
3
Aortic Cross-Clamping
CSFP
SCPP
CSFP
SCPP
8.2 2 3.1 10.7 2 2.6 10.72 4.5
96.8 f 9.1 95.0 2 9.2 96.62 8.1
12.2 f 3.6 14.2 2 4.6 0.4? 0.4b
-3.9 2 4.8 -5.7 f 4.8 3.3 2 1.1'
Data are shown as mean values k standard deviation in millimeters of mercury. Significance: p < 0.05 versus baseline. 'Significance: p < 0.05 versus groups 1 and 2.
a
CSFP = cerebrospinal fluid pressure; pressure.
SCPP
=
spinal cord perhsion
80
WOLOSZYN ET AL STEROIDS WITH CSFD
Ann Thorac Surg 1990;49:78-83
PRESSURE (mm Hg)
PRESSURE (rnm Hg)
1
20-!
1
T
I
l5
i
0 S/D
1
I
I
8 STEROIDS 0 CONTROL
15
I
10
*p
0.05) (Fig 2). The three groups received similar total amounts of SNP: 27 & 12, 24 -t 9, and 31 ? 8 mg, respectively, to control P-BP ( p > 0.05). The total amount of blood removed was also similar among the three groups: 933 & 216,993 & 124, and 975 ? 173 mL, respectively ( p > 0.05).
rological outcome, with only 3 of 7 animals suffering severe neurological injury. Statistical analysis showed a significantly better neurological outcome of group 3 animals as compared with controls ( p < 0.05). No difference was noted when group 2 was compared with either group 1 or 3 ( p > 0.05). In addition, a comparison between animals with positive SCPP and those with negative SCPP showed that 2 of 9 animals with positive SCPP were neurologically injured, whereas 7 of 10 animals with negative SCPPs suffered paraplegia ( p < 0.05).
Neurological Injury For statistical analysis, animals were subdivided into animals with minimal or no injury (Tarlov grades 3 and 4) and those unable to ambulate (Tarlov grades 0-2). There was a significant difference in the neurological injury sustained in the three groups (Fig 3). All but 1 animal in group 3, with positive SCPP, were graded as Tarlov 3 or 4, whereas 5 of 6 animals in group 1 suffered spastic paraplegia. Animals in group 2 had a somewhat varied neu-
PRESSURE (mm Hg)
5
Somatosensory Evoked Potential Changes Time until SEP loss (Fig 4) was significantly longer at 11 minutes 16 seconds in group 3 (positive SCPPs) as compared with groups 1 and 2 (4 minutes 31 seconds and 4 minutes 18 seconds, p < 0.05); however, there was no significant difference between groups 1 and 2 ( p > 0.05).
*-
-1
TIME (min)
*
15 -,
1I 10 -
*p
i
0.05) but enabled successful reduction of CSFP by CSFD in group 3. Of greater importance in this group, CSFP could be maintained at near zero levels throughout the cross-clamp interval, whereas this could not be accomplished by CSFD in our previous study. Because we did not use steroids and CSFD in a group in which P-BP was controlled with SNP alone, we cannot isolate the factor (steroids or exsanguination) responsible for the prolonged reduction of CSFP. Steroids alone appeared to have some protective effect on the spinal cord in the presence of negative SCPP: 4 of 7 animals in this group did not suffer neurological injury. There was no statistical difference, however, between the steroid-treated group and untreated controls (4 of 7 versus 5 of 6 injured, p < 0.05). This differs from the results of Laschinger and co-workers [7], which showed that all steroid-treated animals had complete recovery after an 18.33-minute ischemic interval. The shorter duration of warm ischemia with a higher P-BP (120 to 140 mm Hg) may account for the different results. A higher P-BP implies a higher D-BP, hence a better SCPP. Because less SNP may have been required to maintain a mean P-BP of 120 to 140 mm Hg, the detrimental effects of SNP may have not been apparent. Our results showed that 7 of 9 animals with positive SCPP were neurologically intact; in contrast, only 3 of 10 animals with negative SCPP were intact postoperatively ( p < 0.05). The two paralyzed animals in the group with positive SCPP had pressures ranging from 5 to 10 mm Hg; paradoxically, the remaining 7 uninjured animals had SCPPs ranging from 0 to 5 mm Hg. In a previous study [12], we showed that an SCPP of 10 mm Hg was the threshold value for maintaining spinal cord integrity for a 40-minute cross-clamp interval, and with a shorter cross-clamp of 30 minutes SCPP as low as 2.3 mm Hg was associated with spinal cord integrity. In the present study, 3 animals with negative SCPP scored Tarlov grade 3 (unsteady gait), in contrast to our previous observations. None of the animals with negative SCPPs had complete neurological recovery (Tarlov grade 4). Based on the present findings regarding SCPP, as it relates to postoperative neurological outcome, a critical threshold of SCPP appears not to be an absolute error-free measurement. We believe that the individual variability may be based on the degree and type of collateralization, which cannot be properly measured by SCPP alone.
