- Email: [email protected]
Estimation of cerebral blood flow during cardiopulmonary resuscitation in humans
Resuscitation, 19 (1990) 115-123 Elsevier Scientific Publishers Beland Ltd.
115
Estimation of cerebral blood flow during cardiopulmonary resuscitation in humans
Soren Fischer Christensena, Carsten Stadeager” and Eugeniusz Siemkowiczbp* ‘Hvidovre University Hospital, Hvidovre (Denmark) and %eparttnent of Medicine. King Fahal Specialist Hospita1 and Research Centre, Riyadh (Saudi Arabia) (Received May lst, 1989; revision received September 17th, 1989; accepted September 24th, 1989)
Cerebral blood flow (CBF) and cardiac output (CO) were measured during cardiopulmonary resuscitation in patients who were unsuccessfully resuscitated by use of Cl4iodoantipyrine injected into the left ventricle. CO varied between 1.3 and 2.2 l/min with mean 1.8 + 0.6 l/min (+ S.D.) (28 ml/kg/min). The cortical CBF was found between 14 and 211 ml 100 g-’ * min-’ with mean 42 ml 100 g-’ * minc’ and mean white matter CBF equal to 27 ml 100 g-’ * min-‘. It is suggested that the external cardiac massage in humans may be of poor efficacy in terms of brain revival. Cortical CBF after long-lasting cardiopulmonary resuscitation showed signs of maldistribution suggestive of a patchy and incomplete perfusion. Cardiac massage - Cardiac output - CBF
INTRODUCTION
The purpose of cardiopulmonary resuscitation (CPR) is to deliver oxygenated blood to the body and to re-establish its spontaneous function, thereby preventing development of organ damage, particularly brain damage. The practice of closed-chest resuscitation for more than two decades has provided evidente that only one out of over 5-6 cardiac victims subjected to CPR survives [1,2]. Failure to survive can be ascribed in most cases to poor efficacy in producing blood flow during closed chest resuscitation. This explains why most of the successfully resuscitated patients develop anoxic brain damage after CPR lasting more than 10 minutes [3]. New reports on the cardiac output and carotid blood flows during CPR have been published [4-61. To our knowledge, the injection sites of the applied indicators were either v. cava or the right atrium. As the blood flows forward in the aorta and backward in the w cavas during sternal compression, it is possible that a large pro*To whom correspondence should be sent at: Consultant Intensivist, Ring Faisal Specialist Hospital and Research Centre, P.O. Box 3354, Riyadh 11211, Saudi Arabia. 0300-9572/90/$03.50 0 1990 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland
116
portion of the indicator injected into the venous blood is distributed multidirectionally and therefore contributes to an erroneous estimation of blood flow. In the case of arterial injection of the indicator, the propelling of the indicator is undirectional because the arteries are less compressible than the veins [4]. During CPR the entire aortic pressure is transmitted peripherally while the intrathoracic venous pressure is not. One can assume therefore that only intraarterial injections of the indicator wil1 result in reliable estimations of blood flow under CPR. In order to explain the increased frequency of brain damage after successful CPR, we estimated cardiac output (CO) and cerebral blood flow (CBF) in patients suffering from cardiac arrest. Our results strongly suggest that development of brain damage in association with CPR might be explained by a low rate of cortical CBF produced by cardiac massage. MATERIALS AND METHODS
The patients studied were brought to our hospital immediately after cardiac arrest had occurred at home or in the street. Cardiac massage was instituted within 3-4 min after collapse and continued during transportation to the hospital. On arrival, asystole was observed in four patients and ventricular fibrillation in the remaining two patients (No. 1 and 6). Al1 patients in ventricular fibrillation received electrical defibrillation as soon as the condition was obvious. The patients were intubated and ventilated with pure oxygen. A venous line was established and 8.4% bicarbonate infused at the rate of 100 ml/10 min of CPR additionally to a rapid infusion of 5% glucose (usually > 1000 ml in total). The patients received medication including basic cardiac drugs used during the CPR (adrenalin, atropine, lidocaine, calcium and betablockers). After approximately 10 min of CPR, cardiac massage was discontinued and the patient’s status evaluated. In case of asystole no further attempts of CPR were undertaken and the patient pronounced dead (by an independent team of doctors on ER calls) basically due to the following signs: asystole, no blood pressure, no pulse and fixed, dilated pupils. After pronouncement of death the subclaviar vein was catheterized with a 15gauge catheter which was introduced into the right atrium. Another Sgauge catheter was, after an incision in the femoral triangle, introduced into the femoral artery and advanced about 20-25 cm in a proximal direction. This catheter was connected to a withdrawing pump in order to collect blood at a constant rate for calculating the integrals of arterial isotope concentration. Position of catheters was not visualized by X-ray. The catheterizations took approximately 3 min. Cardiac massage was performed according to the standards of American Heart Association. Cardiac massage and ventilation were then resumed. Twenty seconds later the pump withdrawing arterial blood was started and simultaneously, a bolus of 100 &Ji [14C]iodoantipyrine was injected into the left ventricle. The injection site was identified by the presence of a well-oxygenated blood and, retrospectively, by the distribution pattern of collected activities in the venous and arterial blood. Cardiac massage was then continued for another 60 s. Thereafter cardiac massage was discontinued
