- Email: [email protected]
Cerebral Perfusion during Nonpulsatile Cardiopulmonary Bypass
Cerebral Perfusion during Nonpulsatile Cardiopulmonary Bypass Tryggve Lundar, M.D., Karl-Fredrik Lindegaard, M.D., Tor Fr~rysaker,M.D., Rune Aaslid, Ph.D., Jan Wiberg, M.D., and Helge Nornes, M.D. ABSTRACT The recording of middle cerebral artery (MCA) flow velocity by the transcranial Doppler method offers a new, noninvasive, continuous technique for studies of cerebral circulation. Comparative studies of electromagnetic internal carotid artery (ICA) flowmetry and MCA flow velocity by the transcranial Doppler technique have demonstrated that observed changes in MCA flow velocities reflect concomitant changes in cerebral circulation. Eleven high-risk patients undergoing cardiopulmonary bypass (CPB) procedures were included in a pilot study. Arterial blood pressure (BP), central venous pressure, and epidural intracranial pressure (EDP) were recorded during CPB. Cerebral electrical activity was recorded by a cerebral function monitor. Flow velocity in the MCA was increased during nonpulsatile CPB in 10 of the 11 patients. This increase was related to the degree of hemodilution, and the flow velocity during steady-state CPB was 80 to 300% of the prebypass value. The MCA flow velocity changed, however, in a pressure-passive manner with the cerebral perfusion pressure (CPP = BP - EDP) in the individual patient, which indicates that cerebral autoregulation was not operative. During the first 15 minutes after termination of bypass, the MCA flow velocity was reduced, but remained higher than the prebypass level, 110 to 210% of the level during the last 5 minutes preceding CPB.
Permanent brain damage is among the most feared complications following cardiac operations and can spoil an otherwise excellent result. Today, open-heart operations seem to be associated with a 12% risk of diffuse encephalopathy as well as a 2 to 5% risk of stroke [l].Although a left ventricular mural thrombus or intraoperative "accidents" may explain the stroke in selected instances [2], more often the reason remains obscure in the individual patient . Compared with the rapid developments in cardiac Surgery, knowledge oibrain pathophysiology, including cerebral circulation, during cardiopulmonary bypass (CPB) is surprisingly scarce. Indirect studies on cerebral
circulation during CPB based on arteriovenous oxygen differences are hard to accept because they are based on the coupling of cerebral metabolism and flow. This coupling seems to be lost during extracorporeal circulation, at least during nonpulsatile CPB [3]. Recent studies of cerebral blood flow during nonpulsatile CPB using the xenon washout technique have also yielded conflicting results [4-61. In a previous study involving 5 patients, we [7] recorded internal carotid artery (ICA) flow (electromagnetic flowmeter) and cerebral perfusion pressure (CPP = mean arterial blood pressure [BPI - mean intracranial epidural pressure [EDP]) during CPB. The record demonstrated increased ICA bulk flow down to CPP levels in the range of 20 mm Hg during CPB. Transcranial Doppler studies offer a new, noninvasive method for continuous evaluation of cerebral circulation [8-lo]. During the last two years, we have used this technique to perform studies of flow velocity in the basal cerebral arteries. Such studies proved to be a useful tool in our management of patients with subarachnoid hemorrhage [lo] as well as patients with occlusive extracranial artery disease [ l l ] . The diameter of the proximal middle cerebral artery (MCA) is supposed to remain relatively constant (12, 131 as regulation of cerebrovascular resistance acts on the arteriolar level, both in autoregulation of cerebral blood flow during changes in CPP as well as during carbon dioxideinduced changes in the cerebral perfusion. Combined studies have demonstrated that changes in MCA flow velocity correspond well with ICA flow changes recorded with an electromagnetic flowmeter and with ICA flow velocities ( r = 0.92).* This article presents our experience with recording MCA flow velocity and monitoring CPP in 11 patients during nonpulsatile CPB. In 1 patient we were able to perform MCA flow velocity and ipsilateral electromagnetic ICA flow recordings simultaneously.
Material and Method
Accepted for puhlication Nov 16, 1984.
Eleven patients undergoing valve replacements and aortocoronary bypass procedures were included in the present study. They were considered to be at high risk because of multiple valve replacements, combined valve and aortocoronary bypass procedures, or a history of stroke prior to the heart operations. Age and sex distribution, CPB times, and operative procedures are summarized in Table 1. Intracranial epidural pressure was recorded by a
Address reprint requests to Dr. Lundar, Dcpartmcnt of Neurosurgery, Rikshospitalet, 0027 Oslo 1, Norway.
'Lundar T, ct al: Unpublished data, 1985
From the Dcpartmcnts ot Surgery and Neurosurgery, Rikshospitalet, Oslo, Norway.
144
145 Lundar et al: Cerebral Perfusion during Nonpulsatile CPB
Table I. Summary of Patient Data Patient No., Age (yr), Sex
Duration of CPB (min)
M M M F M 6. 63, F
111 105 101 106 155
7. 43, M 8. 70, M 9. 60, F 10. 57, F 11. 52, M
210 106 94 57 72
1. 51, 2. 65, 3. 68, 4. 62, 5. 65,
83
CPB = cardiopulmonary bypass; AVR internal carotid artery.