82
WOLOSZYN ET AL STEROIDS WITH CSFD
Paraplegia results from spinal cord ischemia, which in turn is caused by inadequate spinal cord blood flow: spinal cord blood flow as low as 4 mL/100 g per minute may prevent neurological injury in primates [17]. Monitoring techniques that yield on-line information on spinal cord blood flow are needed, because they will enable prediction of neurological outcome without error. Measurement of SCPP is now the only available method of assessing spinal cord blood flow indirectly. Spastic paraplegia developed in 2 of 9 animals with positive SCPP postoperatively. This corroborates clinical findings reported by Crawford and associates [ 181 that showed that patients with high distal aortic pressure (80 to 116 mm Hg) while on bypass, at flow rates less than 2,000 mL per minute, had a high incidence of paraplegia. We must identify the role of regional vascular resistance and the factors affecting it to understand why high SCPPs do not always generate the spinal cord blood flow required to prevent paraplegia. As in a previous study [7], we confirmed that time to SEP loss is not an accurate predictor of neurological outcome in animals treated with steroids. Loss of SEP in the group treated by steroids alone occurred at a time similar to that of controls (4 minutes 18 seconds versus 4 minutes 31 seconds, p > 0.05); however, although all control animals were paralyzed, 4 of 7 steroid-treated animals were spared neurological injury. Animals with loss of electrophysiological conduction of the spinal cord at a statistically significant later time (group 3 versus groups 1 and 2, 11 minutes 16 seconds versus 4 minutes 18 seconds and 4 minutes 31 seconds, p < 0.05) had a better neurological outcome. This study shows that SEP loss occurring later, independent of treatment, is more predictive of neurological outcome than an early loss occurring in animals treated with steroids. Based on the results of the present study, we believe that in cases in which mechanical control of central hypertension cannot be achieved for either anatomical or technical reasons and in which SNP or any other nitrate vasodilators must be used, steroids in conjunction with cerebrospinal fluid drainage may provide better spinal cord protection than either treatment alone. At present we are unable to identify an intrinsic protective role of partial exsanguination, but we can infer that exsanguination may mitigate the deleterious effect of SNP on SCPP by decreasing the amount of SNP required to control proximal hypertension in a canine model. Further studies are required to elucidate the effects of partial exsanguination on hemodynamics and cerebrospinal fluid dynamics. We gratefully acknowledge C. Bertuglia and P. Damiani for technical operative assistance and R. Robertazzi and M. Bottali for manuscript preparation.
Ann Thorac Surg 1990;49:78-83
References 1. Crawford ES, Fenstermacher JM, Richardson W, et al. Reappraisal of adjuncts to avoid ischemia in the treatment of thoracic aneurysms. Surgery 1970;67182-96. 2. Najafi H, Javid H, Hunter J, et al. Descending aortic aneurysmectomy without adjuncts to avoid ischemia. Ann Thorac Surg 1980;30:326-35. 3. Marini CP, Grubbs PE, Toporoff B, et al. Effect of sodium nitroprusside on spinal cord perfusion and paraplegia during aortic cross-clamping. Ann Thorac Surg 1989;47379-83. 4. McCullough JL, Hollier LH, Nugent M. Paraplegia after thoracic aortic occlusion: influence of cerebrospinal fluid drainage. J Vasc Surg 1988;7:153-60. 5. Blaisdell FW, Cooley DA. The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid pressure. Surgery 1962;51:351-5. 6. Woloszyn TT, Marini CP, Coons MS, Nathan IM, Jacobowitz IJ, Cunningham JN. Cerebrospinal fluid drainage does not counteract the negative effect of sodium nitroprusside on spinal cord perfusion during aortic crossclamping. Curr Surg (in press). 7. Laschinger JC, Cunningham JN, Cooper MM, Krieger K, Nathan IM, Spencer FC. Prevention of ischemic spinal cord injury following aortic cross-clamping: use of corticosteroids. Ann Thorac Surg 1984;38:50&7. 8. Hall ED, Braughler JM. Effects of intravenous methylprednisolone on spinal cord lipid peroxidation and (Na++ K+)ATPase activity. J Neurosurg 1982;57247-53. 9. Schumer W, Erve PR. Effect of dexamethasone on the level of serotonin in the blood of endotoxified monkeys [Abstract]. Clin Res 1976;24:560A. 10. Nelson SR, Dick AR. Steroids in the treatment of brain edema. In: Azarnoff DL, ed. Steroid therapy. Philadelphia: WB Saunders, 1975:313-24. 