J Neurosurg 48:390-401, 1978
Arterial blood gases during raised intracranial pressure in anesthetized cats under controlled ventilation SHEILA JENNETT, M.D., PH.D., AND JULIAN T. HOFF, M.D.
Department of Physiology, University of Glasgow, Glasgow, Scotland, and the Department of Neurological Surgery, San Francisco General Hospital, University of California School of Medicine, San Francisco, California v" In anesthetized paralyzed cats ventilated with air, blood gases were analyzed repeatedly before and during episodes of raised intracranial pressure (ICP). The ICP was raised by infusion via the lumbar subarachnoid or intraventricular route, and increases were maintained for at least 30 minutes. A minor degree of hypoxemia commonly developed, but was always associated with hypercapnia; normoxia was restored by increasing the ventilation sufficiently to restore normocapnia. Relative underventilation is thus liable to develop if the minute volume is maintained constant when ICP is raised, probably because of increased metabolic rate which may be associated with a rise in temperature; there is no evidence to implicate more obscure causes of hypoxemia in this circumstance. Pulmonary hemorrhage and edema were found post mortem in nine of 20 animals, but only two of these had developed greater hypoxemia than could be accounted for by under-ventilation. Phrenic electroneurograms were recorded, and respiratory activity was shown to persist during prolonged periods of cerebral ischemia. KEY WORDS intracranial pressure 9 cerebral ischemia 9
p
ULMONARY h e m o r r h a g e and edema m a y be associated with brain d a m a g e induced clinically or experimentally by a variety of agents; also some degree of arterial hypoxemia is c o m m o n in patients with brain damage. S o m e experimental evidence suggests that not only edema 18,22 but also v a r y i n g degrees of v e n t i l a t i o n / p e r f u s i o n ( g A / Q ) imbalance 2 m a y result from a direct neural effect on the lungs. Both factors could contribute to hypoxemia. Raised intracranial pressure ( I C P ) commonly accompanies brain damage. Some investigators have linked experimentally maintained increases in I C P in dogs with progressive arterial hypoxemia or intra390
9 hypoxia
9 pulmonary edema
pulmonary shunt, 2,13,17 but other workers have not confirmed this relationship. TM A previous study in spontaneously breathing cats, in which the I C P was increased for 10minute periods, indicated that hypoxemia developed only to an extent accounted for by ventilatory depression? 2 It was suggested to us that the development of hypoxemia might require m o r e prolonged exposure to raised ICP, although a neurally-mediated effect on the lungs could be expected to develop as rapidly as, and concurrently with, the effect on systemic circulation. Accordingly, in the experiments described here increases in I C P were maintained for at least 30 minutes to establish whether or not
J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure hypoxemia developed during a longer period when ventilation was controlled with air as the inspired gas. In order to estimate brainstem respiratory function in the absence of spontaneous breathing, phrenic nerve activity was recorded. Materials and Methods
Two series of experiments were conducted separately, the first by both authors, the second by J.H. alone. Animal Groups
minute by Rotameters, and being within 5% of linearity at the extremes of this range; instantaneous heart rate (HR), with the BP transducer output to trigger a ratemeter via an oscilloscope.* For the phrenic electroneurogram, the right phrenic nerve was exposed in the neck and divided; the distal end was stripped and placed on electrodes under paraffin for audiomonitoring and to record integrated electrical activity by means of an AC integrator assembled for the purpose (amplifier, integrator, and a low-pass filter). The peak output obtained in each burst of inspiratory impulses measured by this instrumentation correlates well with tidal volume during spontaneous breathing. In addition to these continuous records, arterial blood was sampled intermittently, and immediately analyzed for pH, pCO2, and pO2 on the Radiometer system, calibrated by gas mixtures from a Wosthoff pump. Measurement temperature was 37 ~ C, and the values were corrected to the animal's rectal temperature at the time of each sample, using the Severinghaus blood gas calculator, t Rectal temperature was recorded at intervals of 15 minutes throughout, and the heated operating table was adjusted to correct deviations from normal. Mock CSF ~9at 37 ~ C was infused to raise ICP; infusion was controlled
Series 1. Ten cats, weighing 2 to 4 kg, were anesthetized with halothane and nitrous oxide, and anesthesia was maintained with intravenous pentobarbitone (30 mg/kg, with later supplementary doses of 5 to 10 mg/kg hourly). For the period of controlled ventilation, the cats were paralyzed with 2 ml gallamine (Flaxedil), repeated after about 2 hours as necessary. Air was inspired throughout. After tracheal and femoral arterial and venous cannulation, a lumbar laminectomy was performed and a cannula passed up alongside the spinal cord intrathecally for several centimeters. It was secured and sealed in place by an extradural circumferential ligature. A parietal burr hole was made, and a cannula with multiple punctures at the tip was inserted into the subdural plane. The skull *UV chart recorder (S.E. 2005) manufactured was sealed with dental acrylic. This cannula by SE Laboratory (EMI) Ltd., North Feltham was attached to a pressure transducer and the Trading Estate, Feltham, Essex, England. output checked for normal pulsations; the Photoelectric transducers manufactured by Merpresence of ~hese, and an immediate rise of cury Electronics (Scotland) Ltd., Pollock Castle ICP in response to a small infusion, allowed Estate, Newton Mearns, Glasgow, Scotland. repeated confirmation of free communica- Rapid infrared analyzer (URAS CO2 analyzer) manufactured by Godart/Statham, Jan van tion and patency, indicating valid ICP Eycklaan 2, Bilthoven, Holland. Lloyd Haldane measurements. apparatus manufactured by Gallenkamp & Co., The following continuous measurements PO Box 290, Technico House, Christopher Street, were established on an ultraviolet chart London, EC2P 2ER, England. Computing recorder: pressures from the intracranial can- Spirometer CS1 manufactured by Mercury Elecnula (ICP) and from the femoral artery (BP) tronics (Scotland) Ltd., Pollock Castle Estate, measured by photoelectric transducers, Newton Mearns, Glasgow, Scotland. Rotameters calibrated in parallel against a mercury manufactured by GEC-Elliott Process Instrumanometer; %CO2 in inspired-expired air ments Ltd., Rotameter Works, 330 Purley Way, from the tracheal cannula, for calculation of Croydon, Surrey, England. Ratemeter manufactured by Devices Instruments Ltd., Welwyn end-tidal pCO2 by rapid infra-red analyzer Garden City, London, England. CRO oscillocalibrated by gas mixtures analyzed on the scope manufactured by Tektronix, Guernsey, Ltd., Lloyd Haldane apparatus; integrated inspired Guernsey, Channel Islands, England. volume (VI) by pneumotachograph with a tSeveringhaus blood gas calculator manufacflow head attached to the tracheal cannula tured by Radiometer A/S 72 Emdrupvej, DK calibrated for flow rates of 2 to 6 liters per 2400, Copenhagen, Denmark. J. Neurosurg. / Volume 48 / March, 1978
391
S. Jennett and J. T. Hoff
Fro. 1. Example of changes in one animal in Series 1, during the sequence of alterations in intracranial pressure (ICP). Open symbols = PaO2: closed symbols = PaCO2. Between the vertical lines, ICP (shaded) was held at 100, 150, and 200 mm Hg, respectively. Broken line = mean arterial blood pressure (BP). Mean BP and ICP are shown graphically as a constant average: the values below show the range of fluctuations of cerebral perfusion pressure (CPP = B P - I C P ) during each run. At each arrow, the ventilation was increased so as to restore the PaCO2 as nearly ~is possible to the value before the ICP was raised. During the run at 200 mm Hg, BP fell below ICP and the experiment was terminated. manually during pressure increases and a rate of rise of 1 m m / s e c was achieved by reference to the display of the I C P transducer on the recording apparatus; thereafter the raised pressure was maintained by means of a pressurized W o u l f f bottle.:~ Series 2. Ten cats, weighing 2.5 to 4 kg, were anesthetized and monitored according to the method described for the animals in Series 1. M o c k C S F was infused not into the lumbar subarachnoid space, but into the right lateral ventricle through a No. 20 needle placed stereotaxically and secured with dental acrylic. Phrenic activity was not recorded.