Operative Procedures; Risk Factors
Clinical Result
AVR; previous stroke MVR + CABG x 2 AVR + CABG x 2; occluded left ICA
Good
MVR, replacement of 14-yr-old malfunctioning Beall valve AVR + MVR AVR + prosthesis to aortic arch + CABG x 3; previous AVR, preop Cniididn sepsis AVR + MVR + CABG AVR MVR + CABG MVR; anuria (hemodialysis) R ICA endarterectomy + CABG x 2; previous stroke =
aortic valve replacement; MVR
minitransducer (AME, Horten, Norway) implanted through a burr hole in the right frontal region. The procedure has been described previously [ 141. Informed consent was obtained from each patient. Arterial blood pressure and central venous pressure were recorded with fluid-pressure transducers (AME, Horten, Norway). Cerebral electrical activity was recorded by a cerebral function monitor (Cerebral Function Monitor 4640, Devices Limited, Welwyn Garden City, Hertfordshire, England). Flow velocity in the right MCA was recorded continuously throughout CPB by the transcranial Doppler method, which has previously been described in detail [ 8 ] . Taking advantage of the relatively thin skull bone in the temporal region above the zygomatic arch, Doppler signals from the MCA were obtained using a laboratory prototype, range-gated, 2 MHz, pulsed-wave Doppler instrument with acoustically focusing transducers. In Patient 11, ICA flow was recorded during CPB with an electromagnetic flowmeter (Nycotron, Lier, Norway). All variables were recorded on a six-channel pen recorder (W & W Electronic AG, Basel, Switzerland). Anesthesia was induced with nitrous oxide, fentanyl, diazepam, and pancuronium bromide, and maintained on the same regimen. A Rygg-Kyvsgaard heart-lung machine with a Polystan Venotherm bubble oxygenator was used (Polystan, Copenhagen, Denmark). The regimen for extracorporeal perfusion was standardized to generalized hypothermia (28"to 30°C), flow rite of 1.5 L/min/m', a d a m e a n systemic blood pressure between 40 and 60 mm Hg. Carbon dioxide was added to the oxygenator in concentrations of 2.5 to 5% to obtain temperature-corrected values for arterial CO? tension (PaC0') of about 37.5 mm Hg during CPB. Hematocrit was determined every fifteen seconds during the first 5 minutes of CPB (until stabilization) and thereafter every 15 minutes. In Patient 11, carotid endarterectomy was performed an hour before CPB because of severe right-sided ICA
Good Good Good Died of prolonged pump failure Died of acute pump failure 24 hr postop Good
Good Good Fair
Good
mitral valve replaccnicnt; CABC = coronary artery bypass grafting; ICA
=
stenosis and a previous left-sided stroke. A Pruitt shunt was used during the endarterectomy, and electromagnetic ICA flow was recorded during CPB, as previously described [71.
Results Middle cerebral artery flow velocity increased at the introduction of CPB in all 11 patients (Table 2). As the diameter of the MCA is unknown in the individual patient, it is meaningless to compare the absolute velocity values (centimeters per second) in these patients. When compared with the prebypass value, however, the percent change in velocity will reflect changes in flow in the individual. Therefore, all velocity values in Table 2 are given with the percentage of the prebypass value in parentheses. At the maximum of this initial increase, MCA flow velocities were in the range of 140 to 260% of the prebypass values. The initial increase in MCA flow velocity was followed by a reduction of short duration. This reduction corresponds to the low CPP state that occurs during the wellknown drop in BP at the introduction of CPB. The lowest value of this "dip" in flow velocity as well as the corresponding CPP dip (initial drop in CPP) also are listed in Table 2. The range was 72 to 180%of the prebypass value of 100. During steady-state CPB, MCA flow velocity was markedly increased in 10 of the 11 patients (range, 80 to 300% of the prebypass values). In comparison, the CPP range during steady-state CPB is shown for the individual patient in Table 2, as is a hemodilution ratio (prebypass hematocrit divided by steady-state hematocrit during bypass). Flow velocities in the MCA were obtained during the first period after termination of CPB in 10 patients. In 9 of them, flow velocity was reduced at the conclusion of bypass but remained higher than the prebypass value (see Table 2). Figure 1 demonstrates a typical record of
146 The Annals of Thoracic Surgery Vol 40
No 2 August 1985
Table 2. Data on Cerebral Perfusion during Nonpulsatile CPB
Patient No.
MCA Flow Velocity, Initial Increase (cm/sec)”,b
MCA Flow Velocity during Dip (cm/sec)”
CPP Dip (mm Hg)
25/65 (260)
45 (180) 33 (100)
25 20
60 (150) 36 (100) 36 (150) 60 (150) 60 (120) 35 (117)
20-25 26 22 30 24 15
30 (120) 18 (72) 35 (117)
35 20 20
33/60 (180) 40/100 (250)
9 10 11
36/60 (167) 24/60 (250) 40/60 (150) 50/100 (200) 30/45/60 (150/200) 25/50 (200) 25/35 (140) 30155 (183)
MCA Flow Velocity, Steady-State CPB (cm/sec)”
CPP, SteadyState CPB (mm Hg)
50 30-50 40 50 40
54/22 (2.5) 46/21 (2.2) 54/27 (2.0) 43/21 (2.0) 43/21 (2.0) 41/21 (2.0) 36/17 (2.1) 43/23 (1.9)
40 20-25 20-40
45/21 (2.1) 27/19 (1.4) 42/24 (1.8)
20-40 50 40
70 (280) 55 (167) 110 (275) 60 (167) 55 (229) 80 (200) 120 (240) 90 (300) 50 (200) 20 (80) 70 (233)
Hernodilution‘
MCA Flow Velocity after End of CPB (cm/sec)a
...
CFM Response
60 (150) 80 (160) 50 (167)
Biphasic No change Biphasic Fall Biphasic Fall Biphasic Fall
35 (140) 40 (160) 50 (167)
Biphasic Fall Fall
40 (121) 60 (150) 40 (111) 50 (208)
“Numbers in parentheses are percentage of the prebypass value. bThe first number represents the prebypass value and the second, the initial increase. ‘Numbers are shown for hematocrit measured before/during CPB; the number in parentheses is the ratio
CPB
= cardiopulmonary bypass; MCA = middle cerebral artery; CPP = cerebral perfusion pressure;
CFM
=
cerebral function monitor.
UV
100YCA flow velocity 6 0 .
cm/rec .I
I
. Fig 1 . (Patient 5.) Recording of middle cerebral artery (MCA)flow velocity, blood pressure (BP), epidural intracranial pressure (EDP), and cerebral function monitor (CFM)at the start of cardiopulmonary bypass. Repeat hematocrit (Hct) values illustrate the progress and stabilization of hemodilution. As the priming solution was introduced (arrow a), the MCA flow velocity increased rapidly. A biphasic re-
.
I
minutom
sponse in the activity of the cerebral function monitor was observed. The well-known drop in BP seen at the‘iritroduction of bypass (arrow b) caused the dip in cerebral perfusion pressure fCPP) fBP - EDP) and MCA flow velocity. As BP and CPP were restored (arrow c), MCA flow velocity again increased, to about twice the prebypass value.