11. Cunningham JN Jr, Laschinger JC, Merkin HA, et al. Measurement of spinal cord ischemia during operations upon the thoracic aorta. Ann Surg 1982;196:285-96. 12. Grubbs PE Jr, Marini C, Toporoff 8, et al. Somatosensory evoked potentials and spinal cord perfusion pressure are significant predictors of postoperative neurologic dysfunction. Surgery 1988;104:216-23. 13. Tarlov IM. Spinal cord compression: mechanism of paralysis and treatment. Springfield, IL: Charles C Thomas, 1957147. 14. 1986 Report of the AVMA Panel on Euthanasia. J Am Vet Med Assoc 1986;188:252-68. 15. Crawford ES, Crawford JL, Safi HJ, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J Vasc Surg 1986;3:389404. 16. Gelman S, Reves JG, Fowler K, Samuelson PN, Lell WA, Smith LR. Regional blood flow during cross-clamping of the thoracic aorta and infusion of sodium nitroprusside. J Thorac Cardiovasc Surg 1983;85:287-91. 17. Svensson LG, Von Ritter CM, Groeneveld HT, et al. Crossclamping of the thoracic aorta. Ann Surg 1986;204:38-47. 18. Crawford ES, Mizrahi EM, Hess KR, et al. The impact of distal aortic perfusion and somatosensory evoked potential monitoring on prevention of paraplegia after aortic aneurysm operation. J Thorac Cardiovasc Surg 1988;95:357-67.
WOLOSZYN ET AL STEROIDS WITH CSFD
Ann Thorac Surg 1990;49:7843
83
DISCUSSION DR DENTON A. COOLEY (Houston, TX): I've enjoyed hearing this paper and the paper by Svensson and associates [l]and 1 think they are certainly very important. Some of them bring back memories of the work we did more than 20 years ago with spinal fluid drainage. From a practical standpoint, one wonders whether this work is worthwhile. I hope it does not get into the literature that CSFD is going to be consistently protective in this problem of paraplegia, which is not only a medical problem but has become a serious legal problem for the clinical surgeon. Dr Blaisdell and I used spinal cord drainage clinically in a small series of patients in the past and found very little effect. Another problem was the high incidence of serious spinal headaches after use of this technique. That was a disabling feature. In our own laboratory, we have tried to d o some local hypothermic experiments in animals. Dr Rolondo Colon had a very interesting paper recently in which he tried to segmentally cool the spinal cord during periods of cross-clamping. It was very effective. I am sure that all cardiovascular surgeons are impressed with the excellent results that have been achieved by inducing general body hypothermia for cerebral protection during arch replacement. Hopefully we could use this technique in protecting the rest of the central nervous system, namely, the spinal cord. It is not often practical to try to reimplant segmental vessels. There may be no patent segmental vessels visible. All methods of protecting against paraplegia are interesting, but the problem is far from solved. No method yet developed gives complete and dependable prevention of paraplegia in extensive resection of the distal thoracic aorta. DR WOLOSZYN: Thank you, Dr Cooley, for bringing to our attention all of the stated facts.
We do not want to give the impression that this is a cure-all. What we are looking for is a combined modality approach to prolong warm ischemia, and by doing separate animal studies, we may be able to isolate different factors that will influence postoperative paraplegia rates. The fact that CSFD has little effect on the outcome of postoperative paraplegia is noted, and it is indeed true. Our intention was to show that drainage is effective if proximal hypertension is not controlled, whereas when the proximal hypertension or central hypertension is controlled with pharmacological agents, specifically the nitrate vasodilators (Nipride), that some of its deleterious effects may be mitigated by CSFD. Drainage alone is not effective, but when combined with steroid administration it may indeed prove beneficial, although not uniformly protective. Other methods will need to supplement this to eventually prevent paraplegia after thoracic aortic cross-clamping. Your second comment on profound hypothermic arrest is indeed interesting. It has been shown that hypothermic arrest is beneficial for maintaining cerebral neurological status. It has not shown any benefit for the spinal cord unless it is administered through the intrathecal space. Third, your mention of reimplanting the critical intercostal vessels is absolutely essential, as illustrated in the paper by Svensson and associates.
Reference 1. Svensson LG, Pate1 VM, Coselli JS, Crawford ES. Localization of spinal cord blood supply during aortic operation. Ann Thorac Surg (in press).
