The arteria radicularis magna anterior as a decisive factor influencing spinal cord damage during aortic occlusion

Volume 88,

Number 1

July 1984

THORACIC AND CARDIOVASCULAR SURGERY The Journal oj

J

THORAC CARDIOVASC SURG

88:1-10, 1984

Original Communications

The arteria radicularis magna anterior as a decisive factor influencing spinal cord damage during aortic occlusion Cross-clamping of the descending aorta immediately belowthe subclavianartery may result in damage to the spinal cord. Despite various protective procedures, the risk of such damage cannot be entirely eliminated. In an experimental study with 47 pigs, the influenceof variousfactors on the genesisof spinal of the cord damage was examined. The pigs were divided into five groups: Groups I to IV -occ~ion descending aorta for 45 minutes;Group I-no reduction in arterial blood pressure proximal to the site of occlusion; Group II-like Group I, plus drainage of the cerebrospinal fluid; Group ill-reduction in arterial blood pressure; Group IV-like Group III, plus drainage of the cerebrospinal fluid; Group V-permanent ligation of the artery of Adamkiewicz. The degree of permanent spinal cord damage was 85.7% (Groups I to IV, six animals) and 71.4% (Group V, five animaIs~ Thus there were no significant differences among the various groups. The frequency of spinal cord damage was independent of arterial blood pressure, intracranial pressure, and intraspinal pressure. The intracranial pressure and the intraspinal pressure were significantly dependent uponthe central venous pressure but were independentof the arterial blood pressure.

Faysal Wadouh, M.D., Eva-Maria Lindemann, D.D.M., Christoph F. Arndt, Cand.Med., Roland Hetzer, M.D., and Hans G. Borst, M.D., Hannover, Federal Republic of Germany

CrOSS-clamping of the descending aorta immediately below the left subclavian artery entails two risks for the from the Divisionof Thoracic and Cardiovascular Surgery, Hannover Medical School, Konstanty-Gutschow-Strasse 8, 3000 Hannover 61, federal Republic of Germany. Received for publication June I, 1983. Accepted for publication Sept. 20, 1983. Address for reprints: f. Wadouh, wissenschaftlicher Assistenzarzt, Abteilung fur Thorax-, Herz- und Gefasschirurgie, Medizinische Hochschule Hannover, Konstanty-Gutschow-Strasse 8, 3000 Hannover 61, federal Republic of Germany.

organism. First, the acute rise in blood pressure proximal to the point of occlusion leads to a volume burden on the brain and heart resulting in cerebrovascular damage and cardiac insufficiency. Second, the decrease in blood pressure distal to the occlusion induces organ damage, especially to the spinal cord. For organs possessing an anatomically defined blood supply, for example, the heart, kidneys, and brain, a tolerance time for ischemia can be determined to a limited extent. This has not been possible in the case of the spinal cord. Consequently, controversial and contradictory assertions

The Journal of Thoracic and Cardiovascular Surgery

2 Wadouh et al.

have been made. Carrel! was the first author to report paraplegia in dogs after cross-elamping the aorta for 10 to 15 minutes. Similar results were achieved by Tureen,' and Gross and Hufnagel' reported spinal cord damage after only 4Jh minutes of occlusion. Van Harreveld and Schade" found no spinal cord damage after asphyxia lasting for 15 to 20 minutes. After 25 to 35 minutes, they observed asphyxia-related damage to the spinal cord. On the basis of numerous studies, Adams and Van Geertruyden' demonstrated that a minimum occlusion time of 18 minutes, 40 seconds could be tolerated without any consequences. This clearly demonstrates that so far neither animal experiments nor clinical observations allow for a defmed period of tolerance of the spinal cord for ischemia. The cause and the variable conditions resulting in paraplegia after cross-elamping of the descending aorta remain unclear and continue to be the subject of numerous clinical and experimental studies. The resulting uncertainties forced many scientists to provide an explanation for these discrepancies: Miyamoto and associates" maintained that the cause of paraplegia in animal experiments is the sudden increase in the intracranial pressure during the occlusion phase, which results in compression of the spinal vessels. Blaisdell and Cooley' came to a similar conclusion, although Crawford and Rubio" claimed the cause of paraplegia was either the drop in blood pressure before, during, and after the operation or else the ligature of vital intercostal or lumbar arteries. The clinical studies of Katz and associates," showed clearly that in patients operated upon without use of a shunt the rate of paraplegia increased with the occlusion time. With a shunt, however, no relationship could be determined between the rate of paraplegia and the occlusion time. Despite these fmdings, further clinical experience could not prevent damage to the spinal cord, even though certain preventive measures were introduced. Although a shunt during aortic cross-elamping prevented or minimized the drop in blood pressure distal to the site of cross-elamping, spinal cord damage was still observed in 3.6%10 to 17.7%11 of the cases. DeBakey and colleagues," in clinical studies, found the incidence of paraplegia to be similar with and without the shunt. Crawford and Rubio" reported an even higher paraplegia rate with the application of a shunt (7.8% with compared to 2.2% without a shunt), and Katz and associates? observed a 15.8% rate of paraplegia despite use of a shunt. These results justified researching this problem in larger animals and closely examining the various possible causes of spinal cord damage during cross-elamping of the aorta. The following premises were examined and are discussed in this paper: (1) the

