- Email: [email protected]
Electroencephalographic changes and their regional differences during pediatric cardiovascular surgery with hypothermia
Brain & Development 23 (2001) 115±121
www.elsevier.com/locate/braindev
Original article
Electroencephalographic changes and their regional differences during pediatric cardiovascular surgery with hypothermia Tomoyuki Akiyama a,*, Katsuhiro Kobayashi a, Tomoyuki Nakahori a, Harumi Yoshinaga a, Tatsuya Ogino a,b, Yoko Ohtsuka a, Mamoru Takeuchi c, Kiyoshi Morita c, Shunji Sano d, Eiji Oka a a
Department of Child Neurology, Okayama University Medical School, 2-5-1 Shikata-cho, Okayama 700-8558, Japan b Department of Pediatrics, Mabi Central Hospital, Okayama, Japan c Department of Anesthesiology & Resuscitology, Okayama University Medical School, Okayama, Japan d Department of Cardiovascular Surgery, Okayama University Medical School, Okayama, Japan Received 11 August 2000; received in revised form 11 December 2000; accepted 25 January 2001
Abstract Monitoring brain function by EEG is an important means of preventing cerebral insults in pediatric cardiovascular surgery. We studied intraoperative EEG changes and their regional differences associated with hypothermia and brain ischemia. The subjects of this study consisted of 13 children ranging in age from 4 months to 4 years and 6 months. Multi-channel EEGs were recorded using a portable digital EEG system, and the EEG changes were examined by visual inspection and computerized analyses. The results were as follows. (1) During cooling, a discontinuous EEG pattern was transiently observed in four patients, and this phenomenon indicated rapid suppression of cerebral function and subsequent adaptation. (2) Regarding the patterns of change in equivalent potentials induced by hypothermia, there were two different patterns depending on the degree of hypothermia, and the borderline rectal temperature was found to be around 328C. (3) During cooling, regional differences in the changes in equivalent potentials were observed in nine patients. A decrease in slow waves was marked in the occipital head area, and a decrease in fast waves was prominent in the anterior head area. (4) Arterial hypotension caused transient EEG abnormalities. Of them, bilaterally synchronous rhythmic high voltage slow waves were remarkable and exhibited bifrontal or bicentral dominance. (5) The EEG changes induced by hypothermia were in¯uenced not only by the rectal temperature itself, but also by the rate of change in rectal temperature, and we speculated that this phenomenon was a result of adaptation. In intraoperative EEG monitoring, these ®ndings constitute the basis for early detection of a cerebral hypoxic±ischemic state during pediatric cardiovascular surgery. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Pediatric cardiovascular surgery; Hypothermia; Extracorporeal circulation; Electroencephalographic monitoring; Electroencephalographic analysis
1. Introduction
2. Subjects and methods
Currently, operations are performed for severe congenital heart diseases, even in patients in early infancy, and it is especially critical in these operations to prevent postoperative neurological complications. Monitoring the brain function by EEG is important for this purpose during pediatric cardiovascular surgery involving extracorporeal circulation and hypothermia. In this study, we elucidated the characteristics of EEG changes associated with hypothermia and transient ischemia. These EEG changes were examined in detail, especially in regard to the morphology of waveforms, frequency properties, and regional differences.
The subjects consisted of 13 children (nine boys and four girls) who underwent open-heart surgery involving extracorporeal circulation and hypothermia at the Okayama University Hospital. Their ages at the time of operation ranged from 4 months to 4 years and 6 months (mean, 2 years and 2 months) and their weights ranged from 4.6 to 16.8 kg (mean, 9.8 kg). Patients who had gross neurological abnormalities before surgery or chromosomal anomalies were excluded from the study. While the patient was under anesthesia, the EEG was recorded via 21 scalp and earlobe electrodes of the 10±20 electrode system. A portable digital electroencephalograph system, Ceegraph E w (Bio-logic, USA), was used. The mean arterial blood pressure, central venous pressure, rectal temperature, O2 saturation, and expiratory CO2 tension were monitored simultaneously.
* Corresponding author. Fax: 181-86-235-7377. E-mail address: [email protected] (T. Akiyama).
0387-7604/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0387-760 4(01)00192-9
116
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
Anesthesia was performed by intravenous injection of fentanyl. Nitrous oxide, iso¯urane, and sebo¯urane were added if necessary. The administration of these inhalation anesthetics was completely stopped before the start of extracorporeal circulation. Midazolam and chlorpromazine were used at the beginning of extracorporeal circulation for maintenance of the anesthetic level and vasodilation, respectively. Cooling was started after the establishment of extracorporeal circulation. The degree of hypothermia was mild or moderate. The EEGs were analyzed by self-made programs written using Visual Basic w and Visual C11 w (Microsoft, USA). Averaged power spectra were computed by fast Fourier transform. Equivalent potentials, i.e. the square root of power values, were also computed to build deviation ratio topographic maps [1]. In the analysis of the EEG during cooling or rewarming, a 60-s segment immediately before cooling or rewarming was used as the baseline. Equivalent potentials and deviation ratio topographic maps were sequentially obtained every 60 s, and the changes in equivalent potentials and their regional differences were examined. Averaged power spectra were computed for every 60-s segment of EEG data during extracorporeal circulation in order to build compressed spectral arrays and to calculate the spectral edge frequency (SEF) which indicates the 95thpercentile power frequency. Relationships between the changes in SEF and those in the mean arterial blood pressure, central venous pressure, and rectal temperature were evaluated. For statistical evaluation, a multiple regression analysis was performed using JMP w (SASS, USA) with the signi®cance level set at 0.05.