117
and the pump stopped. Venous and arterial blood samples were collected for measuring gas tensions and isotope concentrations. The patient’s status was once again checked and death confirmed. A burrhole was drilled in the tempora1 region and the brain exposed was obtained by using a needle with 3 mm internal diameter. The biopsies were taken vertically and consisted of grey and white matter. The length of a single biopsy ranged from 4 to 7 cm. The samples were divided horizontally, starting at the brain surface. Each brain sample weighed approximately 20 mgs and measured 4-6 mm in length. Cerebral blood flow rates were calculated from tissue and arterial blood samples of labelled iodoantipyrine 60 s (T) after injecting the isotope. Cardiac output was calculated according to: CO =
F; x Qt/Q,(t)
Where Qt is the total injected activity, Q,,(t) the amount of radioactivity present in the whole arterial sample, collected at a constant and known rate (Ft’). f =
l/W)
x Ff x QJWQ,(T)
(2)
The equation used relates CBF per unit weight of brain v), the amount of isotope present in unit weight of brain (QJT)) and the amount of iodoantipyrine present in arterial sample (Qo(T collected at a constant and known rate (0’). E(T) is the net extraction fraction in the time (T). Since E(T) varies chiefly with the cerebral blood flow, as explained and validated in detail previously [7], the equation could be solved by an iterative procedure. The blood and tissue samples were transferred rapidly to - 25OC and later processed as follows: the samples were transferred to preweighed counting vials containing 1.5 ml of a Soluene (Packard) and isopropanol mixture (1: 1). The vials and samples were then reweighed, blanched with 0.5 ml hydrogen peroxide and stored overnight at 5OOC. The next morning preparation was completed by the addition of 20 ml of a mixture of Instagel (Packard) and 0.5 N HCl(9: 1). The vials were allowed to stabilize at constant temperature for 24 h prior to counting in a Packard TriCarb 2454 liquid scintillation spectrometer. Iodoantipyrine was purchased from New England Nuclear Corporation and its specific activity was 52.89 mCi mmol/l. The project was evaluated and approved by The Danish Medical Ethica1 Committee. In al1 cases the pronounciation of death was done by a team of doctors not involved with this study directly. RESULTS
The mean age of patients was 67 f 7 ( f S.D.) years and the mean weight 65 + 3 kg. Four patients were males and two females. The results of blood gas analysis during CPR are depicted in Table 1. None of the patients studied showed any evidente of arterial hypoxia. However, the arterio-venous oxygen and carbon dioxide gradients were increased in al1 cases suggesting a hypoxic condition át the tissue level. In al1 cases the arterial Pao, concentration was higher than that of the venous blood. In
118 Table 1. The individual values of arterial (a) and venous (v) blood gas analysis dwing CPR in mmHg in six patients included in the study. No.
Pao,
Pvo,
Pvco,
Paco,
pHa
PH”
1 2 3 4 5 6
321 131 65 63 285 106
21 31 18 23 20 22
75 42 85 51 26 41
119 80 127 114 86 126
1.24 7.56 7.30 7.54 7.31 7.00
7.04 7.19 6.90 7.16 7.05 6.74
4 cases the arterial pH was acidotic and in 2 alkalotic. The venous pH was severely acidotic in al1 cases. Cardiac output varied between 1.3 and 2.2 1 * min-’ (mean 1.8 f 0.6 1. min-’ or 28 ml kg-‘). The results are shown in Table 11. The cortical CBF (Table 11) varied from 14 to 211 ml * 100 g-’ min-’ with the mean 42 f 38 for al1 cortical samples. In 3 out of 6 sampled brains there were found regions with cortical CBF lower or equal to 16 ml * 100 g-i. In these brains the overall levels of cortical CBF were also relatively low. In the remaining 3 brains, however, the levels of cortical CBF were higher and more uneven. The ratio highest/lowest local CBF (H/L ratio) was > 3. In one brain even 2 regions with hyperemie values were found (120 and 211 ml 100 g-l min-‘). The mean cortical CBF did not correlate to the cardiac output values (linear regression, r = 0.53, Fig. 1). The CBF values in the white matter were lower and more even than in the cortical samples (H/L ratio < 3). However, in 2 cases they were higher than in the corresponding cortical CBF values. The mean white matter CBF was 28 ml - 100 g-1 amin-1, which corresponds to 66% of the mean cortical level. The white matter CBF values showed no linear correlation to the cardiac output values. The lowest values of white matter CBF were found in 3 brains and were between 12 and 15 ml - 100 g-’ * min-‘. In one patient (outside the protocol) the brain: venous blood concentration ratio for iodoantipyrine was estimated after 20 min of continuous CPR. It was homogenous as sampled and averaged 0.75 f 0.02 (N = 6) for the gray and 0.70 f 0.08 (N = 6) for the white matter. ??
??
??
??
