Hypertension With Hemodilution Prevents Multifocal Cerebral Hypoperfusion After Cardiac Arrest in Dogs

Transcription

1 45 Hypertension With Hemodilution Prevents Multifocal Cerebral Hypoperfusion After Cardiac Arrest in Dogs Yuval Leonov, MD; Fritz Sterz, MD; Peter Safar, MD; David W. Johnson, MD; Samuel A. Tisherman, MD; and Ken-Ichi Oku, MD We postulate that achieving optimal neurological recovery after cardiac arrest requires a multifaceted treatment to mitigate Negovsky's postresuscitation syndrome, which we believe consists of four components: 1) perfusion failure, 2) reoxygenation cascades leading to lipid peroxidation, 3) self-intoxication from postischemic extracerebral organs, and 4) blood dyscrasias from stasis. This study concerns the first component, perfusion failure, which consists of four stages: 1) no reflow, 9 2) transient hyperemia, 3) protracted hypoperfusion, 10 and 4) resolution. The mechanism of postarrest cerebral hypoperfusion is not known. Probably it is multifactorial, including vasospasm, tissue edema, and cell aggregates. In healthy dogs, a reduction of hematocrit (Hct) to 25% does not reduce systemic arterial O 2 transport, because of compensatory increase in cardiac output. 11 Hemodilution or hypertension can improve microcirculatory blood flow and O 2 delivery in extra- Background: Improved neurological outcome with postarrest hypertensive hemodilution in an earlier study could be the result of more homogeneous cerebral perfusion and improved O 2 delivery. We explored global, regional, and local cerebral blood flow by stable xenon-enhanced computed tomography and global cerebral metabolism in our dog cardiac arrest model. Methods: Ventricular fibrillation cardiac arrest of 12.5 minutes was reversed by brief cardiopulmonary bypass, followed by life support to 4 hours postarrest We compared control group I (n=5; mean arterial blood pressure, mm Hg; hematocrit, ^35%) with immediately postarrest reflow-promoted group II (n=5; mean arterial blood pressure, mm Hg; hypervolemic hemodilution with plasma substitute to hematocrit, -25%). Results: After initial hyperemia in both groups, during the "delayed hypoperfusion phase" at 1-4 hours postarrest, global cerebral blood flow was 51-60% of baseline in group I versus 85-% of baseline in group II (/><6.01). Percentages of brain tissue voxels with no flow, trickle flow, or low flow were lower (/?<0.01) and mean regional cerebral blood flow values were higher in group II (p<0.0l). Global cerebral oxygen uptake recovered to near baseline values at 3-4 hours postarrest in both groups. Postarrest arterial O 2 content, however, in hemodiluted group II was -50% of that in group I. Thus, the O 2 uptake/delivery ratio was increased (worsened) in both groups at 2-4 hours postarrest. Conclusions: After prolonged cardiac arrest, immediately induced moderate hypertensive hemodilution to hematocrit -25% can normalize cerebral blood flow patterns (improve homogeneity of cerebral perfusion), but does not improve cerebral O 2 delivery, since the flow benefit is offset by decreased arterial O 2 content Individualized titration of hematocrit or hemodilution with acellular O 2 carrying blood substitute (stroma-free hemoglobin orfluorocarbonsolution) would be required to improve O 2 uptake/delivery ratio. (Stroke 1992^3:45-53) We previously reported that after cardiac arrest of 12 minutes (no blood flow) in dogs, immediate postarrest hypertension with hemodilution improves cerebral outcome. 12 Since 1985, we have studied cerebral blood flow (CBF) with the stable xenon-enhanced computed tomography (CT) method 3 in our dog models of cardiac arrest. 4 " 8 From the International Resuscitation Research Center (Y.L., F.S., P.S-, S.A.T., K.O.), the Departments of Anesthesiology/ Critical Care Medicine (Y.L., F.S., P.S., K.O.), Radiology (DJ.), and Surgery (S.A.T.), and Presbyterian-University Hospital, University of Pittsburgh School of Medicine, Pittsburgh, Pa. Supported by National Institutes of Health grant NS and the Asmund S. Laerdal Foundation. The Schroedinger Foundation of Austria and the Fulbright Foundation supported F. Sterz. Address for correspondence: Peter Safar, MD, International Resuscitation Research Center, University of Pittsburgh, 3434 Fifth Avenue, Pittsburgh, PA Received May 16, 1991; accepted August, 1991.

