Brain hyperperfusion during cardiac operations

Brain hyperperfusion during cardiac operationsBrain hyperperfusion during cardiac operations
Cerebral blood flow measured in man by intra-arterial injection of xenon 133: Evidence suggestive of intraoperative microembolism
Cerebral blood flow (CBF) was measured by intra-arterial injection of xenon 133 in 29 patients during cardiac operations. Marked changes occurred in all patients. A normal and significant correlation with temperature and plasma Pco, (p < 0.01) support the reliability of the method. Mean CBF measured between sternotomy and the onset of extracorporeal circulation (ECC) was 38 mlllOO gm . min. The first minute of ECC was associated with a decrease in CBF in nine of 12 patients (p < 0.02-). During steady-state hypothermic ECC (temperature 29° C), CBF increased unexpectedly to 64 mlliOO gm . min (p < 0.01). Following rewarming during steady-state normothermic ECC, mean CBF decreased to 42 mlllOO gm . min with signs of impairment of cerebral autoregulation. Ten and 20 minutes after. termination of ECC, mean CBF was 40 and 41 mlllOO gm . min, respectively. Arterial Pco, was found to be important in regulating CBF. The cerebral autoregulation maintained CBF down to arterial pressures of around 55 mm Hg. Below this level, CBF was significantly correlated with perfusion pressure (p < 0.01). Multiple small emboli with a hyperemic border zone could cause a brain hyperperfusion, as seen in our patients during bypass. Measurements of CBF during ECC hold promise as a guide toward safer cardiac operations.
Leif Henriksen, M.D.,* Erik Hjelms, M.D.,** and Thimm Lindeburgh, M.D.,*** Copenhagen. Denmark
DeathS, neurologic deficits, and mental impairment are widely known complications following cardiac oper- ations. However, evidence is accumulating that brain damage follows most and maybe all cardiac proce- dures.!" Methods for assessing and quantifying brain damage during and after bypass are urgently needed. Very few methods exist that can measure brain function during the cardiac repair, and those available (e.g., electroencephalography, evoked potentials) are impre-
From Rigshospitalet (State University Hospital). Copenhagen. Den- mark.
Supported by a grant from The Johann and Hanne Weimann Foundation.
Received for publication Sept. 22. 1982.
Accepted for publication Jan. 25. 1983.
Address for reprints: Leif Henriksen. M.D.. Department of Neurolo- gy. Rigshospitalet, Blegdamsvej 9. DK-2100 Copenhagen. Den- mark.
*Department of Neurology.
202
cise and cannot be used to select the optimal bypass technique.
Cerebral blood flow (CBF) is normally maintained at a constant level within a wide range of perfusion pressures,' but during extracorporeal circulation (ECC), factors such as reduced blood pressure, hemodilution, nonpulsatile flow, hypothermia, and changes in blood gas tensions may influence this level. The assumption that cerebral perfusion and perfusion pressure are adequate is based on animal studies and empiric consid- erations, and to our knowledge no direct measurements of CBF in man during cardiac operations have been published.
Adequate CBF, being related to the functional level of brain tissue, is of major importance if the frequency of brain injuries is to be minimized. The present study measured CBF by intra-arterially injected xenon 133 during the different periods of the cardiac operation.
Patients and methods
CBF was studied in 29 adult patients undergoing cardiac operations for acquired or congenital heart
Volume 86
Number 2
August, 1983 CBF during cardiac operations 2 0 3
Table I. Cerebral blood flow (CBF) in 29 patients between sternotomy and extracorporeal circulation.