Procedure All recording systems were established and the anesthetic level was stabilized with the animal breathing spontaneously, so that its arterial blood gases were within the normal range for the cat. 7,~ The animal was then paralyzed and the level of mechanical ventila-
:~Pressurized Woulff bottle manufactured by Gallenkamp & Co., PO Box 290, Technico House, Christopher Street, London EC2P 2ER, England. 392
tion adjusted until the end-expired pCO2 was close to the previous level during spontaneous breathing. When the end-tidal pCO~ and ventilation had remained steady for at least 5 minutes, a baseline arterial sample was analyzed; intracranial pressure was then raised.
Pattern o f lntracranial Pressure Increase Series 1. A sequence of I C P elevations was imposed starting at 80 to 100 mm Hg. This level was chosen because it had been shown earlier to be associated with effects on respiration and on arterial blood pressure, 1~ and with increased sympathetic activity. 1~ Raised I C P was maintained for 30 minutes; pressure was then lowered and all recorded variables were allowed to return to normal; I C P was then raised to about 150 m m Hg, and again m a i n t a i n e d for 30 minutes; recovery was allowed once more, after which I C P was raised to 200 m m Hg, and held there for as long as the arterial BP remained elevated. I f BP began to decrease, indicating imminent circulatory failure, C S F infusion was stopped, I C P was allowed to fall, and the animal was immediately killed by an overdose of barbiturate. J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure Series 2. These animals were subjected to a single increase of I C P of 150 m m Hg lasting 30 to 60 minutes (six cats) or of 200 m m Hg lasting 30 minutes or as long as the animal tolerated it (four cats). A 30-minute recovery period followed and the animal was then killed by an overdose of barbiturate. The volume of m o c k C S F required to maintain I C P elevations never exceeded 40 c c / k g / h r , an amount shown in other studies to have no effect on arterial blood gases. 11 Ventilation and Blood Gas M e a s u r e m e n t s
Cumulative inspired volume and tidal CO2 were recorded continuously. Arterial blood gases were measured immediately before each increase of I C P (in a period of stable ventilation and end-expired pCO2) and at 10-minute intervals during raised pressure. When the period at 200 m m Hg was less than 10 minutes, a final sample was taken immediately before lowering ICP. During the first minutes of raised ICP, the end-expired pCO2 usually increased despite constant ventilation. In Series 1 experiments, the p u m p - s t r o k e volume was increased sufficiently to restore the end-expired pCO2 to its previous value before the final blood sample was taken at any given ICP. In Series 2, ventilation was not adjusted when pCO2 increased. The lungs were examined macroscopically post mortem. Results
Arterial Blood Gases and End-Expired pCO~
When I C P was raised, a small increase in the end-expired pCO2 usually began within the first minute and was progressive, while the recorded inspired minute volume was unchanged. The arterial blood sample taken at 10 minutes showed a parallel small rise in the arterial CO2 and a small decrease in PaO2 c o m p a r e d with the baseline measurement (Fig. 1). The final sample at raised pressure showed that the increased ventilation which corrected the hypercapnia had, in most instances, also corrected the hypoxemia. This was true for seven of nine animals at 150 m m Hg I C P (Fig. 2), and suggested that the changes had been due to relative underventilation. In three of the seven, there was a small and i n c o n s i s t e n t rise in b o d y J. Neurosurg. / Volume 48 / March, 1978
TABLE 1 Mean valuesfor blood gases before raising intracranial pressure (ICP) and after 30 minutes at 150 mm Itg ICP
Variables PaO2 (torr) initial final change PaCO2 initial final change
Series 1 Ventilation Adjusted
Series 2 Ventilation Not Adjusted
(mean a= SE)
(mean =~ SE)
88.9 • 2.6* 87.4 :~ 2.7 --1.5 • 2.6t
99.7 a= 5.1" 85.3 a= 5.8 --14.7 =~ 2.6t
32.6 • 0.9 30.2 ~= 0.8 --2.4 • 1.2:[:
35.3 • 1.1 40.9 • 1.4 5.6 • 1.1~;
n =9
n =7
*The difference in initial PaO.2 between the two groups is not statistically significant: it suggests, however, a better state of pulmonary gas exchange in "normal" cats in San Francisco than in Glasgow. tDifference between series: p < 0.001. ~Difference between series: p < 0.01.
temperature during raised I C P ; in the other four animals, t e m p e r a t u r e was steady. The two exceptions showed hemorrhagic edematous lungs post m o r t e m ; in one, the hypoxemia progressed until the end of the experiment but the other regained n o r m o x i a before and during the subsequent period at an I C P of 200 m m Hg. Figure 3 contrasts these results for Series 1 with those for animals at 150 m m H g I C P in Series 2 that were exposed to the s a m e I C P for the same period, but without the adjustment to normocapnia. The latter group showed a persistent, usually progressive, minor degree of hypercapnia and hypoxia, which resolved after release of pressure; the most m a r k e d changes occurred in three animals which developed a rise in body temperature. The mean changes in PaO2 f r o m baseline to 30 minutes at I C P 150 m m H g were significantly different between the two groups; there was no significant m e a n change in either PaCO2 or PaO2 f r o m the baseline measurements to that at 30 minutes of I C P 150 m m H g in Series 1 where ventilation was adjusted (Table 1). The graphic data refer only to periods at I C P 150 m m Hg because full data were available for every animal after 30 minutes at 393
S. Jennett and J. T. Hoff I
30 rnin
I
+
I00 OaO 2
90
I I I
80 7O
torr
PoCO 2
4O 3O
,I
I
~o./" i~ o
OI
2O
rlCP
100
Pa 0 2
J
O-,
90
I
150mmHg
+
i
+
o--~~176
(
I
88 I
I
70
f I I
/
Pa CO 2
I
e~.-- ef -'~11e
O--
I
I
I I
I
I
80
3
i
+
f I
I
+
+
I
o..l__o~" +.,~.o I
(E)
I I I
a
(E)
(E)
FIG. 2. Changes in PaO2 and PaCO2 for each animal of Series 1, during the period at ICP 150 mm Hg. Ventilation increased at the arrow. Same data as Fig. 3 left. The two animals represented below right, were those in which normoxia was not restored at normocapnia. These and one other showed frank pulmonary edema (E).
this pressure. At 100 m m Hg I C P , similar changes occurred but were less consistent and tended to be smaller; at 200 m m Hg, some animals tolerated only shorter periods, and final blood gas samples were complicated by circulatory failure. In Series 1, the animals, which at I C P 150 m m H g were restored to n o r m o x i a by increasing the ventilation to r e s t o r e n o r m o c a p n i a , showed a similar
TABLE 2 Values for intracranial pressure (ICP), mean arterial blood pressure ( MBP), and cerebral perJJtsion pressure (CPP) in Series 1
Run 1 2 3 394
ICP MBP CPP DuraRange Range Mean Range tion (mm Hg)(mm Hg)(mm Hg)(mm Hg) (rain) 80-100 120-170 140-160 155-210 180-220 190-220
45 38 10
25-80 10-70 0-25
30 30 3-25
change at I C P 200 m m Hg, which was likewise corrected. Two animals in Series 2 showed the typical rise in pCO2 and fall in pO2 at 200 m m H g but two others had no significant change in blood gases; these latter showed a fall in body temperature during the period of raised I C P . A r t e r i a l B l o o d P r e s s u r e a n d Cerebral Perfusion P r e s s u r e
In all instances arterial blood pressure increased p r o m p t l y as I C P was raised (Figs. 4, 5, 7), and remained steadily elevated. Extent of increase varied f r o m animal to animal, so that any one level of raised I C P represented a different cerebral perfusion pressure (CPP = mean BP - m e a n I C P ) in each animal (Table 2). In m a n y instances, a C P P of only 10 to 25 m m Hg was maintained for 30 minutes without a sign of deterioration in cardiovascular function (Fig. 5). This is considerably less than the value for C P P below which cerebral vasodilatation can no longer mainJ. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure
2~I
Series 1
Series 2
15 I0
ZX,oCO2
5 A PaO 2 torr
O~ -5 -10 -15 -20
o
I
after 30rain at ICP 150mmHg
after 30rain at ICP 150mmHg
FIG. 3. Changes in blood gases over 30 minutes at an ICP of 150 mm Hg, compared between Series 1 animals in which ventilation was adjusted to correct the PaCO2 (left), and Series 2 animals in which ventilation was kept constant. Closed circles and broken lines = PaCO~; open circles and unbroken lines = PaO2. On left, the broken lines show range of PaCO2 changes, which were on average nonsignificant. Mean APaO2 and APaCO2 were significantly different between the two series (Table 1).