3. Results 3.1. EEG changes during cooling The summaries of EEG changes during cooling are shown in Table 1. Visual inspection of the EEGs revealed that fast waves induced by the anesthetics decreased in 12 of 13 patients. In the remaining patient, patient 13, in whom the rectal temperature at the end of cooling was 34.08C and the highest of all patients, a decrease in the frequency of fast waves was observed. During the process of cooling (rectal temperature, 30.0± 36.58C), a discontinuous EEG pattern was observed in ®ve patients. In this pattern, the low amplitude phase lasting 2±13 s alternated with the phase composed of 10±100 mV, 2±12 Hz waves lasting 3±6 s (Fig. 1). In four of ®ve patients (patients 3, 6, 8, and 9), this pattern was a transient phenomenon and disappeared at 28.6±30.48C. In patient 4, this pattern continued during the period of hypothermia and disappeared after the start of rewarming. The changes in equivalent potentials were studied in
11 patients with EEGs which were almost free of artifacts and high-amplitude abnormalities. In nine patients with rectal temperatures below 32.18C at the end of cooling (patients 1±4 and 7±11: rectal temperature range, 25.3± 32.18C), a decrease in equivalent potentials was seen in all frequency bands, especially in the b1 and b2 bands, except for the u band in patient 10. In the remaining two patients with rectal temperatures of 32.78C or higher at the end of cooling (patients 12 and 13: rectal temperatures, 32.7 and 34.08C, respectively), a decrease in the equivalent potential in the b2 band was observed. An increase in equivalent potentials in lower frequency bands was also observed, especially in patient 13. Thus, the patterns of changes in equivalent potentials appeared to differ depending on the rectal temperature at the end of cooling, and in the present patients, the borderline temperature was around 328C. A decrease in equivalent potentials was noted in the posterior head area in the d1 to u bands, and in the anterior head area, in the a to b2 bands. 3.2. EEG changes during rewarming The summaries of EEG changes during rewarming are shown in Table 1. An increase in fast waves was observed in 12 of 13 patients. In patient 13, who showed persisting fast waves during cooling, an increase in the frequency of fast waves was observed. At the end of rewarming in all patients, the EEGs showed low voltage fast waves and appeared to be the same as those before the start of cooling. Changes in equivalent potentials were examined in ten patients with EEGs which were almost free of artifacts and high-amplitude abnormalities. An increase in equivalent potentials in the b2 band was observed in all patients except for patient 13. The changes in equivalent potentials in the other frequency bands were inconsistent. No remarkable regional differences were seen in the changes of equivalent potentials during rewarming. 3.3. EEG changes in compressed spectral arrays Power spectra were obtained from C3 and C4 where contamination with artifacts was minimal. The relations of rectal temperature, its rate of change, mean arterial blood pressure, and central venous pressure to SEF were studied using multiple linear regression analysis. Since the O2 saturation was kept at 100% and the expiratory CO2 tension was not monitored during extracorporeal circulation, they were not included in the analysis. The results are shown in Table 2. In ten of 13 patients (patients 2, 3, and 6±13), the SEF correlated signi®cantly with both rectal temperature and its rate of change. The SEF correlated with only rectal temperature in one patient, and only with the rate of change in rectal temperature in another patient. In ®ve patients (patients 2, 8, 9, 11 and 13), the rate of change in rectal temperature had a greater effect on SEF than did rectal temperature itself.
b
a
1 2 3 4 5 6 7 8 9 10 11 12 13
1 1 1 1 1 1 1 1 1 1 1 1 Decrease in frequency of fast waves 1 1
1
1 1 # #P # # #P " "
#P #P #C #P
d1
# #P #P # # " !
#P #P # #
d2
# #A #P ! # ! "
# #P # #
u
# #A #A #A #A # "
#P # #A #A
a
# #A #A #A # # "
#P # # #
b1
Changes in equivalent potentials
# #A #A # # # #A
#P # # #A
b2 25.9 25.8 27.8 28.1 28.2 29.9 30.2 29.9 30.1 30.1 32.1 34.3 33.9
Rectal temperature at the start of rewarming (8C)
Rewarming
P, posterior dominance; C, central dominance; A, anterior dominance; ( " ), increase; ( ! ), no remarkable change; ( # ), decrease. d1, 1.0±1.5 Hz; d2, 2.0±3.5 Hz; u, 4.0±7.5 Hz; a, 8.0±12.5 Hz; b1, 13.0±19.5 Hz; b2, 20.0±29.5 Hz.
25.3 27.2 27.9 28.1 28.1 29.9 30.0 30.1 30.3 30.3 32.1 32.7 34.0
Rectal temperature Visual evaluation at the end of cooling (8C) Decrease in fast Discontinuous waves pattern
Patient Cooling
Table 1 EEG changes during cooling and rewarming a,b
1 1 1 1 1 1 1 1 1 1 1 1 Increase in frequency of fast waves
Increase in fast waves
Visual evaluation
! " " # # !
"P ! #
"
d1
# " " # ! !
" ! #
"
d2
# ! # # " #
# ! #
#
u
# "A # #A " #
" # #
#
a
" "A " ! ! #
" # "
"
b1
Changes in equivalent potentials
" " " " " #
" " "
"
b2
T. Akiyama et al. / Brain & Development 23 (2001) 115±121 117
118
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
Fig. 1. Discontinuous EEG pattern transiently observed during cooling (patient 3). The rectal temperature was 30.38C. An average of the potentials at electrodes A1 and A2 was used as the reference.