DISCUSSION
We believe that this is the first report of CBF during CPR in humans. There are many suitable methods for CBF estimation in humans during normal conditions. Their use during CPR is not applicable because of different haemodynamics. Assuming that the cardiac valves during CPR are intact and that the arterial circuit of circulation is a one-way system, our method of CBF and CO measuring should be, in theory, accurate. The use of microsphere technique during CPR in a swine model has been validated by Taylor et al. [ 131. Our method resembles the microsphere technique but has
119 Table II.
The individual values of cortical, white matter CBF (in ml - 100 g-’- min-‘) and cardiac out-
put (CO,in 1. min-‘) in 6 patients included in the study. H/L ratio = highest to lowest flow ratios. No.
coltex
H/L ratio
White matter
8.8
43 19 27 36 15
3.6
29 42 28 38
24
1
2
120 68 59 211 16 28 54 15 18 20 19 18 17 15 56 29 30 39 38 45 21 26 36 14 31
6
Mean f S.D.
38 59 37 22 65 43 42 f 38
H/L ratio
2.1 2.9
1.3 1.2
15 16 15 18
1.8
1.2
1.9 22 21 30 30 24 1.9
1.5
1.4 16 12 22 17
2.6
1.9 1.4
43 52 43
3.0 3.5 f 2.7
CO
2.2
1.2 27 i: 11
1.6 f 0.6
1.8 f 0.6
the advantage of no circulatory perturbation related to micro embolism caused by microspheres. The method has been validated in animal studies and is frequently used. We injected the indicator into the left ventricle and collected by a femoral catheter which was directed towards the heart. This should decrease errors due to prolonged circulatory time and ensure that the collected blood was from a one-way flow system, i.e., arterial circuit.
J’
-
0
;0
40
50 -
CBF (ml loog’ mitï’) Fig. 1. sage.
io
60
Cortical CBF (mean) plotted against cardiac output values in six patients dwing cardiac mas-
Also the circulatory time for the isotope was only 60 s, optimally short for the brain to receive the arterial supply of isotope and to prevent the isotope being washed out by the chaotic venous flow with no isotope. Therefore, this unique method should be suitable for CBF estimation under CPR conditions in humans. CBF in man during CPR has not been measured to be compared with. However, our results could be attributed to long-lasting CPR (more likely direct cardiac compression), average body weight of the patients, effective cardiac massage, hemodilution, increasing with duration of CPR, and others [ 14,16-181. The main finding in this study is that during external cardiac massage in unsuccessfully resuscitated patients the levels of CBF and CO are too low. We found, on average, a 50% depression of the mean cortical CBF and a 30% of the mean white matter CBF in comparison to the control conditions, i.e., the normal cortical CBF equals - 80 and that of the white matter - 40 ml 100 g-l min-’ [8,9]. However, the most interesting finding is that the cortical CBF was maldistributed (Table 11) with very uneven distribution pattern (H/L ratio > 3). This patchy perfusion may explain the high incidence of anoxic brain damage after CPR. The levels of CBF found in our patients are not compatible with brain survival in connection with 20-30 min lasting CPR and prove that incomplete ischemia was prevailing during CPR. During a long-lasting CPR, cortical CBF compatible with brain revival should probably be above 50% of normal levels and not below 20% as found in some regions in our
121
study. It is very likely that a study of CBF at a lower grade of sampling would discover areas of no-flow, found in animal studies [ 191. Our results, therefore, could suggest that external cardiac massage, sufficient to revive circulation, is not always effective in restoring cerebral circulation. In several cortical regions the CBF was below 20% of the normal cortical flow, i.e., 14-16 ml 100 g-’ min-’ Such levels were shown to be associated with a disintegration of membrane function and eventual brain damage [ 101. The cerebrovascular resistance and CBF threshold for reperfusion increase during ischemia [20]. This is in agreement with the fact that the post ischemic brain is very sensitive to hypotension [12,15]. What we see occurring in the hypoxic brain is a vicieus circle where the CBF wil1 not increase until the membrane function is restored. To restore this function oxygen is required but as its transportation in low CBF conditions is slowed there is, therefore, a delay in resuming membrane function and normalizing CVR until eventually hemodynamic condition has recovered. If oxygenation and blood pressure are normal or above normal levels, the importante of the vicieus circle is negligible. However, under poor hemodynamic condition of hypoxia and hypotension, which is encountered during and after CPR in humans, it may be of a paramount importante and the main reason for extremely poor results of CPR. It is very likely that patients, who respond to CPR do so because CPR is commenced prior to membrane failure of the brain neurons. Once membrane failure occurs the CVR wil1 increase and the CPR wil1 be ineffectual in providing the brain with enough oxygen to restore cerebral circulation. This i’ssupported by Siemkowicz et al. [ 111, whose study of post CPR patients showed that the normal brain function was regained only if the CSF [K]’ in the suboccipital CSF immediately after cardiac resuscitation was normal. Once membrane failure occurs and [K]’ increases during brain ischemia, it is unlikely that CPR wil1 be successful in preventing anoxic brain damage. In this study we used 100 mCi of the isotope, a relatively low dose for