influence of arterial blood pressure on the intracranial pressure (ICP) and intraspinal pressure (ISP) and on the rate of paraplegia; (2) the influence of the reduction in arterial blood pressure on the ICP, ISP, and rate of paraplegia; (3) the advantages of cerebrospinal fluid (CSF) drainage in reducing paraplegia; and (4) the influence of the permanent ligature of the arteria radicularis magna anterior (artery of Adamkiewicz) on the rate of paraplegia during a normal hemodynamic state.

Materials and methods Forty-seven pigs (25 to 30 kg body weight) were divided into five groups. The animals were intubated to receive nitrous oxide and oxygen (3: 1) and were given stresnil* and hypnodilt as needed. An arterial catheter was inserted into the common carotid artery via a cervical side branch to measure the arterial pressure in the aortic arch (proximal arterial pressure) and a venous catheter was inserted into the external jugular vein to measure the central venous pressure (CVP) in the superior vena cava. A second arterial catheter was inserted into the femoral artery to monitor the distal arterial pressure. In Groups I to IV, an occlusion catheter (Fogarty occlusion catheter 8-14 OC):j: was inserted into the femoral artery, and the balloon was positioned under radiologic control immediately distal to the origin of the left subclavain artery. In Group V, either the lumbar vessels (L1-S1) or the artery of Adamkiewicz was ligated by means of a left lateral retroperitoneal approach. In all animals, the cisterna cerebellomedullaris was punctured after the dorsal cervical musculature had been split in the midline, and the ICP was measured. After the spinous process at the level of L 4-Ls had been resected, the subarachnoid space was punctured in Groups I, III, and V and the ISP was measured. In Groups II and IV, the CSF was drained by a wide laminectomy (L 4-Ls); in these animals only ICP could be determined reliably; ISP became unmeasurable because of technical errors, such as air entering the spinal canal at the site of laminectomy. All catheters were connected to electromagnetic pressure transducers,§ and continuous monitoring was done by a six-ehannel recorder ~I The electrocardiogram and rectal temperature were monitored continually. Blood gases, electrolytes, and standard *Azaperon, Janssen Pharmaceutica, Inc., New Brunswick, N. J. tMetomidate, Janssen Pharmaceutica, Inc., New Brunswick, N. J. :j:American Edwards Laboratories, Irvine, Calif. §Gould-Statham P23ID, Gould, Inc., Oxnard, Calif. IjHellige & Co. GmbH., Federal Republic of Germany.

Volume 88

Spinal cord damage during aortic occlusion

Number 1

3

July, 1984

Table I. Group I (seven pigs): No reduction in arterial pressure and no drainage of cerebrospinal fluid During occlusion

Before occlusion Animals No.

AP prox (mm Hg)

CVP (mm Hg)

I 2 3 4 5 6 7 Mean ±SEM

105 95 70 110 115 70 lOS 95.71 7.02

7.5 11.5 6.5 6 6 4 5 6.64 0.91

ICP/ISP (mm Hg)

AP prox (mm Hg)

CVP (mmHg)

ICP/ISP (mm Hg)

10.5

145 132.5 125 ISS 155 125 155 141.86 5.29

9 14.5 10 8 9 6.5 8.5 9.36 0.95

14.5 13.5 14 10.5 14 11.5 13.5 13.07 0.56

II

11 7.5 10 8.5 6.5 9.29 0.68

Legend. AP prox, Arterial pressure proximally. AP dist, Arterial pressure distally. CVP, Central venous pressure. ICP, Intracranial pressure. ISP, Intraspinal pressure. SEM, Standard error of the mean.

Table Il, Group II (seven pigs): No reduction in arterial pressure; drainage of cerebrospinal fluid (CSF) Before occlusion

During occlusion

Animal No.