Compressed spectral arrays of the representative patients (patients 4 and 11) are shown in Fig. 2 with correspondence to rectal temperature and its rate of change. The mean arterial blood pressure and central venous pressure were not correlated with SEF, or had very little effect on SEF in comparison with rectal temperature and its rate of change.
reduction in the mean arterial blood pressure at the time of arterial and venous cannulation, immediately after the start of extracorporeal circulation, and during the period of reduction of pump ¯ow for the convenience of operational procedures (Table 3). A decrease in fast waves, voltage attenuation, and the emergence of diffuse irregular high voltage slow waves and bilaterally synchronous rhythmic high voltage slow waves were noted as abnormal EEG patterns. Bilaterally synchronous rhythmic high voltage slow waves were observed only at the time of cannulation and immediately after the start of extracorporeal circulation, and they showed bifrontal or
3.4. Transient EEG abnormalities In ten of 13 patients, transient EEG abnormalities suddenly appeared, and they temporally coincided with a
Table 2 Multiple linear regression analysis: spectral edge frequency ®tted by rectal temperature and the rate of change in rectal temperature a,b Patient
1 2 3 4 5 6 7 8 9 10 11 12 13 a
R 2c
0.667 0.719 0.629 0.592 0.053 0.842 0.333 0.779 0.638 0.800 0.910 0.506 0.412
P
,0.001 ,0.001 ,0.001 ,0.001 NS ,0.001 0.010 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001
Rectal temperature
Rate of change in rectal temperature
Mean arterial blood pressure
Central venous pressure
bt
P
br
P
bm
P
bc
P
0.070 0.550 0.634 0.803 0.193 0.961 0.728 0.576 0.707 0.639 0.314 0.598 0.423
NS ,0.001 ,0.001 ,0.001 NS ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 0.002 ,0.001 0.004
0.713 0.776 0.268 20.235 0.010 0.518 0.353 0.850 0.884 0.459 0.968 0.232 0.592
,0.001 ,0.001 ,0.001 NS NS ,0.001 0.009 ,0.001 ,0.001 ,0.001 ,0.001 0.045 ,0.001
0.089 0.229 0.120 0.145 20.059 20.072 0.180 0.231 20.011 0.121 0.036 0.174 0.060
NS ,0.001 NS NS NS NS NS 0.030 NS NS NS NS NS
20.192 20.125 0.106 N/A 20.029 20.105 20.213 20.001 20.151 0.126 0.078 20.231 0.072
0.008 0.041 NS N/A NS NS NS NS NS 0.039 NS NS NS
Spectral edge frequency indicates the 95th-percentile power frequency. b t, b r, b m and b c, standard partial regression coef®cients corresponding to rectal temperature, the rate of change in rectal temperature, mean arterial blood pressure, and central venous pressure, respectively; NS, no signi®cance; N/A, not available. c R 2, coef®cient of determination. b
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
119
Fig. 2. Relationship between power spectrum and body temperature. The spectral edge frequency is indicated by black dots on the spectrum. (a) Patient 4: the spectral edge frequency changed in parallel to the rectal temperature. (b) Patient 11: the spectral edge frequency was related not only to the rectal temperature, but also to the rate of change in rectal temperature. RT, rectal temperature.
bicentral dominance (Fig. 3). The duration of bilaterally synchronous rhythmic high voltage slow waves was briefer than 20 s. The other types of EEG abnormalities lasted for up to 274 s. Every EEG abnormality was transient and reversible, and the EEGs recovered following the recovery of mean arterial blood pressure. 4. Discussion In pediatric heart surgery involving hypothermic anesthe-
sia and extracorporeal circulation, many factors may in¯uence brain function, which can be monitored by EEG. Therefore, it is very important to have a thorough knowledge of the EEG changes associated with hypothermia and abnormal ischemic±hypoxic states. Regarding the previous reports of EEG changes induced by hypothermia, Pagni et al. [2] stated the following: EEGs changed little until the esophageal temperature reached 288C, then slowing proceeded below 28±258C, paroxysmal high voltage discharges appeared below 20±158C, and ®nally, a suppression±burst pattern and ¯at pattern were
Table 3 Transient EEG abnormalities and their corresponding conditions a Condition
Cannulation Start of ECC Pump ¯ow down Clamping of aorta Declamping of aorta End of ECC Others a
Number of patients
Abnormal EEG pattern Bilaterally synchronous rhythmic HVS
Diffuse irregular HVS
Decrease in fast waves
Voltage attenuation
4 7
3/4 6/7
3/4 5/7
4/4 7/7
1/4 1/7
3 3 1 3
0/3 0/3 0/1 0/3
3/3 2/3 1/1 2/3
3/3 3/3 1/1 2/3
0/3 1/3 0/1 0/3
ECC, extracorporeal circulation; HVS, high voltage slow waves.
120
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
observed at temperatures lower than 158C. Hicks et al. [3] reported progressive slowing of the dominant rhythms when the nasopharyngeal temperature was between 35 and 308C, a decrease in faster rhythms and an increase in slow activity between 29 and 248C, the appearance of periodic bursts below 248C, and a ¯at pattern below 188C. A decrease in fast waves accompanying a reduction in rectal temperature was also observed in our study. However, a discontinuous EEG pattern was observed during the process of cooling. This pattern was different from paroxysmal discharges or the suppression±burst pattern which reportedly appeared during deep hypothermia, in that the high-amplitude phase lasted for up to 6 s and EEG activity did not completely vanish during the low amplitude phase. Generally, periodic patterns in EEGs are observed under deep anesthesia [4], in cases of severe brain damage [5], or as trace alternant during the quiet sleep of newborns. Therefore, periodic patterns are thought to be based on a profound loss of function or prematurity of the brain. A discontinuous EEG pattern transiently appeared during cooling and disappeared even when the hypothermic state continued. It is possible that rapid cooling caused dysfunction of the brain which induced the discontinuous pattern, and that the brain function recovered afterwards by adaptation. We utilized a digital EEG system for the analysis of EEG changes induced by hypothermia in order to clarify the
following points. First, we identi®ed two patterns of EEG slowing with different changing modes of equivalent potentials during hypothermia. In mild hypothermia, with the lowest rectal temperature being 32.78C or above, a decrease in equivalent potentials of fast waves and an increase in those of slow waves were observed. In moderate hypothermia with a lower rectal temperature, equivalent potentials of slow waves as well as those of fast waves decreased, and the apparent EEG slowing was due to a relative increase in slow waves in comparison with fast waves, since fast waves decreased more markedly than slow waves. Secondly, compressed spectral arrays revealed that EEG changes during cooling and rewarming were in¯uenced not only by the rectal temperature itself, but also by the rate of change in rectal temperature. This phenomenon may be the result of adaptation, as in the case of the discontinuous EEG pattern. In previous studies on the power spectral analysis of intraoperative EEGs [6±8], there was only a general consensus regarding frequency slowing in association with hypothermia. However, the reported patterns of changes in total power and frequency parameters during cooling and rewarming disagreed considerably. In these studies, only body temperature was taken into account to explain the EEG changes, while the rate of change in temperature was not. We showed that rectal temperature is not the only factor
Fig. 3. Bilaterally synchronous rhythmic high voltage slow waves (patient 9). A burst of diffuse 2.5±3 Hz rhythmic high voltage slow waves appeared with bifrontal predominance 15 s after the start of extracorporeal circulation. The high voltage slow waves subsided 30 s after the start of extracorporeal circulation. An average of the A1 and A2 potentials was used as the reference.