2 46 Stroke Vol 23, No 1 January 1992 cerebral organs. The results of several acute animal studies on global or focal brain ischemia by others concerning benefit from postischemic hypertension, or hemodilution, are consistent with our findings in cardiac arrest outcome studies. 1-2 We hypothesize that increasing perfusion pressure and decreasing blood viscosity postarrest can mitigate the cerebral hypoperfusion and thereby improve brain O 2 uptake/supply relationships and outcome. In this study, 6 we used our established dog model to monitor for the first time the effect of postarrest hypertensive hemodilution on global, regional, and local CBF to 4 hours.* In exploratory subgroups, we also studied global cerebral oxygen uptake (CMRO 2 ) in relation to cerebral arterial O 2 delivery. Materials and Methods This project was approved by the Animal Use Committee of the University of Pittsburgh (Pa.). We used 10 healthy, custom-bred male coon hounds from the same breeding colony, mean age 10 (range, 8-12) months and mean weight 21 (range, 16-30) kg. Control group I (n=5) also served other studies It was compared with a reflow-promoted group (n=5) with immediate hypertension plus delayed hypervolemic hemodilution. Both groups were studied in by the same team over 3 weeks in randomized sequence. Placebo control of treatment was not possible. Anesthesia was induced with ketamine (10 mg/kg i.m.) and maintained with 50:50% N 2 O/O 2 plus % halothane by endotracheal intermittent positive-pressure ventilation, with paralysis by 0.2 mg/kg i.v. pancuronium as needed, and controlled normoxia and normocarbia. Sterile cutdowns were performed for monitoring and cardiopulmonary bypass Control of extracerebral variables prearrest and postarrest, from start of reperfusion with bypass of <5 minutes to resuscitation time at 4 hours, were the same in both groups. We continuously monitored the electrocardiogram, heart rate, mean arterial blood pressure (MABP), central venous pressure, end-tidal CO^ and central venous (core) temperature. We intermittently monitored arterial blood gases and Hct. Ringer's solution without glucose was infused intravenously. We controlled MABP at ±10 mmhg (mean±sd) by adjusting the halothane concentration before and by using norepinephrine or trimethaphan after cardiac arrest, central venous pressure at 5-15 mm Hg, Pac^ at ^ mm Hg, PacOz at mm Hg, base excess at ±7 meq/1, blood glucose concentration at -175 mg/dl before arrest, and central venous temperature at 37.5±0.5 c C. Cerebral (tympanic membrane) T, not monitored here because it would interfere with CT, was equal to central venous temperature in other experiments. 22 In preparation for the insult and during paralysis with pancuronium and intermittent positive-pres- Tne data, too extensive for publication here, are available upon request to interested investigators, by writing directly to Dr. Safar. sure ventilation, inhalation anesthesia was reduced with % O 2 for 1 minute, followed by room air for 4 minutes. Ventricular fibrillation was induced by external transthoracic AC electric shock. At the end of ventricular fibrillation (no blood flow) of 12.5 minutes, resuscitation was begun with intra-arterial epinephrine and cardiopulmonary bypass (resuscitation time, 0 minutes). Bypass was primed with 0 ml dextran in 0.9% NaCl and Ringer's solution, 50:50. After 2 minutes of bypass at ml/kg/min, external countershocks were started and repeated until defibrillation and restoration of spontaneous heartbeat. All dogs were weaned from bypass before 5 minutes. Intermittent positive-pressure ventilation and control of physiological variables (above) were continued to 4 hours. Cardiopulmonary bypass did not produce motion artifacts for CBF determinations, 5 as did external cardiopulmonary resuscitation. 4 After arrest, intermittent positive-pressure ventilation was continued until resuscitation time 4 hours, with Pacc^ controlled at mm Hg, PaOj at > mm Hg, central venous temperature at 37.5±0.5 C, and Hct as prescribed below. In control group I, cardiopulmonary bypass during ventricular fibrillation was to generate a MABP of mm Hg by the initial epinephrine dose of mg/kg i.a., and additional doses if necessary. In group II, cardiopulmonary bypass was to generate a MABP of 1 mm Hg by an initial epinephrine dose of mg/kg i.a., and additional doses if necessary. Immediately after restoration of spontaneous heartbeat, transient MABP peaks were expected to vary between dogs in both groups, some to exceed the above levels. Mean arterial blood pressure was then to be controlled with a titrated intravenous infusion of norepinephrine: in group I at mm Hg throughout, and in group II at 1 mm Hg for the first hour, 130 for the second, 1 for the third, and 110 for the fourth hour. In both groups, the bypass priming volume induced slight hemodilution (to Hct approximately 30%) for the first 5 minutes of reperfusion. In group I at resuscitation time 5 minutes, the bypass blood (about 15 ml/kg) was returned intravenously, and Hct was restored to above 35%. In group II, the bypass blood was not returned, and additional hypervolemic hemodilution was performed by bleeding arterially 35 ml/kg and simultaneously infusing 50 ml/kg i.v. dextran in 0.9% NaCl. This resulted in a Hct of about -25% until resuscitation time 2 hours, when the shed blood was reinfused. Cerebral bloodflowwas determined with the novel noninvasive stable xenon-enhanced CT method, 3 adapted to cardiac arrest experiments with dogs. 4 " 8 All CBF values were obtained from a computerized analysis of CT voxel density changes during inhaled xenon washin over 4.5 minutes and are expressed as milliliters per cubic centimeters (or gram) brain tissue per minute. The xenon washin technique is based on the Fick principle and the Kety-Schmidt equation, using the CT scanner as a densitometer. Six xenon-enhanced images were obtained, beginning