Temperature
Mean: 35.5" C Mean: 36.2" C Mean: 36.8° C Range: 35.0°-35.4° C Range: 35.7"-36.5" C Range: 36.6°-37.4" C n=8 n= 21 n = 15
Cerebral blood flow" (ml/lOO gm- min) Mean ± SO 19.3 ± 4.9 27.0 ± 10.9 43.7 ± 25.4 Range 11-26 14-52 18-107
Arterial Pco, (mm Hg) Mean ± SO 3\.6 ± 3.4 33.3 ± 4.5 39.6 ± 12 Range 27-36 25-43 19-61
Cerebral blood flowt (ml/100 grrr-min) Mean ± SO 25.4 ± 4.6 33.1 ± 13.3 40.7 ± 15.0 Range 15-30 18-72 26-83
Mean arterial pressure (mm Hg) Mean ± SD 84 ± 13 77 ± 15 78 ± 15 Range 66-95 45-105 47-105
Hematocrit (%)
Mean ± SD 36 ± 5.4 36 ± 4.9 37 ± 6.4 Range 29-42 27-47 25-47
'CBF was significantly correlated to temperature (p < 0.01; r = 0.535) and Pco, (p < 0.01; r = 0.586).
tCBF corrected to Pco, = 40 mm Hg (p <0.01; r = 0.462).
disease. There were 14 male and 15 female patients with a mean age of 54 years (range 27 to 72). Eleven patients had aortic valve replacement, 10 mitral valve replace- ment, six coronary bypass, one closure of an atrial septal defect, and one mitral plus aortic valve replacement. None of the patients had any history or clinical evidence of cerebral vascular disease. Informed consent was obtained in every case.
In 24 patients there were five time periods examined: (1) between sternotomy and initiation of ECC, (2) during hypothermic ECC, (3) during normothermic ECC, and (4) 10 and 20 minutes after termination of ECC. In 12 patients additional measurements were made during initial normothermic perfusion prior to systemic cooling, viz., tl1e first I to 2 minutes of ECC.
The following variables were recorded: CBF, mean arterial blood pressure (M~), hematocrit value, hemo- globin value, arterial partial pressure of oxygen and plasma carbon dioxide (Paoz and Pacoz' respectively), arterial pH, and rectal and arterial temperatures. The blood gases were measured at 37 0 C and corrected to
'the estimated brain temperature. Before ECC the rectal temperature was used, and during ECC the arterial line temperature close to the heart was used.
CBF method. CBF was measured from the clearance of the gamma emitting radioisotope 133Xe, which was injected as a bolus (1 second) into the ascending aorta (10 patients) or the brachiocephalic or common carotid artery (19 patients). The isotope clearance was mea-
sured by 16 small scintillation detectors placed over one hemisphere in 12 cases and with a single large detector (inner diameter 50 mm, length 45 mm) in 17 cases. The CBF was calculated from the tenth to the fortieth seconds of the slope of the logarithmically recorded clearance curve.' An estimate of the average gray matter flow was obtained a few minutes following injection. A typical sum curve would give 10,000 cps at peak count rate. A value for A (blood-tissue partition coefficient for 133Xe) was chosen to accord with temper- ature and hemoglobin.' The main disadvantage of injecting into the ascending aorta, the brachiocephalic artery, or the common carotid artery is that 133Xe enters the external carotid circulation, which supplies the scalp muscles and bone underlying the detector(s), and results in a slight underestimation of CBF. The amount of 133Xe in the scalp of a normal subject 1 minute after injection is approximately 7% of that in the brain." CBF was corrected for the activity remaining from the previous measurements by retraction of its monoexponential interpolated curve. Incremental doses of 133Xe (total dose from 2 to 4 mCi) dissolved in 1 rn1 of isotonic saline were used to keep the remaining activity below 10% of peak count rate. The observed relationship between CBF and Pco, during the different periods of the operation were used to correct CBF for Pco, changes, giving a correction from 1% to 3% per 1 mm Hg change in Pco, (periods 1 to 5: y = 1.51x - 21; y = O.44x + 47; Y=0.57x + 23; Y= 1.67x - 26; Y= 1.06x - 2).
204 Henriksen. Hjelms, Lindeburgh The Journal of
Thoracic and Cardiovascular Surgery
Table n. Cerebral and systemic hemodynamic measurements before. during, and after extracorporeal circulation (ECC) in 24 patients having cardiac operations.
