EurJ VascSurg 4, 19-24 (1990)
Does Prostaglandin E 1 and Superoxide Dismutase Prevent Ischaemic Spinal Cord Injury A f t e r Thoracic Aortic Cross-Clamping?* K. Grabitz, E. Freye, R. Prior, K. Schr6r and W. S a n d m a n n University of Dusseldorf, West Germany The beneficial use of prostaglandin E1 (PGE1) and superoxide dismutase (SOD) on the tolerance to ischaemia of the spinal cord was evaluated following thoracic aortic cross-clamping in dogs. Aside from spinally evoked somatosensory potential (SEP) by means of a bipolar epidural catheter, postoperative evaluation of motor deficits was used to determine the efficiency of pharmacological protection when compared with controls. The animals were divided into four groups. Group I (n = 12) served as controls. The dogs of Group II (n = 12) were treated with PGE I (100 ng/kg/min) during clamping and the first hour after &clamping. In the third group (n = 12) SOD was given as an intra-arterial bolus (1 mg/kg) prior to &clamping which was followed by a continuous perfusion (0.4 mg/kg/min) into the carotid artery for 25 rain. In Group IV (n = 12) the dogs were treated with a combination of PGE~ and SOD in the same manner as in Groups 3 and 4. Results after pharmacological protection were significantly better than controls. In Group I all animals but one (92%) were paraplegic, as were five in Group II (42%) and eight in Group III (67%). In contrast no dog in Group IV developed paraplegia. There was a close correlation of SEP and postoperative recovery. The group with combination therapy (PGE~plus SOD) was characterised by a loss of the evoked potential for a mean of 15 rain, the PGE~ group for 45.8 rain and the SOD group for 58.5 rain. While the control group was characterised by a loss of 72.7 rain. Histological examination demonstrated extensive grey matter necrosis in all control cases, whereas in animals treated by PGE1 plus SOD, histopathological changes were absent or of minor degree without evidence of spinal cord infarction. Clinical, neurophysiological and neuropathological data present substantial evidence for a protective effect of PGE1 and SOD during aortic cross-clamping and post ischaemia. It can be concluded that the best protection was achieved with PGE I and SOD given together and it remains to be seen if this proves to be the case in man. Key Words: Spinal cord ischaemia; Thoracic aortic cross-clamping; Prostaglandin El; Superoxide dismutase; Somatosensory evoked potential.
Introduction Paraplegia continues to be a significant complication of thoraco-abdominal reconstructions. The incidence of paraplegia following thoraco-abdominal aneurysm repair, due to the extension of the aneurysm, ranges between 4 and 30%. 4 In spite of various surgical techniques, prolonged hypoxia of the lower parts of the spinal cord cannot always be avoided. Although barbiturates, 18 corticosteroids, 12 calcium channel blockers 8 and the *Presented at the 3rd Annual Meeting of the European Society for Vascular Surgery, Maim6, September 1989.
Please address all correspondence to: Klaus Grabitz, Medizinische Einrichtungen der University Dusseldorf, Mooreustrasse 5, 4000 Dusseldorf 1, West Germany. 0950-821X/90/010019+06 $03.00/0 © 1990Grune&StrattonLtd
opiate antagonist naloxone 9 are known to protect spinal nervous tissues during stages of temporary ischaemia, prostaglandin E1 (PGEO and superoxide dismutase (SOD) seem to be promising new candidates. It has been observed, that PGE1 protects the cell by means of precapillary sphincter relaxation, is, 16 an increase in 02 and glucose supply to the cell, 2'23 as well as by the inhibition of the post ischaemic release of lysosomal enzymes. 19,2s Superoxide dismutase on the other hand is considered to act as a scavenger of toxic superoxide radicals in the post ischaemic period. Therapy with antioxidants supposed to improve the tolerance of cells against reperfusion injury, w In this study the effectiveness of PGE1 and SOD in the prevention of post ischaemic paraplegia was tested under near clinical conditions using thoracic aortic cross-
20
K. Grabitz e t al.
clamping in dogs tbr 1 h. During the experiments spinal cord function was measured by means of somatosensory evoked potentials (SEP). Additionally, the neurological outcome and the histological findings were examined and compared with a control group. The following questions were posed: 1. Does pharmacological treatment (PGE1, SOD and their combination) improve the tolerance of the spinal cord to ischaemia? 2. Can a potentially protective effect be determined from clinical findings, results of SEP neuromonitoring and histological examinations?