3 Leonov et al CBF With Hypertensive Hemodilution After Cardiac Arrest seconds after the start of xenon inhalation. Constancy of the CT slices was ascertained from the consistent position of bony landmarks. Resolution power in monkeys and humans seems to be 5x5x5=125 mm 3 regions With the dog's head rigidly immobilized, we studied global CBF for the tissue volumes of two 5-mm-thick coronal slices located 10 mm apart. The anterior slice included the hippocampus and thalamus, and the posterior slice included the brain stem. Computer programs provided one CBF value for each voxel of 1x1x5=5 mm 3. The CBF values of all voxels in each slice and in each selected anatomic region (-0 mm 3 volume for each region) were averaged to determine global and regional CBF, respectively. We calculated local CBF as the percentage of voxels of each CT slice (percentage area) having specific flow ranges 5 : no flow, 0-5 ml/ cm 3 /min; trickle flow, 6-10 ml/ cmvmin; low flow, 11- ml/ cm 3 /min; normal flow for white matter, 21- ml/ cm 3 /min; normal flow for gray matter, 41-1 ml/ cmvmin; and hyperemic flow, >1 ml/ cm3/min. We arbitrarily considered gray matter with a local CBF of < ml/ cm3/min as having "inadequate CBF." Because it interferes with xenon monitoring, N 2 O was replaced by N 2 and the analgesic effect was replaced by fentanyl given as a continuous intravenous infusion of 10 /ig/kg/hr, starting 1-2 hours before baseline CBF. About 10 minutes before starting xenon washin of the baseline CBF, halothane (<0.5%) was discontinued. No halothane was given during CBF measurements before cardiac arrest, and none at all after cardiac arrest. The inhaled gas delivered contained 67% 02/33% N 2, which was switched to 33% Xe/67% O 2 during CBF measurements. Two baseline CBF studies were performed before cardiac arrest. Additional CBF studies were performed at 10 and 30 minutes and 1, 2, 3, and 4 hours after reperfusion. Before each CBF measurement, MABP, PaO2, Paco2, arterial ph, base excess, Hct, and central venous temperature were controlled. Two sham experiments without cardiac arrest reported in another article 5 revealed stable reproducible global, regional, and local CBF values over time, autoregulation, and responsivity of CBF to lowered The CMRO 2 was determined in three dogs of each group. Through a sterile 2-cm craniotomy, a PE 50 nonocclusive catheter was inserted into the sagittal sinus with the catheter tip placed 1 cm rostral to the confluence of the sinuses. For each CMRO 2 determination, the catheter deadspace was cleared and arterial and sagittal sinus blood samples were drawn at a constant slow rate into heparinized syringes, cooled, and analyzed within 2 hours for O 2 content, using a Co-oximeter. Samples for arterial and sagittal sinus O 2 contents (Cac^ and Cssc^, respectively) were taken just before each xenon inhalation without halothane. Global CMRO 2 was calculated as the global CBF of the posterior CT slice times the CaOj content gradient (i.e., global f The cerebral O 2 utilization coefficient was calculated as global CMRO 2 /arterial O 2 delivery (i.e., global CBFx[Cao 2 -Csso 2 ]/global CBFxCao 2 ; i.e., [CaOz-CssOjJ/CaOj). Although global CBF was calculated from only one CT slice and CaOj-Cssc^ from the entire cerebrum, relative changes of CMRO 2 were considered valid. Data were analyzed for group differences within times and group interaction over time using a univariate repeated-measures analysis of variance. Scheffe's post hoc procedure was used for analyzing changes within each group over time (baseline versus postarrest). The CMRO 2 data were not statistically analyzed because of small numbers. Results are reported as mean±sd. Results All 10 dogs followed protocol to 4 hours postarrest. Before cardiac arrest there was no statistically significant difference between groups in central venous temperature, arterial ph, PacOz, PaOz, Hct, MABP, and central venous pressure. At 10 minutes postarrest, hemodilution fluid at room temperature in group II caused mean central venous temperature to decrease from 37.5 C (baseline) to 36.2 C. Spontaneous normotension was restored within a mean of 2 minutes and 37 seconds in group I, and a mean of 1 minute and 50 seconds in group II (NS). In group I, mild hemodilution from Hct 42±3% at baseline to 30±8% at 10 minutes postarrest, caused by cardiopulmonary bypass, was transient; Hct was 34 ±4% at 2 hours and 39 ±4% at 4 hours. In group II, Hct was reduced according to protocol from 45 ±3% at baseline to 22±1% at 10 minutes postarrest; Hct was 19±2% at 30 minutes, 17±1% at 1 hour and 22±2% at 4 hours (p<0.01). The hypertensive bout, immediately after restoration of spontaneous heartbeat, varied in both groups. According to protocol, it was higher in group II ( mmhg; range, 1-250) than group I (112±24 mmhg; range, 90-0) (/?<0.01). At 30 minutes postarrest, MABP was 158 ±9 mm Hg in group II versus 110±12 in group I (p<0.01); thereafter, the difference was only numerical (NS). Central venous pressure was 3±1 mmhg prearrest in group I and 3±3 in group II; central venous pressure was 14 ±2 mm Hg at 10 minutes postarrest in group I versus 13±7 in group II; and at 4 hours, central venous pressure was 5±3 mm Hg in group I versus 8±4 in group II (NS). In each dog, the two baseline global CBF measurements were similar (approximately 50 ml/ cm 3 / min); there was a small SD within each group, and no significant difference between groups (Table 1 and Figure 1). Baseline regional CBF values (Figure 2) also were similar in both groups. Baseline local CBF values were also the same between groups, with essentially no voxels with local CBF <10 ml/ cnrvmin (Figure 1).