147 Lundar et al: Cerebral Perfusion during Nonpulsatile CPB
A
B
Fig 2. (Patient 7.) Spectral display of the Doppler signal recorded ( A ) shortly before introduction of cardiopulmonary bypass (CPB), ( B ) during steady-state CPB, and (C) after termination of CPB. A realtime 64-point spectral analysis of the Doppler-shifted frequencies is performed by a 6502A microprocessor and a fast Fourier transform routine designed by one of us ( R . A.). A special software subroutine translates the spectral envelope into an instantaneous analog flowvelocity signal.
MCA flow velocity, BP, EDP, and the cerebral function monitor during CPB. The initial increase in flow velocity at the introduction of the priming solution was followed by the dip when CPP (BP - EDP) fell to about 22 mm Hg. As BP and CPP recovered, MCA flow velocity once again increased. Hematocrit was markedly decreased immediately after the introduction of the priming solution, and a response on the cerebral function monitor was observed at this point in all but 1 patient. This response was either biphasic (increase followed by depression of the cerebral function monitor activity [as in Fig 1, arrow a ] ) or monophasic (fall in activity). The initial cerebral function monitor responses are also listed in Table 2. Figure 2 demonstrates the MCA flow velocity as it appears on the Doppler monitor before bypass, during steady-state CPB, and after bypass. Although MCA flow velocities generally were markedly increased during steady-state CPB, concomitant changes in flow velocities were observed when CPP changed, indicating a pressure-passive situation during nonpulsatile CPB. This is clearly demonstrated in Figure 3. Cerebral autoregulation is obviously not operative during the spontaneous fall in CPP. The CPP range, however, was very low-40 to 20 mm Hg. Figure 4 shows data obtained from Patient 11 during steady-state hemodilution (hematocrit, 24%), PaCO2 (temperature correlated, 38.25 mm Hg), and temperature (28°C).This was the patient in whom simultaneous recordings of MCA flow velocity and ipsilateral ICA flow were obtained. At arrow a , the heart-lung machine flow was increased from 1.5 Wmin/m2 to 2.0 Wmin/m2. At
C
arrow b, machine flow was reduced back from 2.0 to 1.5 Wmin/m2.The increase in BP and the concomitant slight increase in EDP caused a profound increase in CPP, resulting in markedly increased perfusion. The curves in Figure 4 show the close relationship between the changes in MCA flow velocity and ICA flow. There was no evidence of autoregulation during this period, even when CPP reached levels higher than 60 mm Hg. Between arrows u and b, "vasomotor waves" occurred spontaneously in BP, which are also reflected in EDP, MCA flow velocity, and ICA flow. The regimen for CPB aimed at keeping temperaturecorrected PaCO2 values at 37.5 mm Hg by adjusting the C02 admixture between 2.5 and 5%. It was necessary to keep PaCO2 constant to be able to study the effect of changes in CPP on flow velocity. The single patient in whom MCA flow velocity was reduced throughout CPB (Patient 10) differed from the others in having reduced renal function, low hematocrit prior to bypass, and, thus, relatively less additional hemodilution as CPB was established (see Table 2). Ultrafiltration was performed before CPB was terminated, and MCA flow velocity actually increased at the termination of bypass. The monitoring of BP, EDP, and cerebral function was discontinued one to four days after operation depending on the individual need for prolonged central nervous system supervision. The results of postoperative monitoring are outside the scope of this report and will be presented elsewhere. A short description of the clinical results is provided in Table 1. There were no complications associated with the monitoring procedures.
Comment The study of cerebral circulation during CPB is hampered by methodological difficulties. There is an obvious need for on-line information on cerebral perfusion during cardiac operations. In a previous study of electromagnetic ICA flowmetry during CPB [7],we found increased ICA flow during bypass. Carotid endarterectomy shortly before the open-heart procedure might in-
1.18 The Annals of Thoracic Surgery Vol 40
No 2 August 1985
Fig 3. (Patient 1.) Records of middle cerebral artery (MCA) Pmc, z&xily, Mood pressure (BP), and epidural intracranial pressure (EDP)during steadystate cardiopulmonary lypass. A spontaneous fall in cerebral perfusion pressure caused profound reduction in MCA florc, d o c i t y .
terfere with the cerebral circulation during the subsequent CPB procedure. However, for ethical reasons, ICA flowmetry records are not obtainable for other kinds of cardiac procedures. Transcranial Doppler recording of MCA flow velocity offers new opportunities to investigate the blood supply of the brain during CPB. The basal cerebral arteries are not involved in the regulatory mechanisms elicited by changes in CPP and COz (12, 131. Simultaneous recordings of MCA flow velocity and ICA flow or ICA flow velocity changes in response to changes in CPP have demonstrated that observed changes in MCA flow velocity reflect concomitant and closely related changes in the cerebral circulation (r = 0.92).* These data agree with our observations in Patient 11 (see Fig 4). A recent report on the effect of changes in COz on MCA flow velocity in healthy volunteers demonstrates the classic effect of COz on cerebral circulation and supports the theory that MCA flow velocities reflect changes in MCA volume flow [15]. The present study confirms our previous observations of increased perfusion during nonpulsatile CPB [7]. The regimen for extracorporeal perfusion was changed somewhat from that study: from 2.5 L/min/m2 and morphine anesthesia to 1.5 L/min/m' and fentanyl anesthesia. Nevertheless, MCA flow veloc'Lundar T, et al: Unpublished data, 1985.