Materials and Methods
Forty-eight beagle dogs (body weight 1 5 - 2 5 kg) were included in this study. After premedication with azopromazinc (7 mg i.m.) and i.v. induction of anaesthesia with a barbiturate (thalmylar sodium 1 7 m g / k g i.v.) the dogs were ventilated with a N20/O2 mixture (3:1). The minute volume was arranged for a PaCO2 level of around 35 torr. Anaesthesia was maintained with enflurane using a m e a n inspiratory concentration of 6% vol (Drager Vapor) which was sufficient to render the animal unresponsive to surgical stimulation, as monitored by continuous blood pressure and heart rate recording. Blood pressure readings anterior and posterior to the occlusion site were taken continuously, using the left carotid and left femoral artery respectively. In order to determine the function of the afferent sensory nervous pathways of the spinal cord, somatosensory evoked potentials (SEP) were recorded. With the aid of a peridurally attached bipolar catheter 1° of 22 gauges (Bipacing-Cath, Vigon Company, Aachen, West Germany), evoked potentials were initiated in the lower lumbar area using a rectangular impulse with an intensity between O. 5-0.1 mA (supramaximal stimulus), 5 Hz frequency and a duration of 0.2 ms (Digi Stim II, Houston, Texas, U.S.A.). For recording of SEPs from the underlying sensory cortical areas the negative Ag/AgCI cupelectrode was fastened to the midline of the cranium, 3 cm behind the intra-auricular plane, and the positive on the proximal part of the snout using collodium (impedance < 2 kfl). An earth electrode was fastened to the cleaned back covering the lower thoracic spinal cord. A total of 256 sweeps were picked up by a preamplifier (range 48 dB, impute resistance 100 Mfl), fed into a computer (Lifescan, Diatek, San Diego, U.S.A.) and averaged. The resulting signal was computed with regard to amplitude and of a peak, centering around 50 ms post stimulus Eur l Vasc Surg Vol 4, February 1990
(peak to peak amplitude). With a frequency range between 3 0 - 1 5 0 0 Hz, and a sampling rate of 3.1 kHz the device seemed sensitive enough to detect any functional changes in the ascending nervous pathways of the spinal cord following ischaemia. Evoked potentials were recorded at fixed intervals: (1) before aortic cross-clamping (control period); (2) at 3 min intervals following aortic cross-clamping in order to determine the time of potential loss; (3) at 3 min intervals after aortic declamping to determine the time of potential regeneration; (4) 60 rain post aortic declamping. Following a left-sided thoracotomy in the fourth intercostal space distal to branching of the left subclavian artery, the aorta was dissected and a clamp positioned. A total of 48 dogs were randomly divided into four groups: Group I, (control; n = 12) received no pharmacological protection. Group II, (PGE1 group; n = 1 2 ) received prostaglandin E1 (Prostavasin). A previously established dose of 100 ng/kg/min was considered sufficient w h e n arterial pressure dropped by at least 15 torr. The infusion was started 15 min prior to cross-clamping and maintained over the whole ischaemic period (60 min) and for 60 min after declamping. Group III, (superoxide dismutase; n = 12). Treatment was similar to the control group, except that an intraarterial bolus of SOD 1 mg/kg (Peroxinorm, Grunenthal, Stolberg) was given 5 min prior to declamping. This was followed by the intra-arteriat infusion of SOD (0.4 mg/kg/ min) for the next 25 min. Group IV, (combination of PGE1 and SOD; n = 12). The animals in this group were treated with PGEa and SOD in the same m a n n e r as Group II and III. Post-operatively the animals were evaluated for 7 days with regard to motor function of the hind limbs. Tarlov's score 22 was used to differentiate between animals with good, poor and/or insufficient motor function: Grade O: Spastic paralysis with no active movements of the lower extremities. Grade 1: Paraplegia with little active movements of the lower extremities. Grade 2: Paraparesis with good active movements of the lower extremities but unable to stand. Grade 3: Paraparesis with some problems in standing and walking. Grade 4: Animals with normal motor function. In order to evaluate the significant difference in motor function a m o n g the four groups, the Fischer exact test was used. Differences were evaluated using a simple yes or no response regarding normal or abnormal motor function. For the evaluation of significance of evoked potentials a m o n g the four groups, the Kruskal-Wallis test was used taking the following end points: (a) the time of onset of cross-clamping to the time of total loss of the SEP.