We investigated whether intravenous methylprednisolone (30 mg/kg) before 30 minutes of aortic crossclamping and after 4 hours could enhance the effects of cerebrospinal fluid drainage on spinal cord perfusion pressure and postoperative paraplegia when proximal blood pressure was controlled with sodium nitroprusside and partial exsanguination. Dogs were randomized into three groups: group 1 (n = 61, control; group 2 (n = 7), steroids; and group 3 (n = 61, steroids with cerebrospinal fluid drainage. During aortic cross-clamping, blood pressure proximal to the clamp decreased significantly in each group compared with baseline ( p < 0.051, but did not differ among groups (group 1 = 82.2, group 2 = 82.1, group 3 = 86.6 mm Hg, p > 0.05). Mean distal pressure decreased from systemic values to 8.4, 8.5, and 3.7 mm Hg, respectively, after aortic cross-clamping ( p < 0.05); these values did not differ from one another ( p > 0.05). During aortic cross-clamping, cerebrospinal fluid pressure in groups 1 and 2 did not differ significantly compared with baseline (12.2 versus 8.2,14.2 versus 10.7
mm Hg, p > 0.05), whereas in group 3 the baseline cerebral spinal fluid pressure of 10.7 mm Hg decreased to 0.4 mm Hg ( p < 0.05). Spinal cord perfusion pressure in group 3 was significantly higher than in groups 1 and 2 (3.3 versus -3.9 and -5.7 mm Hg, p < 0.05), but did not differ between groups 1 and 2 ( p > 0.05). Somatosensory evoked potential loss occurred significantly earlier in groups 1 and 2 than in group 3 (4 minutes 31 seconds and 4 minutes 18 seconds vs 11 minutes 16 seconds, p < 0.05). No significant difference in the time to somatosensory evoked potential loss was noted between groups 1 and 2 ( p > 0.05).Paraplegia rates differed significantly between groups 3 and 1 (1/6 versus 5/6 paralyzed, p < 0.05), but were not different when group 3 was compared with group 2 (1/6versus 3/7paralyzed, p > 0.05). This study shows that steroids with cerebrospinal fluid drainage provide spinal cord protection during aortic crossclamping, whereas steroids alone are ineffective.
C
stabilizing membranes [9], and reducing spinal cord edema [lo], but they have not been used either experimentally or clinically in association with CSFD. Because the postulated mechanism of CSFP increase during SNP infusion is a loss of compliance of the spinal canal secondary to edema of the spinal cord [3], and the ineffectiveness of CSFD may be related to the edema of the spinal cord, we designed a study to evaluate whether steroids could enhance the effects of CSFD on cerebrospinal fluid dynamics and neurological outcome in a canine model in which SNP and partial exsanguination were used to control proximal hypertension.
ross-clamping of the thoracic aorta during resection of thoracic or thoracoabdominal aneurysms causes proximal hypertension that must be controlled either pharmacologically or with use of shunts or bypasses. Sodium nitroprusside (SNP) remains the most widely used agent for control of blood pressure during aortic operations [l, 21. It has been experimentally shown to have detrimental effects on spinal cord perfusion, however, because it reduces distal aortic blood pressure (DBP) and causes an increase in cerebrospinal fluid pressure (CSFP) [3]. Cerebrospinal fluid drainage (CSFD) improves spinal cord perfusion pressure (SCPP) and decreases paraplegia rates after aortic operations [4, 51, but it is ineffective in an experimental model when SNP is used to control proximal hypertension [6]. Steroids can lengthen the warm ischemic time of the spinal cord [7] by acting as free radical scavengers [8], Presented at the Twenty-fifth Anniversary Meeting of The Society of Thoracic Surgeons, Baltimore, MD, Sep 11-13, 1989. Address reprint requests to Dr Cunningham, Department of Surgely, Maimonides Medical Center, 4802 Tenth Ave, Brooklyn, NY 11219.