tain constant blood flOW; 14'24 therefore the cerebral oxygen supply must have been severely diminished. At the higher I C P levels (200 and sometimes 150 m m Hg), BP rose dramatically (to a mean as high as 180 to 220 m m Hg) and showed wide fluctuations (Fig. 6); these were reflected in the I C P , which consequently could not be stabilized. Animals survived for varying periods with C P P near zero (Table 3). Respiratory A c t i v i t y
The phrenic electroneurogram was recorded successfully in seven Series 1 animals and was of interest in the following three respects: 1. During the rise in I C P , with mechanical ventilation of the lungs maintained constant, inspiratory activity in m a n y instances increased (Fig. 7), a p h e n o m e n o n which confirms the earlier finding that respiration is stimulated in s p o n t a n e o u s l y breathing animals in otherwise similar conditions. TM J. Neurosurg. / Volume 48 / March, 1978
2. As C P P decreased, there were various irregularities in phrenic activity, including a Cheyne-Stokes pattern. Such periodicity in the spontaneously breathing animal might be thought to be related to fluctuating blood gas levels but in this instance PaO2 and PaCO~ were constant and normal. The relationship
TABLE 3 Phrenic nerve discharge during reduced cerebral perfusion pressure ( CPP) Cat No. 2 4 5 6 7 9 10
Discharge Present at CPP (mm Hg)
after (min)
20 25-35 20-25 25-30 0 5-20 0
10 20 30 30 10 30 30 395
S. Jennett and J. T. Hoff
FIG. 4. Example of recorded variables, during a period at ICP approximately 100 mm Hg. Reading from top down: Airway CO~; heart rate (HR); arterial blood pressure (BP); intracranial pressure (ICP); cerebral perfusion pressure ( C P P ) = 25 mm Hg (difference between mean BP and mean ICP); cumulative minute volume ("staircase" trace - - constant ventilation); cerebral venous pressure (CVP); integrated phrenic electroneurogram, showing "breaths" that are irregular both in size and frequency.
FIG. 5. Same recorded variables as in Fig. 4. An intracranial pressure (ICP) of approximately 150 mm Hg had been maintained in this animal for 20 minutes at the time of this trace. The ICP was falling below the required level and was deliberately increased twice during the period shown (at A and at B); the immediate blood pressure (BP) response can be seen, and also a stimulation of "breathing." 396
J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intraeranial pressure
FIG. 6. Example of low perfusion pressure (varying between zero and 30 mm Hg) maintained for 30 minutes at ICP 150 mm Hg. The blood pressure (BP) and intracranial pressure (ICP) traces overlap and each was separately interrupted at the dotted lines. The phrenic activity shows an irregular pattern of "breaths" with a tendency to periodicity. The peak activity coincides with the lowest perfusion pressure. of respiratory activity to cerebral ischemia was not clear, because in some cases maximal activity appeared to coincide with minimal perfusion pressure, rather than the reverse (Fig. 6). However, when animals were virtually "apneic," the occasional " b r e a t h " occurred only when BP exceeded ICP. 3. Inspiratory activity survived remarkably prolonged low C P P in several instances (Table 3). In such conditions spontaneously breathing animals would be expected to become apneic.
Appearance o f Lungs Series 1. The lungs appeared normal immediately after death in seven of the 10 animals. The remaining three showed gross edema and patchy hemorrhage; two of these had been slightly hypoxemic (PaO2 82, 77 torr, respectively) and the third was normoxic at normocapnia 5 minutes before sacrifice, after 30 minutes with ICP at 200 mm Hg. Series 2. The lungs were grossly normal in four of the six animals which had a sustained J. Neurosurg. / Volume 48 / March, 1978
ICP of 150 mm Hg for 30 minutes; the other two showed gross edema and hemorrhage. All four animals exposed to an ICP of 200 mm Hg for 20 to 30 minutes had gross hemorrhagic change with edema.
Discussion
Pulmonary Edema and Hypoxemia Increased s y m p a t h e t i c outflow with systemic vasoconstriction can shift blood into the pulmonary circulation; this shift, in conjunction with alterations in cardiac function, could indirectly affect oxygenation by altering the rate of fluid extravasation a n d / o r the distribution of ~A/O r a t i o s . 5'6'13'23 In addition, sympathetic stimulation or autonomic imbalance might act directly on the lungs: pulmonary vasoconstriction could increase the capillary pressure; 15'18'22 the ~A/Q distribution might be affected so as to cause an effective increase in shunting and thence hypoxemia; 2,3 surfactant might be depleted, 397
S. Jennett and J. T. Hoff
FIG. 7. Example of increases in phrenic activity and of blood pressure (BP) during rise in intracranial pressure (ICP). The BP starts to rise 20 seconds after the beginning of ICP elevation, when the ICP is 80 to 90 mm Hg. The beginning of an increase in end-tidal pCO2 can be seen, top right.
resulting in decreased complianceI and a tendency to extravasation. 2~ Grossly evident pulmonary edema was found in three animals out of 10 in Series 1; a minor degree of hypoxemia occurred in two of these, and the other was normoxic after ventilation was adjusted. Six animals in Series 2 showed similar changes but in no instance did PaO2 decrease below 70 torr during the 30 minutes after ICP elevation was stopped and before sacrifice. Our results provide no particular evidence for or against a direct neural effect on the lungs causing congestion and edema, as opposed to the alternative explanation in terms of the consequences of systemic effects. Elevation of blood pressure in all animals was severe and prolonged, implying that there must have been widespread systemic vasoconstriction and some shift of blood into the lesser circulation. The late and inconstant development of edema is compatible with an earlier increase in extravasation rate, which would not necessarily exceed the capacity of lymphatic drainage, or if it did, would take 398
time to cause accumulation of fluid to a grossly recognizable extent. It is quite possible that lesser degrees of interstitial edema took place in the other animals in our two series, and could have been detected by appropriate methods for estimating lung water, xl It is known that considerable increases in interstitial fluid can occur without impairing diffusion,8,21 so the absence of hypoxemia in this series, except in association with frank lung changes, is understandable. In another series of experiments using a similar preparation for raised ICP, lung water has been measured and a significant increase was usual; again hypoxemia did not occur? ~ The persistent normoxia (after correction of the mild hypercapnia) in the face of prolonged increase in ICP, in arterial BP, and presumably in sympathetic outflow, is evidence against any direct or indirect pulmonary hemodynamic effect causing "v'A/Q maldistribution. Correction of PaO2 might have been achieved in such circumstances but only by hyperventilation to hypocapnia. J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure 60 - -
o
o 40
--
Pa C O 2
~' ~~ , , ~ o
(o'
o
o
o
o
~ " ~ 9 ~, Oo ' ~ - - o , ~
,~o~
~,,~
o o Oo 9
9
9
9
torr
20--
o-.//
I
60
I
80
I
Pa 0 2 torr
100
I
120
FIG. 8. Diagram showing the interrelationship between PaO2 and PaCO2 for cats under pentobarbitone anesthesia. The regression lines refer to an earlier series of spontaneously breathing animals; ~2the solid line relates PaO2 to PaCO2 in different animals at different levels of anesthesia, at normal intracranial pressure (ICP); the broken line relates values in the same group at raised ICP, showing varying degrees of ventilatory depression. Blood gas data from the present experiments on ventilated animals have been superimposed. Open symbols represent estimations at raised ICP, triangles with, and circles without adjustment of ventilation; and closed symbols, at normal ICP.