a patient. However, our samples showed the radioactivity levels which were 15 times higher than the background radiation. Therefore, only a smal1 error in estimation could be ascribed to this relatively low dose. As we used 60 s circulation time (CT) for the isotope, we fee1 that the time employed was optimal to estimate CBF in humans if the isotope is injected into the left ventricle. During normal conditions, the CT from the subclaviar vein injection site to the femoral artery is less than 45 s [12]. During CPR, the CT must be prolonged, this is supported by lower than normal CO values. However, the circulation of blood from the left ventricle to the brain is much shorter than 45 s as the venous and pulmonary circulation are the slowest paths of the circuit. After the intraventricular injections of isotope, we found only, on average, 13% and never detected more radioactivity in the centra1 venous blood than 30% of the lowest arterial end-concentration, a proof that recirculation and retrograde blood flow were of little importante. Our results of CO could be overestimated if significant regurgitation of blood from the left ventricle took place. The results of CBF,
122
however, should be correct even in case of blood regurgitation as the collected blood was exposed to similar profiles of isotope concentrations as the bram. The hemodynamics of CPR are not quite understood. There is, however, no reason to suspect that, in the context of the CO levels of this study, any significant error due to the above factors could invalidate our results. The same applies to the question of homogeneity of blood flow within the arterial circulation during CPR. This is supported by the following facts: (a) aorta and arteries are little compressible, (b) aortic blood flows in one direction during CPR [4], (c) no isotope was found in the vena cava blood and (d) physiologically directed 0, and CO, concentrations. Accepting our results as representative of CPR conditions in humans, we conclude that the CPR as employed at present is not an absolutely effective method of preventing brain damage after an extended period of cardiac arrest. Marginal levels of CO and CBF during CPR in man may wel1 explain inefficiency of closed CPR. It seems reasonable to suggest that revival of the ‘old’ open cardiac massage method, at least in a good proportion of cardiorespiratory resuscitations, should be encouraged as it is more effective [16] than the closed cardiac massage. ACKNOWLEDGEMENT
We wish to thank Ms. Judith L. Sosa for providing secretarial assistance in the preparation of this manuscript. REFERENCES 1 2 3
10 11 12 13
G.J. Farha, R.J. Capehart and P.N. Barker, Cardiopulmonary resuscitation, J. Kans. Med. Sec., 73 (1972) 406. J.A.W. Wildsmith, W.G. Dennyson and K.W. Meyers, Results of resuscitation following cardiac arrest, Br. J. Anaesth., 44 (1972) 716. J.J. Caronna and S. Finklestein, Neurological syndromes after cardiac arrest, Stroke, 9 (1978) 517 -510. N. Chandra, M. Rudikoff and M.L. Weisfeldt, Simultaneous chest compression and ventilation at high airway pressure during cardiopulmonary resuscitation, Lancet, i (1980) 175-178. A. Oriol and H.J. Smith, Hemodynamic observations during cardiac massage, J. Can. Med. Assoc., 98 (1%8) 84. L.R.M. Guercio, R.P. Coomaraswarmy and D. State, Cardiac output and other hemodynamic variables during external cardiac massage in man, N. Engl. J. Med., (1963) 1398-1404. A. Gjedde, A.J. Hansen and E. Siemkowicz, Simultaneous determination of regional blood flow and blood brain glucose transfer in rat brain, Acts Physiol. Stand., 108 (1980) 321-330. A.M. Harper, General physiology of the cerebral circulation, Int. Anesthesiol. Clin., 7 (1969) 473506. C. Eintrei, W. Leszniewski and C. Carlsson. Local application of 133 Xenon for measurement of regional and isoflurane anesthesia in humans, Anesthesiology, 63 (1985) 391-394. J. Astrup, Energy-requiring cell functions in the ischemic bram, J. Neurosurg., 56 (1982) 482497. E. Siemkowicz, 1. Christiansen and S.C. Sorensen, Changes in cisternal fluid potassium concentration following cardiac arrest, Acts Neurol. Stand., 55 (1977) 173. B.K. Siesjo. Cell damage in the bram: A speculative synthesis, J. Cereb. Blood Flow Metabol., 1 (1981) 155-185. B.B. Taylor, CC. Brown, T. Bridges et al., A model for regional blood flow measurements during cardiopulmonary resuscitation in a swine model, Resuscitation, 16 (1988) 107-118.
123 14 15 16 17
18 19 20
L.M.R. Guercio, R. Coomenaswary and D. State, Cardiac output and other neurodynamic variables during external massage in man, N. Engl. J. Med., 269 (1%3) 1398-1401. G.S. Krause. K. Kumar. B.C. White et al., Ischemia. resuscitation and reperfussion; Mechanisms of tissue typing and prospects for protection, Am. Health J., April (1986) 768-780. L.M.R. Guercio, Open chest cardiac massage: An overview, Resuscitation, 15 (1987) 9-11. K.H. Lindner, F.W. Ahnefeld and I.M. Boudler, The effect of epinephrine on hemodynamics, acid bone status and potassium during spontaneous circulation and cardiopulmonary resuscitation, Resuscitation. 16 (1988) 251-261. D.K. Alifimoff, Open versus closed chest cardiac massage in non-traumatic cardiac arrest, Resuscitation, IS (1987) 13-21. E. Kogstrom, M.L. Smith and B.K. Siesjo, Local cerebral blood flow in the recovery period following complete cerebral ischemia in the rat, J. Cereb. Blood Flow Metab., 3(1983) 170-182. E. Siemkowicz, Cerebrovascular resistance in ischemia, Arch. Pflug., 388 (1980) 243-247.