4 48 Stroke Vol 23, No 1 January 1992 TABLE 1. Global (Slice) Cerebral Blood Flow Before and After Cardiac Arrest in Dogs Control group I (dogs 13, 16,, 33, 34) gcbf' % of baseline Hypertensive hemodilution group II (dogs 15, 19, 24, 28, 29) gcbf* % of baseline Before arrest (-30 Min) 47±6 43 ±5 10 Min 106± ± Min 48± ±9 87 1Hr 28± ±7 90 After arrest 2Hr 26±3 56 ±8 93 3Hr 24± ±6 4Hr 26± ± Hr 26±0 55 ±7» 92* Values are mean±sd in milliliters per cm 3 per minute for posterior coronal slices 5 mm thick. Values from anterior slices were similar. gcbf, global cerebral blood flow. *A11 values at 1-4 hours postarrest are different between groups (p<0.01). During the initial transient hyperemia, neither group showed evidence of sustained no reflow foci. In both groups, hyperemia was more pronounced and lasted longer in midbrain and thalamus and less in neocortex, hippocampus, and white matter. During the hypoperfusion phase at 1-4 hours postarrest (Table 1 and Figure 1), mean global CBF was 51-60% of baseline in group I versus 85-% of baseline in group II (p<0.01). In group I (Reference 7, Table 2), the first postarrest CBF measurement (at 3-22 minutes after restoration of heart beat) showed diffuse hyperemia with global CBF of about two times baseline (Table 1) and no voxels with CBF < ml/ cnrvmin (Figure 1A). During the hypoperfusion phase at 1-4 hours postarrest, mean global CBF was 55% of baseline. The percentage of voxels with a local CBF of 0-10 ml/ cnrvmin (no flow or trickle flow) was about 15%, whereas % of voxels had a local CBF of 0- ml (worse than baseline) (p<0.0l), 44% had a local CBF of - ml, and 16% had a local CBF of -1 ml/ cm 3 /min. In group II, the first postarrest CBF measurement (at 8- minutes after restoration of heart beat) showed the same diffuse hyperemia with global CBF of about two to three times baseline (Table 1) and also with no voxels with CBF < ml/ cmvmin (Figure IB). During the hypoperfusion phase at 1-4 hours postarrest, however, global and local CBF values were significantly higher than in group I (p<0.01) and not significantly different from group II baseline values. At 1-4 hours postarrest, mean global CBF was 92% baseline (versus 55% in group I) (Table 1). The percentage of voxels with a local CBF of 0-10 ml/ cm 3 /min was 4% (versus 15% in group I), while 15% of voxels had a local CBF of 0- ml (versus % in group I), % had a local CBF of - ml (similar to the 44% in group I); and 44% had a local CBF of -1 ml/ cm 3 /min (versus 16% in group I). After blood was reinfused 2 hours postarrest, there was no worsening of CBF patterns, except in dog 29, which had a slight decrease in global CBF and an increase in low-flow voxels at 4 hours postarrest (Figure IB). Regional CBF values for all regions studied revealed the same flow advantages for group II as did global CBF values (/xo.ol; for hippocampus p<0.05) (Figure 2). All regional CBF values (except for the white matter in group I) were > ml/ cm 3 /min. Throughout 1-4 hours postarrest, all regional CBF values for gray matter in both groups were > ml/ cnrvmin. In group II at 1-4 hours postarrest, all CBF values of predominantly gray matter regions were significantly higher than in group I (p<0.01) and not significantly different from group II baseline values. The CaO2 and CssO2 values (Table 2) varied considerably between animals, but followed a consistent pattern within each dog over time. During the first postarrest CBF measurement ("10 minutes postarrest"), Cac>2 decreased as a result of transient hemodilution by cardiopulmonary bypass. At 1-4 hours postarrest, with group II further hemodiluted by protocol, CaO2 in group II was about one half of that in group I. The baseline CaO2-CssO2 was 8.0±2.7 ml/dl in group I and 9.0±2.1 in group II; after transient decrease early postarrest, it increased (worsened) in group I to 12.2±2.3 at 4 hours and remained near baseline values of 7.5 ±1.3 at 4 hours in group II. The calculated O 2 utilization coefficient also increased (worsened) similarly in both groups, from 0.4 ±0.1 prearrest to 0.7 ±0.1 at 3-4 hours postarrest. Thus, in group II the higher CBF was offset by the reduced CaO2- Arterial O 2 delivery (global CBFxCao 2 ) (Table 2) postarrest was approximately the same in both groups, since in the hemodiluted dogs (group II) the normalized global CBF was offset by a CaOj of 60% of that in the control dogs (group I). The CMRO 2 values (Table 2) prearrest in both groups combined ranged between 2.3 and 4.5 ml/ cnrvmin. Although global CBF at 1-4 hours postarrest was about 50% of baseline values in group I and similar to baseline values in group II (Table 1), CMRO 2 values were similar in both groups at 1-4 hours (Table 2), recovering from very low levels immediately postarrest to near baseline values 4 hours

