|
|
| |||||||||||||||||||||||||||||||||||||||||
Background Hypothermic circulatory arrest is a widely used support technique during heart surgery in infants, but its effects on neurologic outcome have been controversial. An alternative method, low-flow cardiopulmonary bypass, maintains continuous cerebral circulation but may increase exposure to known pump-related sources of brain injury, such as embolism or inadequate cerebral perfusion.
Methods We compared the incidence of perioperative brain injury after deep hypothermia and support consisting predominantly of total circulatory arrest with the incidence after deep hypothermia and support consisting predominantly of low-flow cardiopulmonary bypass in a randomized, single-center trial. The criteria for eligibility included a diagnosis of transposition of the great arteries with an intact ventricular septum or a ventricular septal defect and a planned arterial-switch operation before the age of three months.
Results Of 171 patients with D-transposition of the great arteries, 129 (66 of whom were assigned to circulatory arrest and 63 to low-flow bypass) had an intact ventricular septum, and 42 (21 assigned to circulatory arrest and 21 to low-flow bypass) had a ventricular septal defect. After adjustment for diagnosis, assignment to circulatory arrest as compared with low-flow bypass was associated with a higher risk of clinical seizures (odds ratio, 11.4; 95 percent confidence interval, 1.4 to 93.0), a tendency to a higher risk of ictal activity on continuous electroencephalographic (EEG) monitoring during the first 48 hours after surgery (odds ratio, 2.5; 95 percent confidence interval, 1.0 to 6.4), a longer recovery time to the first reappearance of EEG activity (only in the group with an intact ventricular septum, P<0.001), and greater release of the brain isoenzyme of creatine kinase in the first 6 hours after surgery (P = 0.046). Analyses comparing durations of circulatory arrest produced results similar to those of analyses comparing treatments.
Conclusions In heart surgery in infants, a strategy consisting predominantly of circulatory arrest is associated with greater central nervous system perturbation in the early postoperative period than a strategy consisting predominantly of low-flow cardiopulmonary bypass. Assessment of the effect of these findings on later outcomes awaits follow-up of this cohort.
In a randomized clinical trial, we compared the incidence of brain injury after assignment to a strategy consisting predominantly of total circulatory arrest with the rate after assignment to a strategy consisting predominantly of low-flow bypass during repair of D-transposition of the great arteries within the first three months of life. A related goal was to examine the effect of the duration of circulatory arrest on outcome.
Methods
Enrollment of Patients
We enrolled patients between April 1988 and February 1992 at Children's Hospital, Boston. Of 191 eligible infants, 180 (94 percent) were enrolled. The arterial-switch operation was performed in 171 infants (95 percent of those enrolled); the remaining 9 patients were found at operation to have coronary-artery anatomy unsuitable for the arterial-switch operation. The criteria for eligibility included a diagnosis of D-transposition of the great arteries with either an intact ventricular septum or a ventricular septal defect, repair scheduled to occur by three months of age, and coronary-artery anatomy considered suitable for the arterial-switch operation. The criteria for exclusion included a birth weight of less than 2.5 kg, a recognizable syndrome of congenital anomalies, associated extracardiac anomalies of more than minor severity, previous cardiac surgery, and associated cardiovascular anomalies requiring aortic-arch reconstruction or other open surgical procedures. Informed consent was obtained from the parents of all the infants according to the guidelines of the institutional Human Investigation Committee.
Study Design
We randomly assigned the participating patients to receive vital-organ support consisting predominantly of either total circulatory arrest or low-flow bypass, with stratification according to diagnosis and surgeon. Randomization schemes were developed with a permuted-blocks design; the method of support was assigned immediately before surgery. Neurologic-outcome data were obtained by investigators unaware of each patient's treatment assignment. Surgeons and anesthesiologists were kept unaware of the interim results during the study.