115
Estimation of cerebral blood flow during cardiopulmonary resuscitation in humans
Soren Fischer Christensena, Carsten Stadeager” and Eugeniusz Siemkowiczbp* ‘Hvidovre University Hospital, Hvidovre (Denmark) and %eparttnent of Medicine. King Fahal Specialist Hospita1 and Research Centre, Riyadh (Saudi Arabia) (Received May lst, 1989; revision received September 17th, 1989; accepted September 24th, 1989)
Cerebral blood flow (CBF) and cardiac output (CO) were measured during cardiopulmonary resuscitation in patients who were unsuccessfully resuscitated by use of Cl4iodoantipyrine injected into the left ventricle. CO varied between 1.3 and 2.2 l/min with mean 1.8 + 0.6 l/min (+ S.D.) (28 ml/kg/min). The cortical CBF was found between 14 and 211 ml 100 g-’ * min-’ with mean 42 ml 100 g-’ * minc’ and mean white matter CBF equal to 27 ml 100 g-’ * min-‘. It is suggested that the external cardiac massage in humans may be of poor efficacy in terms of brain revival. Cortical CBF after long-lasting cardiopulmonary resuscitation showed signs of maldistribution suggestive of a patchy and incomplete perfusion. Cardiac massage - Cardiac output - CBF
INTRODUCTION
The purpose of cardiopulmonary resuscitation (CPR) is to deliver oxygenated blood to the body and to re-establish its spontaneous function, thereby preventing development of organ damage, particularly brain damage. The practice of closed-chest resuscitation for more than two decades has provided evidente that only one out of over 5-6 cardiac victims subjected to CPR survives [1,2]. Failure to survive can be ascribed in most cases to poor efficacy in producing blood flow during closed chest resuscitation. This explains why most of the successfully resuscitated patients develop anoxic brain damage after CPR lasting more than 10 minutes [3]. New reports on the cardiac output and carotid blood flows during CPR have been published [4-61. To our knowledge, the injection sites of the applied indicators were either v. cava or the right atrium. As the blood flows forward in the aorta and backward in the w cavas during sternal compression, it is possible that a large pro*To whom correspondence should be sent at: Consultant Intensivist, Ring Faisal Specialist Hospital and Research Centre, P.O. Box 3354, Riyadh 11211, Saudi Arabia. 0300-9572/90/$03.50 0 1990 Elsevier Scientific Publishers Ireland Ltd. Printed and Published in Ireland
116
portion of the indicator injected into the venous blood is distributed multidirectionally and therefore contributes to an erroneous estimation of blood flow. In the case of arterial injection of the indicator, the propelling of the indicator is undirectional because the arteries are less compressible than the veins [4]. During CPR the entire aortic pressure is transmitted peripherally while the intrathoracic venous pressure is not. One can assume therefore that only intraarterial injections of the indicator wil1 result in reliable estimations of blood flow under CPR. In order to explain the increased frequency of brain damage after successful CPR, we estimated cardiac output (CO) and cerebral blood flow (CBF) in patients suffering from cardiac arrest. Our results strongly suggest that development of brain damage in association with CPR might be explained by a low rate of cortical CBF produced by cardiac massage. MATERIALS AND METHODS
The patients studied were brought to our hospital immediately after cardiac arrest had occurred at home or in the street. Cardiac massage was instituted within 3-4 min after collapse and continued during transportation to the hospital. On arrival, asystole was observed in four patients and ventricular fibrillation in the remaining two patients (No. 1 and 6). Al1 patients in ventricular fibrillation received electrical defibrillation as soon as the condition was obvious. The patients were intubated and ventilated with pure oxygen. A venous line was established and 8.4% bicarbonate infused at the rate of 100 ml/10 min of CPR additionally to a rapid infusion of 5% glucose (usually > 1000 ml in total). The patients received medication including basic cardiac drugs used during the CPR (adrenalin, atropine, lidocaine, calcium and betablockers). After approximately 10 min of CPR, cardiac massage was discontinued and the patient’s status evaluated. In case of asystole no further attempts of CPR were undertaken and the patient pronounced dead (by an independent team of doctors on ER calls) basically due to the following signs: asystole, no blood pressure, no pulse and fixed, dilated pupils. After pronouncement of death the subclaviar vein was catheterized with a 15gauge catheter which was introduced into the right atrium. Another Sgauge catheter was, after an incision in the femoral triangle, introduced into the femoral artery and advanced about 20-25 cm in a proximal direction. This catheter was connected to a withdrawing pump in order to collect blood at a constant rate for calculating the integrals of arterial isotope concentration. Position of catheters was not visualized by X-ray. The catheterizations took approximately 3 min. Cardiac massage was performed according to the standards of American Heart Association. Cardiac massage and ventilation were then resumed. Twenty seconds later the pump withdrawing arterial blood was started and simultaneously, a bolus of 100 &Ji [14C]iodoantipyrine was injected into the left ventricle. The injection site was identified by the presence of a well-oxygenated blood and, retrospectively, by the distribution pattern of collected activities in the venous and arterial blood. Cardiac massage was then continued for another 60 s. Thereafter cardiac massage was discontinued