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
related to EEG power, and that the rate of change in temperature is also important in explaining EEG changes. In the current study, multi-channel EEG recording enabled us to discover regional differences in the EEG changes during cooling. It is well known that various brain regions have different degrees of tolerance to hypoxia [9]. Similarly, hypothermia may have different degrees of effects depending on the brain regions. In our study, a decrease in fast waves during cooling was prominent in the frontal regions and suppression of cerebral function might be more profound in the regions where fast waves decrease in the early stage of cooling. There were no consistent ®ndings in terms of the EEG changes during rewarming. This was probably due to considerable differences in conditions among subjects, including the rectal temperature at the beginning of rewarming, the duration of time spent rewarming, and the anesthetic level. Concerning transient EEG abnormalities, Salerno et al. [10] found slow waves and voltage attenuation mostly at the onset of extracorporeal circulation. Hicks et al. [3] observed a decrease in fast waves and an emergence of slow waves and epileptiform discharges at the beginning of extracorporeal circulation. We noted diffuse irregular high voltage slow waves, a decrease in fast waves, voltage attenuation and bilaterally synchronous rhythmic high voltage slow waves as the transient EEG abnormalities. All of these ®ndings appeared immediately after the reduction of mean arterial blood pressure. Bilaterally synchronous rhythmic high voltage slow waves, the most remarkable pattern among the transient EEG abnormalities, were not observed when the pump ¯ow was carefully reduced. They appeared only when the mean arterial blood pressure dropped uncontrollably at the time of cannulation or at the start of extracorporeal circulation. In addition, they emerged during the normothermic phase prior to cooling. Bilaterally synchronous rhythmic high voltage slow waves may indicate transient but considerable brain dysfunction, although postoperative neurological complications were not observed in any patients in this study. Bilaterally synchronous rhythmic high voltage slow waves have not been reported in any other studies so far. This pattern was a clearly recognizable EEG abnormality with bifrontal or bicentral dominance, but it might have been regarded as similar to the ordinary slow waves in the previous reports describing 2±4 channel recordings. Regarding the pathological basis of bilaterally synchronous discharges, Gloor et al. [11] reported that bilaterally synchronous paroxysmal discharges were observed in
121
patients with diffuse dysfunction of both cortical and subcortical gray matter, and that they were not prominent in patients with dysfunction of the cerebral cortex alone. Therefore, bilaterally synchronous rhythmic high voltage slow waves indicate dysfunction of the subcortical gray matter, as well as of the cerebral cortex. EEG abnormalities may happen to be localized as mentioned above. When multi-channel recording is impossible for an intraoperative EEG, we recommend monitoring of the bifrontal or bicentral regions, because the chance to record EEG abnormalities is higher in these regions than in others. In cardiovascular surgery, early detection of brain ischemia and hypoxia is a critical issue. Intraoperative EEG monitoring based on the knowledge of these ®ndings is very important for the prevention of postoperative neurological complications. References [1] Shichijo F, Matsumoto K. Functional EEG mapping: clinical application of deviation ratio topography (DRT). Rinsho Noha 1994;36:629± 633 in Japanese. [2] Pagni CA, Courjon J. Electroencephalographic modi®cations induced by moderate and deep hypothermia in man. Acta Neurochir Suppl 1964;13:35±49. [3] Hicks RG, Poole JL. Electroencephalographic changes with hypothermia and cardiopulmonary bypass in children. J Thorac Cardiovasc Surg 1981;76:781±786. [4] Faulconer Jr A. Correlation of concentrations of ether in arterial blood with electroencephalographic patterns occurring during ether±oxygen and during nitrous oxide, oxygen and ether anesthesia of human surgical patients. Anesthesiology 1952;13:361±369. [5] Hockaday JM, Potts F, Epstein E, Bonazzi A, Schwab RS. Electroencephalographic changes in acute cerebral anoxia from cardiac or respiratory arrest. Electroenceph clin Neurophysiol 1965;18:575± 586. [6] Levy WJ. Quantitative analysis of EEG changes during hypothermia. Anesthesiology 1984;60:291±297. [7] Russ W, Kling D, Sauerwein G, Hempelmann G. Spectral analysis of the EEG during hypothermic cardiopulmonary bypass. Acta Anaesthesiol Scand 1987;31:111±116. [8] Bashein G, Nessly ML, Bledsoe SW, Townes BD, Davis KB, Coppel DB, et al. Electroencephalography during surgery with cardiopulmonary bypass and hypothermia. Anesthesiology 1992;76:878±891. [9] Volpe JJ. Brain injury and infant cardiac surgery: overview. In: Jonas RA, Newburger JW, Volpe JJ, editors. Brain injury and pediatric cardiac surgery. Boston, MA: Butterworth±Heinemann, 1996. pp. 1±9. [10] Salerno TA, Lince DP, White DN, Lynn RB, Charrette EJP. Monitoring of electroencephalogram during open-heart surgery. J Thorac Cardiovasc Surg 1978;76:97±100. [11] Gloor P, Kalabay O, Giard N. The electroencephalogram in diffuse encephalopathies: electroencephalographic correlates of grey and white matter lesions. Brain 1968;91:779±802.