0 1990 by The Society of Thoracic Surgeons
(Ann Thorac Surg 1990;49:78-83)
Material and Methods Nineteen mongrel dogs weighing 25 to 35 kg were anesthetized with intravenous pentobarbital sodium (20 mg/ kg) and fentanyl (4 pg/mL). Animals were intubated and placed on a Harvard ventilator using room air. Respirator settings included a tidal volume of 12 mL/kg, respiratory rate to maintain the partial pressure of carbon dioxide between 25 and 35 mm Hg, and 5 cm H,O of positive end-expiratory pressure. 0003-4975/90/$3.50
79
WOLOSZYN ET AL STEROIDS WITH CSFD
Ann Thorac Surg 1990:49:78-83
Pressure-monitoring catheters were inserted into the left carotid and femoral arteries and left jugular vein to monitor aortic pressures proximal and distal to the crossclamp and central venous pressure. A 20-gauge, 1.5-inch (37.5 mm) spinal needle was inserted percutaneously into the cisterna cerebellomedullaris to monitor CSFP. All pressures were monitored with Bentley Trantec physiological pressure transducers (model 60-800; Santa Ana, CA). Intravenous fluids (lactated Ringer's solution) were administered to maintain a central venous pressure of 3 to 5 mm Hg. Somatosensory evoked potentials (SEPs) were monitored using previously described techniques [111; SEP waveforms were recorded every minute during the crossclamp interval. The time required to identify a morphologically flat waveform (on two consecutive measurements) was designated as SEP loss. Through a left thoracotomy in the fourth intercostal space, the aorta was isolated 1 cm distal to the left subclavian artery. Baseline SEPs, proximal aortic blood pressure (P-BP)and D-BP, and CSFP were measured after a ten-minute stabilization period. Proximal blood pressure was controlled to a mean of 85 mm Hg by removing 40% of the estimated circulating blood volume through the left carotid arterial catheter before application of the aortic cross-clamp. In addition, SNP (50 mg in 250 mL of 5% dextrose in water) was infused at a rate of 15 to 45 pglkglmin throughout the experimental interval. Methylprednisolone sodium succinate (30 mg/kg) was administered intravenously ten minutes before crossclamping and repeated after four hours in groups 2 (n = 7) and 3 (n = 6). In addition, in group 3 cerebrospinal fluid was drained, using previously described techniques [ 121. Group 1 (n = 6) controls received no treatment. The aorta was cross-clamped at the previously isolated site for 30 minutes. All variables were monitored every minute until termination of the experiment. One minute before release of the aortic cross-clamp, 44.6 mEq of sodium bicarbonate was administered intravenously, and the shed blood was slowly retransfused. The animals were allowed to recover, and at 48 hours postoperatively their neurological status was evaluated by an observer unaware of the experimental protocol using Tarlov's criteria [13]: grade 0, spastic paraplegia and no movement of the lower limbs; grade 1, spastic paraplegia and slight movement of the lower limb joint; grade 2, good movement of the lower limbs, but unable to stand; grade 3, able to stand but not able to walk normally; grade 4, complete recovery. Animals were then killed in compliance with the guidelines of the 1986 report of the American Veterinary Medical Association Panel on Euthanasia [14]. All animals received humane care and treatment in compliance with the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH Publication No. 85-23, revised 1985). All data are reported as the mean the standard deviation, and were analyzed using analysis of variance, Student-Newman-Keuls test, and Fisher's exact test. Sta-
*
Table 1. Hemodynamics" Baseline Group
1 2 3
P-BP
Aortic Cross-Clamping D-BP
P-BP
D-BP
104 f 15.2 105 2 15.2 82.2 2 5.4b 8.4 f 6.0b 104.6 f 15.7 105.7 2 15.8 82.1 2 5.0b 8.5 f 4.1b 107.2 2 11.8 107.3 2 14.6 86.6 2 5.0b 3.7 rt lBb
a Data are shown as mean values plus or minus standard deviation in Significance: p < 0.05 versus baseline. millimeters of mercury. P-BP = proximal blood pressure. D-BP = distal blood pressure;
tistical significance was accepted at a p value of less than 0.05.
Results Hemodynamics and Cerebrospinal Fluid Dynamics There was no significant difference in mean P-BP, D-BP and CSFP among the three groups at baseline. After aortic cross-clamping, while SNP was being infused, P-BP decreased significantly from systemic baseline values to 82.2 f 5.4 mm Hg in group 1, 82.1 & 5 mm Hg in group 2, and 86.6 ? 5 mm Hg in group 3 (p < 0.05 versus baseline). These values were not significantly different from each other (p > 0.05). Mean D-BP decreased from systemic values to 8.4 ? 6 mm Hg, 8.5 & 4.1 mm Hg, and 3.7 1.8 mm Hg, respectively, in the three groups ( p < 0.05 versus baseline, p > 0.05 versus each other) (Table 1). Cerebrospinal fluid pressure did not change significantly from its baseline value while the aorta was crossclamped in groups 1 and 2; CSFP was significantly reduced to 0.4 ? 0.4 mm Hg by CSFD in group 3 ( p < 0.05 versus groups 1and 2) (Table 2). This difference remained statistically significant throughout the aortic cross-clamp interval ( p < 0.05). The CSFPs at each five-minute interval were not significantly different within or between groups 1 and 2 (Fig 1).No significant change at similar intervals was noted in group 3 after the initial statistically significant decrease (p < 0.05) caused by CSFD. Spinal cord perfusion pressure (defined as D-BP minus 5.8 and -5.7 5.5 mm Hg during the CSFP) was -3.9 aortic cross-clamp interval in groups 1and 2, respectively, whereas it remained significantly higher at 3.3 & 1.9 mm
*
*
*
Table 2. Cerebrospinal Fluid Dynamics" Baseline Group
1 2
3
Aortic Cross-Clamping
CSFP
SCPP
CSFP
SCPP
8.2 2 3.1 10.7 2 2.6 10.72 4.5
96.8 f 9.1 95.0 2 9.2 96.62 8.1
12.2 f 3.6 14.2 2 4.6 0.4? 0.4b
-3.9 2 4.8 -5.7 f 4.8 3.3 2 1.1'
Data are shown as mean values k standard deviation in millimeters of mercury. Significance: p < 0.05 versus baseline. 'Significance: p < 0.05 versus groups 1 and 2.