117
and the pump stopped. Venous and arterial blood samples were collected for measuring gas tensions and isotope concentrations. The patient’s status was once again checked and death confirmed. A burrhole was drilled in the tempora1 region and the brain exposed was obtained by using a needle with 3 mm internal diameter. The biopsies were taken vertically and consisted of grey and white matter. The length of a single biopsy ranged from 4 to 7 cm. The samples were divided horizontally, starting at the brain surface. Each brain sample weighed approximately 20 mgs and measured 4-6 mm in length. Cerebral blood flow rates were calculated from tissue and arterial blood samples of labelled iodoantipyrine 60 s (T) after injecting the isotope. Cardiac output was calculated according to: CO =
F; x Qt/Q,(t)
Where Qt is the total injected activity, Q,,(t) the amount of radioactivity present in the whole arterial sample, collected at a constant and known rate (Ft’). f =
l/W)
x Ff x QJWQ,(T)
(2)
The equation used relates CBF per unit weight of brain v), the amount of isotope present in unit weight of brain (QJT)) and the amount of iodoantipyrine present in arterial sample (Qo(T collected at a constant and known rate (0’). E(T) is the net extraction fraction in the time (T). Since E(T) varies chiefly with the cerebral blood flow, as explained and validated in detail previously [7], the equation could be solved by an iterative procedure. The blood and tissue samples were transferred rapidly to - 25OC and later processed as follows: the samples were transferred to preweighed counting vials containing 1.5 ml of a Soluene (Packard) and isopropanol mixture (1: 1). The vials and samples were then reweighed, blanched with 0.5 ml hydrogen peroxide and stored overnight at 5OOC. The next morning preparation was completed by the addition of 20 ml of a mixture of Instagel (Packard) and 0.5 N HCl(9: 1). The vials were allowed to stabilize at constant temperature for 24 h prior to counting in a Packard TriCarb 2454 liquid scintillation spectrometer. Iodoantipyrine was purchased from New England Nuclear Corporation and its specific activity was 52.89 mCi mmol/l. The project was evaluated and approved by The Danish Medical Ethica1 Committee. In al1 cases the pronounciation of death was done by a team of doctors not involved with this study directly. RESULTS
The mean age of patients was 67 f 7 ( f S.D.) years and the mean weight 65 + 3 kg. Four patients were males and two females. The results of blood gas analysis during CPR are depicted in Table 1. None of the patients studied showed any evidente of arterial hypoxia. However, the arterio-venous oxygen and carbon dioxide gradients were increased in al1 cases suggesting a hypoxic condition át the tissue level. In al1 cases the arterial Pao, concentration was higher than that of the venous blood. In
118 Table 1. The individual values of arterial (a) and venous (v) blood gas analysis dwing CPR in mmHg in six patients included in the study. No.
Pao,
Pvo,
Pvco,
Paco,
pHa
PH”
1 2 3 4 5 6
321 131 65 63 285 106
21 31 18 23 20 22
75 42 85 51 26 41
119 80 127 114 86 126
1.24 7.56 7.30 7.54 7.31 7.00
7.04 7.19 6.90 7.16 7.05 6.74
4 cases the arterial pH was acidotic and in 2 alkalotic. The venous pH was severely acidotic in al1 cases. Cardiac output varied between 1.3 and 2.2 1 * min-’ (mean 1.8 f 0.6 1. min-’ or 28 ml kg-‘). The results are shown in Table 11. The cortical CBF (Table 11) varied from 14 to 211 ml * 100 g-’ min-’ with the mean 42 f 38 for al1 cortical samples. In 3 out of 6 sampled brains there were found regions with cortical CBF lower or equal to 16 ml * 100 g-i. In these brains the overall levels of cortical CBF were also relatively low. In the remaining 3 brains, however, the levels of cortical CBF were higher and more uneven. The ratio highest/lowest local CBF (H/L ratio) was > 3. In one brain even 2 regions with hyperemie values were found (120 and 211 ml 100 g-l min-‘). The mean cortical CBF did not correlate to the cardiac output values (linear regression, r = 0.53, Fig. 1). The CBF values in the white matter were lower and more even than in the cortical samples (H/L ratio < 3). However, in 2 cases they were higher than in the corresponding cortical CBF values. The mean white matter CBF was 28 ml - 100 g-1 amin-1, which corresponds to 66% of the mean cortical level. The white matter CBF values showed no linear correlation to the cardiac output values. The lowest values of white matter CBF were found in 3 brains and were between 12 and 15 ml - 100 g-’ * min-‘. In one patient (outside the protocol) the brain: venous blood concentration ratio for iodoantipyrine was estimated after 20 min of continuous CPR. It was homogenous as sampled and averaged 0.75 f 0.02 (N = 6) for the gray and 0.70 f 0.08 (N = 6) for the white matter. ??
??
??
??