ICP'* (mm Hg)

8 9

5 4.5 5.5

10 II 12 13 14 Mean ±SEM

4

3.5 6 7

5.07 0.46

AP prox (mm Hg)

145 140

150 120 125 150 110 134.28 6.02

ICP"* (mmHg)

6.5 7 7

6 6 8

9 7.07 0.41

Legend:AP prox, Arterial pressure proximally. AP dist, Arterial pressure distally. CVP, Central venous pressure. ICP, Intracranial pressure. SEM, Standard error of the mean. *'CP after CSF drainage. No positive intraspinal pressure was measurable due to wide laminectomy and CSF drainage.

bicarbonate were controlled periodically. The period of aortic occlusion was 45 minutes for Groups I to IV. In Groups III and IV, the proximal arterial pressure was reduced to preoperative values with sodium nitroprusside* during the period of occlusion. Shortly before and after the blood flow was released into the aorta, sodium bicarbonate was administered to balance the expected acidosis, The time during which the arterial pressure had fallen was at the most 1 minute. The animals were examined neurologically on the first and second postoperative days. The characteristics of the five groups were as follows: Group L' Forty-five minutes' occlusion time of the descending aorta. Pressure recordings included proximal and distal arterial pressures, CVP, ICP, and ISP. Group II: Forty-five minutes' occlusion time of the *Nipruss, Schwarz GmbH., Federal Republic of Germany.

descending aorta. The CSF was drained after laminectomy (L4- L) and opening of the subarachnoid space. Pressure recordings included proximal and distal arterial pressures, CVP, and ICP. Group IlL' Forty-five minutes' occlusion time of the descending aorta, and controlled reduction of proximal arterial pressures, to preoperative values with sodium nitroprusside during occlusion. Pressure recordings included proximal and distal arterial pressures, CVP, ICP, and ISP. Group IV: Forty-five minutes' occlusion time of the descending aorta. The CSF was drained after laminectomy (L4-Ls) and opening of the subarachnoid space. Controlled reduction of proximal arterial pressure to preoperative values was effected with sodium nitroprusside during occlusion. Pressure recordings included proximal and distal arterial pressures, CVP, and ICP. Group v.. No occlusion of the descending aorta. Ligation of the lumbar vessels (L1-S 1) or artery of

4

The Journal of Thoracic and Cardiovascular Surgery

Wadouh et al.

Table ill. Group III (seven pigs): Controlled arterial pressure reduction with sodium nitroprusside to almost the initial value before occlusion Before occlusion Animals No.

AP prox (mmHg)

15 16 17 18 19 20 21 Mean ±SEM

115 80 80 80 105 95 85 91.43 5.31

During occlusion CVP (mm Hg)

ICP/ISP (mm Hg)

AP prox (mm Hg)

CVP (mm Hg)

ICP/ISP (mmHg)

7.5 6.5 7 6.5 12 8.5 7 7.86 0.74

115 85 90 90 95 100 90 95.00 3.78

6 7.5 10.5 12

8.5 7.5

5 6.5 10 8.5

5 5.5 5 6.50 0.76

5.5 7 6.5 7.86 0.92

8 10 15 11

9 9.8 0.9

Legend: AP prox Arterial pressure proximally. AP dist, Arterial pressure distally. CVP, Central venous pressure. ICP. Intracranial pressure. ISP. Intraspinal pressure. SEM, Standard error of the mean.

Table IV. Group IV (seven pigs): Controlled reduction of arterial pressure with sodium nitroprusside to almost the initial value before occlusion with cerebrospinal fluid (CSF) drainage During occlusion

Before occlusion Animal No. 22 23 24 25 26 27 28 Mean ±SEM

80 90 110 70 100 80 60 84.28 6.49

85 90 110 80 105 85 70 89.28 5.28

7 6 8 6 5 10 5 6.71 0.68

9 8 10 7 7 11 8 8.57 0.57

ICP'* (mm Hg)

AP prox (mmHg)

4 4 5 3 4 6 4.5 4.30 0.36

95 95 100 80 100 90 70 90.00 4.23

ICP" * (mm Hg) 20 20 18 18 22 20 16 19.14 0.74

9 10 9 8 6.5 11 6 8.50 0.68

6 5 7 5 5 7 5 5 0

Legend: AP prox, Arterial pressure proximally. AP dist, Arterial pressure distally. CVP, Central venous pressure. ICP, Intracranial pressure. SEM, Standard error of the mean. *ICP after CSF drainage. No positive intraspinal pressure measurable due to wide laminectomy and CSF drainage.

Adamkiewicz. Pressure recording included proximal and distal arterial pressures, CVP, ICP, and ISP. Only 35 animals that survived for at least 48 hours (seven animals in each group) were included in the evaluation of paraplegia. Twelve additional animals died during the early postoperative period (24 hours): four animals in Group I, five in Group II, one in Group III, two in Group IV, and none in Group V. All animals received humane care in complicance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the "Guide for the Care and Use of Laboratory Animals" prepared by the National Academy of Sciences and published by the National Institutes of Health (NIH publication No. 80-23, revised 1978). Statistical methods. For the descriptive data analy-

sis, contingency table analysis using chi square statistics, unpaired and paired t test, analysis of variance, and correlation and regression analysis have been included. All computations were performed by means of the Statistical Package for the Social Sciences (CDC version 7.0).13 Results Occlusion of the thoracic aorta: Groups I to IV. The measured pressures before and during occlusionare demonstrated in Tables I to IV and Figs. 1 and 2, which also show pressure changes caused by occlusion. In Groups I and II the proximal arterial pressures rose significantly (p < 0.001) after occlusion,whereas in Groups III and IV these pressures were maintained at almost their preocclusion levels.