The existing evidence in favor of such an effect derives from experiments in dogs breathing 100% oxygen. Berman and Ducker 2 reported an increase in alveolar-arterial pO2 difference during raised ICP, and Maxwell and Goodwin is showed an increase in the calculated percentage shunt, which was related to the severity of applied ICP. The magnitude of the shunt increase in the latter study (7% to 11%) was such that, if an animal were breathing air, the PaO2 could decrease from, say, 100 to 75 torr. Some cats in our present series showed a comparable decrease in arterial pO2. Had this hypoxemia been due to shunt, it would not have been corrected by ventilating to normocapnia. Thus our results do not require the hypothesis that raised ICP, by virtue of increasing sympathetic outflow, induces shunting in the lungs. The reason for the difference between these and other findings is not clear. J. Neurosurg. / Volume 48 / March, 1978
Increased Metabolic Rate During Raised Intracranial Pressure The mild hypoxemia and hypercapnia that occurred with unchanged minute volume, suggested an increase in metabolic activity and therefore relative hypoventilation. In some animals this could be accounted for by an increase in body temperature, but in others in which temperature was better controlled the phenomenon still occurred. This would be compatible with an increase in catecholamine release. 4 A possible alternative explanation of the rise in pCO2 might have been that effective alveolar ventilation was reduced because of an increase in alveolar dead space; this would have caused a widening of the end-expired arterial difference for pCO2, which did not occur. Rise in body temperature could increase the ventilation requirement, and it is perhaps 399
S. Jennett and J. T. Hoff significant that those animals that developed the greatest degree of hypoxia and hypercapnia were those in which the temperatures increased noticeably during the relevant period. Conversely, two animals which showed no change in blood gases during a period at 200 mm Hg ICP had a fall in body temperature at that time. Data from previous experiments in spontaneously breathing cats 12 plotted on an O2-CO2 diagram, served as the basis for predicting the likely change in PaO~ relative to the change in PaCO2, when ventilation was depressed either by raised ICP or by incremental barbiturates; in that study, the APaCO2/APaO2 relationship was similar in the two circumstances, making it unnecessary to postulate a specific effect of raised ICP on the lungs. TM In the present series, the data fitted this same relationship (Fig. 8).
with a rise in body temperature. We conclude that, in this preparation, there are no grounds for implicating a specific effect of raised int r a c r a n i a l pressure on the p u l m o n a r y vascular bed that leads to hypoxemia. A proportion of animals develop pulmonary edema and hemorrhage, but even this is not necessarily associated with arterial hypoxemia, despite being sufficient to be recognizable macroscopically. Unless there is a species difference, therefore, raised ICP per se is probably not the factor in clinical brain damage responsible for hypoxemia until or unless it leads to frank pulmonary edema.
Respiratory Activity During Severely Decreased Cerebral Perfusion Pressure
1. Beckman DL, Bean JW, Baslock DR: Sympathetic influence on lung compliance and surface forces in head injury. J Appi Physiol 30:394-399, 1971 2. Berman IR, Ducker TB: Pulmonary, somatic and splanchnic circulatory responses to increased intracranial pressure. Ann Surg 169: 210-216, 1969 3. Brackett CE: Respiratory complications of head injury, in International Symposium on Head Injuries. Edinburgh/London: Churchill Livingstone, 1971, pp 255-265 4. Brashear RE, Ross JC: Circulating beta adrenergic stimulator during elevated cerebrospinal fluid pressure. Arch Intern Med 127:748-753, 1971 5. Chen HI, Chai CY: Pulmonary edema and hemorrhage as a consequence of systemic vasoconstriction. Am J Physiol 227:144-15 l, 1974 6. Chen HI, Sun SC, Chai CY: Pulmonary edema and haemorrhage resulting from cerebral compression. Am J Physiol 224: 223-229, 1973 7. Dejours P, Lacaisse A: Arterial blood pH and partial pressures of 02 and CO2 in normal awake cats. J Physioi 68:87-90, 1971 8. Fishman AP" Pulmonary edema. The water exchanging function of the lung. Circulation 46:390-408, 1972 9. Herbert DA, Mitchell RA: Blood gas tensions and acid-base balance in awake cats. J Appl Physiol 30:434-436, 1971 10. Hoff JT, Mitchell RA: The effect of hypoxia on the Cushing response, in Brock M, Dietz H (eds): Intraeranial Pressure. Berlin/Heidelberg/New York: Springer-Verlag, 1972, pp 205-209
The persistence of inspiratory impulse activity during prolonged increases in ICP, when the spontaneously breathing animal would almost certainly have been apneic, ~2 requires explanation. H y p e r c a p n i a and hypoxemia develop when respiration begins to diminish in the breathing animal. A decrease in PaO2 in the presence of cerebral ischemia will further threaten the activity or respiratory neurons. Thus, a positive feedback mechanism is operative and tends to worsen the situation. In the artificially ventilated animals, however, arterial blood gases were normal. It remains remarkable that regular phrenic activity persisted for so long in some instances. Conclusions
Anesthetized cats, breathing air, commonly develop a slight increase in PaC02 and a concomitant decrease in PaO~ during periods of severely increased intracranial pressure, if the ventilation is kept constant at the level that maintained normocapnia at normal ICP. However, if the ventilation is adjusted to correct the pCO2, the pO2 is also c o r r e c t e d , implying t h a t the alteration resulted from inadequate alveolar ventilation. This could be due to increased metabolic rate related to increased sympathetic activity, and is sometimes, but not necessarily, associated 400
Acknowledgments
We are grateful for the technical assistance of G. Burnside, Glasgow, and M. Nishimura, San Francisco. References
J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure 11. Hoff J, Nishimura M, Pitts L: A quantitative method for neurogenic pulmonary edema in cats. Chicago Conference on Neural Trauma (In press) 12. Jennett S, North JB: Effect of intermittently raised intracranial pressure on breathing pattern, ventilatory response to CO2, and blood gases in anesthetized cats. J Neurosurg 44:156-167, 1976 13. Maxwell JA, Goodwin JW: Neurogenic pulmonary shunting. J Trauma 13:368-373, 1973 14. Miller JD, Stanek AE, Langfitt TW: A comparison of autoregulation to changes in intracranial and arterial pressure in the same preparation, in Fieschi C (ed): Cerebral Blood Flow and Intracranial Pressure. Basel: S Karger, 1972, pp 34-38 15. Moss G, Staunton C, Stein AA: The centrineurogenic etiology of the acute respiratory distress syndromes. Universal, species-independent phenomenon. Am J Surg 126: 37-41, 1973 16. Moss G, Staunton C, Stein AA: Cerebral etiology of the "shock lung syndrome." J Trauma 12:885-890, 1972 17. Moss IR, Lisbon A, Levine JF, et al: The effects of increased intracranial pressure on respiratory functions, in Lundberg N, Pont6n U, Brock M (eds): Intracranial Pressure II. Berlin/Heidelberg/New York: SpringerVerlag, 1975, pp 315-318 18. Pitts LH, Severinghaus JW, Mitchell RA, et al: The role of increased intracranial pressure in the production of neurogenic pulmonary edema, in Lundberg N, Pont6n U, Brock M
J. Neurosurg. / Volume 48 / March, 1978
19.
20. 21. 22.
23. 24.
(eds): Intracranial Pressure II. Berlin/ Heidelberg/New York: Springer-Verlag, 1975, pp 319-323 Schlaefke ME, See WR, Loeschcke HH: Ventilatory response to alterations of H + ion concentration in small areas of the ventral medullary surface. Respir Physiol 10:198-212, 1970 Staub NC: Effects of alveolar surface tension on the pulmonary vascular bed. Jpn Heart J 7:386-399, 1966 Staub NC: Pulmonary edema. Physiol Rev 54:678-811, 1974 Sugg WL, Craver WD, Webb WR, et al: Pressure changes in the dog lung secondary to hemorrhagic shock: protective effect of pulmonary reimplantation. Ann Surg 169:592-598, 1969 Theodore J, Robin ED: Pathogenesis of neurogenic pulmenary oedema. Lancet 2: 749-751, 1975 Zwetnow N: CBF autoregulation to blood pressure and intracranial pressure variations. Seand J Clin Lab Invest Suppl 102:V:A, 1968
This work was supported in part by a grant from the Scottish Hospital Endowments Research Trust and Trauma Center Program (Project Grant GM 18470). Dr. Hoff is recipient of Teacher-Investigator Award NS 11051 NSRB. Address reprint requests to: Sheila Jennett, M.D., Ph.D., Department of Physiology, University of Glasgow, Glasgow G12 8QQ, Scotland.