ity is clearly increased during CPB compared with the prebypass level, and the increase is related to the degree of hemodilution (see Table 2). It is impressive how well MCA flow velocity is maintained during CPB, considering the low CPP levels observed. It is obvious, however, that autoregulation is not operative during the CPP levels observed during nonpulsatile CPB in this series (20 to 60 mm Hg). This is clearly shown in Figures 3 and 4, which are observations made during steady-state hemodilution, steady-state PaC02, and constant temperature. Several cerebral function monitor studies during CPB have been published [16-181. Both depression of the activity of the cerebral function monitor and the biphasic response at the introduction of CPB are well-known patterns. More profound changes, with depression of the activity toward the zero line, were not observed in this series. The short-lasting depression of activity seen during the initial increase in flow velocity, occurring during maximum hemodilution, can be interpreted as insufficient oxygenation despite increased MCA flow velocity, and is suggestive of a qualitatively insufficient perfusion. During CPB, temperature-corrected PaCOz values were maintained at 37.5 mm Hg by adjusting the COz admixture. Further manipulations with COz were avoided so as not to interfere with the study of changes in MCA flow velocity relative to changes in CPP. Observations during COZ manipulations are required, however, to determine to what extent COr reactivity is maintained during CPB. Such studies are now in progress. Previous clinical studies during nonpulsatile CPB using xenon washout or arteriovenous 0, difference tech-
149 Lundar et al: Cerebral Perfusion during Nonpulsatile CPB
Fig 4. (Patient 11.)Siniultaiieous recvrdings of middle cerebral artery (MCA) pow wlocity, ipsilateral internal camtid artery (ICA) pocv (electroniagnetic),blood pressure (BP),epidural intmcranial pressure (EDP),and cerebral futiction monitor (CFM) during established cardiopulmonary @pas.fncrrasing machine floui from 1.5 to 2.0 Uniinl m2 (arrow a) caused a tnarhd increase iri cerebral perfusion pressure (CPP) a d a corresponding increase iti MCA j 7 o u g wlm*tyand ICA flow. Note the close refatiorishipbetriven MCA pout wlocity and ICA pow changes. As nrachine pow was reduced from 2.0 to 1.5 Uminltn' (arrow b), subsequent reductions iti CPP, MCAfloup velocity, and ICA pour were observed. Diiring the i n c r d machine flow (bettiwen arrows a and b), "vasontotor tuoves" in BP were seen. Gnicontitant changes in EDP, MCA pow twlocity, and fCAflori~are also demonstrated. Temperature (28°C). hemudilution (hematmit, 24%), and partial pressure of arterial C 0 2 (teniperatirre corrected, 38.25 nnn Hg)uwre constant during the recordings. niques have indicated preserved COz reactivity [4,61. Because our n u m b e r of observations during CPP levels greater than 50 m m Hg is limited, extended studies are needed to determine whether autoregulation is completely lost during nonpulsatile CPB or if CPP is simply below t h e individual lower limit of autoregulation during t h e actual degree of hemodilution a n d regimen for extracorporeal perfusion. Such studies cannot be performed without intracranial pressure monitoring, because BP alone can be a poor indicator of the CPP state in such situations [ 191. Comparative studies on nonpulsatile a n d pulsatile CPB a n d on different CPB regimens a r e needed. W e believe that further studies of cerebral hemodynamics during CPB procedures may have considerable clinical implications.
References 1. Furlan AJ, Breuer AC: Central nervous system complications of open heart surgery. Stroke 15:912, 1984 2. Breuer AC, Franco I, Mariewski D, Soto-Velasco J: Left
3.
4.
5. 6.
ventricular thrombi seen by ventriculography are a significant risk factor for stroke in open heart surgery. Ann Neurol 10:103, 1981 Andersen K, Waaben J, Husum 6, et al: Non-pulsatile cardiopulmonary by-pass disrupts flow-metabolism couple in the brain. Presented at the Annual Meeting of the Scandinavian Association for Thoracic and Cardiovascular Surgery, Turku, Finland, Aug 24-26, 1983 Henriksen L, Hjems E, Lindeburgh T: Brain hvperperfusion during cardiac operations: cerebral blood flow measured in man by intra-arterial injection of xenon 133: evidence suggestive of intraoperative microembolism. J Thorac Cardiovasc Surg 86202, 1983 McKay RD, Reves JG, Karp RB, et al: Effects of cardiopulmonary bypass on cerebral blood flow. J Cereb Blood Flow Metab 3:Suppl 1:119, 1983 Govier AV, Reves JG, McKay RD, et al: Factors and their influence on regional cerebral blood flow during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 38592,
1984 7. Lundar T,Freysaker T,Lindegaard KF, et al: Some observa-
tions on cerebral perfusion during Cardiopulmonary bypass. Ann Thorac Surg 39:318, 1985 8. Aaslid R, Marcwalder TM, Nornes H: Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 57:769, 1982 9. Aaslid R, Nornes H: Musical murmurs in human cerebral arteries after subarachnoid hemorrhage. J Neurosurg 60:32, 1984 10. Aaslid R, Huber P, Nornes H: Evaluation of cerebrovascu-
5
150 The Annals of Thoracic Surgery Vol 40 No 2 August 1985
11.
12. 13.
14.
15.
lar spasm with transcranial Doppler ultrasound. J Neurosurg 60:37, 1984 Lindegaard KF, Grolimund P, Bakke SJ, et al: Carotid artery disease: assessment of intracranial hemodynamic pattern by noninvasive transcranial Doppler ultrasound. J Neurosurg (in press, 1985) Strandgaard S, Paulson OB: Cerebral autoregulation. Stroke 15:413, 1984 Huber P, Handa J: Effect of contrast material, hypercapnia, hyperventilation, hypertonic glucose and papaverine on the diameter of the cerebral arteries: angiographic determination in man. Invest Radio1 2:17, 1967 Lundar T, Fraysaker T, Nornes H, Lilleaasen P: Aspects of cerebral perfusion on open-heart surgery. Scand J Thorac Cardiovasc Surg 16:217, 1982 Markwalder TM, Grolimund P, Seiler RW, et al: Depen-
16.
17.
18. 19.
dency of blood flow velocity in the middle cerebral artery on end-tidal carbon dioxide partial pressure: a transcranial ultrasound Doppler study. J Cereb Blood Flow Metab 4:368, 1984 Schwartz MS, Colvin MP, Prior PF, et al: The cerebral function monitor: its value in predicting the neurological outcome in patients undergoing cardiopulmonary bypass. Anaesthesia 29:135, 1973 Kritikou PE, Branthwaite MA: Significance of changes in cerebral electrical activity at onset of cardiopulmonary bypass. Thorax 32:534, 1977 Prior PF: Monitoring Cerebral Function. Amsterdam, ElseviedNorth Holland, 1979 Lundar T, Freysaker T, Nornes H: The clinical significance of changes in cerebral perfusion pressure during openheart surgery. Scand J Thorac Cardiovasc Surg 17163, 1983
Notice from the American Board of Thoracic Surgery The Part I (written) examination will be held at the Amfac Hotel, DallasFort Worth Airport, Dallas, TX, in February, 1986. The closing date for registration was August 1, 1985. To be admissible for the Part I1 (oral) examination, a candidate must have successfully completed the Part I (written) examination.