Volume 88 Number 1

Spinal cord damage during aortic occlusion 5

July, 1984

mm Hg

100

50

tITJr ~I

-

[JI

o.

IIInr Dy

20'

I

mm Hg

I 10

o Fig. 1. Changes in the measured pressures. APPb. Mean arterial pressure proximal to the site of occlusionbefore occlusion. APPa. After occlusion. APDb. Mean arterial pressure distal to the site of occlusion-before occlusion. APDa. After occlusion. CVPb. Mean central venous pressure-before occlusion. CVPa, After occlusion. [CPb. Mean intracranial pressure-before occlusion. [CPa, After occlusion. [CP, Mean intracranial pressure after drainage of cerebrospinal fluid. [CP", Mean intracranial pressure after drainage of cerebrospinal fluid and after occlusion. V. No occlusion of the aorta; no drainage of cerebrospinal fluid; no changes in arterial, central venous, intracranial, and intraspinal pressures; only ligation of the arteria radicularis magna anterior.

The distal arterial pressure fell significantly (p < 0.(01) in all groups after occlusion to approximately 20 mm Hg. The CVP rose considerably (p < 0.001 to 0.014) after occlusion; however, because of treatment of Groups III and IV with sodium nitroprusside, the rise in CVP was less significant than in Groups I and II (p < 0.009), i.e., sodium nitroprusside lowered the CVP after occlusion, but not to the level before occlusion. In Groups II and IV, the.ICP was significantly reduced (p < 0.(02) after CSF drainage, but always remained positive. After occlusion the ICP rose markedly (p < 0.001 to 0.026) in all groups; however, in Groups II and IV it remained below the initial values. Even though the proximal arterial pressure in Group III was maintained during the occlusion phase at its initial value, the CSF pressure nevertheless rose significantly (p < 0.(03). It can be concluded, therefore, that even a controlled reduction in blood pressure during the occlusion phase could not prevent a rise in the CSF pressure.

The covariance, regression, and partial correlation analyses all congruously demonstrated that the change in the ICP is significantly dependent upon the change in the CVP (p < 0.02), whereas no dependency on the change in proximal arterial pressure could be demonstrated. Ligature of the artery of Adamkiewicz. The measured pressures are demonstrated in Table V and Figs. I and 2. In this group, the pressures remained constant. Experimental results. At 48 hours postoperatively, 85.7% of the animals in Groups I to IV and 71.4% of the pigs in Group V had spinal cord damage (Table VI and Fig. 3). Comparing all five groups by means of the chi square test revealed no significant differences in the rates of paraplegia. Thus neither CSF drainage nor controlled blood pressure afforded any protection against damage to the spinal cord. Through controlled reduction of blood pressure after occlusion of the descending aorta, the operative mortality (24 hours) was reduced from 39.1% (Groups I and II) to 17.6% (Groups III and IV) (Table IV). There were no deaths in Group V.

The Journal of

6 Wadouh et al.

Thoracic and Cardiovascular Surgery

I

I

I

:Ill

• •

1 - -_ _

• •

11

_

6APD.~

6ICP

"

• •

.. ..L-

_

Fig. 2. Pressurechangesin percentafter occlusion of the thoracicaorta in Groups I to IV. V. no occlusion of the aorta; no changes in pressures; no drainage of cerebrospinal fluid; only ligation of the arteria radicularis magna anterior. Initial pressures beforeocclusion of the descending thoracicaorta. APP, Arterial pressure proximal to site of occlusion. APD, Arterial pressuredistal to site of occlusion. CVP, Central venous pressureafter occlusion of the descending thoracic aorta. fCP. Intracranial pressureafter drainage of cerebrospinal fluid, beforeocclusion of the descending thoracicaorta. fCP', Intracranial pressureafter drainage of cerebrospinal fluid, beforeocclusion of the descending thoracic aorta. fCP' " Intracranial pressureafter drainage of cerebrospinal fluid and occlusion of the descending thoracicaorta. None, No occlusion of the descending thoracicaorta, no drainage of cerebrospinal fluid, and no changes in pressure.