401
Arterial blood gases during raised intracranial pressure in anesthetized cats under controlled ventilation SHEILA JENNETT, M.D., PH.D., AND JULIAN T. HOFF, M.D.
Department of Physiology, University of Glasgow, Glasgow, Scotland, and the Department of Neurological Surgery, San Francisco General Hospital, University of California School of Medicine, San Francisco, California v" In anesthetized paralyzed cats ventilated with air, blood gases were analyzed repeatedly before and during episodes of raised intracranial pressure (ICP). The ICP was raised by infusion via the lumbar subarachnoid or intraventricular route, and increases were maintained for at least 30 minutes. A minor degree of hypoxemia commonly developed, but was always associated with hypercapnia; normoxia was restored by increasing the ventilation sufficiently to restore normocapnia. Relative underventilation is thus liable to develop if the minute volume is maintained constant when ICP is raised, probably because of increased metabolic rate which may be associated with a rise in temperature; there is no evidence to implicate more obscure causes of hypoxemia in this circumstance. Pulmonary hemorrhage and edema were found post mortem in nine of 20 animals, but only two of these had developed greater hypoxemia than could be accounted for by under-ventilation. Phrenic electroneurograms were recorded, and respiratory activity was shown to persist during prolonged periods of cerebral ischemia. KEY WORDS intracranial pressure 9 cerebral ischemia 9
p
ULMONARY h e m o r r h a g e and edema m a y be associated with brain d a m a g e induced clinically or experimentally by a variety of agents; also some degree of arterial hypoxemia is c o m m o n in patients with brain damage. S o m e experimental evidence suggests that not only edema 18,22 but also v a r y i n g degrees of v e n t i l a t i o n / p e r f u s i o n ( g A / Q ) imbalance 2 m a y result from a direct neural effect on the lungs. Both factors could contribute to hypoxemia. Raised intracranial pressure ( I C P ) commonly accompanies brain damage. Some investigators have linked experimentally maintained increases in I C P in dogs with progressive arterial hypoxemia or intra390
9 hypoxia
9 pulmonary edema
pulmonary shunt, 2,13,17 but other workers have not confirmed this relationship. TM A previous study in spontaneously breathing cats, in which the I C P was increased for 10minute periods, indicated that hypoxemia developed only to an extent accounted for by ventilatory depression? 2 It was suggested to us that the development of hypoxemia might require m o r e prolonged exposure to raised ICP, although a neurally-mediated effect on the lungs could be expected to develop as rapidly as, and concurrently with, the effect on systemic circulation. Accordingly, in the experiments described here increases in I C P were maintained for at least 30 minutes to establish whether or not
J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure hypoxemia developed during a longer period when ventilation was controlled with air as the inspired gas. In order to estimate brainstem respiratory function in the absence of spontaneous breathing, phrenic nerve activity was recorded. Materials and Methods
Two series of experiments were conducted separately, the first by both authors, the second by J.H. alone. Animal Groups
minute by Rotameters, and being within 5% of linearity at the extremes of this range; instantaneous heart rate (HR), with the BP transducer output to trigger a ratemeter via an oscilloscope.* For the phrenic electroneurogram, the right phrenic nerve was exposed in the neck and divided; the distal end was stripped and placed on electrodes under paraffin for audiomonitoring and to record integrated electrical activity by means of an AC integrator assembled for the purpose (amplifier, integrator, and a low-pass filter). The peak output obtained in each burst of inspiratory impulses measured by this instrumentation correlates well with tidal volume during spontaneous breathing. In addition to these continuous records, arterial blood was sampled intermittently, and immediately analyzed for pH, pCO2, and pO2 on the Radiometer system, calibrated by gas mixtures from a Wosthoff pump. Measurement temperature was 37 ~ C, and the values were corrected to the animal's rectal temperature at the time of each sample, using the Severinghaus blood gas calculator, t Rectal temperature was recorded at intervals of 15 minutes throughout, and the heated operating table was adjusted to correct deviations from normal. Mock CSF ~9at 37 ~ C was infused to raise ICP; infusion was controlled
Series 1. Ten cats, weighing 2 to 4 kg, were anesthetized with halothane and nitrous oxide, and anesthesia was maintained with intravenous pentobarbitone (30 mg/kg, with later supplementary doses of 5 to 10 mg/kg hourly). For the period of controlled ventilation, the cats were paralyzed with 2 ml gallamine (Flaxedil), repeated after about 2 hours as necessary. Air was inspired throughout. After tracheal and femoral arterial and venous cannulation, a lumbar laminectomy was performed and a cannula passed up alongside the spinal cord intrathecally for several centimeters. It was secured and sealed in place by an extradural circumferential ligature. A parietal burr hole was made, and a cannula with multiple punctures at the tip was inserted into the subdural plane. The skull *UV chart recorder (S.E. 2005) manufactured was sealed with dental acrylic. This cannula by SE Laboratory (EMI) Ltd., North Feltham was attached to a pressure transducer and the Trading Estate, Feltham, Essex, England. output checked for normal pulsations; the Photoelectric transducers manufactured by Merpresence of ~hese, and an immediate rise of cury Electronics (Scotland) Ltd., Pollock Castle ICP in response to a small infusion, allowed Estate, Newton Mearns, Glasgow, Scotland. repeated confirmation of free communica- Rapid infrared analyzer (URAS CO2 analyzer) manufactured by Godart/Statham, Jan van tion and patency, indicating valid ICP Eycklaan 2, Bilthoven, Holland. Lloyd Haldane measurements. apparatus manufactured by Gallenkamp & Co., The following continuous measurements PO Box 290, Technico House, Christopher Street, were established on an ultraviolet chart London, EC2P 2ER, England. Computing recorder: pressures from the intracranial can- Spirometer CS1 manufactured by Mercury Elecnula (ICP) and from the femoral artery (BP) tronics (Scotland) Ltd., Pollock Castle Estate, measured by photoelectric transducers, Newton Mearns, Glasgow, Scotland. Rotameters calibrated in parallel against a mercury manufactured by GEC-Elliott Process Instrumanometer; %CO2 in inspired-expired air ments Ltd., Rotameter Works, 330 Purley Way, from the tracheal cannula, for calculation of Croydon, Surrey, England. Ratemeter manufactured by Devices Instruments Ltd., Welwyn end-tidal pCO2 by rapid infra-red analyzer Garden City, London, England. CRO oscillocalibrated by gas mixtures analyzed on the scope manufactured by Tektronix, Guernsey, Ltd., Lloyd Haldane apparatus; integrated inspired Guernsey, Channel Islands, England. volume (VI) by pneumotachograph with a tSeveringhaus blood gas calculator manufacflow head attached to the tracheal cannula tured by Radiometer A/S 72 Emdrupvej, DK calibrated for flow rates of 2 to 6 liters per 2400, Copenhagen, Denmark. J. Neurosurg. / Volume 48 / March, 1978
391
S. Jennett and J. T. Hoff
Fro. 1. Example of changes in one animal in Series 1, during the sequence of alterations in intracranial pressure (ICP). Open symbols = PaO2: closed symbols = PaCO2. Between the vertical lines, ICP (shaded) was held at 100, 150, and 200 mm Hg, respectively. Broken line = mean arterial blood pressure (BP). Mean BP and ICP are shown graphically as a constant average: the values below show the range of fluctuations of cerebral perfusion pressure (CPP = B P - I C P ) during each run. At each arrow, the ventilation was increased so as to restore the PaCO2 as nearly ~is possible to the value before the ICP was raised. During the run at 200 mm Hg, BP fell below ICP and the experiment was terminated. manually during pressure increases and a rate of rise of 1 m m / s e c was achieved by reference to the display of the I C P transducer on the recording apparatus; thereafter the raised pressure was maintained by means of a pressurized W o u l f f bottle.:~ Series 2. Ten cats, weighing 2.5 to 4 kg, were anesthetized and monitored according to the method described for the animals in Series 1. M o c k C S F was infused not into the lumbar subarachnoid space, but into the right lateral ventricle through a No. 20 needle placed stereotaxically and secured with dental acrylic. Phrenic activity was not recorded.