DISCUSSION
We believe that this is the first report of CBF during CPR in humans. There are many suitable methods for CBF estimation in humans during normal conditions. Their use during CPR is not applicable because of different haemodynamics. Assuming that the cardiac valves during CPR are intact and that the arterial circuit of circulation is a one-way system, our method of CBF and CO measuring should be, in theory, accurate. The use of microsphere technique during CPR in a swine model has been validated by Taylor et al. [ 131. Our method resembles the microsphere technique but has
119 Table II.
The individual values of cortical, white matter CBF (in ml - 100 g-’- min-‘) and cardiac out-
put (CO,in 1. min-‘) in 6 patients included in the study. H/L ratio = highest to lowest flow ratios. No.
coltex
H/L ratio
White matter
8.8
43 19 27 36 15
3.6
29 42 28 38
24
1
2
120 68 59 211 16 28 54 15 18 20 19 18 17 15 56 29 30 39 38 45 21 26 36 14 31
6
Mean f S.D.
38 59 37 22 65 43 42 f 38
H/L ratio
2.1 2.9
1.3 1.2
15 16 15 18
1.8
1.2
1.9 22 21 30 30 24 1.9
1.5
1.4 16 12 22 17
2.6
1.9 1.4
43 52 43
3.0 3.5 f 2.7
CO
2.2
1.2 27 i: 11
1.6 f 0.6
1.8 f 0.6
the advantage of no circulatory perturbation related to micro embolism caused by microspheres. The method has been validated in animal studies and is frequently used. We injected the indicator into the left ventricle and collected by a femoral catheter which was directed towards the heart. This should decrease errors due to prolonged circulatory time and ensure that the collected blood was from a one-way flow system, i.e., arterial circuit.
J’
-
0
;0
40
50 -
CBF (ml loog’ mitï’) Fig. 1. sage.
io
60
Cortical CBF (mean) plotted against cardiac output values in six patients dwing cardiac mas-
Also the circulatory time for the isotope was only 60 s, optimally short for the brain to receive the arterial supply of isotope and to prevent the isotope being washed out by the chaotic venous flow with no isotope. Therefore, this unique method should be suitable for CBF estimation under CPR conditions in humans. CBF in man during CPR has not been measured to be compared with. However, our results could be attributed to long-lasting CPR (more likely direct cardiac compression), average body weight of the patients, effective cardiac massage, hemodilution, increasing with duration of CPR, and others [ 14,16-181. The main finding in this study is that during external cardiac massage in unsuccessfully resuscitated patients the levels of CBF and CO are too low. We found, on average, a 50% depression of the mean cortical CBF and a 30% of the mean white matter CBF in comparison to the control conditions, i.e., the normal cortical CBF equals - 80 and that of the white matter - 40 ml 100 g-l min-’ [8,9]. However, the most interesting finding is that the cortical CBF was maldistributed (Table 11) with very uneven distribution pattern (H/L ratio > 3). This patchy perfusion may explain the high incidence of anoxic brain damage after CPR. The levels of CBF found in our patients are not compatible with brain survival in connection with 20-30 min lasting CPR and prove that incomplete ischemia was prevailing during CPR. During a long-lasting CPR, cortical CBF compatible with brain revival should probably be above 50% of normal levels and not below 20% as found in some regions in our
121
study. It is very likely that a study of CBF at a lower grade of sampling would discover areas of no-flow, found in animal studies [ 191. Our results, therefore, could suggest that external cardiac massage, sufficient to revive circulation, is not always effective in restoring cerebral circulation. In several cortical regions the CBF was below 20% of the normal cortical flow, i.e., 14-16 ml 100 g-’ min-’ Such levels were shown to be associated with a disintegration of membrane function and eventual brain damage [ 101. The cerebrovascular resistance and CBF threshold for reperfusion increase during ischemia [20]. This is in agreement with the fact that the post ischemic brain is very sensitive to hypotension [12,15]. What we see occurring in the hypoxic brain is a vicieus circle where the CBF wil1 not increase until the membrane function is restored. To restore this function oxygen is required but as its transportation in low CBF conditions is slowed there is, therefore, a delay in resuming membrane function and normalizing CVR until eventually hemodynamic condition has recovered. If oxygenation and blood pressure are normal or above normal levels, the importante of the vicieus circle is negligible. However, under poor hemodynamic condition of hypoxia and hypotension, which is encountered during and after CPR in humans, it may be of a paramount importante and the main reason for extremely poor results of CPR. It is very likely that patients, who respond to CPR do so because CPR is commenced prior to membrane failure of the brain neurons. Once membrane failure occurs the CVR wil1 increase and the CPR wil1 be ineffectual in providing the brain with enough oxygen to restore cerebral circulation. This i’ssupported by Siemkowicz et al. [ 111, whose study of post CPR patients showed that the normal brain function was regained only if the CSF [K]’ in the suboccipital CSF immediately after cardiac resuscitation was normal. Once membrane failure occurs and [K]’ increases during brain ischemia, it is unlikely that CPR wil1 be successful in preventing anoxic brain damage. In this study we used 100 mCi of the isotope, a relatively low dose for