The Prevention of Ischaemic Spinal Cord Injury
21
4.0 :5.5 ¢-
E 5.0 E 2.5 PGEI
0
~
2.0
0
~~S
1.5
~
1.0
. . . .
SOD
• . . . . .
1:3
0.5
T ....
0
I
2
5
4 5 Postoperative day
t
t c°n''°'
6
7
8
Fig, 1. Motor deficit following cross-clamping with different treatments (Tarlov score; m e a n :t: SEM). A s u m m a r y of postoperative neurologic status in the different experimental groups. Grade 0 = paraplegia; Grade 4 = normal neurologic function. - ~ - control.
(b) the time of declamping to the time of potential recovery during reperfusion. (c)the total time of potential loss ( = a + b ) . (d)the mean amplitude before clamping and 60 rain after declamping. Animals were sacrificed at 7 days postoperatively, the spinal cord removed and the lower lumbar regions studied by light microscopy.
Results
At the end of the postoperative period (day 7) only one animal (9%) out of 12 in the control group was able to walk. In the prostaglandin group seven animals out of 12 (58.3%) and four out of ] 2 in the SOD group (33.8%) showed no motor deficiency following the seventh post-
1//@Imcoo,,o, •~
]
III
]
g
~Tzz.
20-
O-
Loss of potential
Regeneration of potential
Duration of potential loss
Fig. g. Loss and recovery of somatosensory-evoked potentials (SEPs) during and after aortic cross-clamping in animals treated with different pharmacological agents. Time of lass, time of recovery, as well as total time of ]ass of the evoked potentia] response (mean + SD). Eur ] Vasc Surg Vol 4, February 1990
22
K. Grabitz e t al.
5 -/
4 i
~%~
SOD
/
r...................... ,~'~\\\'~
PGE I
T
~#~;:~{i~{~,~ ~
PGE~ plus SOD
-
T
/
-
. . . . . . .
................
~\\\\\\"
E ,/
z/////,
:;~;............
0
---
SEP before clamping
SEP 60 min post declamping
Fig. 3. Amplitude height of the somatosensory evoked potential (SEP) before cross-clamping and 60 min after declamping, in animals having received different protection (mean 4- sn).
operative day. In contrast no animal of the combination group sustained paraplegia (Fig. 1). The difference a m o n g the four groups is statistically significant with a value of P < 0 . 0 1 in favour of PGE1 plus SOD w h e n compared with controls. With regard to evoked potential measurements animals in all four groups showed a loss of amplitude of the SEP with the onset of cross-clamping. While in
o
the control and the SOD group no potentials could be derived after a mean of 15 min, loss of sensory function in the PGE] as well as the combination group (SOD plus PGE1) was delayed with a m e a n of 32 min (Fig. 2), this delayed loss of amplitude during prostaglandin treatment is statistically significant (P < 0.05). The time after which evoked potentials recovered in
4-
E~ t~
E
/
to u) /
/
2 -/ Q)
(..9
0 Animal in each group Fig. 4. Neuropathology: The histological grading of histopathological alterations of the spinal cord in the different experimental groups. Grade of tissue damage: 0 = normal tissue; 1 =cellular infiltration; 2 =microinfarction within one segment; 3 = microinfarction within several segments; 4 = total necrosis of the grey matter within one segment; 5 = total necrosis of the grey matter within several segments, One animal with meningitis was excluded from the control group. EurJ Vasc Surg Vol 4, February 1 9 9 0
The Prevention of Ischaemic Spinal Cord Injury
the reperfusion period was 3 6 m i n in the control, 12.5min in the PGE1 and 3.Smin in the combination group (PGE1 plus SOD) (P