During ECC After ECC
Before ECC Hypothermia Normothermia Mean 10 min Mean 23 min
Baseline mean Mean ± D.SD Mean ± D.SD Mean ± D.SD Mean ± D.SD Friedman's (range) (range) (range) (range) (range) test
Cerebral blood flow" (mil 100 grn- min)
Mean arterial pressure (mm Hg)
Hematocrit (%)
39.7 ± 10 (24-51)
29.3 ± 1.8 (27-31)
47.0 ± 12 (26-68)
36.9±1.5 (34-38)
40.0 :t-7 (26-51)
36.7 ± 1.2 (36-37)
39.5 ± 9 (21-53)
36.6 ± 1.2 (36-37)
p <om
P <0.001
p < 0.001
P <0.20
p < 0.001
Legend:~D, Standard deviation of change from baseline. 'Cerebral blood flow values are corrected to Peo, = 40 mm Hg.
tMeasured at 37" C but all values are corrected to the actual temperature.
Table m. Cerebral and systemic hemodynamic measurements in /2 patients before and during the first minute of extracorporeal circulation (ECC) before initiation of cooling
Baseline before ECC First minute of ECC
Cerebral blood flow' (mil I00 gm- min) Arterial Peo, (mm Hg) Mean arterial pressure (mm Hg) Hematocrit (%) Temperature (0C)
Legend:~D. Standard deviation of change from baseline.
'Cerebral blood flow corrected for changes in Pco, from 40 mm Hg.
tp < 0.02.
+p <0.05.
§p <0.01.
I Range
35.0-36.3
The patients were premedicated with 5 to 10 mg of morphine or 10 to 15 mg of diazepam plus scopolamine 0.2 to 0.4 mg, given intramuscularly. General anesthesia was initiated with thiopental 4 rug/kg intravenously in 26 patients and with etomidate 0.3 mg/kg intravenously in three patients. Pancuronium and suxamethonium were given intravenously for muscle relaxation. The patients received 1.5% enflurane in pure oxygen as a volatile anesthetic, which was discontinued at the time of bypass. During ECC, anesthesia was maintained with fentanyl 0.2 mlj 10 kg intravenously and with 0.3 to 0.6 mg of scopolamine intravenously.
The Rygg-Kyvsgaard heart-lung machine and bubble oxygenator were used for the extracorporeal circulation. The prime consisted of Ringer's lactate, 2.0 to 2.5 L. During ECC extracorporeal flow rate was maintained at
2.0 to 2.4 L/m . m', ECC was initiated with gradually increasing flow, full flow being established within the first minute. Cooling was then effected until an esopha- geal temperature of 28 to 30° C had been reached. Temperature was kept at this level during the cardiac repair. During rewarming the temperature gradient was not allowed to exceed 12° C. Rewarming was usually facilitated by active vasodilation with chlorpromazine in a dosage up to 0.5 rug/kg.
Statistics. The material was considered to have a non-Gaussian distribution, so that nonparametric statis- tics were preferred.' Analysis of variance was made with Friedman's two-way test. Differences between two re- lated or unrelated groups were evaluated by the Wilcox- on and Mann-Whitney rank sum tests, respectively. Relationships between two variables were tested with
Volume 86
90 CEREBRAL BLOOD FLOW ml/l00g·min
140
CEREBRAL BLOOD FLOW ml/l00g.min
Fig. 1. Cerebral blood flow in 12 patients between sternoto- my and the onset of extracorporeal circulation (ECC;, and in the initial phase (1st min) of ECC. In the patient with no flow in the second measurement, arterial pressure fell from 101 to 41 mm Hg.
Spearman's rank correlation test. A parametric linear regression analysis between CBF and Pco, was used to define the sensitivity to Pco;
The significance level was p < 0.05. Mean values ± standard deviation, range, and ASD (standard deviation of change from baseline) have been used to summarize data in the tables.