Perfusion Methods
The perfusion methods were identical in both groups, except as described below. Surface cooling was instituted with a low ambient room temperature, a cooling mattress, and ice packs to the head; the mean (±SD) tympanic-membrane temperature at the onset of bypass was 32.7 ±1.3 °C. We used the alpha-stat method of acid-base management. An ascending-aorta arterial cannula and a single right atrial venous cannula were used. Cardiopulmonary bypass and core cooling were begun as soon as the cannulas were in place. When the rectal temperature reached 18 °C or lower and the necessary preliminary surgical dissection had been completed, low-flow bypass or circulatory arrest was begun. Circulatory arrest lasted until continuity was established from the left ventricle to the aorta, the coronary arteries were reimplanted, and the atrial septal defect or ventricular septal defect, if any, was repaired. In the group assigned to low-flow bypass, the perfusion rate was reduced to 50 ml per minute per kilogram of body weight (approximately 0.7 liter per minute per square meter of body-surface area) for the duration of the aortic and coronary repairs. The infants in both groups underwent repair of atrial or ventricular septal defects with circulatory arrest. After the period of circulatory arrest in both groups, cardiopulmonary bypass resumed as rewarming was carried out and continuity was established between the right ventricle and the pulmonary artery. Perfusion pressures and flow rates during rewarming were the same in both groups.
Surgery and Anesthesia
The four study surgeons used standardized techniques for all aspects of the arterial-switch procedure, as described elsewhere16,17.
Anesthetic management was standardized for all patients. The intraoperative administration of intravenous fluids was limited to lactated Ringer's solution (10 to 20 ml per kilogram per hour) unless the blood glucose level was under 50 mg per deciliter (2.8 mmol per liter). Anesthesia was induced with fentanyl (50 µg per kilogram) and pancuronium (100 µg per kilogram). Plasma levels of fentanyl and pancuronium were maintained during bypass with an additional 25 µg of fentanyl per kilogram and 100 µg of pancuronium per kilogram. All the patients received methylprednisolone (30 mg per kilogram) at the onset of bypass. Thiopental (10 mg per kilogram) was given when the tympanic temperature reached 18 °C. During rewarming, an additional 25 µg of fentanyl per kilogram and 100 µg of pancuronium per kilogram were administered.
Clinical Seizures
Definite clinical seizures in the first week after surgery were recorded by the nurses or physicians caring for the infant. The criteria for a definite clinical seizure included the occurrence of a single or recurrent motor event, with tonic or clonic movements of an extremity or cranial muscle that were associated with an alteration of consciousness and were not interruptable by manipulation of the body part involved. Isolated abnormalities involving apnea, tachycardia, eye movements, sucking movements, or tongue movement were considered indeterminate. In one infant, seizures followed a prolonged cardiopulmonary arrest on the day of anticipated discharge from the hospital, nine days after surgery; this outcome was not included in our analyses of seizures in the first week after surgery, although its inclusion would not have altered our conclusions.
Electroencephalography
We monitored the electroencephalogram (EEG) continuously for at least 2 hours before surgery, during surgery, and for 48 hours after surgery by video EEG (Telefactor, Modac, West Conshohocken, Pa.). The EEG data were interpreted by one of four pediatric electroencephalographers according to predetermined criteria18 and were then reviewed by the group to develop a consensus on the final interpretation.
The recovery of cerebral function was assessed as the time in minutes from the onset of rewarming to the occurrence of several EEG activities, defined as follows: first activity, the reappearance of activity in channels FZ-CZ, F3-C3, or F4-C4; close bursts, the first 60-second period when the interval between bursts (measured from mid-burst to mid-burst) was less than 15 seconds; relative continuous pattern, the first 60-second period in which the intervals between bursts (less than 15 micro V in voltage) were shorter than 6 seconds; and continuous pattern, the point at which EEG activity contained no intervals between bursts longer than 1 second and remained continuous for at least 60 seconds.
Rhythmic paroxysmal activity detected on continuous video EEG during the first 48 hours after surgery was classified as ictal (i.e., as EEG seizure activity) if the duration of the discharge was more than 5 seconds.
Brain Isoenzyme of Creatine Kinase
The brain isoenzyme of creatine kinase (BB isoenzyme) was measured in serum at the induction of anesthesia, at the attainment of a rectal temperature of 32 °C during rewarming, and then 1.5, 3, and 6 hours after the resumption of cardiopulmonary bypass. These measurements were performed by International Immunoassay Laboratories (Santa Clara, Calif.) and were expressed in international units per liter.