Does Prostaglandin E 1 and Superoxide Dismutase Prevent Ischaemic Spinal Cord Injury A f t e r Thoracic Aortic Cross-Clamping?* K. Grabitz, E. Freye, R. Prior, K. Schr6r and W. S a n d m a n n University of Dusseldorf, West Germany The beneficial use of prostaglandin E1 (PGE1) and superoxide dismutase (SOD) on the tolerance to ischaemia of the spinal cord was evaluated following thoracic aortic cross-clamping in dogs. Aside from spinally evoked somatosensory potential (SEP) by means of a bipolar epidural catheter, postoperative evaluation of motor deficits was used to determine the efficiency of pharmacological protection when compared with controls. The animals were divided into four groups. Group I (n = 12) served as controls. The dogs of Group II (n = 12) were treated with PGE I (100 ng/kg/min) during clamping and the first hour after &clamping. In the third group (n = 12) SOD was given as an intra-arterial bolus (1 mg/kg) prior to &clamping which was followed by a continuous perfusion (0.4 mg/kg/min) into the carotid artery for 25 rain. In Group IV (n = 12) the dogs were treated with a combination of PGE~ and SOD in the same manner as in Groups 3 and 4. Results after pharmacological protection were significantly better than controls. In Group I all animals but one (92%) were paraplegic, as were five in Group II (42%) and eight in Group III (67%). In contrast no dog in Group IV developed paraplegia. There was a close correlation of SEP and postoperative recovery. The group with combination therapy (PGE~plus SOD) was characterised by a loss of the evoked potential for a mean of 15 rain, the PGE~ group for 45.8 rain and the SOD group for 58.5 rain. While the control group was characterised by a loss of 72.7 rain. Histological examination demonstrated extensive grey matter necrosis in all control cases, whereas in animals treated by PGE1 plus SOD, histopathological changes were absent or of minor degree without evidence of spinal cord infarction. Clinical, neurophysiological and neuropathological data present substantial evidence for a protective effect of PGE1 and SOD during aortic cross-clamping and post ischaemia. It can be concluded that the best protection was achieved with PGE I and SOD given together and it remains to be seen if this proves to be the case in man. Key Words: Spinal cord ischaemia; Thoracic aortic cross-clamping; Prostaglandin El; Superoxide dismutase; Somatosensory evoked potential.
Introduction Paraplegia continues to be a significant complication of thoraco-abdominal reconstructions. The incidence of paraplegia following thoraco-abdominal aneurysm repair, due to the extension of the aneurysm, ranges between 4 and 30%. 4 In spite of various surgical techniques, prolonged hypoxia of the lower parts of the spinal cord cannot always be avoided. Although barbiturates, 18 corticosteroids, 12 calcium channel blockers 8 and the *Presented at the 3rd Annual Meeting of the European Society for Vascular Surgery, Maim6, September 1989.
Please address all correspondence to: Klaus Grabitz, Medizinische Einrichtungen der University Dusseldorf, Mooreustrasse 5, 4000 Dusseldorf 1, West Germany. 0950-821X/90/010019+06 $03.00/0 © 1990Grune&StrattonLtd
opiate antagonist naloxone 9 are known to protect spinal nervous tissues during stages of temporary ischaemia, prostaglandin E1 (PGEO and superoxide dismutase (SOD) seem to be promising new candidates. It has been observed, that PGE1 protects the cell by means of precapillary sphincter relaxation, is, 16 an increase in 02 and glucose supply to the cell, 2'23 as well as by the inhibition of the post ischaemic release of lysosomal enzymes. 19,2s Superoxide dismutase on the other hand is considered to act as a scavenger of toxic superoxide radicals in the post ischaemic period. Therapy with antioxidants supposed to improve the tolerance of cells against reperfusion injury, w In this study the effectiveness of PGE1 and SOD in the prevention of post ischaemic paraplegia was tested under near clinical conditions using thoracic aortic cross-
20
K. Grabitz e t al.
clamping in dogs tbr 1 h. During the experiments spinal cord function was measured by means of somatosensory evoked potentials (SEP). Additionally, the neurological outcome and the histological findings were examined and compared with a control group. The following questions were posed: 1. Does pharmacological treatment (PGE1, SOD and their combination) improve the tolerance of the spinal cord to ischaemia? 2. Can a potentially protective effect be determined from clinical findings, results of SEP neuromonitoring and histological examinations?