A candidate applying for admission to the certifying examination must fulfill all the requirements of the Board in force at the time the application is received. Please address all communications to the American Board of Thoracic Surgery, 14640 E Seven Mile Rd, Detroit, MI 48205.
Permanent brain damage is among the most feared complications following cardiac operations and can spoil an otherwise excellent result. Today, open-heart operations seem to be associated with a 12% risk of diffuse encephalopathy as well as a 2 to 5% risk of stroke [l].Although a left ventricular mural thrombus or intraoperative "accidents" may explain the stroke in selected instances [2], more often the reason remains obscure in the individual patient . Compared with the rapid developments in cardiac Surgery, knowledge oibrain pathophysiology, including cerebral circulation, during cardiopulmonary bypass (CPB) is surprisingly scarce. Indirect studies on cerebral
circulation during CPB based on arteriovenous oxygen differences are hard to accept because they are based on the coupling of cerebral metabolism and flow. This coupling seems to be lost during extracorporeal circulation, at least during nonpulsatile CPB [3]. Recent studies of cerebral blood flow during nonpulsatile CPB using the xenon washout technique have also yielded conflicting results [4-61. In a previous study involving 5 patients, we [7] recorded internal carotid artery (ICA) flow (electromagnetic flowmeter) and cerebral perfusion pressure (CPP = mean arterial blood pressure [BPI - mean intracranial epidural pressure [EDP]) during CPB. The record demonstrated increased ICA bulk flow down to CPP levels in the range of 20 mm Hg during CPB. Transcranial Doppler studies offer a new, noninvasive method for continuous evaluation of cerebral circulation [8-lo]. During the last two years, we have used this technique to perform studies of flow velocity in the basal cerebral arteries. Such studies proved to be a useful tool in our management of patients with subarachnoid hemorrhage [lo] as well as patients with occlusive extracranial artery disease [ l l ] . The diameter of the proximal middle cerebral artery (MCA) is supposed to remain relatively constant (12, 131 as regulation of cerebrovascular resistance acts on the arteriolar level, both in autoregulation of cerebral blood flow during changes in CPP as well as during carbon dioxideinduced changes in the cerebral perfusion. Combined studies have demonstrated that changes in MCA flow velocity correspond well with ICA flow changes recorded with an electromagnetic flowmeter and with ICA flow velocities ( r = 0.92).* This article presents our experience with recording MCA flow velocity and monitoring CPP in 11 patients during nonpulsatile CPB. In 1 patient we were able to perform MCA flow velocity and ipsilateral electromagnetic ICA flow recordings simultaneously.
Material and Method
Accepted for puhlication Nov 16, 1984.
Eleven patients undergoing valve replacements and aortocoronary bypass procedures were included in the present study. They were considered to be at high risk because of multiple valve replacements, combined valve and aortocoronary bypass procedures, or a history of stroke prior to the heart operations. Age and sex distribution, CPB times, and operative procedures are summarized in Table 1. Intracranial epidural pressure was recorded by a
Address reprint requests to Dr. Lundar, Dcpartmcnt of Neurosurgery, Rikshospitalet, 0027 Oslo 1, Norway.
'Lundar T, ct al: Unpublished data, 1985
From the Dcpartmcnts ot Surgery and Neurosurgery, Rikshospitalet, Oslo, Norway.
144
145 Lundar et al: Cerebral Perfusion during Nonpulsatile CPB
Table I. Summary of Patient Data Patient No., Age (yr), Sex
Duration of CPB (min)
M M M F M 6. 63, F
111 105 101 106 155
7. 43, M 8. 70, M 9. 60, F 10. 57, F 11. 52, M
210 106 94 57 72
1. 51, 2. 65, 3. 68, 4. 62, 5. 65,
83
CPB = cardiopulmonary bypass; AVR internal carotid artery.
Operative Procedures; Risk Factors
Clinical Result
AVR; previous stroke MVR + CABG x 2 AVR + CABG x 2; occluded left ICA
Good
MVR, replacement of 14-yr-old malfunctioning Beall valve AVR + MVR AVR + prosthesis to aortic arch + CABG x 3; previous AVR, preop Cniididn sepsis AVR + MVR + CABG AVR MVR + CABG MVR; anuria (hemodialysis) R ICA endarterectomy + CABG x 2; previous stroke =
aortic valve replacement; MVR
minitransducer (AME, Horten, Norway) implanted through a burr hole in the right frontal region. The procedure has been described previously [ 141. Informed consent was obtained from each patient. Arterial blood pressure and central venous pressure were recorded with fluid-pressure transducers (AME, Horten, Norway). Cerebral electrical activity was recorded by a cerebral function monitor (Cerebral Function Monitor 4640, Devices Limited, Welwyn Garden City, Hertfordshire, England). Flow velocity in the right MCA was recorded continuously throughout CPB by the transcranial Doppler method, which has previously been described in detail [ 8 ] . Taking advantage of the relatively thin skull bone in the temporal region above the zygomatic arch, Doppler signals from the MCA were obtained using a laboratory prototype, range-gated, 2 MHz, pulsed-wave Doppler instrument with acoustically focusing transducers. In Patient 11, ICA flow was recorded during CPB with an electromagnetic flowmeter (Nycotron, Lier, Norway). All variables were recorded on a six-channel pen recorder (W & W Electronic AG, Basel, Switzerland). Anesthesia was induced with nitrous oxide, fentanyl, diazepam, and pancuronium bromide, and maintained on the same regimen. A Rygg-Kyvsgaard heart-lung machine with a Polystan Venotherm bubble oxygenator was used (Polystan, Copenhagen, Denmark). The regimen for extracorporeal perfusion was standardized to generalized hypothermia (28"to 30°C), flow rite of 1.5 L/min/m', a d a m e a n systemic blood pressure between 40 and 60 mm Hg. Carbon dioxide was added to the oxygenator in concentrations of 2.5 to 5% to obtain temperature-corrected values for arterial CO? tension (PaC0') of about 37.5 mm Hg during CPB. Hematocrit was determined every fifteen seconds during the first 5 minutes of CPB (until stabilization) and thereafter every 15 minutes. In Patient 11, carotid endarterectomy was performed an hour before CPB because of severe right-sided ICA
Good Good Good Died of prolonged pump failure Died of acute pump failure 24 hr postop Good
Good Good Fair
Good
mitral valve replaccnicnt; CABC = coronary artery bypass grafting; ICA
=
stenosis and a previous left-sided stroke. A Pruitt shunt was used during the endarterectomy, and electromagnetic ICA flow was recorded during CPB, as previously described [71.