www.elsevier.com/locate/braindev
Original article
Electroencephalographic changes and their regional differences during pediatric cardiovascular surgery with hypothermia Tomoyuki Akiyama a,*, Katsuhiro Kobayashi a, Tomoyuki Nakahori a, Harumi Yoshinaga a, Tatsuya Ogino a,b, Yoko Ohtsuka a, Mamoru Takeuchi c, Kiyoshi Morita c, Shunji Sano d, Eiji Oka a a
Department of Child Neurology, Okayama University Medical School, 2-5-1 Shikata-cho, Okayama 700-8558, Japan b Department of Pediatrics, Mabi Central Hospital, Okayama, Japan c Department of Anesthesiology & Resuscitology, Okayama University Medical School, Okayama, Japan d Department of Cardiovascular Surgery, Okayama University Medical School, Okayama, Japan Received 11 August 2000; received in revised form 11 December 2000; accepted 25 January 2001
Abstract Monitoring brain function by EEG is an important means of preventing cerebral insults in pediatric cardiovascular surgery. We studied intraoperative EEG changes and their regional differences associated with hypothermia and brain ischemia. The subjects of this study consisted of 13 children ranging in age from 4 months to 4 years and 6 months. Multi-channel EEGs were recorded using a portable digital EEG system, and the EEG changes were examined by visual inspection and computerized analyses. The results were as follows. (1) During cooling, a discontinuous EEG pattern was transiently observed in four patients, and this phenomenon indicated rapid suppression of cerebral function and subsequent adaptation. (2) Regarding the patterns of change in equivalent potentials induced by hypothermia, there were two different patterns depending on the degree of hypothermia, and the borderline rectal temperature was found to be around 328C. (3) During cooling, regional differences in the changes in equivalent potentials were observed in nine patients. A decrease in slow waves was marked in the occipital head area, and a decrease in fast waves was prominent in the anterior head area. (4) Arterial hypotension caused transient EEG abnormalities. Of them, bilaterally synchronous rhythmic high voltage slow waves were remarkable and exhibited bifrontal or bicentral dominance. (5) The EEG changes induced by hypothermia were in¯uenced not only by the rectal temperature itself, but also by the rate of change in rectal temperature, and we speculated that this phenomenon was a result of adaptation. In intraoperative EEG monitoring, these ®ndings constitute the basis for early detection of a cerebral hypoxic±ischemic state during pediatric cardiovascular surgery. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Pediatric cardiovascular surgery; Hypothermia; Extracorporeal circulation; Electroencephalographic monitoring; Electroencephalographic analysis
1. Introduction
2. Subjects and methods
Currently, operations are performed for severe congenital heart diseases, even in patients in early infancy, and it is especially critical in these operations to prevent postoperative neurological complications. Monitoring the brain function by EEG is important for this purpose during pediatric cardiovascular surgery involving extracorporeal circulation and hypothermia. In this study, we elucidated the characteristics of EEG changes associated with hypothermia and transient ischemia. These EEG changes were examined in detail, especially in regard to the morphology of waveforms, frequency properties, and regional differences.
The subjects consisted of 13 children (nine boys and four girls) who underwent open-heart surgery involving extracorporeal circulation and hypothermia at the Okayama University Hospital. Their ages at the time of operation ranged from 4 months to 4 years and 6 months (mean, 2 years and 2 months) and their weights ranged from 4.6 to 16.8 kg (mean, 9.8 kg). Patients who had gross neurological abnormalities before surgery or chromosomal anomalies were excluded from the study. While the patient was under anesthesia, the EEG was recorded via 21 scalp and earlobe electrodes of the 10±20 electrode system. A portable digital electroencephalograph system, Ceegraph E w (Bio-logic, USA), was used. The mean arterial blood pressure, central venous pressure, rectal temperature, O2 saturation, and expiratory CO2 tension were monitored simultaneously.
* Corresponding author. Fax: 181-86-235-7377. E-mail address: [email protected] (T. Akiyama).
0387-7604/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0387-760 4(01)00192-9
116
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
Anesthesia was performed by intravenous injection of fentanyl. Nitrous oxide, iso¯urane, and sebo¯urane were added if necessary. The administration of these inhalation anesthetics was completely stopped before the start of extracorporeal circulation. Midazolam and chlorpromazine were used at the beginning of extracorporeal circulation for maintenance of the anesthetic level and vasodilation, respectively. Cooling was started after the establishment of extracorporeal circulation. The degree of hypothermia was mild or moderate. The EEGs were analyzed by self-made programs written using Visual Basic w and Visual C11 w (Microsoft, USA). Averaged power spectra were computed by fast Fourier transform. Equivalent potentials, i.e. the square root of power values, were also computed to build deviation ratio topographic maps [1]. In the analysis of the EEG during cooling or rewarming, a 60-s segment immediately before cooling or rewarming was used as the baseline. Equivalent potentials and deviation ratio topographic maps were sequentially obtained every 60 s, and the changes in equivalent potentials and their regional differences were examined. Averaged power spectra were computed for every 60-s segment of EEG data during extracorporeal circulation in order to build compressed spectral arrays and to calculate the spectral edge frequency (SEF) which indicates the 95thpercentile power frequency. Relationships between the changes in SEF and those in the mean arterial blood pressure, central venous pressure, and rectal temperature were evaluated. For statistical evaluation, a multiple regression analysis was performed using JMP w (SASS, USA) with the signi®cance level set at 0.05.