Discussion The causes of spinal cord damage after replacement of the descending aorta have not yet been defined beyond doubt and remain a topicof numerous sicentific studies. Paraplegia is still reported in 2.3% to 24% of patients."12 A varietyof animal experiments and clinical studies resulted in controversial and contradictory conclusions. In animal experiments, ischemic tolerancetime for the spinal cord after aortic occlusion varies between 4~ and 15 minutes.1-3 Von Harreveld and Schade" induced spinal cord asphyxia by elevating the ISP over the arterial blood pressure and observed spinal cord damage after an asphyxia time of 25 to 35 minutes. After an occlusion time of 60 minutes, Pontius and associates" found no paraplegia in 70%, Cooley and Delsakey" in 36%, and Miyamoto and associates" in 33%of their animals. In our ownexperiments, 8%of the animals did not have paraplegia after an occlusion time of 45 minutes; in an additional 9%, completely reversible symptoms resulted. Thus the total percentage of animals with no permanent damage was 17%.

The multifactorial clinical study of Katz and colleagues? demonstrated that with an occlusion time of over 30 minutes (without shunt), the risk of paraplegia in man increases. Other authors observed paraplegia after an occlusion time of 45 minutes," and othersafter only 18 minutes.' In addition, numerous authors havereportedparaplegia in patients without aortic occlusion.v" Thus it appears to be difficult to define a tolerance time of the spinal cord for ischemia. Certain protective procedures havebeenintroduced in an effort to preventinjury to the spinalcord after cross-elamping of the aorta.15,25-28 Even the use of a shunt or bypass procedure could not prevent damage to the spinal cord, with the incidence of paraplegia reported to range from 3.8% to 17.7%.8-12,19,26,29-33 However, the use of a shunt has produced contradictory results." 9,12 During the past 30 years, several scientists have attempted to clarifythe genesis of paraplegia after aorta cross-e1amping. Crawford and Rubio" believed that the answer lay either in a hypotensive phase before, during,

Volume 88 Number 1

Spinal cord damage during aortic occlusion

July, 1984

7

Table V. Group V (seven pigs): ligation of the lumbar arteries L/ through S/ only Animal No. 29 30 31 32 33 34 35 Mean ±SEM

AP prox (mm Hg)

AP dist (mm Hg)

CVP

ICP/ISP

(mmHg)

(mmHg)

70 110 80

75 110 80

90

90

60

70

8 8.5 8.5 II 6

90

90 90

7 5

86.43 4.97

7.71 0.74

80 82.86 6.06

7.5

9 6 10.5 8

7 6.5 7.71

0.59

Legend: AP prox, Arterial pressure proximally. AP dist, Arterial pressure distally. CYP, Central venous pressure. ICP, Intracranial pressure. lSP, Intraspinal pressure. SEM, Standard error of the mean.

Table VI. Incidence of spinal cord damage and death Total

Group I

After 24 hours Paralysis Paresis No spinal damage Total Cause of death Cardiac failure Subacute cerebral damage Total Grand total After 48 hours Paralysis Paresis No spinal damage Total

1/

11/

IV

V

No.

%

1 5 I 7

2 5 0 7

5 1 I 7

5 2 0 7

3 3 I 7

16 16 3 35

45.7 45.7 8.6 100.0

1 3 4 11

1 4 5 12

1 0 I 8

2 0 2 9

0 0 0 7

12 47

25.5 100.0

1 5 I 7

2 4 I 7

3 3 I 7

3 3 I 7

2 3 2 7

II 18 6 35

31.4 51.4 17.2 100.0

or after operation or in the ligation of the intercostal or the lumbar arteries, which are of vital importance for the blood supply to the spinal cord. However, they could not explain why in some cases paraplegia occurred with a chronic supradiaphragmatic aneurysm with supposedly well-developed collateral circulation to the spinal cord and did not occur in other cases with an aneurysm of identical location despite more extensive lesions. Some researchers maintained that the reduction in distal arterial pressure was the underlying cause," However, as reported earlier, a shunt procedure which may prevent a drastic drop in pressure cannot be regarded as an infallible measure against paraplegia. Miyamoto and associates" and Blaisdell and Cooley? maintained that the cause of paraplegia lay in the elevation of the ICP as the result of the acute rise in proximal arterial pressure during the occlusion phase. Whereas Miyamoto's group" said the rise in the ICP led to compression of the spinal

vessels, Blaisdell and Cooley? saw the cause of paraplegia as the pressure gradient between the abdominal aorta and the spinal canal. A reduction in the CSF decompressed the spinal vessels," or the draining of the CSF guaranteed a sufficient minimal perfusion by means of the pressure gradient.' In their animal experiments, drainage of the CSF reduced the rate of paraplegia from 67% and 64% to 0% after an occlusion time of 60 minutes. In their studies, the proximal arterial pressure rose an average of 25% to 55%, the CVP in the superior vena cava rose 3 to 4 mm Hg, and the ICP rose 2 to 4 mm Hg, whereas the distal arterial pressure dropped about 75%. Drainage of the CSF reduced the ICP between 5 and 10 mm Hg. In our studies, we obtained similar pressure changes; however, despite laminectomy, the ICP fell a maximum of only 5 to 7 mm Hg during the occlusion phase, i.e., about 30% to 60% of the original value. The CSF, venous, and arterial