a patient. However, our samples showed the radioactivity levels which were 15 times higher than the background radiation. Therefore, only a smal1 error in estimation could be ascribed to this relatively low dose. As we used 60 s circulation time (CT) for the isotope, we fee1 that the time employed was optimal to estimate CBF in humans if the isotope is injected into the left ventricle. During normal conditions, the CT from the subclaviar vein injection site to the femoral artery is less than 45 s [12]. During CPR, the CT must be prolonged, this is supported by lower than normal CO values. However, the circulation of blood from the left ventricle to the brain is much shorter than 45 s as the venous and pulmonary circulation are the slowest paths of the circuit. After the intraventricular injections of isotope, we found only, on average, 13% and never detected more radioactivity in the centra1 venous blood than 30% of the lowest arterial end-concentration, a proof that recirculation and retrograde blood flow were of little importante. Our results of CO could be overestimated if significant regurgitation of blood from the left ventricle took place. The results of CBF,
122
however, should be correct even in case of blood regurgitation as the collected blood was exposed to similar profiles of isotope concentrations as the bram. The hemodynamics of CPR are not quite understood. There is, however, no reason to suspect that, in the context of the CO levels of this study, any significant error due to the above factors could invalidate our results. The same applies to the question of homogeneity of blood flow within the arterial circulation during CPR. This is supported by the following facts: (a) aorta and arteries are little compressible, (b) aortic blood flows in one direction during CPR [4], (c) no isotope was found in the vena cava blood and (d) physiologically directed 0, and CO, concentrations. Accepting our results as representative of CPR conditions in humans, we conclude that the CPR as employed at present is not an absolutely effective method of preventing brain damage after an extended period of cardiac arrest. Marginal levels of CO and CBF during CPR in man may wel1 explain inefficiency of closed CPR. It seems reasonable to suggest that revival of the ‘old’ open cardiac massage method, at least in a good proportion of cardiorespiratory resuscitations, should be encouraged as it is more effective [16] than the closed cardiac massage. ACKNOWLEDGEMENT
We wish to thank Ms. Judith L. Sosa for providing secretarial assistance in the preparation of this manuscript. REFERENCES 1 2 3
10 11 12 13
G.J. Farha, R.J. Capehart and P.N. Barker, Cardiopulmonary resuscitation, J. Kans. Med. Sec., 73 (1972) 406. J.A.W. Wildsmith, W.G. Dennyson and K.W. Meyers, Results of resuscitation following cardiac arrest, Br. J. Anaesth., 44 (1972) 716. J.J. Caronna and S. Finklestein, Neurological syndromes after cardiac arrest, Stroke, 9 (1978) 517 -510. N. Chandra, M. Rudikoff and M.L. Weisfeldt, Simultaneous chest compression and ventilation at high airway pressure during cardiopulmonary resuscitation, Lancet, i (1980) 175-178. A. Oriol and H.J. Smith, Hemodynamic observations during cardiac massage, J. Can. Med. Assoc., 98 (1%8) 84. L.R.M. Guercio, R.P. Coomaraswarmy and D. State, Cardiac output and other hemodynamic variables during external cardiac massage in man, N. Engl. J. Med., (1963) 1398-1404. A. Gjedde, A.J. Hansen and E. Siemkowicz, Simultaneous determination of regional blood flow and blood brain glucose transfer in rat brain, Acts Physiol. Stand., 108 (1980) 321-330. A.M. Harper, General physiology of the cerebral circulation, Int. Anesthesiol. Clin., 7 (1969) 473506. C. Eintrei, W. Leszniewski and C. Carlsson. Local application of 133 Xenon for measurement of regional and isoflurane anesthesia in humans, Anesthesiology, 63 (1985) 391-394. J. Astrup, Energy-requiring cell functions in the ischemic bram, J. Neurosurg., 56 (1982) 482497. E. Siemkowicz, 1. Christiansen and S.C. Sorensen, Changes in cisternal fluid potassium concentration following cardiac arrest, Acts Neurol. Stand., 55 (1977) 173. B.K. Siesjo. Cell damage in the bram: A speculative synthesis, J. Cereb. Blood Flow Metabol., 1 (1981) 155-185. B.B. Taylor, CC. Brown, T. Bridges et al., A model for regional blood flow measurements during cardiopulmonary resuscitation in a swine model, Resuscitation, 16 (1988) 107-118.
123 14 15 16 17
18 19 20
L.M.R. Guercio, R. Coomenaswary and D. State, Cardiac output and other neurodynamic variables during external massage in man, N. Engl. J. Med., 269 (1%3) 1398-1401. G.S. Krause. K. Kumar. B.C. White et al., Ischemia. resuscitation and reperfussion; Mechanisms of tissue typing and prospects for protection, Am. Health J., April (1986) 768-780. L.M.R. Guercio, Open chest cardiac massage: An overview, Resuscitation, 15 (1987) 9-11. K.H. Lindner, F.W. Ahnefeld and I.M. Boudler, The effect of epinephrine on hemodynamics, acid bone status and potassium during spontaneous circulation and cardiopulmonary resuscitation, Resuscitation. 16 (1988) 251-261. D.K. Alifimoff, Open versus closed chest cardiac massage in non-traumatic cardiac arrest, Resuscitation, IS (1987) 13-21. E. Kogstrom, M.L. Smith and B.K. Siesjo, Local cerebral blood flow in the recovery period following complete cerebral ischemia in the rat, J. Cereb. Blood Flow Metab., 3(1983) 170-182. E. Siemkowicz, Cerebrovascular resistance in ischemia, Arch. Pflug., 388 (1980) 243-247.