3. Results 3.1. EEG changes during cooling The summaries of EEG changes during cooling are shown in Table 1. Visual inspection of the EEGs revealed that fast waves induced by the anesthetics decreased in 12 of 13 patients. In the remaining patient, patient 13, in whom the rectal temperature at the end of cooling was 34.08C and the highest of all patients, a decrease in the frequency of fast waves was observed. During the process of cooling (rectal temperature, 30.0± 36.58C), a discontinuous EEG pattern was observed in ®ve patients. In this pattern, the low amplitude phase lasting 2±13 s alternated with the phase composed of 10±100 mV, 2±12 Hz waves lasting 3±6 s (Fig. 1). In four of ®ve patients (patients 3, 6, 8, and 9), this pattern was a transient phenomenon and disappeared at 28.6±30.48C. In patient 4, this pattern continued during the period of hypothermia and disappeared after the start of rewarming. The changes in equivalent potentials were studied in
11 patients with EEGs which were almost free of artifacts and high-amplitude abnormalities. In nine patients with rectal temperatures below 32.18C at the end of cooling (patients 1±4 and 7±11: rectal temperature range, 25.3± 32.18C), a decrease in equivalent potentials was seen in all frequency bands, especially in the b1 and b2 bands, except for the u band in patient 10. In the remaining two patients with rectal temperatures of 32.78C or higher at the end of cooling (patients 12 and 13: rectal temperatures, 32.7 and 34.08C, respectively), a decrease in the equivalent potential in the b2 band was observed. An increase in equivalent potentials in lower frequency bands was also observed, especially in patient 13. Thus, the patterns of changes in equivalent potentials appeared to differ depending on the rectal temperature at the end of cooling, and in the present patients, the borderline temperature was around 328C. A decrease in equivalent potentials was noted in the posterior head area in the d1 to u bands, and in the anterior head area, in the a to b2 bands. 3.2. EEG changes during rewarming The summaries of EEG changes during rewarming are shown in Table 1. An increase in fast waves was observed in 12 of 13 patients. In patient 13, who showed persisting fast waves during cooling, an increase in the frequency of fast waves was observed. At the end of rewarming in all patients, the EEGs showed low voltage fast waves and appeared to be the same as those before the start of cooling. Changes in equivalent potentials were examined in ten patients with EEGs which were almost free of artifacts and high-amplitude abnormalities. An increase in equivalent potentials in the b2 band was observed in all patients except for patient 13. The changes in equivalent potentials in the other frequency bands were inconsistent. No remarkable regional differences were seen in the changes of equivalent potentials during rewarming. 3.3. EEG changes in compressed spectral arrays Power spectra were obtained from C3 and C4 where contamination with artifacts was minimal. The relations of rectal temperature, its rate of change, mean arterial blood pressure, and central venous pressure to SEF were studied using multiple linear regression analysis. Since the O2 saturation was kept at 100% and the expiratory CO2 tension was not monitored during extracorporeal circulation, they were not included in the analysis. The results are shown in Table 2. In ten of 13 patients (patients 2, 3, and 6±13), the SEF correlated signi®cantly with both rectal temperature and its rate of change. The SEF correlated with only rectal temperature in one patient, and only with the rate of change in rectal temperature in another patient. In ®ve patients (patients 2, 8, 9, 11 and 13), the rate of change in rectal temperature had a greater effect on SEF than did rectal temperature itself.
b
a
1 2 3 4 5 6 7 8 9 10 11 12 13
1 1 1 1 1 1 1 1 1 1 1 1 Decrease in frequency of fast waves 1 1
1
1 1 # #P # # #P " "
#P #P #C #P
d1
# #P #P # # " !
#P #P # #
d2
# #A #P ! # ! "
# #P # #
u
# #A #A #A #A # "
#P # #A #A
a
# #A #A #A # # "
#P # # #
b1
Changes in equivalent potentials
# #A #A # # # #A
#P # # #A
b2 25.9 25.8 27.8 28.1 28.2 29.9 30.2 29.9 30.1 30.1 32.1 34.3 33.9
Rectal temperature at the start of rewarming (8C)
Rewarming
P, posterior dominance; C, central dominance; A, anterior dominance; ( " ), increase; ( ! ), no remarkable change; ( # ), decrease. d1, 1.0±1.5 Hz; d2, 2.0±3.5 Hz; u, 4.0±7.5 Hz; a, 8.0±12.5 Hz; b1, 13.0±19.5 Hz; b2, 20.0±29.5 Hz.
25.3 27.2 27.9 28.1 28.1 29.9 30.0 30.1 30.3 30.3 32.1 32.7 34.0
Rectal temperature Visual evaluation at the end of cooling (8C) Decrease in fast Discontinuous waves pattern
Patient Cooling
Table 1 EEG changes during cooling and rewarming a,b
1 1 1 1 1 1 1 1 1 1 1 1 Increase in frequency of fast waves
Increase in fast waves
Visual evaluation
! " " # # !
"P ! #
"
d1
# " " # ! !
" ! #
"
d2
# ! # # " #
# ! #
#
u
# "A # #A " #
" # #
#
a
" "A " ! ! #
" # "
"
b1
Changes in equivalent potentials
" " " " " #
" " "
"
b2
T. Akiyama et al. / Brain & Development 23 (2001) 115±121 117
118
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
Fig. 1. Discontinuous EEG pattern transiently observed during cooling (patient 3). The rectal temperature was 30.38C. An average of the potentials at electrodes A1 and A2 was used as the reference.
Compressed spectral arrays of the representative patients (patients 4 and 11) are shown in Fig. 2 with correspondence to rectal temperature and its rate of change. The mean arterial blood pressure and central venous pressure were not correlated with SEF, or had very little effect on SEF in comparison with rectal temperature and its rate of change.