The Journal of Thoracic and Cardiovascular Surgery

8 Wadouh et al.

24 hours after operation II

III

IV

V

ooo®o

48 hours after operation I

II

III

IV

V

00000

@O • • O O@ • • @


••••• o

no spinal cord damage

CD

paresis

•

paralysIs

Fig. 3. Neurological results 24 and 48 hours after operation.

components have the greatest effect on the ICP. CSF drainage before and after occlusion would hardly influence the venous and arterial components; a positive ICP of at least 5 mm Hg would persist despite CSF drainage. It is not explicable how such small pressure differences could reduce the rate of paraplegia to 0% when even the physiological CSF pressure is 5 to 11 mm Hg in both animals and man. 35.36 In our studies, CSF drainage offered no protection against spinal cord damage. The rise in CSF pressure correlated significantly with the rise in venous pressure but was independent of the proximal arterial pressure. Even a reduction in the arterial pressure below the level before occlusion did not reduce the rate of paraplegia but did reduce the mortality rate, especially the mortality rate owing to subacute brain damage, from 39.1% to 17.6%. Only Miyamoto and associates" have reported on CSF drainage in man, describing only two patients operated upon with the use of CSF drainage who did not have paraplegia. Since the paraplegia rate after aortic occlusion varies between 2.3% and 24%, results in only two patients do not allow any conclusions. On the basis of our results, the effectiveness of this procedure is questionable. Even if the CSF pressure were reduced to o mm Hg, a distal arterial pressure of 15 to 20 mm Hg, as measured by Blaisdell and Cooley,' would not be a great enough pressure gradient between the abdominal aorta and spinal vessels to perfuse the spinal cord and thereby prevent paraplegia. One might assume that, by means of collateral circulation, the pressure in the spinal vessels is equal to or higher than in the aorta distal to the point of occlusion and the blood flow is in the direction

from spinal vessels to aorta. In our study on the anatomic peculiarities of the spinal vascular supply, the pressures in the collateral circulation, and the oxygen concentration of the spinal cord after aortic occlusion, the pressure in the spinal arteries was almost always higher than in the aorta distal to the point of occlusion (results in preparation). Even the isolated ligature of the artery of Adamkiewicz without aortic occlusion or alteration in hemodynamics resulted in a 72% rate of paraplegia in our experiments; similar results were obtained after occluding the aorta for 45 minutes and draining CSF by constant proximal arterial pressure. One might conclude that the development of paraplegia after aortic occlusion is decisively dependent upon the artery of Adamkiewicz. The somewhat higher rate of paraplegia after 45 minutes of aortic occlusion might be explained by the additional inhibition of the collateral circulation. Our results in pigs may be applied to man only when the discrepancies in structure of the spinal cords are considered. Whereas the artery of Adamkiewicz in the pig always enters the spinal cord at the level of L,-S, (usually L 4),37 in man it enters the spinal cord at the level of T,.T 8 in 12% to 15% of cases, T 9-T12 in 60%, L, in 14%, L 2 in 10%, L 3 in 1.4%, and L 4-L5 in 0.2%.38.39 The circulation of the spinal cord in man is paucisegmental, i.e., less than five radicular arteries in 45% of the cases"; however, the spinal circulation is always plurisegmental in the pig, a fact which has a decisive effect on the collateral circulation. In 1889, Kadyi" put forward the theory that with a reduction in the number of radicular arteries, the caliber of the anterior spinal artery increases. This was confirmed by other authors.39.41.42 Among others,": 40. 43 Adamkiewicz'" drew attention to the importance of the arteria radicularis magna anterior. The branching pattern of the artery of Adamkiewicz is of particular interest. Although the descending branch exhibits the largest caliber of all the radicular arteries, the ascending branch is always the smaller of the two. The diameter differs in each individual case, and in some cases this branch can even be absent." According to these anatomic and clinical details, three different possibilities must be taken into consideration when replacing the descending aorta. I. The artery of Adamkiewicz branches within the area of aortic occlusion. In this case, spinal cord damage is primarily due to the lack of collateral circulation. The shunt procedure would offer no protection. Not only will an arterio-arterial shunt fail to perfuse the artery of Adamkiewicz, but it could also contribute to the disruption of the subclavian-mammarian circulation, especial-