Neurologic Examination
Neurologic examinations were performed by a pediatric neurologist according to a uniform, predetermined protocol before surgery and again before hospital discharge. The neurologist was unaware of each infant's treatment assignment and clinical course (e.g., the occurrence of seizures). We classified the results of the neurologic examination as normal, possibly abnormal, or definitely abnormal; abnormalities were subclassified according to specific type (e.g., abnormalities of cranial nerves or the motor system)19,20,21,22.
Cranial Ultrasound Studies
Cranial ultrasound examinations were performed before surgery and again before discharge from the hospital. In September 1990, the investigators and the Safety and Data Monitoring Committee agreed to remove cranial ultrasound examinations from the protocol, because the very low incidence of abnormalities detected by this test precluded its ability to contribute importantly to the conclusions of the study.
Statistical Analysis
Treatment groups were compared in intention-to-treat analyses, in which the strategy consisting predominantly of total circulatory arrest was compared with the strategy consisting predominantly of low-flow bypass. Secondary analyses examined the effect of the duration (in minutes) of circulatory arrest on outcome. Except where noted, all tests of hypotheses and regression analyses were adjusted for diagnosis. The study conclusions were not altered after adjustment for surgeon or interactions between surgeon and treatment. All P values are two-tailed.
Perioperative outcomes included continuous and categorical variables. We used a natural logarithm transformation of serum creatine kinase BB isoenzyme levels, integrated over a six-hour period, and of the times to recovery of EEG activities in order to normalize their distributions before analysis. Multiple linear regression was used to analyze continuous outcome variables. Stratified exact tests and exact Wilcoxon tests,23 as well as logistic-regression methods, were used to analyze categorical outcome variables. The estimated treatment effects within each diagnostic group were similar to those for the study population as a whole, except as noted.
Results
Comparability of Treatment Groups
Among the 171 infants with D-transposition of the great arteries, 129 (75 percent) had an intact ventricular septum, and 42 (25 percent) had a ventricular septal defect (Table 1). As anticipated, the infants with a ventricular septal defect were older and less acutely ill at the time of enrollment than those with an intact ventricular septum. Treatment assignments were balanced within the randomization strata of diagnostic group and surgeon. The infants within each diagnostic group who were randomly assigned to the two support techniques were similar at enrollment with regard both to preoperative variables that might influence neurologic outcome and to preoperative neurologic assessments.
|
In accordance with the randomized assignments, the treatment groups differed significantly with regard to the duration of total circulatory arrest and low-flow bypass, as well as to total bypass time (Table 2). Surgical circumstances sometimes required the use of a strategy consisting predominantly of circulatory arrest in infants randomly assigned to a strategy of low-flow bypass. There were no significant differences between the treatment groups in the total time spent receiving support (i.e., total bypass time plus circulatory-arrest time) or in cross-clamping time (i.e., the duration of myocardial ischemia), demonstrating that the choice of support method did not affect the difficulty of the surgery.
|
The infants in the two treatment groups were similar with respect to surgical mortality, hospital course, and the incidence of non-neurologic events. Three infants (2 percent) died within one month of surgery, two in the early postoperative period and one shortly after discharge from the hospital; one of the three was assigned to circulatory arrest and two to low-flow bypass. In the combined groups, infants were intubated for a median of 3 days (range, 1 to 61) and were discharged from the hospital a median of 9 days (range, 5 to 71) after surgery.
Neurologic Outcomes
Definite clinical seizures occurred more frequently among the infants randomly assigned to the circulatory-arrest group (exact P = 0.009; odds ratio, 11.4; 95 percent confidence interval, 1.4 to 93.0) (Table 3). Similarly, definite seizures were strongly associated with a longer period of circulatory arrest in the logistic-regression model in which the number of minutes of circulatory arrest and diagnosis were both used as predictor variables (P = 0.004) (Figure 1A). All the infants with definite clinical seizures had at least 35 minutes of total circulatory arrest. Using multivariate logistic-regression techniques, we excluded the possibility that preoperative, potentially confounding variables could explain this effect of treatment assignment or duration of circulatory arrest.