20
40
80
60
120
100
Discussion
Autoregulation of the cerebral circulation maintains a relatively constant CBF within a wide range of perfusion pressures (60 to 150 mm Hg in normal subjects). The
I+- ECC .... I before ECC normothermia 2nd after
hypothermia 1st after
Fig. 2. Cerebral blood flow in 24 patients having cardiac operations. Measurements were performed (I) between ster- notomy and the onset of extracorporeal circulation (ECC); (2) during hypothermic ECC; (3) during normothermic ECC; (4) 10 minutes after ECC; and (5) 20 minutes after ECC. Cerebral blood flow values are corrected for changes in Pco, to a mean Pco,of 40 mm Hg. Median values are indicated in the figure as solid dots.
when a steady-state temperature of 37° C had been reached, CBF fell to 42 mI/lOO gm . min.
CBF fell in nine patients during ECC (Fig. 2), probably because MAP was reduced below the lower limit of autoregulation (Fig. 3). In these patients mean MAP ± I SD was 45 ± 7 mm Hg, compared to 57 ± 14 mm Hg in the other 15 patients, of whom all had an increased CBF during ECC.
After ECC. CBF measured twice after the termina- tion of ECC (10 and 20 minutes) was 40 and 41 mI/lOO gm . min, respectively.
An overall significant correlation between plasma Pco, and CBF was found (p < 0.(01). During ECC the sensitivity to PC02 was markedly reduced compared to values obtained before and after ECC (p < 0.01; Fig. 4). That CBF decreased in a manner related to MAP below a perfusion pressure of 55 mm Hg indicates that the cerebral autoregulation was lost (Fig. 3; P < 0.01; r = 0041).
1st min. of ECCbefore ECC
50
80
20
10
30
70
40
60
Results
Before ECC. The combination of low temperature and low Pco, reduced CBF markedly before ECC (Table I). CBF correlated significantly with Pco, and temperature (p < 0.01).' Mean CBF in all patients (n = 29) was 31 mI/lOO gm . min at a temperature of 36.3° C and a Pco, of 34.1. mm Hg. Corrected to a temperature of 37° C and a Pco, of 40 mm Hg, it would correspond to 42 mI/IOO gm . min. Mean CBF in the 24 patients who were studied five times during the operation was 38 ml/ I00 gm . min prior to ECC (Ta- ole II).
During ECC. CBF fell by a mean of 33% (p < 0.02) during the initial phase of ECC (Fig. I, Table III). In one patient CBF fell to zero (Fig. I) as MAP was reduced from 101 to 41 mm Hg during this initial phase. During steady-state hypothermic ECC (temperature 29° C), CBF increased unexpectedly from 38 up to 64 ml/IOO gm . min (Fig. 2, Table II). UJXln rewarming,
206 Henriksen, Hjelms, Lindeburgh The Journal of
Thoracic and Cardiovascular Surgery
Fig. 3. Cerebral blood flow plotted versus mean arterial blood pressure gives an indication of the critical level of perfusion pressure. Autoregulation seem to be impaired at the end of bypass during normothermic extracorporeal circulation (ECC), and it is lost with perfusion pressures around or below 55 mm Hg (p < 0.01; r = 0.40). Vertical bars indicate stan- dard error of mean.
CEREBRAL BLOOD FLOW ml/l00g.min
• 12 • 0
• 0 0
0 10 20 30 4CJ 50 60 mmHg
Fig. 4. A significant overall relationship was found between Pco, and cerebral blood flow (Spearman's test: p < 0.00I; r = 0.44). The relationship between Pco, and cerebral blood flow was significantly changed (p < 0.0 I) during extracorpo- real circulation (black circles; y = 0.02x + 57; r = 0.56) com- pared to the relationship before and after bypass (open circles; y = 1.36x - 15; r = 0.01).
MEAN ARTERIAL BLOOD PRESSURE
....._.----.--- 3
CEREBRAL BLOOD FLOW ml/l00g·min 1Hypothermic ECC 2Normothermic ECC 3 Before and after ECC
30
20'
10
90
80
70
60
50
40
constancy of CBF is achieved by dilatation or constric- tion of cerebral resistance vessels in response to changes in perfusion pressure. However, there are pressures at which the autoregulatory dilatation or constriction is maximal (termed the lower and upper limits of autoreg- ulation, respectively). Activation of the sympathetic nervous system shifts the upper and lower limits toward higher absolute levelsof systemic arterial blood pressure. Fitch, Mackenzie, and Harper" have demonstrated that alpha adrenergic blockade reduced the lower limit of autoregulation to approximately 40 mm Hg. This finding indicates that the sympathoadrenal discharge that accompanies hemorrhagic hypotension curtails the ability of cerebral resistance vessels to dilate at low systemic arterial pressure. When this discharge is removed, the arterioles can then dilate further. This aspect of the cerebral circulation might be of importance during ECC, since an MAP below 60 mm Hg is not a rare occurrence.