a
CSFP = cerebrospinal fluid pressure; pressure.
SCPP
=
spinal cord perhsion
80
WOLOSZYN ET AL STEROIDS WITH CSFD
Ann Thorac Surg 1990;49:78-83
PRESSURE (mm Hg)
PRESSURE (rnm Hg)
1
20-!
1
T
I
l5
i
0 S/D
1
I
I
8 STEROIDS 0 CONTROL
15
I
10
*p
0.05) (Fig 2). The three groups received similar total amounts of SNP: 27 & 12, 24 -t 9, and 31 ? 8 mg, respectively, to control P-BP ( p > 0.05). The total amount of blood removed was also similar among the three groups: 933 & 216,993 & 124, and 975 ? 173 mL, respectively ( p > 0.05).
rological outcome, with only 3 of 7 animals suffering severe neurological injury. Statistical analysis showed a significantly better neurological outcome of group 3 animals as compared with controls ( p < 0.05). No difference was noted when group 2 was compared with either group 1 or 3 ( p > 0.05). In addition, a comparison between animals with positive SCPP and those with negative SCPP showed that 2 of 9 animals with positive SCPP were neurologically injured, whereas 7 of 10 animals with negative SCPPs suffered paraplegia ( p < 0.05).
Neurological Injury For statistical analysis, animals were subdivided into animals with minimal or no injury (Tarlov grades 3 and 4) and those unable to ambulate (Tarlov grades 0-2). There was a significant difference in the neurological injury sustained in the three groups (Fig 3). All but 1 animal in group 3, with positive SCPP, were graded as Tarlov 3 or 4, whereas 5 of 6 animals in group 1 suffered spastic paraplegia. Animals in group 2 had a somewhat varied neu-
PRESSURE (mm Hg)
5
Somatosensory Evoked Potential Changes Time until SEP loss (Fig 4) was significantly longer at 11 minutes 16 seconds in group 3 (positive SCPPs) as compared with groups 1 and 2 (4 minutes 31 seconds and 4 minutes 18 seconds, p < 0.05); however, there was no significant difference between groups 1 and 2 ( p > 0.05).
*-
-1
TIME (min)
*
15 -,
1I 10 -
*p
i
0.05) but enabled successful reduction of CSFP by CSFD in group 3. Of greater importance in this group, CSFP could be maintained at near zero levels throughout the cross-clamp interval, whereas this could not be accomplished by CSFD in our previous study. Because we did not use steroids and CSFD in a group in which P-BP was controlled with SNP alone, we cannot isolate the factor (steroids or exsanguination) responsible for the prolonged reduction of CSFP. Steroids alone appeared to have some protective effect on the spinal cord in the presence of negative SCPP: 4 of 7 animals in this group did not suffer neurological injury. There was no statistical difference, however, between the steroid-treated group and untreated controls (4 of 7 versus 5 of 6 injured, p < 0.05). This differs from the results of Laschinger and co-workers [7], which showed that all steroid-treated animals had complete recovery after an 18.33-minute ischemic interval. The shorter duration of warm ischemia with a higher P-BP (120 to 140 mm Hg) may account for the different results. A higher P-BP implies a higher D-BP, hence a better SCPP. Because less SNP may have been required to maintain a mean P-BP of 120 to 140 mm Hg, the detrimental effects of SNP may have not been apparent. Our results showed that 7 of 9 animals with positive SCPP were neurologically intact; in contrast, only 3 of 10 animals with negative SCPP were intact postoperatively ( p < 0.05). The two paralyzed animals in the group with positive SCPP had pressures ranging from 5 to 10 mm Hg; paradoxically, the remaining 7 uninjured animals had SCPPs ranging from 0 to 5 mm Hg. In a previous study [12], we showed that an SCPP of 10 mm Hg was the threshold value for maintaining spinal cord integrity for a 40-minute cross-clamp interval, and with a shorter cross-clamp of 30 minutes SCPP as low as 2.3 mm Hg was associated with spinal cord integrity. In the present study, 3 animals with negative SCPP scored Tarlov grade 3 (unsteady gait), in contrast to our previous observations. None of the animals with negative SCPPs had complete neurological recovery (Tarlov grade 4). Based on the present findings regarding SCPP, as it relates to postoperative neurological outcome, a critical threshold of SCPP appears not to be an absolute error-free measurement. We believe that the individual variability may be based on the degree and type of collateralization, which cannot be properly measured by SCPP alone.