Volume BB Number 1

Spinal cord damage during aortic occlusion 9

July, 1984

ly in a left supply modus. With an insufficient collateral circulation, the artery of Adamkiewicz as well as several intercostal arteries should be reimplanted into the aortic graft. If this is not possible within a reasonable period of time, then hypothermia should be employed. The use of hypothermia without reimplantation of the artery of Adamkiewicz would have the same effect as shunt or bypass procedures. 2. The artery of Adamkiewicz branches inferior to the area of occlusion. When the collateral circulation is well developed, as in the paucisegmental type, a protective procedure is not necessary. However, when the collateral circulation is poorly developed, as in the plurisegmental type, hypothermia or a bypass along with preservation of the subclavian-mammary circulation should be employed. 3. The artery of Adamkiewicz branches above the area of occlusion. In this case, ischemia of the spinal cord is not expected. Therefore, protective procedures are superfluous. We conclude that spinal cord damage is solely related to a deficiency in oxygen supply during the occlusion phase and not to changes in the ambient pressures whereby the anatomic location of the artery of Adamkiewicz and the efficiency of the collateral circulation play the decisive roles. Spinal cord damage as a complication during replacement of the descending aorta can be avoided only if, before operation, information on the vascular morphology of the spinal cord, especially the position of the artery of Adamkiewicz, is available. We gratefully acknowledge the cooperation of consultant statistician Dr. Ing. U. Ranft, Department of Biometry, Hannover Medical School, Hannover, Federal Republic of Germany, Professor Dr. B. Schneider, Director.

2

3

4

5 6

REFERENCES Carrel A: Report on the experimental surgery of the thoracic aorta and heart. Ann Surg 52:83-95, 1910 Tureen LL: Circulation of the spinal cord and the effect of vascular occlusion, The Circulation of the Brain and Spinal Cord: A Symposium on Blood Supply, Vol 18, Baltimore, 1938, The Williams & Wilkins Company, pp 394-437 Gross RE, Hufnagel CA: Coarctation of aorta: Experimental studies regarding its surgical correction. N Engl J Med 233:287-293, 1945 Van Harreveld R, Schade JP: Nerve cell destruction by asphyxiation of the spinal cord. J Neuropathol Exp Neurol 21:410-423,1962 Adams HD, Van Geertruyden HH: Neurologic complication of aortic surgery. Ann Surg 144:574-610, 1956 Miyamoto 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 16:188-199, 1960 7 Blaisdell FW, Cooley DA: The mechanism of paraplegia after temporary thoracic aortic occlusion and its relationship to spinal fluid pressure. Surgery 51:351-355, 1962 8 Crawford ES, Rubio PA: Reappraisal of adjuncts to avoid ischemia in the treatment of aneurysms of descending thoracic aorta. J THORAC CARDIOVASC SURG 66:693-704, 1973 9 Katz NM, Blackstone EH, Kirklin JW, Karp RB: Incremental risks factors for spinal cord injury following operation for acute aortic traumatic transection. J THORAC CARDIOVASC SURG 81:669-674, 1981 10 Bloodwell RD, Hallman GL, Cooky DA: Partial cardiopulmonary bypass for pericardiectomie and resection of descending thoracic aortic aneurysms. Ann Thorac Surg 6:46-56, 1968 11 Kahn DR, Vathayanon S, Sloan H: Resection of descending thoracic aneurysms without left heart bypass. Arch Surg 97:336-340, 1968 12 DeBakey ME, McCollum CH, Graham JM: Surgical treatment of aneurysms of the descending thoracic aorta. Long term results in 500 patients. J Cardiovasc Surg 19:571-576,1978 13 Nie NH, Hull CH, Jenikins JG, Steinbrenner K, Bent DH: SPAA-Statistical Package for the Social Sciences, ed 2, New York, 1975, McGraw-Hili Book Company, Inc. 14 Pontius RG, Brockman HL, Hardy EG, Cooley DA, DeBakey ME: The use of hypothermia in the prevention of paraplegia following temporary aortic occlusion. Experimental observations. Surgery 36:33-38, 1954 15 Cooley DA, DeBakey ME: Resection of the thoracic aorta with replacement by homograft for aneurysms and constrictive lesions. J THORAC SURG 29:66-102, 1955 16 Vasko JS, Spencer FC, Bahnson HT: Aneurysms of the aorta treated by excision. Review of 237 cases followed up to 7 years. Am J Surg 105:793-801, 1963 17 Rubio R: Complicationes neurologicas en la chirurgua de la hipertension arterial. Bol Soc Uruguay 23:439-449, 1952 18 Hughes JT, McIntry AC: Spinal cord infarction occurring during thoraco-Iumbar sympathectomie. J Neurol Neurosurg Psychiatry 26:418-421, 1963 19 Rouques L, Passelecq A: Syndrome de Brown-Sequard apres thoracoplastie. Rev Neurol 97:146-148, 1957 20 Moore DC: Complications following use of efocaine. Surgery 35:109-114,1954 21 Garcin R, Goldlewski S, Rondot P: Etude c1inique des medullopathies d' origine vasculair. Rev Neurol 106:558585, 1962 22 Bartsch W, Hopf HC: Neue Beobachtungen iiber die Beziehung zwischen Herzleistung und Ruckenmarkskreislauf. Dtsch Z Nervenheilk 184:288-307, 1963 23 Ziilch KJ: Die spinale Mangeldurchblutung und ihre Folgen. Verh Dtsch Ges Inn Med 72:1007-1059,1967

I0

The Journal of Thoracic and Cardiovascular Surgery

Wadouh et al.