5 Leonov et al CBF With Hypertensive Hemodilution After Cardiac Arrest 49 (A): DOG CT16 (NORMOTENSION) % VOXELS gcbf ml/g/min FLOW RANOE8 m m > h MAP 95 PaCO HCT 39-30min min min min h h h h (B): DOG CT29 (HYPERTENSION & HEMODILUTION) % VOXELS uu H FLOWRANOES nil > h MAP PaCO2 31 HCT 43 ff o>:':'>:':-:o:*:o -30min min min h h30min h h gcbf ml/g/min FIGURE 1. Two ventricularfibrillationcardiac arrest (VFCA) experiments to resuscitation time 4 hours. Dots are global cerebral blood flow (gcbf) of posterior computed tomographic (CT) slice. At same times as for gcbf, local CBF ranges in percent of voxels is quantitatively presented. Prearrest, there were essentially no voxels with local CBF <10 ml/cm 3 /min. Panel A: Representative dog (CT 16) of control group I. Postarrest normotension and normal hematocrit (Hct). ThefirstCBF determination postarrest was at resuscitation time 3 minutes, immediately after defibrillation and restoration of spontaneous circulation; gcbf was about two times baseline with increased hyperemic voxels and no voxels with local CBF < ml/ cm 3 /min. Up to % of voxels with no flow or trickle flow and 50% of voxels with < ml low flow, which were not present prearrest, appeared at resuscitation time 25 minutes. At resuscitation time 1-4 hours, there was continued gcbf at about 50% of baseline and increased percent voxels with low flow ranges. Panel B: Representative dog (CT 29) of reflow-promoted group II. Postarrest hypertension and hemodilution. Baseline values are the same as control dog (panel A). At resuscitation time 10 minutes, hyperemia with gcbf was two times baseline. At resuscitation time 30 minutes, CBF returned to baseline values. At resuscitation time 1-4 hours, gcbf and local CBF patterns were the same as baseline. There was no postischemic hypoperfusion. No local CBF voxels in no-flow, trickle-flow, or low-flow ranges were in excess of baseline values. At resuscitation time 4 hours, reinfusion of shed blood caused a reduction ofgcbf and increase in low-flow voxels. 4h postarrest. Postarrest CMRO 2 values were similar in both groups, ranging between 1.4 and 2.1 ml/ cm 3 /min at 1 hour and between 2.2 and 4.2 at 4 hours postarrest. There was no evidence of postarrest hypermetabolism in either group. Discussion In this study of post-cardiac arrest CBF reduction to 50% baseline with standard life support to 4 hours, we achieved normalization of global, regional, and

6 50 Stroke Vol 23, No 1 January 1992 rotic flush with plasma substitute, or both. A similar intracarotid flush occurred in both groups in this study and in our previous outcome study with the same cardiopulmonary bypass model. 2 The dog model and the xenon-enhanced CT method are discussed in detail elsewhere. Because of the computational and physical limitations of the xenon-enhanced CT method, the local CBF data on percentage of voxels in slice area with no-flow and trickle-flow ranges are only suggestive. The system noise and computational and physical limitations of the method are amplified when CBF is low. For regions smaller than 5x5x5 mm 3, CBF values of <10 ml/ cm3/min might be in part accounted for by methodological error Available evidence on the method, however, suggests that the regional CBF values in the range of 10- ml/ cmvmin are accurate. In previous studies 4-5 we also found early postarrest normal global CBF accompanied by increased percent of trickle flow voxels. The rationale for inducing hypertension includes the likelihood that the initial no-reflow phenomenon 9 is in part caused by capillary compression and increased blood viscosity, 25 both of which require initial high perfusion pressure to ameliorate. Outcome benefit from an induced hypertensive bout early postarrest had experimental support not only from our studies, 1-2 but from others. 12 " 15 When in an additional pilot experiment (dog 14), after normotensive reper- TABLE 2. Cerebral Metabolic Variables Before and After Cardiac Arrest in Dogs Arterial O2 content (ml/dl) Group I Group II Sagittal sinus Oj content (ml/dl) Group I Group II Cerebral arterial O 2 delivery (ml/ cnrvmin) Group I Group II Cerebral O 2 uptake (ml/ cnrvmin) (i.e., gcmrch) Before arrest After arrest (-30 Min) 30 Min 1 Hr 2 Hr 3Hr 4Hr.2± ± ± ± ± ±1.6 (n2: 16.2, 15.5) (n2: 15.0, 18.3) 21.5± ± ± ± ± ±0.8 (n2: 10.1, 9.7) (n2: 9.7, 7.1) 12.2± ± ± ± ± ±1.6 (n2: 9.0, 13.1) (n2: 5.1, 8, 6) 12.1± ± ± ± ± ±0.6 (n2: 7.7, 0.3) (n2: 2.2, 1.9) 8.7± ± ±03 4.2± (n2: 6.8, 9.4) (n2: 4.2, 4.2) 8.6± ± ± ± ± ±0.1 (n2: 2.8, 4.6) (n2: 5.6, 3.1) 33± ± ± ± ± ±0.8 Group I (n2: 3.0, 1.4) (n2: 2.7, 2.2) 3.5± ± ± ± ± ±0.2 Group II (n2: 2.7, 4.3) (n2: 0.6, 1.1) (n2: 4.3, 2.3) Values are mean±sd. Group I, control dogs (nos., 33, 34); Group II, dogs treated postarrest with hypertension and hemodilution (Hct -25%) (nos. 24, 28, 29); n2, values for two dogs. Statistical analysis not appropriate (numbers per group too small). local (multifocal) CBF values postarrest with a combination of immediately postarrest hypertension plus hypervolemic hemodilution from Hct % to %. In the 5-30 minutes postarrest CBF determinations there was no sign of protracted multifocal no reflow, with or without hypertensive hemodilution. Exploratory monitoring of CMRO 2 and related variables revealed that the flow-promoting treatment studied did not result in increased arterial O 2 delivery, because hemodilution reduced CaO2 to less than 50% baseline. This offset the increase in CBF. We had found the same combination treatment in the same model to improve cerebral outcome 2 ; an early hypertensive bout seemed more responsible for the improved outcome than the hemodilution, which gave only numerical benefit. Hypertension then might have improved outcome not via protracted improvement in O 2 delivery, but perhaps by overcoming the initial no reflow phenomenon. 