reduction in the mean arterial blood pressure at the time of arterial and venous cannulation, immediately after the start of extracorporeal circulation, and during the period of reduction of pump ¯ow for the convenience of operational procedures (Table 3). A decrease in fast waves, voltage attenuation, and the emergence of diffuse irregular high voltage slow waves and bilaterally synchronous rhythmic high voltage slow waves were noted as abnormal EEG patterns. Bilaterally synchronous rhythmic high voltage slow waves were observed only at the time of cannulation and immediately after the start of extracorporeal circulation, and they showed bifrontal or
3.4. Transient EEG abnormalities In ten of 13 patients, transient EEG abnormalities suddenly appeared, and they temporally coincided with a
Table 2 Multiple linear regression analysis: spectral edge frequency ®tted by rectal temperature and the rate of change in rectal temperature a,b Patient
1 2 3 4 5 6 7 8 9 10 11 12 13 a
R 2c
0.667 0.719 0.629 0.592 0.053 0.842 0.333 0.779 0.638 0.800 0.910 0.506 0.412
P
,0.001 ,0.001 ,0.001 ,0.001 NS ,0.001 0.010 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 ,0.001
Rectal temperature
Rate of change in rectal temperature
Mean arterial blood pressure
Central venous pressure
bt
P
br
P
bm
P
bc
P
0.070 0.550 0.634 0.803 0.193 0.961 0.728 0.576 0.707 0.639 0.314 0.598 0.423
NS ,0.001 ,0.001 ,0.001 NS ,0.001 ,0.001 ,0.001 ,0.001 ,0.001 0.002 ,0.001 0.004
0.713 0.776 0.268 20.235 0.010 0.518 0.353 0.850 0.884 0.459 0.968 0.232 0.592
,0.001 ,0.001 ,0.001 NS NS ,0.001 0.009 ,0.001 ,0.001 ,0.001 ,0.001 0.045 ,0.001
0.089 0.229 0.120 0.145 20.059 20.072 0.180 0.231 20.011 0.121 0.036 0.174 0.060
NS ,0.001 NS NS NS NS NS 0.030 NS NS NS NS NS
20.192 20.125 0.106 N/A 20.029 20.105 20.213 20.001 20.151 0.126 0.078 20.231 0.072
0.008 0.041 NS N/A NS NS NS NS NS 0.039 NS NS NS
Spectral edge frequency indicates the 95th-percentile power frequency. b t, b r, b m and b c, standard partial regression coef®cients corresponding to rectal temperature, the rate of change in rectal temperature, mean arterial blood pressure, and central venous pressure, respectively; NS, no signi®cance; N/A, not available. c R 2, coef®cient of determination. b
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
119
Fig. 2. Relationship between power spectrum and body temperature. The spectral edge frequency is indicated by black dots on the spectrum. (a) Patient 4: the spectral edge frequency changed in parallel to the rectal temperature. (b) Patient 11: the spectral edge frequency was related not only to the rectal temperature, but also to the rate of change in rectal temperature. RT, rectal temperature.
bicentral dominance (Fig. 3). The duration of bilaterally synchronous rhythmic high voltage slow waves was briefer than 20 s. The other types of EEG abnormalities lasted for up to 274 s. Every EEG abnormality was transient and reversible, and the EEGs recovered following the recovery of mean arterial blood pressure. 4. Discussion In pediatric heart surgery involving hypothermic anesthe-
sia and extracorporeal circulation, many factors may in¯uence brain function, which can be monitored by EEG. Therefore, it is very important to have a thorough knowledge of the EEG changes associated with hypothermia and abnormal ischemic±hypoxic states. Regarding the previous reports of EEG changes induced by hypothermia, Pagni et al. [2] stated the following: EEGs changed little until the esophageal temperature reached 288C, then slowing proceeded below 28±258C, paroxysmal high voltage discharges appeared below 20±158C, and ®nally, a suppression±burst pattern and ¯at pattern were
Table 3 Transient EEG abnormalities and their corresponding conditions a Condition
Cannulation Start of ECC Pump ¯ow down Clamping of aorta Declamping of aorta End of ECC Others a
Number of patients
Abnormal EEG pattern Bilaterally synchronous rhythmic HVS
Diffuse irregular HVS
Decrease in fast waves
Voltage attenuation
4 7
3/4 6/7
3/4 5/7
4/4 7/7
1/4 1/7
3 3 1 3
0/3 0/3 0/1 0/3
3/3 2/3 1/1 2/3
3/3 3/3 1/1 2/3
0/3 1/3 0/1 0/3
ECC, extracorporeal circulation; HVS, high voltage slow waves.
120
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
observed at temperatures lower than 158C. Hicks et al. [3] reported progressive slowing of the dominant rhythms when the nasopharyngeal temperature was between 35 and 308C, a decrease in faster rhythms and an increase in slow activity between 29 and 248C, the appearance of periodic bursts below 248C, and a ¯at pattern below 188C. A decrease in fast waves accompanying a reduction in rectal temperature was also observed in our study. However, a discontinuous EEG pattern was observed during the process of cooling. This pattern was different from paroxysmal discharges or the suppression±burst pattern which reportedly appeared during deep hypothermia, in that the high-amplitude phase lasted for up to 6 s and EEG activity did not completely vanish during the low amplitude phase. Generally, periodic patterns in EEGs are observed under deep anesthesia [4], in cases of severe brain damage [5], or as trace alternant during the quiet sleep of newborns. Therefore, periodic patterns are thought to be based on a profound loss of function or prematurity of the brain. A discontinuous EEG pattern transiently appeared during cooling and disappeared even when the hypothermic state continued. It is possible that rapid cooling caused dysfunction of the brain which induced the discontinuous pattern, and that the brain function recovered afterwards by adaptation. We utilized a digital EEG system for the analysis of EEG changes induced by hypothermia in order to clarify the
following points. First, we identi®ed two patterns of EEG slowing with different changing modes of equivalent potentials during hypothermia. In mild hypothermia, with the lowest rectal temperature being 32.78C or above, a decrease in equivalent potentials of fast waves and an increase in those of slow waves were observed. In moderate hypothermia with a lower rectal temperature, equivalent potentials of slow waves as well as those of fast waves decreased, and the apparent EEG slowing was due to a relative increase in slow waves in comparison with fast waves, since fast waves decreased more markedly than slow waves. Secondly, compressed spectral arrays revealed that EEG changes during cooling and rewarming were in¯uenced not only by the rectal temperature itself, but also by the rate of change in rectal temperature. This phenomenon may be the result of adaptation, as in the case of the discontinuous EEG pattern. In previous studies on the power spectral analysis of intraoperative EEGs [6±8], there was only a general consensus regarding frequency slowing in association with hypothermia. However, the reported patterns of changes in total power and frequency parameters during cooling and rewarming disagreed considerably. In these studies, only body temperature was taken into account to explain the EEG changes, while the rate of change in temperature was not. We showed that rectal temperature is not the only factor
Fig. 3. Bilaterally synchronous rhythmic high voltage slow waves (patient 9). A burst of diffuse 2.5±3 Hz rhythmic high voltage slow waves appeared with bifrontal predominance 15 s after the start of extracorporeal circulation. The high voltage slow waves subsided 30 s after the start of extracorporeal circulation. An average of the A1 and A2 potentials was used as the reference.