Procedure All recording systems were established and the anesthetic level was stabilized with the animal breathing spontaneously, so that its arterial blood gases were within the normal range for the cat. 7,~ The animal was then paralyzed and the level of mechanical ventila-
:~Pressurized Woulff bottle manufactured by Gallenkamp & Co., PO Box 290, Technico House, Christopher Street, London EC2P 2ER, England. 392
tion adjusted until the end-expired pCO2 was close to the previous level during spontaneous breathing. When the end-tidal pCO~ and ventilation had remained steady for at least 5 minutes, a baseline arterial sample was analyzed; intracranial pressure was then raised.
Pattern o f lntracranial Pressure Increase Series 1. A sequence of I C P elevations was imposed starting at 80 to 100 mm Hg. This level was chosen because it had been shown earlier to be associated with effects on respiration and on arterial blood pressure, 1~ and with increased sympathetic activity. 1~ Raised I C P was maintained for 30 minutes; pressure was then lowered and all recorded variables were allowed to return to normal; I C P was then raised to about 150 m m Hg, and again m a i n t a i n e d for 30 minutes; recovery was allowed once more, after which I C P was raised to 200 m m Hg, and held there for as long as the arterial BP remained elevated. I f BP began to decrease, indicating imminent circulatory failure, C S F infusion was stopped, I C P was allowed to fall, and the animal was immediately killed by an overdose of barbiturate. J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure Series 2. These animals were subjected to a single increase of I C P of 150 m m Hg lasting 30 to 60 minutes (six cats) or of 200 m m Hg lasting 30 minutes or as long as the animal tolerated it (four cats). A 30-minute recovery period followed and the animal was then killed by an overdose of barbiturate. The volume of m o c k C S F required to maintain I C P elevations never exceeded 40 c c / k g / h r , an amount shown in other studies to have no effect on arterial blood gases. 11 Ventilation and Blood Gas M e a s u r e m e n t s
Cumulative inspired volume and tidal CO2 were recorded continuously. Arterial blood gases were measured immediately before each increase of I C P (in a period of stable ventilation and end-expired pCO2) and at 10-minute intervals during raised pressure. When the period at 200 m m Hg was less than 10 minutes, a final sample was taken immediately before lowering ICP. During the first minutes of raised ICP, the end-expired pCO2 usually increased despite constant ventilation. In Series 1 experiments, the p u m p - s t r o k e volume was increased sufficiently to restore the end-expired pCO2 to its previous value before the final blood sample was taken at any given ICP. In Series 2, ventilation was not adjusted when pCO2 increased. The lungs were examined macroscopically post mortem. Results
Arterial Blood Gases and End-Expired pCO~
When I C P was raised, a small increase in the end-expired pCO2 usually began within the first minute and was progressive, while the recorded inspired minute volume was unchanged. The arterial blood sample taken at 10 minutes showed a parallel small rise in the arterial CO2 and a small decrease in PaO2 c o m p a r e d with the baseline measurement (Fig. 1). The final sample at raised pressure showed that the increased ventilation which corrected the hypercapnia had, in most instances, also corrected the hypoxemia. This was true for seven of nine animals at 150 m m Hg I C P (Fig. 2), and suggested that the changes had been due to relative underventilation. In three of the seven, there was a small and i n c o n s i s t e n t rise in b o d y J. Neurosurg. / Volume 48 / March, 1978
TABLE 1 Mean valuesfor blood gases before raising intracranial pressure (ICP) and after 30 minutes at 150 mm Itg ICP
Variables PaO2 (torr) initial final change PaCO2 initial final change
Series 1 Ventilation Adjusted
Series 2 Ventilation Not Adjusted
(mean a= SE)
(mean =~ SE)
88.9 • 2.6* 87.4 :~ 2.7 --1.5 • 2.6t
99.7 a= 5.1" 85.3 a= 5.8 --14.7 =~ 2.6t
32.6 • 0.9 30.2 ~= 0.8 --2.4 • 1.2:[:
35.3 • 1.1 40.9 • 1.4 5.6 • 1.1~;
n =9
n =7
*The difference in initial PaO.2 between the two groups is not statistically significant: it suggests, however, a better state of pulmonary gas exchange in "normal" cats in San Francisco than in Glasgow. tDifference between series: p < 0.001. ~Difference between series: p < 0.01.
temperature during raised I C P ; in the other four animals, t e m p e r a t u r e was steady. The two exceptions showed hemorrhagic edematous lungs post m o r t e m ; in one, the hypoxemia progressed until the end of the experiment but the other regained n o r m o x i a before and during the subsequent period at an I C P of 200 m m Hg. Figure 3 contrasts these results for Series 1 with those for animals at 150 m m H g I C P in Series 2 that were exposed to the s a m e I C P for the same period, but without the adjustment to normocapnia. The latter group showed a persistent, usually progressive, minor degree of hypercapnia and hypoxia, which resolved after release of pressure; the most m a r k e d changes occurred in three animals which developed a rise in body temperature. The mean changes in PaO2 f r o m baseline to 30 minutes at I C P 150 m m H g were significantly different between the two groups; there was no significant m e a n change in either PaCO2 or PaO2 f r o m the baseline measurements to that at 30 minutes of I C P 150 m m H g in Series 1 where ventilation was adjusted (Table 1). The graphic data refer only to periods at I C P 150 m m Hg because full data were available for every animal after 30 minutes at 393
S. Jennett and J. T. Hoff I
30 rnin
I
+
I00 OaO 2
90
I I I
80 7O
torr
PoCO 2
4O 3O
,I
I
~o./" i~ o
OI
2O
rlCP
100
Pa 0 2
J
O-,
90
I
150mmHg
+
i
+
o--~~176
(
I
88 I
I
70
f I I
/
Pa CO 2
I
e~.-- ef -'~11e
O--
I
I
I I
I
I
80
3
i
+
f I
I
+
+
I
o..l__o~" +.,~.o I
(E)
I I I
a
(E)
(E)
FIG. 2. Changes in PaO2 and PaCO2 for each animal of Series 1, during the period at ICP 150 mm Hg. Ventilation increased at the arrow. Same data as Fig. 3 left. The two animals represented below right, were those in which normoxia was not restored at normocapnia. These and one other showed frank pulmonary edema (E).
this pressure. At 100 m m Hg I C P , similar changes occurred but were less consistent and tended to be smaller; at 200 m m Hg, some animals tolerated only shorter periods, and final blood gas samples were complicated by circulatory failure. In Series 1, the animals, which at I C P 150 m m H g were restored to n o r m o x i a by increasing the ventilation to r e s t o r e n o r m o c a p n i a , showed a similar
TABLE 2 Values for intracranial pressure (ICP), mean arterial blood pressure ( MBP), and cerebral perJJtsion pressure (CPP) in Series 1
Run 1 2 3 394
ICP MBP CPP DuraRange Range Mean Range tion (mm Hg)(mm Hg)(mm Hg)(mm Hg) (rain) 80-100 120-170 140-160 155-210 180-220 190-220
45 38 10
25-80 10-70 0-25
30 30 3-25
change at I C P 200 m m Hg, which was likewise corrected. Two animals in Series 2 showed the typical rise in pCO2 and fall in pO2 at 200 m m H g but two others had no significant change in blood gases; these latter showed a fall in body temperature during the period of raised I C P . A r t e r i a l B l o o d P r e s s u r e a n d Cerebral Perfusion P r e s s u r e
In all instances arterial blood pressure increased p r o m p t l y as I C P was raised (Figs. 4, 5, 7), and remained steadily elevated. Extent of increase varied f r o m animal to animal, so that any one level of raised I C P represented a different cerebral perfusion pressure (CPP = mean BP - m e a n I C P ) in each animal (Table 2). In m a n y instances, a C P P of only 10 to 25 m m Hg was maintained for 30 minutes without a sign of deterioration in cardiovascular function (Fig. 5). This is considerably less than the value for C P P below which cerebral vasodilatation can no longer mainJ. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure
2~I
Series 1
Series 2
15 I0
ZX,oCO2
5 A PaO 2 torr
O~ -5 -10 -15 -20
o
I
after 30rain at ICP 150mmHg
after 30rain at ICP 150mmHg
FIG. 3. Changes in blood gases over 30 minutes at an ICP of 150 mm Hg, compared between Series 1 animals in which ventilation was adjusted to correct the PaCO2 (left), and Series 2 animals in which ventilation was kept constant. Closed circles and broken lines = PaCO~; open circles and unbroken lines = PaO2. On left, the broken lines show range of PaCO2 changes, which were on average nonsignificant. Mean APaO2 and APaCO2 were significantly different between the two series (Table 1).