9 The remaining uncertainty about the beneficial effect of moderate hemodilution to Hct -25% after cardiac arrest might also explain why clinical hemodilution after acute stroke has not resulted in improvement of outcome, although CBF was improved. 18 In our first cardiac arrest outcome study in dogs, done in and using a similar model but with external cardiopulmonary resuscitation, the beneficial effect of hypertensive hemodilution plus heparinization could have been the result of a hypertensive bout, the initial intraca-

7 Leonov et al CBF With Hypertensive HemodiluUon After Cardiac Arrest 51 i, GROUP *\ CORTEX \rt L i i i VFCA HIPPOCAMPUS AT J < * XX WHITE MATTER MIDBRAIN r- THALAMU8-00* -ar 10min 30mln 1h 2h 3h 4h fusion, we transiently induced hypertension (MABP, 190 mm Hg) without hemodilution at 1 hour postarrest, we found only transient mitigation of reduced CBF. Immediate postarrest hypotension resulted in more low-flow voxels during the hyperemic phase. 4-5 In patients, immediate postarrest hypotension correlated with worse cerebral outcome. 26 The effect of epinephrine-induced immediate postarrest hypertension on cerebral outcome in patients is now being investigated with our multicenter clinical cardiac arrest study (N. Abramson, P. Safar, K. Detre, unpublished observations). The rationale for inducing hemodilution after cardiac arrest is based on the Hagen-Poiseuille equation: Blood flow is inversely proportional to blood viscosity. Blood viscosity increases in the microcirculation during stasis. 16 Hematocrit is the main factor influencing blood viscosity, particularly in the microcirculation. During initial transient hyperemia, we saw no evidence of a persistent no-reflow phenomenon with or without hypertensive hemodilution. Both FIGURE 2. Graph of regional cerebral blood flow (ordinate) versus time before and after ventricular fibrillation cardiac arrest (VFCA) (abscissa) in control dogs (group I, n=5) and hypertension plus hemodilution-treated dogs (group II, n=4). Values of each region (-0 mm 3 tissue volume) were calculated from local cerebral blood flow values of 1x1x5=5 mm 3 voxels. All regional cerebral blood flow values of group I followed hyperemia-hypoperfusion pattern of global cerebral blood flow (Figure 1 and Table 1). There were significant group differences in regional cerebral blood flow values in cortex, white matter, midbrain and thalamus (p<0.01), and hippocampus (p<0.05). In group II, all postarrest regional cerebral blood flow values were the same as baseline values. Postarrest, regional cerebral blood flow values < ml/ cm 3 /min only in white matter of control (group I). hypertension and hemodilution might increase microcirculatory blood flow if hypoperfusion postarrest is caused by vasospasm, blood sludging, or cell aggregates The possibility of vasospasm as an important factor in the protracted cerebral postarrest hypoperfusion phase is supported by improved outcome with use of calcium entry Mockers The drugs and fluids used in group II could influence CBF and CMRO 2, 31 irrespective of MABP and Hct. These treatments, however, could not be provided clinically without drugs and fluids. The rationale for improving cerebral arterial O 2 delivery during the postischemic hypoperfusion phase is based on our earlier observation of mismatched CBF to CMRO In the present study, the baseline CMRO 2 values of ml/ cm V min were low normal. The low halothane concentration of % used before, but not during, prearrest CBF measurements may have slightly increased CBF; higher concentrations seem required to reduce CMRO In our study, CMRO 2 returned from very low early postarrest