Results Middle cerebral artery flow velocity increased at the introduction of CPB in all 11 patients (Table 2). As the diameter of the MCA is unknown in the individual patient, it is meaningless to compare the absolute velocity values (centimeters per second) in these patients. When compared with the prebypass value, however, the percent change in velocity will reflect changes in flow in the individual. Therefore, all velocity values in Table 2 are given with the percentage of the prebypass value in parentheses. At the maximum of this initial increase, MCA flow velocities were in the range of 140 to 260% of the prebypass values. The initial increase in MCA flow velocity was followed by a reduction of short duration. This reduction corresponds to the low CPP state that occurs during the wellknown drop in BP at the introduction of CPB. The lowest value of this "dip" in flow velocity as well as the corresponding CPP dip (initial drop in CPP) also are listed in Table 2. The range was 72 to 180%of the prebypass value of 100. During steady-state CPB, MCA flow velocity was markedly increased in 10 of the 11 patients (range, 80 to 300% of the prebypass values). In comparison, the CPP range during steady-state CPB is shown for the individual patient in Table 2, as is a hemodilution ratio (prebypass hematocrit divided by steady-state hematocrit during bypass). Flow velocities in the MCA were obtained during the first period after termination of CPB in 10 patients. In 9 of them, flow velocity was reduced at the conclusion of bypass but remained higher than the prebypass value (see Table 2). Figure 1 demonstrates a typical record of
146 The Annals of Thoracic Surgery Vol 40
No 2 August 1985
Table 2. Data on Cerebral Perfusion during Nonpulsatile CPB
Patient No.
MCA Flow Velocity, Initial Increase (cm/sec)”,b
MCA Flow Velocity during Dip (cm/sec)”
CPP Dip (mm Hg)
25/65 (260)
45 (180) 33 (100)
25 20
60 (150) 36 (100) 36 (150) 60 (150) 60 (120) 35 (117)
20-25 26 22 30 24 15
30 (120) 18 (72) 35 (117)
35 20 20
33/60 (180) 40/100 (250)
9 10 11
36/60 (167) 24/60 (250) 40/60 (150) 50/100 (200) 30/45/60 (150/200) 25/50 (200) 25/35 (140) 30155 (183)
MCA Flow Velocity, Steady-State CPB (cm/sec)”
CPP, SteadyState CPB (mm Hg)
50 30-50 40 50 40
54/22 (2.5) 46/21 (2.2) 54/27 (2.0) 43/21 (2.0) 43/21 (2.0) 41/21 (2.0) 36/17 (2.1) 43/23 (1.9)
40 20-25 20-40
45/21 (2.1) 27/19 (1.4) 42/24 (1.8)
20-40 50 40
70 (280) 55 (167) 110 (275) 60 (167) 55 (229) 80 (200) 120 (240) 90 (300) 50 (200) 20 (80) 70 (233)
Hernodilution‘
MCA Flow Velocity after End of CPB (cm/sec)a
...
CFM Response
60 (150) 80 (160) 50 (167)
Biphasic No change Biphasic Fall Biphasic Fall Biphasic Fall
35 (140) 40 (160) 50 (167)
Biphasic Fall Fall
40 (121) 60 (150) 40 (111) 50 (208)
“Numbers in parentheses are percentage of the prebypass value. bThe first number represents the prebypass value and the second, the initial increase. ‘Numbers are shown for hematocrit measured before/during CPB; the number in parentheses is the ratio
CPB
= cardiopulmonary bypass; MCA = middle cerebral artery; CPP = cerebral perfusion pressure;
CFM
=
cerebral function monitor.
UV
100YCA flow velocity 6 0 .
cm/rec .I
I
. Fig 1 . (Patient 5.) Recording of middle cerebral artery (MCA)flow velocity, blood pressure (BP), epidural intracranial pressure (EDP), and cerebral function monitor (CFM)at the start of cardiopulmonary bypass. Repeat hematocrit (Hct) values illustrate the progress and stabilization of hemodilution. As the priming solution was introduced (arrow a), the MCA flow velocity increased rapidly. A biphasic re-
.
I
minutom
sponse in the activity of the cerebral function monitor was observed. The well-known drop in BP seen at the‘iritroduction of bypass (arrow b) caused the dip in cerebral perfusion pressure fCPP) fBP - EDP) and MCA flow velocity. As BP and CPP were restored (arrow c), MCA flow velocity again increased, to about twice the prebypass value.
147 Lundar et al: Cerebral Perfusion during Nonpulsatile CPB
A
B
Fig 2. (Patient 7.) Spectral display of the Doppler signal recorded ( A ) shortly before introduction of cardiopulmonary bypass (CPB), ( B ) during steady-state CPB, and (C) after termination of CPB. A realtime 64-point spectral analysis of the Doppler-shifted frequencies is performed by a 6502A microprocessor and a fast Fourier transform routine designed by one of us ( R . A.). A special software subroutine translates the spectral envelope into an instantaneous analog flowvelocity signal.