tain constant blood flOW; 14'24 therefore the cerebral oxygen supply must have been severely diminished. At the higher I C P levels (200 and sometimes 150 m m Hg), BP rose dramatically (to a mean as high as 180 to 220 m m Hg) and showed wide fluctuations (Fig. 6); these were reflected in the I C P , which consequently could not be stabilized. Animals survived for varying periods with C P P near zero (Table 3). Respiratory A c t i v i t y
The phrenic electroneurogram was recorded successfully in seven Series 1 animals and was of interest in the following three respects: 1. During the rise in I C P , with mechanical ventilation of the lungs maintained constant, inspiratory activity in m a n y instances increased (Fig. 7), a p h e n o m e n o n which confirms the earlier finding that respiration is stimulated in s p o n t a n e o u s l y breathing animals in otherwise similar conditions. TM J. Neurosurg. / Volume 48 / March, 1978
2. As C P P decreased, there were various irregularities in phrenic activity, including a Cheyne-Stokes pattern. Such periodicity in the spontaneously breathing animal might be thought to be related to fluctuating blood gas levels but in this instance PaO2 and PaCO~ were constant and normal. The relationship
TABLE 3 Phrenic nerve discharge during reduced cerebral perfusion pressure ( CPP) Cat No. 2 4 5 6 7 9 10
Discharge Present at CPP (mm Hg)
after (min)
20 25-35 20-25 25-30 0 5-20 0
10 20 30 30 10 30 30 395
S. Jennett and J. T. Hoff
FIG. 4. Example of recorded variables, during a period at ICP approximately 100 mm Hg. Reading from top down: Airway CO~; heart rate (HR); arterial blood pressure (BP); intracranial pressure (ICP); cerebral perfusion pressure ( C P P ) = 25 mm Hg (difference between mean BP and mean ICP); cumulative minute volume ("staircase" trace - - constant ventilation); cerebral venous pressure (CVP); integrated phrenic electroneurogram, showing "breaths" that are irregular both in size and frequency.
FIG. 5. Same recorded variables as in Fig. 4. An intracranial pressure (ICP) of approximately 150 mm Hg had been maintained in this animal for 20 minutes at the time of this trace. The ICP was falling below the required level and was deliberately increased twice during the period shown (at A and at B); the immediate blood pressure (BP) response can be seen, and also a stimulation of "breathing." 396
J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intraeranial pressure
FIG. 6. Example of low perfusion pressure (varying between zero and 30 mm Hg) maintained for 30 minutes at ICP 150 mm Hg. The blood pressure (BP) and intracranial pressure (ICP) traces overlap and each was separately interrupted at the dotted lines. The phrenic activity shows an irregular pattern of "breaths" with a tendency to periodicity. The peak activity coincides with the lowest perfusion pressure. of respiratory activity to cerebral ischemia was not clear, because in some cases maximal activity appeared to coincide with minimal perfusion pressure, rather than the reverse (Fig. 6). However, when animals were virtually "apneic," the occasional " b r e a t h " occurred only when BP exceeded ICP. 3. Inspiratory activity survived remarkably prolonged low C P P in several instances (Table 3). In such conditions spontaneously breathing animals would be expected to become apneic.
Appearance o f Lungs Series 1. The lungs appeared normal immediately after death in seven of the 10 animals. The remaining three showed gross edema and patchy hemorrhage; two of these had been slightly hypoxemic (PaO2 82, 77 torr, respectively) and the third was normoxic at normocapnia 5 minutes before sacrifice, after 30 minutes with ICP at 200 mm Hg. Series 2. The lungs were grossly normal in four of the six animals which had a sustained J. Neurosurg. / Volume 48 / March, 1978
ICP of 150 mm Hg for 30 minutes; the other two showed gross edema and hemorrhage. All four animals exposed to an ICP of 200 mm Hg for 20 to 30 minutes had gross hemorrhagic change with edema.
Discussion
Pulmonary Edema and Hypoxemia Increased s y m p a t h e t i c outflow with systemic vasoconstriction can shift blood into the pulmonary circulation; this shift, in conjunction with alterations in cardiac function, could indirectly affect oxygenation by altering the rate of fluid extravasation a n d / o r the distribution of ~A/O r a t i o s . 5'6'13'23 In addition, sympathetic stimulation or autonomic imbalance might act directly on the lungs: pulmonary vasoconstriction could increase the capillary pressure; 15'18'22 the ~A/Q distribution might be affected so as to cause an effective increase in shunting and thence hypoxemia; 2,3 surfactant might be depleted, 397
S. Jennett and J. T. Hoff
FIG. 7. Example of increases in phrenic activity and of blood pressure (BP) during rise in intracranial pressure (ICP). The BP starts to rise 20 seconds after the beginning of ICP elevation, when the ICP is 80 to 90 mm Hg. The beginning of an increase in end-tidal pCO2 can be seen, top right.
resulting in decreased complianceI and a tendency to extravasation. 2~ Grossly evident pulmonary edema was found in three animals out of 10 in Series 1; a minor degree of hypoxemia occurred in two of these, and the other was normoxic after ventilation was adjusted. Six animals in Series 2 showed similar changes but in no instance did PaO2 decrease below 70 torr during the 30 minutes after ICP elevation was stopped and before sacrifice. Our results provide no particular evidence for or against a direct neural effect on the lungs causing congestion and edema, as opposed to the alternative explanation in terms of the consequences of systemic effects. Elevation of blood pressure in all animals was severe and prolonged, implying that there must have been widespread systemic vasoconstriction and some shift of blood into the lesser circulation. The late and inconstant development of edema is compatible with an earlier increase in extravasation rate, which would not necessarily exceed the capacity of lymphatic drainage, or if it did, would take 398
time to cause accumulation of fluid to a grossly recognizable extent. It is quite possible that lesser degrees of interstitial edema took place in the other animals in our two series, and could have been detected by appropriate methods for estimating lung water, xl It is known that considerable increases in interstitial fluid can occur without impairing diffusion,8,21 so the absence of hypoxemia in this series, except in association with frank lung changes, is understandable. In another series of experiments using a similar preparation for raised ICP, lung water has been measured and a significant increase was usual; again hypoxemia did not occur? ~ The persistent normoxia (after correction of the mild hypercapnia) in the face of prolonged increase in ICP, in arterial BP, and presumably in sympathetic outflow, is evidence against any direct or indirect pulmonary hemodynamic effect causing "v'A/Q maldistribution. Correction of PaO2 might have been achieved in such circumstances but only by hyperventilation to hypocapnia. J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure 60 - -
o
o 40
--
Pa C O 2
~' ~~ , , ~ o
(o'
o
o
o
o
~ " ~ 9 ~, Oo ' ~ - - o , ~
,~o~
~,,~
o o Oo 9
9
9
9
torr
20--
o-.//
I
60
I
80
I
Pa 0 2 torr
100
I
120
FIG. 8. Diagram showing the interrelationship between PaO2 and PaCO2 for cats under pentobarbitone anesthesia. The regression lines refer to an earlier series of spontaneously breathing animals; ~2the solid line relates PaO2 to PaCO2 in different animals at different levels of anesthesia, at normal intracranial pressure (ICP); the broken line relates values in the same group at raised ICP, showing varying degrees of ventilatory depression. Blood gas data from the present experiments on ventilated animals have been superimposed. Open symbols represent estimations at raised ICP, triangles with, and circles without adjustment of ventilation; and closed symbols, at normal ICP.