T. Akiyama et al. / Brain & Development 23 (2001) 115±121
related to EEG power, and that the rate of change in temperature is also important in explaining EEG changes. In the current study, multi-channel EEG recording enabled us to discover regional differences in the EEG changes during cooling. It is well known that various brain regions have different degrees of tolerance to hypoxia [9]. Similarly, hypothermia may have different degrees of effects depending on the brain regions. In our study, a decrease in fast waves during cooling was prominent in the frontal regions and suppression of cerebral function might be more profound in the regions where fast waves decrease in the early stage of cooling. There were no consistent ®ndings in terms of the EEG changes during rewarming. This was probably due to considerable differences in conditions among subjects, including the rectal temperature at the beginning of rewarming, the duration of time spent rewarming, and the anesthetic level. Concerning transient EEG abnormalities, Salerno et al. [10] found slow waves and voltage attenuation mostly at the onset of extracorporeal circulation. Hicks et al. [3] observed a decrease in fast waves and an emergence of slow waves and epileptiform discharges at the beginning of extracorporeal circulation. We noted diffuse irregular high voltage slow waves, a decrease in fast waves, voltage attenuation and bilaterally synchronous rhythmic high voltage slow waves as the transient EEG abnormalities. All of these ®ndings appeared immediately after the reduction of mean arterial blood pressure. Bilaterally synchronous rhythmic high voltage slow waves, the most remarkable pattern among the transient EEG abnormalities, were not observed when the pump ¯ow was carefully reduced. They appeared only when the mean arterial blood pressure dropped uncontrollably at the time of cannulation or at the start of extracorporeal circulation. In addition, they emerged during the normothermic phase prior to cooling. Bilaterally synchronous rhythmic high voltage slow waves may indicate transient but considerable brain dysfunction, although postoperative neurological complications were not observed in any patients in this study. Bilaterally synchronous rhythmic high voltage slow waves have not been reported in any other studies so far. This pattern was a clearly recognizable EEG abnormality with bifrontal or bicentral dominance, but it might have been regarded as similar to the ordinary slow waves in the previous reports describing 2±4 channel recordings. Regarding the pathological basis of bilaterally synchronous discharges, Gloor et al. [11] reported that bilaterally synchronous paroxysmal discharges were observed in
121
patients with diffuse dysfunction of both cortical and subcortical gray matter, and that they were not prominent in patients with dysfunction of the cerebral cortex alone. Therefore, bilaterally synchronous rhythmic high voltage slow waves indicate dysfunction of the subcortical gray matter, as well as of the cerebral cortex. EEG abnormalities may happen to be localized as mentioned above. When multi-channel recording is impossible for an intraoperative EEG, we recommend monitoring of the bifrontal or bicentral regions, because the chance to record EEG abnormalities is higher in these regions than in others. In cardiovascular surgery, early detection of brain ischemia and hypoxia is a critical issue. Intraoperative EEG monitoring based on the knowledge of these ®ndings is very important for the prevention of postoperative neurological complications. References [1] Shichijo F, Matsumoto K. Functional EEG mapping: clinical application of deviation ratio topography (DRT). Rinsho Noha 1994;36:629± 633 in Japanese. [2] Pagni CA, Courjon J. Electroencephalographic modi®cations induced by moderate and deep hypothermia in man. Acta Neurochir Suppl 1964;13:35±49. [3] Hicks RG, Poole JL. Electroencephalographic changes with hypothermia and cardiopulmonary bypass in children. J Thorac Cardiovasc Surg 1981;76:781±786. [4] Faulconer Jr A. Correlation of concentrations of ether in arterial blood with electroencephalographic patterns occurring during ether±oxygen and during nitrous oxide, oxygen and ether anesthesia of human surgical patients. Anesthesiology 1952;13:361±369. [5] Hockaday JM, Potts F, Epstein E, Bonazzi A, Schwab RS. Electroencephalographic changes in acute cerebral anoxia from cardiac or respiratory arrest. Electroenceph clin Neurophysiol 1965;18:575± 586. [6] Levy WJ. Quantitative analysis of EEG changes during hypothermia. Anesthesiology 1984;60:291±297. [7] Russ W, Kling D, Sauerwein G, Hempelmann G. Spectral analysis of the EEG during hypothermic cardiopulmonary bypass. Acta Anaesthesiol Scand 1987;31:111±116. [8] Bashein G, Nessly ML, Bledsoe SW, Townes BD, Davis KB, Coppel DB, et al. Electroencephalography during surgery with cardiopulmonary bypass and hypothermia. Anesthesiology 1992;76:878±891. [9] Volpe JJ. Brain injury and infant cardiac surgery: overview. In: Jonas RA, Newburger JW, Volpe JJ, editors. Brain injury and pediatric cardiac surgery. Boston, MA: Butterworth±Heinemann, 1996. pp. 1±9. [10] Salerno TA, Lince DP, White DN, Lynn RB, Charrette EJP. Monitoring of electroencephalogram during open-heart surgery. J Thorac Cardiovasc Surg 1978;76:97±100. [11] Gloor P, Kalabay O, Giard N. The electroencephalogram in diffuse encephalopathies: electroencephalographic correlates of grey and white matter lesions. Brain 1968;91:779±802.