8 52 Stroke Vol 23, No 1 January 1992 levels to baseline levels by 3-4 hours postarrest, while CBF was still 50% baseline (Table 2). This delayed CMR0 2 to CBF mismatching was missed by Michenfelder and Milde 32 whose observations to only 90 minutes postarrest suggest matching of CBF to CMR0 2. The very low Csso 2 values in our study and very low PSSO2 values in another study 8 also suggest an O 2 demand/supply uncoupling several hours postarrest. More work is needed to resolve this question. For improving O 2 delivery, the hemodilution to Hct % in this study was apparently too severe, since the associated reduction of Cao2 to almost one hah0 baseline offset the improvement in CBF and left arterial O 2 delivery unchanged (Table 2). For clinical application, both MABP and Hct manipulations will have to be titrated against continuously or frequently monitored arterial and cerebral venous (superior jugular bulb) O 2 values. Once postarrest hypoperfusion was established, lowering Hct seemed more effective than hypertension for improving CBF. We conducted pilot experiments at 4-5 hours postarrest. When in four dogs (CT Nos. 9,13,14, and 17) we increased MABP from between 110 and 125 mm Hg to between 160 and 190 mm Hg at 4-5 hours postarrest, global CBF and percent low-flow voxels remained essentially unchanged. When in another four dogs (CT Nos. 22,25, 26, and 27), we lowered Hct from between 35% and 50% to 25% at 4-5 hours postarrest, global CBF increased from -27 ml/ cm 3 /min to ml/ cm 3 /min after hemodilution, which decreased the percentage of voxels with local CBF <10 ml/ cm3/min in these four dogs from 5% to 3%, 15% to 6%, 11% to 2%, and 22% to 4% of voxels, respectively. In other pilot experiments at 4 hours postarrest, we tested acetazolamide (dogs CT 13 and 15) or hypercarbia of PacOj mm Hg (dogs CT 14,17, and 18); both treatments increased global and local CBF transiently. Such acidification, however, could impair chemical recovery of neurons. These observations suggest that the delayed protracted hypoperfusion is not caused by a fixed obstruction, such as edema or thrombosis, but perhaps by mediatorinduced vasospasm or transient blood cell aggregates that yield to flow-promoting manipulations. We conclude that after prolonged cardiac arrest, the combination of immediate hypertension plus hypervolemic hemodilution from Hct 45% to - 25% (a rather aggressive therapy), which improved outcome in previous studies, 1-2 can normalize the otherwise reduced global, regional, and local (multifocal) CBF. This degree of hemodilution, however, does not improve the O 2 uptake/delivery ratio postarrest, because it reduces arterial O 2 content to less than 50% of baseline, which offsets the positive flow effect. Improving the O 2 uptake/delivery ratio would require individualized titration of MABP and Hct to optimize arterial-cerebrovenous O 2 content differences. Postarrest improvement of cerebral O 2 uptake/delivery ratio might be achieved by reducing Hct with an acellular O 2 carrying blood substitute, such as a stroma-free hemoglobin or fluorocarbon solution. Acknowledgments Richard Latchaw, MD; Stephen Hecht, MD; Robert Tarr, MD; and Ann Radovsky, DVM, PhD, helped with local CBF imaging. Walter Obrist, PhD; David Gur, PhD; Sidney Wolfson, MD; Howard Yonas, MD; and Walter Good, PhD, provided valuable advice on CBF calculations. Janine E. Janosky, PhD, helped with statistical analysis. S. William Stezoski and Henry Alexander gave technical assistance. Alan Abraham, Scott Nagel, Annette Pastula, and Mark Ritchey helped with the experiments. Teresa Adams, Pat Colorito, Franca Defelice, Vikie Mavrakis, William Wigginton, and Donna Wojciechowski were radiotechnologists. Lisa Cohn provided editorial assistance. Gale Foster helped prepare the manuscript. References 1. Safar P, Stezoski SW, Nemoto EM: Amelioration of brain damage after 12 minutes cardiac arrest in dogs. Arch Neural 1976^3: Sterz F, Leonov Y, Safar P, Tisherman SA, Oku K, Radovsky A, Alexander H: Hypertension with or without hemodilution for reperfusion after cardiac arrest: Effect on neurologic outcome in dogs. Stroke 1990;21: Gur D, Good WF, Wolfson SK, Yonas H, Shabason L: In vivo mapping of local cerebral blood flow by xenon-enhanced computed tomography. Science 1982;215: Wolfson SK, Safar P, Reich H, Clark J, Gur D: Multifocal, dynamic cerebral hypoperfusion after prolonged cardiac arrest in dogs, measured by the stable xenon-ct technique (abstract). Crit Care Med 1988;16: Safar P, Sterz F, Leonov Y, Johnson D, Oku K, Tisherman SA, Latchaw R, Obrist W, Stezoski SW, Janosky JE: Multifocal cerebral blood flow by stable xenon-computed tomography and global cerebral metabolism, after cardiac arrest in dogs (abstract). / Cereb Blood Flow Metab 1991;ll(suppl 2):S3 6. Sterz F, Safar P, Leonov Y, Johnson D, Latchaw R, Hecht S, Oku K: Cerebral multifocal hypoperfusion after cardiac arrest in dogs, mitigated by hypertension and hemodilution (abstract). Ann EmergMed 1989;18: Sterz F, Safar P, Johnson D, Oku K, Tisherman SA: Effects of U706F on multifocal cerebral blood flow and metabolism after cardiac arrest in dogs. Stroke 1991;22: Oku K, Kuboyama K, Safar P, Johnson D, Sterz F, Obrist W, Leonov Y, Tisherman S: Multifocal cerebral blood flow (CBF) and global metabolism (CMR) after prolonged cardiac arrest in dogs: Effect of mild hypothermia (34 C). Anesthesiology 1990;73:A Ames A III, Wright RL, Kowada M, Thurston JM, Majno G: Cerebral ischemia: II. The no-reflow phenomenon. Am J Pathol 1968;152: Snyder JV, Nemoto EM, Carroll RG, Safar P: Global ischemia in dogs: Intracranial pressures, brain blood flow and metabolism. Stroke 1975;6: Takaori M, Safar P: Treatment of massive hemorrhage with colloid and crystalloid solutions. JAMA I967;199:297-3O2 12. Gaenshirt H, Zylka W: Ueberlebenszeit, Erholungslatenz und Elektrocorticograram des Warmbluetergehirns in ihrer Abhaengigkeit vom Blutdruck. PflugersArch 1952^56: Hossmann K-A: Resuscitation potentials after prolonged global cerebral ischemia in cats. Crit Care Med 1988;16: Nemoto EM, Erdman NW, Strong E, Rao G, Moossy J: Regional brain PO2 after global ischemia in monkeys: Evi-