Materials and Methods
Forty-eight beagle dogs (body weight 1 5 - 2 5 kg) were included in this study. After premedication with azopromazinc (7 mg i.m.) and i.v. induction of anaesthesia with a barbiturate (thalmylar sodium 1 7 m g / k g i.v.) the dogs were ventilated with a N20/O2 mixture (3:1). The minute volume was arranged for a PaCO2 level of around 35 torr. Anaesthesia was maintained with enflurane using a m e a n inspiratory concentration of 6% vol (Drager Vapor) which was sufficient to render the animal unresponsive to surgical stimulation, as monitored by continuous blood pressure and heart rate recording. Blood pressure readings anterior and posterior to the occlusion site were taken continuously, using the left carotid and left femoral artery respectively. In order to determine the function of the afferent sensory nervous pathways of the spinal cord, somatosensory evoked potentials (SEP) were recorded. With the aid of a peridurally attached bipolar catheter 1° of 22 gauges (Bipacing-Cath, Vigon Company, Aachen, West Germany), evoked potentials were initiated in the lower lumbar area using a rectangular impulse with an intensity between O. 5-0.1 mA (supramaximal stimulus), 5 Hz frequency and a duration of 0.2 ms (Digi Stim II, Houston, Texas, U.S.A.). For recording of SEPs from the underlying sensory cortical areas the negative Ag/AgCI cupelectrode was fastened to the midline of the cranium, 3 cm behind the intra-auricular plane, and the positive on the proximal part of the snout using collodium (impedance < 2 kfl). An earth electrode was fastened to the cleaned back covering the lower thoracic spinal cord. A total of 256 sweeps were picked up by a preamplifier (range 48 dB, impute resistance 100 Mfl), fed into a computer (Lifescan, Diatek, San Diego, U.S.A.) and averaged. The resulting signal was computed with regard to amplitude and of a peak, centering around 50 ms post stimulus Eur l Vasc Surg Vol 4, February 1990
(peak to peak amplitude). With a frequency range between 3 0 - 1 5 0 0 Hz, and a sampling rate of 3.1 kHz the device seemed sensitive enough to detect any functional changes in the ascending nervous pathways of the spinal cord following ischaemia. Evoked potentials were recorded at fixed intervals: (1) before aortic cross-clamping (control period); (2) at 3 min intervals following aortic cross-clamping in order to determine the time of potential loss; (3) at 3 min intervals after aortic declamping to determine the time of potential regeneration; (4) 60 rain post aortic declamping. Following a left-sided thoracotomy in the fourth intercostal space distal to branching of the left subclavian artery, the aorta was dissected and a clamp positioned. A total of 48 dogs were randomly divided into four groups: Group I, (control; n = 12) received no pharmacological protection. Group II, (PGE1 group; n = 1 2 ) received prostaglandin E1 (Prostavasin). A previously established dose of 100 ng/kg/min was considered sufficient w h e n arterial pressure dropped by at least 15 torr. The infusion was started 15 min prior to cross-clamping and maintained over the whole ischaemic period (60 min) and for 60 min after declamping. Group III, (superoxide dismutase; n = 12). Treatment was similar to the control group, except that an intraarterial bolus of SOD 1 mg/kg (Peroxinorm, Grunenthal, Stolberg) was given 5 min prior to declamping. This was followed by the intra-arteriat infusion of SOD (0.4 mg/kg/ min) for the next 25 min. Group IV, (combination of PGE1 and SOD; n = 12). The animals in this group were treated with PGEa and SOD in the same m a n n e r as Group II and III. Post-operatively the animals were evaluated for 7 days with regard to motor function of the hind limbs. Tarlov's score 22 was used to differentiate between animals with good, poor and/or insufficient motor function: Grade O: Spastic paralysis with no active movements of the lower extremities. Grade 1: Paraplegia with little active movements of the lower extremities. Grade 2: Paraparesis with good active movements of the lower extremities but unable to stand. Grade 3: Paraparesis with some problems in standing and walking. Grade 4: Animals with normal motor function. In order to evaluate the significant difference in motor function a m o n g the four groups, the Fischer exact test was used. Differences were evaluated using a simple yes or no response regarding normal or abnormal motor function. For the evaluation of significance of evoked potentials a m o n g the four groups, the Kruskal-Wallis test was used taking the following end points: (a) the time of onset of cross-clamping to the time of total loss of the SEP.