82
WOLOSZYN ET AL STEROIDS WITH CSFD
Paraplegia results from spinal cord ischemia, which in turn is caused by inadequate spinal cord blood flow: spinal cord blood flow as low as 4 mL/100 g per minute may prevent neurological injury in primates [17]. Monitoring techniques that yield on-line information on spinal cord blood flow are needed, because they will enable prediction of neurological outcome without error. Measurement of SCPP is now the only available method of assessing spinal cord blood flow indirectly. Spastic paraplegia developed in 2 of 9 animals with positive SCPP postoperatively. This corroborates clinical findings reported by Crawford and associates [ 181 that showed that patients with high distal aortic pressure (80 to 116 mm Hg) while on bypass, at flow rates less than 2,000 mL per minute, had a high incidence of paraplegia. We must identify the role of regional vascular resistance and the factors affecting it to understand why high SCPPs do not always generate the spinal cord blood flow required to prevent paraplegia. As in a previous study [7], we confirmed that time to SEP loss is not an accurate predictor of neurological outcome in animals treated with steroids. Loss of SEP in the group treated by steroids alone occurred at a time similar to that of controls (4 minutes 18 seconds versus 4 minutes 31 seconds, p > 0.05); however, although all control animals were paralyzed, 4 of 7 steroid-treated animals were spared neurological injury. Animals with loss of electrophysiological conduction of the spinal cord at a statistically significant later time (group 3 versus groups 1 and 2, 11 minutes 16 seconds versus 4 minutes 18 seconds and 4 minutes 31 seconds, p < 0.05) had a better neurological outcome. This study shows that SEP loss occurring later, independent of treatment, is more predictive of neurological outcome than an early loss occurring in animals treated with steroids. Based on the results of the present study, we believe that in cases in which mechanical control of central hypertension cannot be achieved for either anatomical or technical reasons and in which SNP or any other nitrate vasodilators must be used, steroids in conjunction with cerebrospinal fluid drainage may provide better spinal cord protection than either treatment alone. At present we are unable to identify an intrinsic protective role of partial exsanguination, but we can infer that exsanguination may mitigate the deleterious effect of SNP on SCPP by decreasing the amount of SNP required to control proximal hypertension in a canine model. Further studies are required to elucidate the effects of partial exsanguination on hemodynamics and cerebrospinal fluid dynamics. We gratefully acknowledge C. Bertuglia and P. Damiani for technical operative assistance and R. Robertazzi and M. Bottali for manuscript preparation.
Ann Thorac Surg 1990;49:78-83
References 1. Crawford ES, Fenstermacher JM, Richardson W, et al. Reappraisal of adjuncts to avoid ischemia in the treatment of thoracic aneurysms. Surgery 1970;67182-96. 2. Najafi H, Javid H, Hunter J, et al. Descending aortic aneurysmectomy without adjuncts to avoid ischemia. Ann Thorac Surg 1980;30:326-35. 3. Marini CP, Grubbs PE, Toporoff B, et al. Effect of sodium nitroprusside on spinal cord perfusion and paraplegia during aortic cross-clamping. Ann Thorac Surg 1989;47379-83. 4. McCullough JL, Hollier LH, Nugent M. Paraplegia after thoracic aortic occlusion: influence of cerebrospinal fluid drainage. J Vasc Surg 1988;7:153-60. 5. Blaisdell FW, Cooley DA. The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid pressure. Surgery 1962;51:351-5. 6. Woloszyn TT, Marini CP, Coons MS, Nathan IM, Jacobowitz IJ, Cunningham JN. Cerebrospinal fluid drainage does not counteract the negative effect of sodium nitroprusside on spinal cord perfusion during aortic crossclamping. Curr Surg (in press). 7. Laschinger JC, Cunningham JN, Cooper MM, Krieger K, Nathan IM, Spencer FC. Prevention of ischemic spinal cord injury following aortic cross-clamping: use of corticosteroids. Ann Thorac Surg 1984;38:50&7. 8. Hall ED, Braughler JM. Effects of intravenous methylprednisolone on spinal cord lipid peroxidation and (Na++ K+)ATPase activity. J Neurosurg 1982;57247-53. 9. Schumer W, Erve PR. Effect of dexamethasone on the level of serotonin in the blood of endotoxified monkeys [Abstract]. Clin Res 1976;24:560A. 10. Nelson SR, Dick AR. Steroids in the treatment of brain edema. In: Azarnoff DL, ed. Steroid therapy. Philadelphia: WB Saunders, 1975:313-24. 11. Cunningham JN Jr, Laschinger JC, Merkin HA, et al. Measurement of spinal cord ischemia during operations upon the thoracic aorta. Ann Surg 1982;196:285-96. 