24 Weingarten K: Cber neurologische Komplikationen nach Aortographie. Wien Z Nervenheilk 20:257-269, 1962 25 Carrel A: Permanent intubation of the thoracic aorta. J Exp Med, 1912, pp 17-24 26 DeBakey ME, Cooley DA, Crawford ES, Morris GC: Aneurysms of thoracic aorta. Analysis of 179 patients treated by resection. J THORAC SURG 36:393-420, 1958 27 Ellis FH, Kirklin JW, Brewer AJ: Surgical experiences in the treatment of aneurysms of the thoracic aorta. Surg Gynecol Obstet 106: 179-192, 1958 28 Borst HG: Neurologische Komplikationen der thorakalen Aortenchirurgie. Thoraxchirurgie 19:408-411, 1971 29 DeBakey ME: Changing concepts in thoracic vascular surgery. J THORAC CARDIOVASC SURG 38:145-165, 1959 30 DeBakey ME, Henly WS, Cooley DA, Crawford ES, Morris GC: Surgical treatment of dissecting aneurysm of the aorta. Analysis of 72 cases. Circulation 24:290-303, 1961 31 Crawford ES, Fenstermacher JM, Richardson W, Sandiford F: Reappraisal of adjunts to avoid ischemia in the treatment of thoracic aortic aneurysms. Surgery 67: 182196, 1970 32 Jaen RJ, Grossmann V, Ortega MA, Bello A: Surgical treatment of the aneurysms of the descending thoracic aorta. J Cardiovasc Surg 12:371-378, 1971 33 Najafi H, Javid H, Hunter J, Serry C, Monson D: Descending aortic aneurysmectomy without adjuncts to avoid ischemis. Ann Thorac Surg 30:326-335, 1980 34 Heberer G: Zur operativen Behandlung traumatischer Rupturen und Aneurysmen der thorakalen Aorta. Langenbecks Arch Klin Chir 301:673-680, 1962 35 Hemmer R: Der Liquordruck, Stuttgart, 1960, GeorgThieme-Verlag, pp 1-12 36 Endler S, Klagge E, Hindersin P, Sedlarick K: Kontinuier-

37

38

39

40

41 42

43

44

45

liche Liquordruckmessung in der neurologischen Diagnostik. Arztliche Fortbildung. A Arztl Fortbild 75:792-795, 1981 Wissdorf H: Die Gefassversorgung der Wirbelsaule und des Riickenmarkes vom Hausschwein, Tierarztliche Hochschule Hannover, Med vet Habilschrift, Berlin, 1970, Patey-Verlag, pp 73-96 Jellinger K: Zur Orthologie und Pathologie der Riickenmarksdurchblutung, Wien, 1966, Springer-Verlag, pp 841, 55-59 Piscol K: Die Blutversorgung des Riickenmarkes und ihre klinische Relevanz, Berlin, 1972, Springer-Verlag, pp 1-77 Kadyi H: Uber die Blutgefasse des menschlichen Riickenmarkes. Nach einer im 14, Bande der Denkschrift der Math Natur Classe tier Akad der Wissenschaften in Krakau, erschienen in Monographie, Lemberg, 1889, Aus dem polnischen iibersetzt vom Verfasser Gubrynowicz and Smith Suh TH, Alexander L: Vascular system of human spinal cord. Arch Neurol Psychiatry 41:659-677, 1939 Bartsch W: Die Durchblutung des Riickenmarkes und ihre klinischen Storungen, Hab Schrift, Wiirzburg 1960, zitiert bei Jellinger (35) und Piscol (36) Noeske K, Roll D: Uber die arterielle Versorgung des menschlichen Riickenmarkes. Cber die Arterien des pars caudalis. Morphol Jb 99:452-497, 1957 Adamkiewicz A: Die Blutgefasse der menschlichen Riickenmarksflache, Sitz Akad Wiss in Wien Math Natur Classe 85:101-130, 1882 Lazorthes G, Zadek 0: La vascularisation arterielle de la moelle epiniere, Etude anatomique et physiologique. Rev Neurol 106:535-557, 1962