The Prevention of Ischaemic Spinal Cord Injury
21
4.0 :5.5 ¢-
E 5.0 E 2.5 PGEI
0
~
2.0
0
~~S
1.5
~
1.0
. . . .
SOD
• . . . . .
1:3
0.5
T ....
0
I
2
5
4 5 Postoperative day
t
t c°n''°'
6
7
8
Fig, 1. Motor deficit following cross-clamping with different treatments (Tarlov score; m e a n :t: SEM). A s u m m a r y of postoperative neurologic status in the different experimental groups. Grade 0 = paraplegia; Grade 4 = normal neurologic function. - ~ - control.
(b) the time of declamping to the time of potential recovery during reperfusion. (c)the total time of potential loss ( = a + b ) . (d)the mean amplitude before clamping and 60 rain after declamping. Animals were sacrificed at 7 days postoperatively, the spinal cord removed and the lower lumbar regions studied by light microscopy.
Results
At the end of the postoperative period (day 7) only one animal (9%) out of 12 in the control group was able to walk. In the prostaglandin group seven animals out of 12 (58.3%) and four out of ] 2 in the SOD group (33.8%) showed no motor deficiency following the seventh post-
1//@Imcoo,,o, •~
]
III
]
g
~Tzz.
20-
O-
Loss of potential
Regeneration of potential
Duration of potential loss
Fig. g. Loss and recovery of somatosensory-evoked potentials (SEPs) during and after aortic cross-clamping in animals treated with different pharmacological agents. Time of lass, time of recovery, as well as total time of ]ass of the evoked potentia] response (mean + SD). Eur ] Vasc Surg Vol 4, February 1990
22
K. Grabitz e t al.
5 -/
4 i
~%~
SOD
/
r...................... ,~'~\\\'~
PGE I
T
~#~;:~{i~{~,~ ~
PGE~ plus SOD
-
T
/
-
. . . . . . .
................
~\\\\\\"
E ,/
z/////,
:;~;............
0
---
SEP before clamping
SEP 60 min post declamping
Fig. 3. Amplitude height of the somatosensory evoked potential (SEP) before cross-clamping and 60 min after declamping, in animals having received different protection (mean 4- sn).
operative day. In contrast no animal of the combination group sustained paraplegia (Fig. 1). The difference a m o n g the four groups is statistically significant with a value of P < 0 . 0 1 in favour of PGE1 plus SOD w h e n compared with controls. With regard to evoked potential measurements animals in all four groups showed a loss of amplitude of the SEP with the onset of cross-clamping. While in
o
the control and the SOD group no potentials could be derived after a mean of 15 min, loss of sensory function in the PGE] as well as the combination group (SOD plus PGE1) was delayed with a m e a n of 32 min (Fig. 2), this delayed loss of amplitude during prostaglandin treatment is statistically significant (P < 0.05). The time after which evoked potentials recovered in
4-
E~ t~
E
/
to u) /
/
2 -/ Q)
(..9
0 Animal in each group Fig. 4. Neuropathology: The histological grading of histopathological alterations of the spinal cord in the different experimental groups. Grade of tissue damage: 0 = normal tissue; 1 =cellular infiltration; 2 =microinfarction within one segment; 3 = microinfarction within several segments; 4 = total necrosis of the grey matter within one segment; 5 = total necrosis of the grey matter within several segments, One animal with meningitis was excluded from the control group. EurJ Vasc Surg Vol 4, February 1 9 9 0
The Prevention of Ischaemic Spinal Cord Injury
the reperfusion period was 3 6 m i n in the control, 12.5min in the PGE1 and 3.Smin in the combination group (PGE1 plus SOD) (P