The existing evidence in favor of such an effect derives from experiments in dogs breathing 100% oxygen. Berman and Ducker 2 reported an increase in alveolar-arterial pO2 difference during raised ICP, and Maxwell and Goodwin is showed an increase in the calculated percentage shunt, which was related to the severity of applied ICP. The magnitude of the shunt increase in the latter study (7% to 11%) was such that, if an animal were breathing air, the PaO2 could decrease from, say, 100 to 75 torr. Some cats in our present series showed a comparable decrease in arterial pO2. Had this hypoxemia been due to shunt, it would not have been corrected by ventilating to normocapnia. Thus our results do not require the hypothesis that raised ICP, by virtue of increasing sympathetic outflow, induces shunting in the lungs. The reason for the difference between these and other findings is not clear. J. Neurosurg. / Volume 48 / March, 1978
Increased Metabolic Rate During Raised Intracranial Pressure The mild hypoxemia and hypercapnia that occurred with unchanged minute volume, suggested an increase in metabolic activity and therefore relative hypoventilation. In some animals this could be accounted for by an increase in body temperature, but in others in which temperature was better controlled the phenomenon still occurred. This would be compatible with an increase in catecholamine release. 4 A possible alternative explanation of the rise in pCO2 might have been that effective alveolar ventilation was reduced because of an increase in alveolar dead space; this would have caused a widening of the end-expired arterial difference for pCO2, which did not occur. Rise in body temperature could increase the ventilation requirement, and it is perhaps 399
S. Jennett and J. T. Hoff significant that those animals that developed the greatest degree of hypoxia and hypercapnia were those in which the temperatures increased noticeably during the relevant period. Conversely, two animals which showed no change in blood gases during a period at 200 mm Hg ICP had a fall in body temperature at that time. Data from previous experiments in spontaneously breathing cats 12 plotted on an O2-CO2 diagram, served as the basis for predicting the likely change in PaO~ relative to the change in PaCO2, when ventilation was depressed either by raised ICP or by incremental barbiturates; in that study, the APaCO2/APaO2 relationship was similar in the two circumstances, making it unnecessary to postulate a specific effect of raised ICP on the lungs. TM In the present series, the data fitted this same relationship (Fig. 8).
with a rise in body temperature. We conclude that, in this preparation, there are no grounds for implicating a specific effect of raised int r a c r a n i a l pressure on the p u l m o n a r y vascular bed that leads to hypoxemia. A proportion of animals develop pulmonary edema and hemorrhage, but even this is not necessarily associated with arterial hypoxemia, despite being sufficient to be recognizable macroscopically. Unless there is a species difference, therefore, raised ICP per se is probably not the factor in clinical brain damage responsible for hypoxemia until or unless it leads to frank pulmonary edema.
Respiratory Activity During Severely Decreased Cerebral Perfusion Pressure
1. Beckman DL, Bean JW, Baslock DR: Sympathetic influence on lung compliance and surface forces in head injury. J Appi Physiol 30:394-399, 1971 2. Berman IR, Ducker TB: Pulmonary, somatic and splanchnic circulatory responses to increased intracranial pressure. Ann Surg 169: 210-216, 1969 3. Brackett CE: Respiratory complications of head injury, in International Symposium on Head Injuries. Edinburgh/London: Churchill Livingstone, 1971, pp 255-265 4. Brashear RE, Ross JC: Circulating beta adrenergic stimulator during elevated cerebrospinal fluid pressure. Arch Intern Med 127:748-753, 1971 5. Chen HI, Chai CY: Pulmonary edema and hemorrhage as a consequence of systemic vasoconstriction. Am J Physiol 227:144-15 l, 1974 6. Chen HI, Sun SC, Chai CY: Pulmonary edema and haemorrhage resulting from cerebral compression. Am J Physiol 224: 223-229, 1973 7. Dejours P, Lacaisse A: Arterial blood pH and partial pressures of 02 and CO2 in normal awake cats. J Physioi 68:87-90, 1971 8. Fishman AP" Pulmonary edema. The water exchanging function of the lung. Circulation 46:390-408, 1972 9. Herbert DA, Mitchell RA: Blood gas tensions and acid-base balance in awake cats. J Appl Physiol 30:434-436, 1971 10. Hoff JT, Mitchell RA: The effect of hypoxia on the Cushing response, in Brock M, Dietz H (eds): Intraeranial Pressure. Berlin/Heidelberg/New York: Springer-Verlag, 1972, pp 205-209
The persistence of inspiratory impulse activity during prolonged increases in ICP, when the spontaneously breathing animal would almost certainly have been apneic, ~2 requires explanation. H y p e r c a p n i a and hypoxemia develop when respiration begins to diminish in the breathing animal. A decrease in PaO2 in the presence of cerebral ischemia will further threaten the activity or respiratory neurons. Thus, a positive feedback mechanism is operative and tends to worsen the situation. In the artificially ventilated animals, however, arterial blood gases were normal. It remains remarkable that regular phrenic activity persisted for so long in some instances. Conclusions
Anesthetized cats, breathing air, commonly develop a slight increase in PaC02 and a concomitant decrease in PaO~ during periods of severely increased intracranial pressure, if the ventilation is kept constant at the level that maintained normocapnia at normal ICP. However, if the ventilation is adjusted to correct the pCO2, the pO2 is also c o r r e c t e d , implying t h a t the alteration resulted from inadequate alveolar ventilation. This could be due to increased metabolic rate related to increased sympathetic activity, and is sometimes, but not necessarily, associated 400
Acknowledgments
We are grateful for the technical assistance of G. Burnside, Glasgow, and M. Nishimura, San Francisco. References
J. Neurosurg. / Volume 48 / March, 1978
Blood gases in increased intracranial pressure 11. Hoff J, Nishimura M, Pitts L: A quantitative method for neurogenic pulmonary edema in cats. Chicago Conference on Neural Trauma (In press) 12. Jennett S, North JB: Effect of intermittently raised intracranial pressure on breathing pattern, ventilatory response to CO2, and blood gases in anesthetized cats. J Neurosurg 44:156-167, 1976 13. Maxwell JA, Goodwin JW: Neurogenic pulmonary shunting. J Trauma 13:368-373, 1973 14. Miller JD, Stanek AE, Langfitt TW: A comparison of autoregulation to changes in intracranial and arterial pressure in the same preparation, in Fieschi C (ed): Cerebral Blood Flow and Intracranial Pressure. Basel: S Karger, 1972, pp 34-38 15. Moss G, Staunton C, Stein AA: The centrineurogenic etiology of the acute respiratory distress syndromes. Universal, species-independent phenomenon. Am J Surg 126: 37-41, 1973 16. Moss G, Staunton C, Stein AA: Cerebral etiology of the "shock lung syndrome." J Trauma 12:885-890, 1972 17. Moss IR, Lisbon A, Levine JF, et al: The effects of increased intracranial pressure on respiratory functions, in Lundberg N, Pont6n U, Brock M (eds): Intracranial Pressure II. Berlin/Heidelberg/New York: SpringerVerlag, 1975, pp 315-318 18. Pitts LH, Severinghaus JW, Mitchell RA, et al: The role of increased intracranial pressure in the production of neurogenic pulmonary edema, in Lundberg N, Pont6n U, Brock M
J. Neurosurg. / Volume 48 / March, 1978
19.
20. 21. 22.
23. 24.
(eds): Intracranial Pressure II. Berlin/ Heidelberg/New York: Springer-Verlag, 1975, pp 319-323 Schlaefke ME, See WR, Loeschcke HH: Ventilatory response to alterations of H + ion concentration in small areas of the ventral medullary surface. Respir Physiol 10:198-212, 1970 Staub NC: Effects of alveolar surface tension on the pulmonary vascular bed. Jpn Heart J 7:386-399, 1966 Staub NC: Pulmonary edema. Physiol Rev 54:678-811, 1974 Sugg WL, Craver WD, Webb WR, et al: Pressure changes in the dog lung secondary to hemorrhagic shock: protective effect of pulmonary reimplantation. Ann Surg 169:592-598, 1969 Theodore J, Robin ED: Pathogenesis of neurogenic pulmenary oedema. Lancet 2: 749-751, 1975 Zwetnow N: CBF autoregulation to blood pressure and intracranial pressure variations. Seand J Clin Lab Invest Suppl 102:V:A, 1968
This work was supported in part by a grant from the Scottish Hospital Endowments Research Trust and Trauma Center Program (Project Grant GM 18470). Dr. Hoff is recipient of Teacher-Investigator Award NS 11051 NSRB. Address reprint requests to: Sheila Jennett, M.D., Ph.D., Department of Physiology, University of Glasgow, Glasgow G12 8QQ, Scotland.
401