12. Grubbs PE Jr, Marini C, Toporoff 8, et al. Somatosensory evoked potentials and spinal cord perfusion pressure are significant predictors of postoperative neurologic dysfunction. Surgery 1988;104:216-23. 13. Tarlov IM. Spinal cord compression: mechanism of paralysis and treatment. Springfield, IL: Charles C Thomas, 1957147. 14. 1986 Report of the AVMA Panel on Euthanasia. J Am Vet Med Assoc 1986;188:252-68. 15. Crawford ES, Crawford JL, Safi HJ, et al. Thoracoabdominal aortic aneurysms: preoperative and intraoperative factors determining immediate and long-term results of operations in 605 patients. J Vasc Surg 1986;3:389404. 16. Gelman S, Reves JG, Fowler K, Samuelson PN, Lell WA, Smith LR. Regional blood flow during cross-clamping of the thoracic aorta and infusion of sodium nitroprusside. J Thorac Cardiovasc Surg 1983;85:287-91. 17. Svensson LG, Von Ritter CM, Groeneveld HT, et al. Crossclamping of the thoracic aorta. Ann Surg 1986;204:38-47. 18. Crawford ES, Mizrahi EM, Hess KR, et al. The impact of distal aortic perfusion and somatosensory evoked potential monitoring on prevention of paraplegia after aortic aneurysm operation. J Thorac Cardiovasc Surg 1988;95:357-67.
WOLOSZYN ET AL STEROIDS WITH CSFD
Ann Thorac Surg 1990;49:7843
83
DISCUSSION DR DENTON A. COOLEY (Houston, TX): I've enjoyed hearing this paper and the paper by Svensson and associates [l]and 1 think they are certainly very important. Some of them bring back memories of the work we did more than 20 years ago with spinal fluid drainage. From a practical standpoint, one wonders whether this work is worthwhile. I hope it does not get into the literature that CSFD is going to be consistently protective in this problem of paraplegia, which is not only a medical problem but has become a serious legal problem for the clinical surgeon. Dr Blaisdell and I used spinal cord drainage clinically in a small series of patients in the past and found very little effect. Another problem was the high incidence of serious spinal headaches after use of this technique. That was a disabling feature. In our own laboratory, we have tried to d o some local hypothermic experiments in animals. Dr Rolondo Colon had a very interesting paper recently in which he tried to segmentally cool the spinal cord during periods of cross-clamping. It was very effective. I am sure that all cardiovascular surgeons are impressed with the excellent results that have been achieved by inducing general body hypothermia for cerebral protection during arch replacement. Hopefully we could use this technique in protecting the rest of the central nervous system, namely, the spinal cord. It is not often practical to try to reimplant segmental vessels. There may be no patent segmental vessels visible. All methods of protecting against paraplegia are interesting, but the problem is far from solved. No method yet developed gives complete and dependable prevention of paraplegia in extensive resection of the distal thoracic aorta. DR WOLOSZYN: Thank you, Dr Cooley, for bringing to our attention all of the stated facts.
We do not want to give the impression that this is a cure-all. What we are looking for is a combined modality approach to prolong warm ischemia, and by doing separate animal studies, we may be able to isolate different factors that will influence postoperative paraplegia rates. The fact that CSFD has little effect on the outcome of postoperative paraplegia is noted, and it is indeed true. Our intention was to show that drainage is effective if proximal hypertension is not controlled, whereas when the proximal hypertension or central hypertension is controlled with pharmacological agents, specifically the nitrate vasodilators (Nipride), that some of its deleterious effects may be mitigated by CSFD. Drainage alone is not effective, but when combined with steroid administration it may indeed prove beneficial, although not uniformly protective. Other methods will need to supplement this to eventually prevent paraplegia after thoracic aortic cross-clamping. Your second comment on profound hypothermic arrest is indeed interesting. It has been shown that hypothermic arrest is beneficial for maintaining cerebral neurological status. It has not shown any benefit for the spinal cord unless it is administered through the intrathecal space. Third, your mention of reimplanting the critical intercostal vessels is absolutely essential, as illustrated in the paper by Svensson and associates.
Reference 1. Svensson LG, Pate1 VM, Coselli JS, Crawford ES. Localization of spinal cord blood supply during aortic operation. Ann Thorac Surg (in press).