The earlier finding of a relationship between MAP and carotid artery blood flow during bypass in dogs9
• 1o
points toward an impairment of cerebral autoregulation, although common carotid artery blood flow was mea- sured and cannot be considered as representative of CBF. In nine patients we observed a fall in CBF during ECC (Fig. 2). In these patients MAP was reduced to 45 ± 7 mm Hg, compared to 57 ± 14 mm Hg in the 15
patients whose CBF increased during ECC. It is most likely that MAP was reduced below the lower limit of autoregulation in these nine patients. At a moderate decrease in CBF, the brain oxygen and glucose uptake is maintained by an increased extraction from blood, but at more marked decreases in CBF, the supply to the cerebral tissue becomes inadequate. Below an MAP of 55 mm Hg, CBF was significantly related to MAP (p < 0.01). This relationship indicates that autoregula- tion was lost below this leveland that a further reduction in MAP would entail a risk for brain damage. The autoregulatory curve constructed during hypothermia appears normal, whereas at the end of rewarming the curve appears to indicate that cerebral autoregulation was impaired (Fig. 3).
The significant relationships prior to ECC between CBF and temperature and between CBF and plasma PC02 support the reliability of the method (Table I). During ECC the sensitivity of carbon dioxide was markedly reduced (Fig. 4), probably because of the decreased MAP, which has a known modulating effect on this relationship." Wollman and colleagues" esti- mated CBF in man during ECC by measuring arterio- venous oxygen difference, and they found a similar importance of Pco, in regulating CBF. However, their results as well as ours do not give any information regarding the optimal level of Pco, during ECC.
Volume 86
Number 2
August, 1983
The finding of a reduced CBF in the majority of patients (Fig. 1) during the early phase of ECC is in accordance with the observation that changes in cortical electrical activity occur in these first minutes of ECC. Boysen13 has shown that the EEG shows a reduced activity at a CBF of around 16 to 22 ml/IOO gm . min. In the present study, although EEG was not recorded, CBF in several patients was at or below the limit at which EEG slowing or extinction occurs (Figs. 1 and 2).
The highest CBF was observed during hypothermic ECC (Fig. 2, Table 11). The mean increase was 67%, but in nine patients the CBF increased more than 100%. Such an increase was unexpected and clearly abnormal and probably reflects an uncoupling of flow and metab- 0Iism,14,15 as metabolism must have been markedly reduced at a temperature of 29° e. Several factors may be responsible for this uncoupling, the most important probably being microvascular blockade because of emboli, thrombocyte aggregation, or stagnant capillary flow, each of which is known to occur in ECC. 16,17
During microvascular blockade no isotope reaches such areas, but the presence of the blockade will be recog- nized by the surrounding luxury perfusion.14, 15 The induced hemodilution, nonpulsatile perfusion, and changes in blood-tissue partition coefficient for I33Xe may contribute to the increased CBF but cannot fully explain the marked changes seen during ECC. The present study givesno indication of the origin behind the brain hyperperfusion, but a diffuse microvascular block- ade could produce such changes. Prosenz" injected microspheresinto the carotid artery in dogs, a procedure producing an effect similar to that of bypass, and observeda hyperperfusion comparable to that which we saw. When local cerebral oxyg~n transport is reduced, local lactic acidosis results. This leads to the develop- ment of tissue hyperemia'hyperoxygenation. The degree of restitution and the time required to recover from a luxury perfusion state" and its accompanying cerebral vasomoter paralysis depend upon the severity and time of prior oxygen deprivation. Thus the reactive hyper- emia is not necessarily bound to a permanent…