MCA flow velocity, BP, EDP, and the cerebral function monitor during CPB. The initial increase in flow velocity at the introduction of the priming solution was followed by the dip when CPP (BP - EDP) fell to about 22 mm Hg. As BP and CPP recovered, MCA flow velocity once again increased. Hematocrit was markedly decreased immediately after the introduction of the priming solution, and a response on the cerebral function monitor was observed at this point in all but 1 patient. This response was either biphasic (increase followed by depression of the cerebral function monitor activity [as in Fig 1, arrow a ] ) or monophasic (fall in activity). The initial cerebral function monitor responses are also listed in Table 2. Figure 2 demonstrates the MCA flow velocity as it appears on the Doppler monitor before bypass, during steady-state CPB, and after bypass. Although MCA flow velocities generally were markedly increased during steady-state CPB, concomitant changes in flow velocities were observed when CPP changed, indicating a pressure-passive situation during nonpulsatile CPB. This is clearly demonstrated in Figure 3. Cerebral autoregulation is obviously not operative during the spontaneous fall in CPP. The CPP range, however, was very low-40 to 20 mm Hg. Figure 4 shows data obtained from Patient 11 during steady-state hemodilution (hematocrit, 24%), PaCO2 (temperature correlated, 38.25 mm Hg), and temperature (28°C).This was the patient in whom simultaneous recordings of MCA flow velocity and ipsilateral ICA flow were obtained. At arrow a , the heart-lung machine flow was increased from 1.5 Wmin/m2 to 2.0 Wmin/m2. At
C
arrow b, machine flow was reduced back from 2.0 to 1.5 Wmin/m2.The increase in BP and the concomitant slight increase in EDP caused a profound increase in CPP, resulting in markedly increased perfusion. The curves in Figure 4 show the close relationship between the changes in MCA flow velocity and ICA flow. There was no evidence of autoregulation during this period, even when CPP reached levels higher than 60 mm Hg. Between arrows u and b, "vasomotor waves" occurred spontaneously in BP, which are also reflected in EDP, MCA flow velocity, and ICA flow. The regimen for CPB aimed at keeping temperaturecorrected PaCO2 values at 37.5 mm Hg by adjusting the C02 admixture between 2.5 and 5%. It was necessary to keep PaCO2 constant to be able to study the effect of changes in CPP on flow velocity. The single patient in whom MCA flow velocity was reduced throughout CPB (Patient 10) differed from the others in having reduced renal function, low hematocrit prior to bypass, and, thus, relatively less additional hemodilution as CPB was established (see Table 2). Ultrafiltration was performed before CPB was terminated, and MCA flow velocity actually increased at the termination of bypass. The monitoring of BP, EDP, and cerebral function was discontinued one to four days after operation depending on the individual need for prolonged central nervous system supervision. The results of postoperative monitoring are outside the scope of this report and will be presented elsewhere. A short description of the clinical results is provided in Table 1. There were no complications associated with the monitoring procedures.
Comment The study of cerebral circulation during CPB is hampered by methodological difficulties. There is an obvious need for on-line information on cerebral perfusion during cardiac operations. In a previous study of electromagnetic ICA flowmetry during CPB [7],we found increased ICA flow during bypass. Carotid endarterectomy shortly before the open-heart procedure might in-
1.18 The Annals of Thoracic Surgery Vol 40
No 2 August 1985
Fig 3. (Patient 1.) Records of middle cerebral artery (MCA) Pmc, z&xily, Mood pressure (BP), and epidural intracranial pressure (EDP)during steadystate cardiopulmonary lypass. A spontaneous fall in cerebral perfusion pressure caused profound reduction in MCA florc, d o c i t y .
terfere with the cerebral circulation during the subsequent CPB procedure. However, for ethical reasons, ICA flowmetry records are not obtainable for other kinds of cardiac procedures. Transcranial Doppler recording of MCA flow velocity offers new opportunities to investigate the blood supply of the brain during CPB. The basal cerebral arteries are not involved in the regulatory mechanisms elicited by changes in CPP and COz (12, 131. Simultaneous recordings of MCA flow velocity and ICA flow or ICA flow velocity changes in response to changes in CPP have demonstrated that observed changes in MCA flow velocity reflect concomitant and closely related changes in the cerebral circulation (r = 0.92).* These data agree with our observations in Patient 11 (see Fig 4). A recent report on the effect of changes in COz on MCA flow velocity in healthy volunteers demonstrates the classic effect of COz on cerebral circulation and supports the theory that MCA flow velocities reflect changes in MCA volume flow [15]. The present study confirms our previous observations of increased perfusion during nonpulsatile CPB [7]. The regimen for extracorporeal perfusion was changed somewhat from that study: from 2.5 L/min/m2 and morphine anesthesia to 1.5 L/min/m' and fentanyl anesthesia. Nevertheless, MCA flow veloc'Lundar T, et al: Unpublished data, 1985.