9 Leonov et al CBF With Hypertensive Hemodilution After Cardiac Arrest 53 dence for regional differences in critical perfusion pressures. Stroke 1979;10: Cole DJ, Drummond JL, Shapiro HM, Hertzog RE, Braner FS: The effect of hypervolemic hemodilution with or without hypertension on cerebral bloodflowfollowing middle cerebral artery occlusion in rats anesthetized with isoflurane. Anesthesiology 1989;71: Hirsch H, Breuer M, Kunzel HP: Ueber die Bildung von Thrombocytenaggregaten und die Aenderung des Haematokrits durch komplette Gehirnischaemie. Dtsch Z Nervenheilk 1964;186: Chan R, Leniger-Follert E: Effect of isovolemic hemodilution on oxygen supply and electrocorticogram in cat brain during focal ischemia and in normal tissue. Int J Microcirc Clin Exp 1983;2: Heros RC, Korosue K: Hemodilution for cerebral ischemia. Stroke 1989;: Hossmann V, Hossmann K-A, Takagi S: Effect of intravascular platelet aggregation on blood recirculation following prolonged ischemia of the cat brain. J Neurol 19;222: Lin S-R: The effect of dextran and streptokinase on cerebral function and blood flow after cardiac arrest: An experimental study on the dog. Neuroradiology 1978;16: Safar P, Abramson NS, Angelos M, Cantadore R, Leonov Y, Levine R, Pretto E, Reich H, Sterz F, Stezoski SW, Tisherman S: Emergency cardiopulmonary bypass for resuscitation from prolonged cardiac arrest. Am J Emerg Med 1990;8: Leonov Y, Sterz F, Safar P, Radovsky A, Oku K, Tisherman S, Stezoski SW: Mild cerebral hypothermia during and after cardiac arrest improves neurologic outcome in dogs. / Cereb Blood Flow Metab 1990;10: Wolfson SK Jr, Clark J, Greenberg JH, Gur D, Yonas H, Brenner RP, Cook EE, Lordeon PA: Xenon-enhanced computed tomography compared with [ 14 C]iodoantipyrine for normal and low cerebral blood flow states in baboons. Stroke 1990;21: Johnson DW, Stringer WA, Marks MP, Yonas H, Good WF, Gur D: Stable xenon CT cerebral blood flow imaging: Rationale for and role in clinical decision making. Am J Neuroradiol 1991;12: Fischer EG, Ames A III, Hedley-Whyte ET, O'Gorman S: Reassessment of cerebral capillary changes in acute global ischemia and their relationship to the "no-reflow phenomenon." Stroke 1977;8: Brain Resuscitation Clinical Trial (BRCT) II Study Group: A randomized clinical study of a calcium-entry blocker (lidoflazine) in the treatment of comatose survivors of cardiac arrest. NEnglJMed 1991;324: Arfors K, Grogaard B, Gerdin B: Neutrophils and delayed hypoperfusion in cerebral cortex after cerebral ischemia in the rat. Fed Proc 1986;45: Schmidt-Schonbein GW: Capillary plugging by granulocytes and the no-reflow phenomenon in the microcirculation. Fed Proc 1987;46: Vaagenes P, Cantadore R, Safar P, Moossy J, Rao G, Diven W, Alexander H, Stezoski W: Amelioration of brain damage by lidoflazine after prolonged ventricular fibrillation cardiac arrest in dogs. Crit Care Med 1984;12: Steen PA, Gisvold SE, Milde JH, Newberg LA, Scheithauer BW, Lanier WL, Michenfelder JD: Nimodipine improves outcome when given after complete cerebral ischemia in primates. Anesthesiology 1985;62: McDowall DG: The effects of clinical concentrations of halothane on the bloodflowand oxygen uptake of the cerebral cortex. BrJAnaesth 1967;39: Michenfelder JD, Milde JH: Postischemic canine cerebral blood flow appears to be determined by cerebral metabolic needs. J Cereb Blood Flow Metab 1990;10:71-76 KEY WORDS dogs heart arrest hemodilution hypertension

10 Hypertension with hemodilution prevents multifocal cerebral hypoperfusion after cardiac arrest in dogs. Y Leonov, F Sterz, P Safar, D W Johnson, S A Tisherman and K Oku Stroke. 1992;23:45-53 doi: /01.STR Stroke is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX Copyright 1992 American Heart Association, Inc. All rights reserved. Print ISSN: Online ISSN: The online version of this article, along with updated information and services, is located on the World Wide Web at: published Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally in Stroke can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office. Once the online version of the published article for which permission is being requested is located, click Request Permissions in the middle column of the Web page under Services. Further information about this process is available in the Permissions and Rights Question and Answer document. Reprints: Information about reprints can be found online at: Subscriptions: Information about subscribing to Stroke is online at: