A study of the electroencephalogram during surgery with deep hypothermia and circulatory arrest in infants Seventeen infants (2!12 to 28 months old) were continuously monitored by six-channel electroencephalography (EEG) during the entire surgical procedure of open-heart repair. They were subjected to surface hypothermia supplemented by cold extracorporeal circulation (ECC) down to an average esophageal temperature of 21 C., to cardiac arrest of 40 minutes average (range 19 to 62 minutes), and to ECC rewarming. Survival time of the EEG was correlated to esophageal temperature at the time of arrest. EEG reappeared an average of 26 minutes (5.30 to 50) after the start of rewarming ECC and became strictly continuous after 44 minutes. Reappearance latency was well correlated with the duration of arrest. Potential normalization was observed in 13 infants, but true normalization was observed in only 2 infants during the 90 to 120 minute period after ECC. By judging the EEG and by comparing this series with two previous series of moderate and deeper hypothermia in older patients, we concluded that the immediate tolerance of the brain to deep hypothermia and circulatory arrest seems no different in infants and in older patients. 0
M. Weiss, J. Weiss, J. Cotton, F. Nicolas, and J. P. Binet, Paris, France
Deep hypothermia combined with extracorporeal circulation (EEC) and circulatory arrest was used rather extensively about 15 years ago following the publications of Drew,9-11 SealY,19 Young," Dubost.v 13 Weiss," and their associates. During this period important problems concerning pathophysiology-" and apparatus (i.e., heat exchangers) were solved. However, the technique was progressively abandoned following reports of neurologic complications. n, 7. 9, 10 As a result, most surgical teams, for many years, used moderate hypothermia only. The advent of cardiac surgery for neonates and infants has prompted the return to circulatory arrest with deep hypothermia for several reasons: (1) to gain access to a completely quiescent and bloodless field for From
the
Department
of
Cardiac
Marie-Lannelongue, Paris, France.
Received for publication Dec. 16, 1974.
316
Surgery,
Hospital
more precise repair; (2) to free the diminutive surgical field from interfering cannulas; and (3) tentatively, to minimize humoral trauma related to ECC by shortening the duration of cardiopulmonary bypass. This revival of deep hypothermia as applied to infants was led by BarrattBoyes, S-5 who attributed this decision to his filiation to Japanese authors." We were then prompted to use the technique, which has now become a routine procedure in our institution. During the first cases, we2, 21, 22 compared electroencephalographic (EEG) changes observed in very young patients with those observed many years ago in older patients in the course of two large and homogenous series. The principal aims of this preliminary study were (1) to judge the tolerance of the brain to hypothermia and to circulatory arrest specifically in these young patients, (2) to define the average patterns of EEG evolution during the pro-
Volume 70 Number 2 August, 1975
3 17
Study of EEG during surgery in infants
10 9
en
-- 8 7 I 6
.-5 ~4
L:i:i 3
3:
x
"t
"
t
o
I£J
.e
~
I
hospital death
'"
ASO YSO FALLOT
o
2
•
x
1
" •
AP
WINDOW
AYC
TAPYR
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
AGE - months
Fig. 1. Age, weight, diagnosis, and hospital mortality rate of the present series, The curved line represents the third percentile weight line. The oldest patient (28 months) has been included on the same diagram. ASD, Atrial septal defect. VSD, Ventricular septal defect. AP, Aortopulmonary. AVe, Atrioventricular canal. T APVR, Total anomalous pulmonary venous return.
cedure, and (3) tentatively, to optimize some parameters. Methods Seventeen babies, all with severe congestive failure, form the basis of this first report. The diagnosis and ranges of age and weight are summarized in Table I and Fig. 1. Significant points are the high frequency of ventricular septal defect associated' with pulmonary hypertension in the preliminary series and the extension of the age range toward 2 years as we gained evidence suggesting good brain tolerance. Hypothermia and cardiopulmonary bypass procedure. Small doses of thiopentone and suxamethonium were given for induction of anesthesia and tracheal intubation, with halothane used thereafter. Catheters were inserted in the radial artery and the basilic vein, and temperature probes in the rectum, esophagus, and nasopharynx. Cooling was then instituted in two stages: 1. Surface cooling down to 25 ° C. was obtained in the first cases by immersion in a cold water bath (10 to 6° C.) with the use of a specially constructed small tub (which remained on the operating table) and a plastic sheet to isolate the baby. More recently we have relied on generous ice packing. The speed of cooling during this
Table I. Clinical data in 17 babies Data
Lesion Single ventricle YSD, pulmonary hypertension ASD,AYC Tetralogy of Fallot Aorta-pulmonary artery fistula TAPYD
I No. ofcases I II
2 I I I
Age Mean: 9'1. months Range: 2'1. to 28 months Weight Mean: 5.5 kilograms Range: 2.4 to 9.4 kilograms Legend: YSO, Ventricular septal defect. ASO, Atrial septal defect. AVe, Atrioventricular canal. anomalous pulmonary venous drainage.
TAPVO,
Total
period is inversely (and very significantly) correlated to the weight of the patient. 2. The thorax is entered through a longitudinal midstemal incision. The aorta and cavae or right atrial appendage are cannulated in conventional fashion, and bypass is started after injection of heparin (3 mg. per kilogram). Blood is cooled to 18 to 17° C. and, when esophageal temperature is in the vicinity of 20° C., BCC is arrested. The speed of cooling during this period is not correlated to weight. The proximal aorta is clamped, the caval
The Journal of
318
Weiss et al.
Thoracic and Cardiovascular Surgery
Tell1p. 'C ceSOphl1gu...
-----,
1 17 cases
BATH
,ECC
i
i
i15
i
1
·1
r
i
j
i
CIRCULATORY
!
15
ECG
ARREST
j
i
f\i i \l
83mn
i
.
i ll i 14mn
i! 7i
19mn to 62mn
1
i
31mn! 8
T rnf rnrte-s Fig. 2. Average time course of the hypothermic procedure. The spread of the duration of successive periods is indicated by the mean ± standard deviation, except for arrest, the range of which is indicated. Similarly, the spread of the esophageal temperatures is indicated by vertical bars (± I standard deviation). ECC, Extracorporeal circulation.
snares are tightened, and cardiotomy and repair are performed; in many cases, the cannulas are removed. Bypass is resumed after an average of 40 minutes, after closure of the cardiotomy and recannulation. Utmost care is taken to prevent air embolism. Arterial blood is progressively warmed, with the difference between blood temperature and any recorded body temperature never allowed to be greater than 8 ° to 9 ° C When "central" temperature is over 33 to 35° C, ECC is stopped. Average durations of the main operative stages are summarized in Fig. 2, along with their standard deviations and ranges and the thermic levels of the esophageal probe. During this early phase, that is, before standardization of the procedure, some variations of the level of hypothermia were accepted or indeed sought; a few patients were cooled to a much lesser degree than the group a.,s~a whole. Fig. 3 shows in the form of a hi;togram the minimum temperature recorded in two locations supposed to reflect central temperatures. Nasopharyngeal temperature was higher than esophageal temperature by an average of 1.8° C Blood samples are taken before ECC and
at least twice during ECC to measure PO z , pH, Pco., base excess, K+, hematocrit, and plasma hemoglobin. Results were in the usual range, provided 10 mEq. (average) of bicarbonate was injected into the oxygenator during the arrest. EEG monitoring. Six simultaneous tracings are obtained continuously from eight needle electrodes inserted symmetrically in the scalp, four on each side (Fig. 4). These needles remain in place from the beginning of anesthesia until the patient leaves the operating room. Standard amplification, 10 p'v per millimeter, and minimum or no filtering of the EEG were used throughout. Arterial pressure and ECG are recorded on the same polygraph apparatus*; this disposition is often useful to help recognize artifacts. Postoperative tracings could not be obtained in this group, as these children were sent to another hospital for the early recovery period. EEG interpretation and nomenclature. For comparison, only quantitative time parameters and one qualitative criterion will • Electroencephalograph. Alvar Electronic Cy, Paris, France.
Volume 70 Number 2 August, 1975
be used in this first study, because a detailed and statistical description of EEG patterns in a larger group of patients is currently in progress. Survival time of EEG is the time (in seconds) elapsed from circulatory arrest (voluntary pump arrest at the end of cooling) until the EEG is flat (null stadium) in all leads (Fig. 5). The occurrence of a flat or isoelectric EEG tracing is judged by careful review and can be determined with excellent reproducibility (± 2 seconds) by two independent observers. No attempt was made to increase the sensitivity of the EEG records for two reasons: I. We believe that the disappearance of EEG waves at the 10 p,v per millimeter recording amplification has definite semeiologic value concerning cortical electro activity (see Discussion). 2. This setting allowed for comparison with previous groups of patients. One must recall that hypothermia by itself induces changes in EEG patterns (slowing, voltage diminution, and characteristic paroxysmal patterns- ~~). However at the thermic levels utilized in this series, disparition of brain waves was never observed before arrest of circulation. At the arrest of bypass, brain waves undergo further and progressive slowing until they disappear completely. Survival time is thought to reflect some kind of relation between the metabolic needs of brain cortical cells (lowered by hypothermia) and metabolite availability (stores), corresponding grossly to the concept of tissue protection. The present study offered the opportunity to obtain comparative data related to age and to temperature at arrest. Other possible factors such as technique of anesthesia, blood-gas status, and previous circulatory state are currently under study in a larger series. Reappearance latency is the time (in minutes) that elapses from the start of rewarming ECC until the first wave (excluding artifacts) reappears on the EEG tracing (Fig. 6). The first waves are always recorded as discontinuous bursts separated by
Study of EEG during surgery in infants
3 19
I 17easesJ
N.c ses Temperature: (),sophageal
r-,-----. MEAN
20.6%2,5
r----
5
-
17
19
21
23
[ MEAN DIFFER.
25
1,8! 11
TO
N. cases
Temperature: Nasopharyngeal
5
17
19
21
23
25
TO
Fig, 3. Temperatures at the beginning of arrest. This figure allows comparison of two probe locations. Esophageal temperature is nearly 1.8° C. colder than nasopharyngeal temperature. The rather important spread of central temperatures in this series should also be noticed.
2
6
1
4
Fig. 4. Location of needle electrodes. The numbers correspond to the electroencephalographic (EEG) channels in Figs. 5, 6,7, and 8.
The Journal of
320
Thoracic and Card iovascular
Weiss et al.
Surgery
CESO 18 ° NP 20,2 ° '11
17 .....
C . l.V. + H .T.P.
-----r - - -'{)V D
.A ,/ ..v
~~
_
_____ -"-'_
v-.--
~ __..r
-
." - ~ ~
. .r ,-.
___v ""V -JV~
0.
- - - - - - - -- - - - - - - - - .......----""'r"-'
.r--
-~-
~
.".
....\. .. ---~-~--~
.'
--'
~
_r_
'- -------
-
_
Fig. 5. Survival time. The tracing is continuous and should be read according to the numbers in the right margin. Each record carries from top to bottom : six EEG channels, arterial pressure, and electrocardiogram. Notice the arrest of cooling ECC , marked by the fall of arterial pressure (fi rst arrow ), the progressive slowing, and the decre asing amplitude in all channels tapering out to a flat line (second arrow). This child had a ventricular septal defect with pulmonary hypertension. The survival time was 127 seconds at an esophageal temperature of 18° C. The next three figures concern the same patient.
periods of flatness which become shorter and shorter. Reappearance latency is thought to be related to the importance of metabolic (oxygen) debt , which in tum is subject to the duration of arrest and probably to the metabolic rate, itself related to the degree of cooling . This latency is consistently longer than 15 minutes , so that central temperature was always higher than 30° C. when
EEG reapp eared and no correlations with temperature could be found. Latency for continuous EEG is the time (in minutes) that elapses from the start of rewarming ECC until the pattern becomes continuous. As a rule this is observed after a period in which EEG activity alternates with pauses, or flat tracings , of progressively shorter durations (see above ). This param-
Volume 70
Study of EEG during surgery in infants
Number 2
321
August, 1975
",.....L
11.1.11 ---r-~
'
_
-
Ik e..
-
-
. .
_.r--~
--..r
Fig. 6. Reappearance latency. This record of the same patient shows the pattern of reappearance of electrocortical activity 19 minutes after the rewarming ECC has been started. Notice the large waves which are quite separated by long periods of isoelectric tracing. Circulatory arrest had lasted 39 minutes. At this moment ECC is stilI in progress and the heartbeat is still irregular and slow, an infrequent occurrence.
LATENC"E POUR E EO C O NTINU 53mn
34' •• 34'
_
ARRET CIRCULA'TOIRE 39 mn .. 19'
Fig. 7. Latency for continuous EEG. Same patient. Although the pattern had shown definite improvement (left), there were stilI periods during which one or more derivations were intermittently flat. After the arrow (fifty-third minute after ECC had been restarted), electrical activity became strictly continuous in all leads. At that time ECC had been discontinued and heart action was good, as shown by arterial pressure. Posterior channels (5 and 6) had become continuous 16 minutes earlier. Such precedence of continuity in the occipital leads is the rule.
eter is thought to have practically the same meaning and same determinants as the reappearance latency (Fig. 7). A fourth criterion was also tentatively evaluated and can best be described as "recovery" or "potential normalization." It is based on three signs: recovery of normal anteroposterior differentiation and decrease of wave voltage in frontal leads, appearance of "alleviation" with fluctuating EEG patterns, and reappearance of theta waves (4
to 5 cis) in a delta background (Fig. 8). Admittedly, this criterion has two main shortcomings. One is related to time constraints, because monitoring was ended when the patient left the operating room, usually between 90 and 120 minutes after ECC. The second is due to the variability of what should be considered a normal EEG in an anesthetized infant after many hours in the operating theater. However, we think that an experienced observer can give judgment
The Journal of
322
Weiss et al.
Thoracic and Cardiovascular Surgery
2.fV'1\.. 3 . /'
~ /~~
·.f \ ./ v.. ~ ~ ; - . .,.N . .
~
4
5 6
....
"....vL, ',
". '
\.
" \. \.
•
~
\ . \. \. . \.. \
e
" •
~
\. "
~
.
..
.
. -..,- ...... \ ,
\ . " \, "
•
\ .
NORMAusAT1OH l'lnt II fin • I·....im :
•
\
". . . ..
,
'
•
•
•
\ \.
\,
.... ",
.
1 .5.72 .
Fig. 8. Normalization of EEG. Same patient. This tracing was taken about 90 minutes after the start of rewarming ECC and was considered practically normal. This kind of observation was rare (only 2 cases), due in part to more delayed recovery and to the shortness of intraoperative monitoring.
on this criterion if he considers the progress and speed of the recovery evolution, apart from the EEG pattern itself. Results Survival time. Rather wide variations were observed in these 17 patients (55 to 250 seconds), probably due mostly to variations in central temperature at the end of ECC cooling. A good correlation was indeed observed with esophageal temperature but, paradoxically, not with nasopharyngeal temperature (Fig. 9). For the group as a whole, mean survival time was 120 seconds at a mean esophageal temperature of 21 0 C. Reappearance latency. As noted previously, two factors are certainly very important: brain temperature and duration of arrest. Brain temperature. It is well known that core cooling induces considerable inequalities of body cooling (so-called thermic gradients) and that there is no exact and unique way of scaling the degree of hypothermia." In a previous large series, temperature was additionally measured with a probe ad-
vanced into the outer auricular conduit, against the eardrum. In 3 neurosurgical cases the temperature measured by a probe at the surface of the cerebral cortex was exactly the same as that of the auricular probe. In. 2:< However, the insertion of an auricular probe was not considered convenient in small babies. Therefore, we obtained a second-best estimation of temperature during circulatory arrest by averaging temperatures, i.e., rectal and esophageal temperatures at the beginning and end of arrest; the mean of these four measurements was called "weighted temperature." (Esophageal temperature usually drifts up one or two degrees whereas rectal drifts down, and their difference is usually small at the end of circulatory arrest.) These weighted temperatures are shown as a histogram in Fig. 10 and are presented as a second-best indicator of central (and hopefully brain) temperature during arrest. The dispersion of weighted temperature is less than the dispersion of esophageal temperature at the initiation of arrest. Duration of arrest. A good correlation was
Volume 70 Number 2 August. 1975
323
Study of EEG during surgery in infants
found between the duration of arrest and the reappearance latency (Fig. 11, a): That is, the longer the arrest, the longer the EEG remained flat. Furthermore, in these 17 first cases there seemed to exist some kind of limit around 37 to 40 minutes of arrest. Arrests of shorter duration all had reappearance latencies shorter than 20 minutes; longer arrests were followed by more variable latencies. However, this series is too small to warrant any firm conclusion. During this period, central temperature had increased considerably due to ECC, and practically all patients had regained near normal esophageal temperature. We feel very strongly that the reappearance of EEG is not related to actual brain temperature but to preceding events. The mean value for reappearance latency was 26 minutes (range 51/2 to 50). However, this parameter warrants a subdivision into two subgroups (see Discussion and Fig. 13) . Latency for continuous EEG. This latency also increased on the whole with the duration of arrest. However, the correlation is not good, indeed not significant with the type of linear correlation tested. Fig. 1I, b, shows that although the trend is rather clear, some cases differ strongly. For example, in Case 2 there was a rather short reappearance latency only 14 minutes longer than that of Case 1. However, the latency for continuous EEG was nearly 60 minutes longer, without any obvious circulatory reason for this discrepancy. (This single case is enough to modify the best-fit straight line in this short series but could lose its significance in a longer series such as the one currently being studied.) Latency for recovery or potential normalization. The results, summarized in Table II, also show a tendency to differences between arrests shorter than 38 minutes and those longer. (This time limit was chosen because of the results concerning reappearance latency.) However, the spread is large, and the statistical significance could
y =-12,6:1:+380
SECONDS 250 . A
r =-0,604 200 150
--
p= very
._- ~ -_._~
A
-;;--
100
A
significant
._A
A --.-
50 yo OESOPHAGEAL
SECONDS
250 200
Y=-12,3X+3l1 r= -0,397
p: non
significant
150 100 50 yo NASOPHARYNGEAL
Fig. 9. Correlations between EEG survival times and two central temperatures. Although most of the nasopharyngeal temperatures are well grouped near the calculated regression line, the over-all statistical significance is weakened by one very divergent measurement, probably because of a technical error. No one probe location is completely reliable, and at least three probes should be used. (Only 15 cases are included in the nasopharyngeal regression diagram.)
not be asserted. The subjective character of this criterion has already been stressed. It is of note that 2 babies regained a normal EEG tracing before the end of the operation. Discussion A semantic point should be raised from the start, i.e., the semeiology and significance of EEG "silence." This is especially important since a conflicting paper has been published recently. Reilly and associates" documented the persistence of EEG waves in 8 of 14 infants subjected to circulatory arrest of up to 46 minutes. They used exactly the same kind of procedure and the same hypothermic levels that we employed.
324
The Journal of Thoracic and Cardiovascular
Weiss et al.
Surgery
Number of eases 6 5 4
3 2
19
20
21
22
23
24
Average T
0
Fig. 10. Weighted average temperature during circulatory arrest. The temperature for each patient results from averaging four temperatures, two initial and two terminal. Notice that the spread is smaller than for temperatures at the initiation of arrest (Fig. 3).
Table II. Recovery, or potential normalization, latency Circulatory arrest period
Less than 35 minutes
No. ofcases 6
Recovery latency 112 min. (65-170)
I normalization More than 35 minutes
11 7 4
141 min. (69-180)
I normalization Sup. to 120 min.
Legend: This stage was reached in all 6 short arrests and in 7 of 11 longer arrests. In the last 4 cases, electroencephalographic monitoring was interrupted before that stage, as the patient had to be taken from the operating theater.
They used skin silver-silver chloride electrodes and an eightfold increase in recording sensitivity (1.25 p. per millimeter), and they found this persistence of EEG activity particularly in the parietal occipital area. In 6 other children the electrical activity disappeared after an average of 19 minutes (4 to 32 minutes). Concurrently, they stated: "Once rewarming by-pass was re-established, the occipital activity (if lost) and fronto-central activity returned and increased in voltage," without any mention of delay or intermittent activity. We believe that the discrepancy between their observations and ours is due to the
nature of central nervous system structures explored. With this interpretation, the EEG activity recorded by Reilly and associates ' S would be related to white matter structures (and maybe the cerebellum) and would become evident because of increased amplification and absence of superimposition of larger waves of cortical origin. Such waves of noncortical origin, which have also been documented by deep electrodes during neurosurgery and experimental studies, have an important physiological signification. Their persistence during arrest of circulation in deep hypothermia (which we have also observed with higher amplification) is by itself an important fact needing further clarification; their relationship with myograms," apparent in Reilly's tracings, should be studied further. On the other hand, the semeiology and time relationships reported in the present paper refer specifically to cortical activity, a major argument resulting from the resemblance of disorganization and reorganization patterns to patterns observed in other, more familiar and more frequent clinical circumstances. The disappearance of rapid rhythms, the invasion of delta waves, and the dedifferentiation of anterior and posterior derivations have distinct cortical connotations; the paroxysmal activity that one
Volume 70 Number
2
Study of EEG during surgery in infants
325
August, 1975
of us- z i first described in 1962 also seems a cortical phenomenon related to the degree of cerebral maturation and therefore to age. It thus seems relevant to analyze our data in terms of "electrocortical silence" rather than "cerebral silence." A more important aspect of the present, time-oriented study lies in two principal directions: immediate prognosis for individual patients and over-all evaluation of brain tolerance for the series. At the completion of this study, we felt that no individual short-term prognosis could be obtained. The reason is that no patient, however short or long the different latencies, showed evidence of permanent brain damage. Admittedly, some of these patients died, but the outcome clearly appeared related to the severity of the clinical condition, to the quality of surgical repair, or to the circulatory state after operation, rather than to clinical brain damage. However, we must also state that complete and systematic postmortem brain examinations were not carried out. Nevertheless, the definition of "normal" duration of EEG disturbances and their spread seems important in view of the increasing number of cases. It is also important to judge the quality of individual anesthesia and the adequacy of so-called minor modifications of the technique, such as the degree of cooling, the duration of rewarming or, perhaps more important, the setting of Peat during the different phases.': 1,. 18, 23 Another important point is that persistence of a flat EEG for up to 30 or 40 minutes after the end of circulatory arrest may be a normal occurrence but that this condition can mask any operative accident, especially air embolism. (In our institution, anesthesiologists and surgeons rely heavily on the EEG for early diagnosis in ordinary ECC.) Brain tolerance to circulatory arrest in young children can best be appraised by comparison with older age groups. Unfortunately, our previous clinical experience did not cover this level of hypothermia but
minuteS
50 REAPPEARANCE LATENCY
40
a)
30 20 10 CIRCULATORY
o
20
40
ARREST
60 minutes
minutes
70 LATENCY \ . FOR 60 CONTINUOUS TRACING
50
b)
40 £
30
Y=~675X+17,152
r=q375
£
10 I
non significant
CIRCULATORY
ARREST
i • -----r!..------.,------r---
LI
o
jp =
60minul••
Fig. 11. Correlations between duration of arrest and the two latencies. Reappearance latency had a much higher statistical significance. Notice that the best-fit slopes are quite similar. a, Notice also that no widely different reappearance latencies were observed for arrests shorter than 38 to 39 minutes, whereas a 49 minute latency was observed for a 40 minute arrest. Two cases in which there were short arrests were singled out (arrows 1 and 2). Both had rather short reappearance latencies, but only I had a short continuous latency; the second had a very prolonged continuous latency. The clinical significance of such divergence is not yet clear.
rather a much lower level, so that comparison can only be indirect. Even so, the juxtaposition of data is indeed rather interesting. EEG during cooling. In variance with what has been reported but not documented by some users of deep hypothermia,"- 20 the EEG never became flat by virtue of hypothermia at 20° C. alone. On the contrary, we would stress the following points: 1. In this group of infants, the EEG re-
The Journal of
326
Weiss et al.
Thoracic and Cardiovascular Surgery
2
".;eeond".;
(Log) 100
50 40 30
20
10
21 0
110
31 0
37 0
Fig. 12. Survival time as a function of esophageal temperature at arrest. The data from Table III have been plotted on a semilog scale, as this was the best manner to obtain a smooth curve.
Although only means from each series have been plotted and although there are important individual variations, this curve shows that the limit to prolongation of survival time is not far away. The difference between 21 and 11 C. is much smaller than the difference between 31 and 21 C. 0
0
Table III. Survival time of electroencephalogram (EEG)
Author
Gastaud et al."
No. of cases Temperature
Survival time ofEEG (sec.)
37
37 0 C.
12
13
31 0 C.
39
Present study
17
21 0 C.
Weiss"
84
11 0 C.
120 ±24 132
(1958)
Weiss'!
0
0
(1962)
(1962)
mained consistently quite active, albeit subdued, down to 21 0 C. 2. From our previous experience, the EEG is practically always still active at 11 0 C. in children, although even more decreased in amplitude and frequency, whereas it is sometimes isoelectric in adults." 3. The hypothermic depression observed in the infant group at 20 to 21 0 C. was not quantitatively different from the depression seen in older patients when they were subjected to this temperature level. Survival time. The present set of measure-
ments can be compared to data obtained in normothermic, unanesthetized patients in the course of a study of the vagovagal reflex by Gastaud-': 1" and in two previous surgical series studied in 1962 21 ; the latter two series employed surface-cooling and core-cooling (so-called Sealy-Dubost) techniques, respectively. The results are summarized in Table III and Fig. 12. The average survival time in the present study (120 seconds for 21 0 C.) is indeed on a smooth curve, which is the best representation that could be plotted. These results are coherent with a law of decreasing oxygen consumption, and the very young children in this group are apparently no different from older age groups. However, such a correlation has not been found by others in studies of moderate hypothermia. H Reappearance latency To evaluate this parameter, we compared the group cooled to 11 0 C. with two subgroups (Fig. 13). As noted previously, the short-arrest subgroup of the present series had short reappearance latency. Furthermore, the spread is small, and the mean is very near that of a
Volume 70 Number 2 August, 1975
Study of EEG during surgery in infants
327
Arrest inferior to 35 rnfmrces IArrest superior to 35Illinutes
21°
80mn
6 cases PRESENT STUDY
50mn
11° 46 cases 1962
48
40mn
30rnn 20mn 10mn
~
..
13;mn
I I I I I I, I
~
14 cases
PRESENT STUDY
1962
.
~
I I I
"-----:19
~
9,40
I 1.40
54
·50
I I
·5,30
11°
21° 11cases
I
39- 62mn
36 -51mn
Fig. 13. Comparison of reappearance latencies at 2 1 and 11 0 (data from Weiss-"). Notice that the means are little different in the short-arrest subgroups and that, in the long-arrest subgroups, the mean for 21 0 C. is nearly double the mean for 11 0 C. However, there is no statistical significance due to wide scatter. Individual values from only the present study are plotted. Values of the means are underlined. The range is indicated by numbers and small arrows. 0
similar subgroup cooled down to 11 0 C. The long-arrest subgroup had a much larger spread. The mean duration of reappearance latency, with admittedly I case of arrest longer than 1 hour, is nearly double the latency of the long-arrest subgroup cooled to 11 0 C. The difference of means is just above significance (p = 0.04). Latency for continuous EEG. The trend in this factor is similar, with the mean duration and spread of both short-arrest subgroups being quite similar. There is a noticeably longer mean latency in the long-arrest subgroup cooled to 21 0 c., but no significant difference (Fig. 14). Normalization of the EEG. The EEG very seldom became normal in this initial short series; however, the probable reason is that recording was not prolonged after the end of the operation. In only 1 case did we observe some undefined abnormal movements. However, they were very transient, lasting only 3 to 4 hours during the first night after the operation. Unfortunately, very few of these patients could be examined later, as their postoperative course was managed in another hospital and they
were then usually sent back to their referring hospital. This preliminary study suggests that deep hypothermia with circulatory arrest does not have a detrimental effect on the infant brain. At least, the effect is no different from that observed in older age groups, as judged by the quantitative chronologie criteria we utilized. However, we believe that two important points need further clarification. First, is there really an initial (and perhaps not really significant) limit to circulatory arrest around 40 minutes at 21 0 C.? In other words, is the sudden prolongation of latencies and their increased scatter the sign of a metabolic or cellular limit being crossed, or is the apparent lack of late ill effects the sign that no physiological barrier has been passed? (Our longest arrest did not exceed 62 minutes, but longer arrests [up to 91 minutes] have been reported.') The answers to such questions will be found only in a careful follow-up of these children, when they become old enough for precise testing of their intelligence quotient. Other authors also stress the tolerance to arrest
The Journal of
328
Weiss et al.
Thoracic and Cardiovascular Surgery
Arrest inferior to 35 rnirurtes 90m
80mn 70mn
PRESENT STUDY
21°
~rrest
97
11°
6 cases
superior to
35ITIinutes
PRESENT STUDY
11°
21°
11cases
46 cases
14 cases
I"
(,962) 70
60mn
(1962) 70
~
50mn 40mn
32mn30sec
30mn 20mn
29mn30sec
~
20
13
10mn 3.40
39-62mn
36- 51mn
Fig. 14. Comparison of continuous latencies at 21 and 11 C. (data from Weiss-"). Differences are less obvious than for reappearance latencies in the long-arrest subgroup. Only the mean value and the range are indicated. 0
longer than 60 minutes. This dilemma is important, because deeper hypothermia and/or short periods of reperfusion might be considered in infants with the most se'vere cardiopathies requiring long repair operations. Second, what is the basis of individual differences, particularly evident in the duration of EEG survival, with apparently equal temperatures? Three hypotheses at least can be raised. They can be studied by increased monitoring in clinical cases or by animal experiments: 1. Assessment of brain temperature is inadequate by present means. 2. Anesthesia plays an important role either to protect the brain or to inhibit EEG cortical activity. 3. The circulatory state of the brain, governed principally by Pco., is a very important factor both at the initiation of arrest' 17. 1~. 23 and later. It controls the state of metabolic stores and the rate of payment of metabolic debt. All these hypotheses may have important practical consequences. We gratefully acknowledge the help of M. Gaillard and A. Masson in the preparation of the manuscript.
0
REFERENCES
2 3
4
5
6 7
8
Almond, C. H., Jones, J. C., Snyder, H. M., Grant, S. M., and Meyer, B. W.: Cooling Gradients and Brain Damage With Deep Hypothermia, 1. THoRAc. CARDIOVASC. SURG. 48: 890, 1964. Arfel, G., and Weiss, J.: Electro-encephalogramme et hypothermie profonde, Ann. Chir, Thorac. Cardiovasc. 1: 666, 1962. Barratt-Boyes, B. G., Simpson, M. M., and Neutze, J. M.: Intracardiac Surgery in Neonates and Infants Using Deep Hypothermia, Circulation 42: 73,1970 (Suppl. III). Barratt-Boyes, B. G., Simpson, M., and Neutze, 1. M.: Intracardiac Surgery in Neonates and Infants Using Deep Hypothermia With Surface Cooling and Limited Cardiopulmonary Bypass, Circulation 43: 25, 1971 (SuppI. I). Barratt-Boyes, B. G.: Complete Correction of Cardiovascular Malformations in the First Two Years of Life Using Profound Hypothermia, in Heart Disease in Infancy, Edinburgh and London, 1973, Churchill-Livingstone. Bjork, V. 0., and Hultquist, G.: Brain Damage in Children After Deep Hypothermia for Open-Heart Surgery, Thorax 15: 284, 1960. Bjork, V. 0., and Hultquist, G.:Contraindications to Profound Hypothermia in Open-Heart Surgery,' J. THoRAc. CARDIOVASC. SURG. 44: 1, 1962. Brechner, W. L., Kavan, E. M., Bethune, R. W., Baver, R. 0., and Dillon, J. R.: The Electro-encephalographic Effects of Arrested Circulation During Hypothermia, Am. Surg. 25: 833, 1959.
Volume 70 Number 2 August, 1975
9 Drew, C. E.: Profound Hypothermia in Cardiac Surgery, Br. Med. Bull. 17: 37, 1961. 10 Drew, C. E.: Hypothermie profonde par quadruple canulation, Ann. Chir. Thorac. Cardiovasc. 1: 4, 1962. I I Drew, C. E., and Anderson, I. M.: Profound Hypothermia in Cardiac Surgery: Report of Three Cases, Lancet 1: 748, 1959. 12 Dubost, Ch., Blondeau, Ph., Passelecq, J., Guery, J., Audrerie, J., Laurent, D., Piwnica, A., Sprovieri, L., and Weiss, M.: L'association du coeur-poumons artificiel et de l'hypothermie profonde dans la chirurgie a coeur ouvert, Acta Cardiol. Intern. 8: 95, 1959. 13 Dubost, Ch., and Blondeau, Ph.: The Association of the Artificial Heart-Lung With Deep Hypothermia in Open-Heart Surgery, J. Cardiovase. Surg. 1: 85, 1960. 14 Gastaud, H., and Gastaud, Y.: La syncope vago-vagale, Paris, 1958, Sandoz. 15 Gastaud, H., Fischer-Williams, M., Lugaresi, J.: Correlations electrocliniques chez 25 sujets enregistres pendant leur syncope, Rev. Neurol. 95: 542, 1956. 16 Guiot, G., Rougerie, J., Dubost, Ch., and Blondeau, Ph.: Le "grand froid" en neurochirurgie, Possibilites et perspectives d'avenir, Neurochirurgie 6: 332, 1960. 17 Mori, A., Muruoka, R., Yokota, Y., and Okamoto, Y.: Deep Hypothermia Combined With Cardiopulmonary Bypass for Cardiac Sur-
Study of EEG during surgery in infants
18
19
20
21 22
23
24
329
gery in Neonates and Infants, J. THoRAc. CARDIOVASC. SURG. 64: 422, 1972. Reilly, E. L., Brunberg, J. A., and Doty, D. B.: The Effect of Deep Hypothermia and Total Circulatory Arrest or in the E.C.C. in Children, Electroencephalogr. Clin. Neurophysiol, 36: 661, 1974. Sealy, W. c., Young, W.O., Brown, I. W., Smith, W. W., and Lesage, A M.: Profound Hypothermia Combined With E.E.G. for OpenHeart Surgery, Surgery 48: 432, 1960. International Symposium: L'hypothermie profonde en chirurgie cardiaque et extracardiaque, Ann. Chir. Thorac. Cardiovasc. 1: 1-2, 1962. Weiss, J.: I.e silence cerebral electrique, Thesis in Medicine, Paris, 1962. Weiss, J., and Arfel, G.: Sequences E.E.G. de I'hypothermie profonde, Rev. Neurol. 103: 220, 1960. Weiss, M., Piwnica, A, Lenfant, C., Sprovieri, L., Laurent, D., Blondeau, Ph., and Dubost, Ch.: Deep Hypothermia With Total Circulatory Arrest, Trans. Am. Soc. Artif. Intern. Organs 6: 227, 1960. Young, W. G., Sealy, W. C., Brown, I. W., Smith, W. W., Callaway, H. A., and Harris, J. S.: Metabolic and Physiologic Observations on Patients Undergoing Extracorporeal Circulation in Conjunction With Hypothermia, Surgery 46: 175, 1959.