ity is clearly increased during CPB compared with the prebypass level, and the increase is related to the degree of hemodilution (see Table 2). It is impressive how well MCA flow velocity is maintained during CPB, considering the low CPP levels observed. It is obvious, however, that autoregulation is not operative during the CPP levels observed during nonpulsatile CPB in this series (20 to 60 mm Hg). This is clearly shown in Figures 3 and 4, which are observations made during steady-state hemodilution, steady-state PaC02, and constant temperature. Several cerebral function monitor studies during CPB have been published [16-181. Both depression of the activity of the cerebral function monitor and the biphasic response at the introduction of CPB are well-known patterns. More profound changes, with depression of the activity toward the zero line, were not observed in this series. The short-lasting depression of activity seen during the initial increase in flow velocity, occurring during maximum hemodilution, can be interpreted as insufficient oxygenation despite increased MCA flow velocity, and is suggestive of a qualitatively insufficient perfusion. During CPB, temperature-corrected PaCOz values were maintained at 37.5 mm Hg by adjusting the COz admixture. Further manipulations with COz were avoided so as not to interfere with the study of changes in MCA flow velocity relative to changes in CPP. Observations during COZ manipulations are required, however, to determine to what extent COr reactivity is maintained during CPB. Such studies are now in progress. Previous clinical studies during nonpulsatile CPB using xenon washout or arteriovenous 0, difference tech-
149 Lundar et al: Cerebral Perfusion during Nonpulsatile CPB
Fig 4. (Patient 11.)Siniultaiieous recvrdings of middle cerebral artery (MCA) pow wlocity, ipsilateral internal camtid artery (ICA) pocv (electroniagnetic),blood pressure (BP),epidural intmcranial pressure (EDP),and cerebral futiction monitor (CFM) during established cardiopulmonary @pas.fncrrasing machine floui from 1.5 to 2.0 Uniinl m2 (arrow a) caused a tnarhd increase iri cerebral perfusion pressure (CPP) a d a corresponding increase iti MCA j 7 o u g wlm*tyand ICA flow. Note the close refatiorishipbetriven MCA pout wlocity and ICA pow changes. As nrachine pow was reduced from 2.0 to 1.5 Uminltn' (arrow b), subsequent reductions iti CPP, MCAfloup velocity, and ICA pour were observed. Diiring the i n c r d machine flow (bettiwen arrows a and b), "vasontotor tuoves" in BP were seen. Gnicontitant changes in EDP, MCA pow twlocity, and fCAflori~are also demonstrated. Temperature (28°C). hemudilution (hematmit, 24%), and partial pressure of arterial C 0 2 (teniperatirre corrected, 38.25 nnn Hg)uwre constant during the recordings. niques have indicated preserved COz reactivity [4,61. Because our n u m b e r of observations during CPP levels greater than 50 m m Hg is limited, extended studies are needed to determine whether autoregulation is completely lost during nonpulsatile CPB or if CPP is simply below t h e individual lower limit of autoregulation during t h e actual degree of hemodilution a n d regimen for extracorporeal perfusion. Such studies cannot be performed without intracranial pressure monitoring, because BP alone can be a poor indicator of the CPP state in such situations [ 191. Comparative studies on nonpulsatile a n d pulsatile CPB a n d on different CPB regimens a r e needed. W e believe that further studies of cerebral hemodynamics during CPB procedures may have considerable clinical implications.
References 1. Furlan AJ, Breuer AC: Central nervous system complications of open heart surgery. Stroke 15:912, 1984 2. Breuer AC, Franco I, Mariewski D, Soto-Velasco J: Left
3.
4.
5. 6.
ventricular thrombi seen by ventriculography are a significant risk factor for stroke in open heart surgery. Ann Neurol 10:103, 1981 Andersen K, Waaben J, Husum 6, et al: Non-pulsatile cardiopulmonary by-pass disrupts flow-metabolism couple in the brain. Presented at the Annual Meeting of the Scandinavian Association for Thoracic and Cardiovascular Surgery, Turku, Finland, Aug 24-26, 1983 Henriksen L, Hjems E, Lindeburgh T: Brain hvperperfusion during cardiac operations: cerebral blood flow measured in man by intra-arterial injection of xenon 133: evidence suggestive of intraoperative microembolism. J Thorac Cardiovasc Surg 86202, 1983 McKay RD, Reves JG, Karp RB, et al: Effects of cardiopulmonary bypass on cerebral blood flow. J Cereb Blood Flow Metab 3:Suppl 1:119, 1983 Govier AV, Reves JG, McKay RD, et al: Factors and their influence on regional cerebral blood flow during nonpulsatile cardiopulmonary bypass. Ann Thorac Surg 38592,
1984 7. Lundar T,Freysaker T,Lindegaard KF, et al: Some observa-
tions on cerebral perfusion during Cardiopulmonary bypass. Ann Thorac Surg 39:318, 1985 8. Aaslid R, Marcwalder TM, Nornes H: Noninvasive transcranial Doppler ultrasound recording of flow velocity in basal cerebral arteries. J Neurosurg 57:769, 1982 9. Aaslid R, Nornes H: Musical murmurs in human cerebral arteries after subarachnoid hemorrhage. J Neurosurg 60:32, 1984 10. Aaslid R, Huber P, Nornes H: Evaluation of cerebrovascu-
5
150 The Annals of Thoracic Surgery Vol 40 No 2 August 1985
11.
12. 13.
14.
15.
lar spasm with transcranial Doppler ultrasound. J Neurosurg 60:37, 1984 Lindegaard KF, Grolimund P, Bakke SJ, et al: Carotid artery disease: assessment of intracranial hemodynamic pattern by noninvasive transcranial Doppler ultrasound. J Neurosurg (in press, 1985) Strandgaard S, Paulson OB: Cerebral autoregulation. Stroke 15:413, 1984 Huber P, Handa J: Effect of contrast material, hypercapnia, hyperventilation, hypertonic glucose and papaverine on the diameter of the cerebral arteries: angiographic determination in man. Invest Radio1 2:17, 1967 Lundar T, Fraysaker T, Nornes H, Lilleaasen P: Aspects of cerebral perfusion on open-heart surgery. Scand J Thorac Cardiovasc Surg 16:217, 1982 Markwalder TM, Grolimund P, Seiler RW, et al: Depen-
16.
17.
18. 19.
dency of blood flow velocity in the middle cerebral artery on end-tidal carbon dioxide partial pressure: a transcranial ultrasound Doppler study. J Cereb Blood Flow Metab 4:368, 1984 Schwartz MS, Colvin MP, Prior PF, et al: The cerebral function monitor: its value in predicting the neurological outcome in patients undergoing cardiopulmonary bypass. Anaesthesia 29:135, 1973 Kritikou PE, Branthwaite MA: Significance of changes in cerebral electrical activity at onset of cardiopulmonary bypass. Thorax 32:534, 1977 Prior PF: Monitoring Cerebral Function. Amsterdam, ElseviedNorth Holland, 1979 Lundar T, Freysaker T, Nornes H: The clinical significance of changes in cerebral perfusion pressure during openheart surgery. Scand J Thorac Cardiovasc Surg 17163, 1983
Notice from the American Board of Thoracic Surgery The Part I (written) examination will be held at the Amfac Hotel, DallasFort Worth Airport, Dallas, TX, in February, 1986. The closing date for registration was August 1, 1985. To be admissible for the Part I1 (oral) examination, a candidate must have successfully completed the Part I (written) examination.
A candidate applying for admission to the certifying examination must fulfill all the requirements of the Board in force at the time the application is received. Please address all communications to the American Board of Thoracic Surgery, 14640 E Seven Mile Rd, Detroit, MI 48205.