|
|
The infants randomly assigned to the circulatory-arrest group had significantly longer recovery times to first EEG activity (P<0.001, in the intact-ventricular-septum group only), close bursts (P<0.001), and relative continuous activity (P = 0.02). Models that included the number of minutes of circulatory arrest as a continuous variable demonstrated similar relations; longer periods of arrest were associated with longer recovery times to first EEG activity (P<0.001) (Figure 2), close bursts (P<0.001), and relative continuous activity (P = 0.02). The effects of treatment were insensitive to adjustment for other potentially explanatory preoperative variables. The time to the recovery of continuous activity was similar in the treatment groups, and none of the infants in either treatment group recovered to the base-line level of activity during the first 48 hours after surgery. The treatment groups did not differ significantly with respect to EEG findings one week after surgery.
|
The results of the neurologic examination at discharge from the hospital were not related to the support method used, in analyses including either treatment assignment or minutes of total circulatory arrest. A large proportion of infants in the combined groups were categorized as having a neurologic examination that was possibly (50 of 160, or 31 percent) or definitely (32 of 160, or 20 percent) abnormal. The most common neurologic abnormalities are shown in Table 4; the majority of infants had mild hypotonia that was consistent with recent systemic illness or the administration of sedative or anticonvulsant medications. Such medications were administered within 24 hours of the neurologic examination in 26 of 32 infants with definite abnormalities (81 percent), 31 of 50 infants with possible abnormalities (62 percent), and 47 of 78 infants with normal examination results (60 percent) (P = 0.08 by exact Wilcoxon test). However, the specific findings of focal or lateralized abnormalities or discrepancies in ability to control flexion and extension of the neck (a measure of the degree of extensor posture) could not be attributed to the effects of medication and occurred in 10 percent and 16 percent, respectively, of the infants in the combined groups.
|
We found that a strategy consisting predominantly of circulatory arrest, as compared with one consisting predominantly of low-flow cardiopulmonary bypass, during open-heart surgery in infancy was associated with a higher likelihood of clinical and EEG seizures, a longer time to the recovery of normal brain activities as assessed by EEG, and a greater release of the BB isoenzyme of creatine kinase over the first six hours after surgery. Infants in both groups received some period of circulatory arrest and some period of low-flow bypass, such that the treatment strategies were not mutually exclusive. Although our primary analyses compared the treatment groups according to the intention to treat, the most important aspect of the treatment assignment appeared to be the duration of total circulatory arrest. Indeed, this variable was at least as sensitive as that of treatment assignment in predicting major outcomes. By the time of hospital discharge, the groups were similar in overall incidence of abnormalities on neurologic examination and EEG.
Inferences about the effect of support techniques during cardiovascular surgery on neurologic and developmental outcomes may be complicated by the presence of many other risk factors for brain injury in children with congenital heart disease24,25,26,27,28,29,30,31,32,33,34,35,36,37. We studied patients in whom potential confounding factors were limited. Infants with D-transposition have a low incidence of coexisting anomalies and infrequently have substantial hemodynamic problems after surgery. The age at which these infants undergo the arterial-switch operation is more uniform than the age at repair for most other forms of congenital heart disease. Furthermore, the arterial-switch procedure requires minimal intracardiac exposure, so the two support techniques could be used with equal facility.
Previous evidence of perioperative cerebral injury during circulatory arrest comes from laboratory and clinical investigations. Studies of brain structure and function in animals after the use of circulatory arrest have suggested that at a core temperature of 15 to 20 °C, a 30-minute total arrest time is safe with respect to central nervous system damage38,39,40,41,42,43. Conflicting results have been reported with circulatory-arrest times of between 45 and 60 minutes40,41,43,44. Electrophysiologic studies have demonstrated an association between EEG recovery times and the duration of circulatory arrest45,46. Serum levels of the BB isoenzyme of creatine kinase have been reported to be significantly higher among children who had surgery using circulatory arrest than among those who underwent closed procedures, with a peak level directly associated with the duration of arrest47; the relation of these levels to long-term neurologic outcome is unknown.
There are fewer investigations of the neurologic and developmental sequelae of low-flow cardiopulmonary bypass. The use of this technique rather than circulatory arrest prolongs the total time of extracorporeal circulation, thereby increasing exposure to pump-related sources of brain injury, including microembolism, macroembolism, and insufficient cerebral perfusion8,9,10,11,12,13,14,15. We used a flow rate of 50 ml per kilogram per minute (approximately 0.7 liter per minute per square meter) in the low-flow group. A bypass flow rate of 0.5 liter per minute per square meter has been shown to support a level of cerebral oxygen consumption that is appropriate for the temperature during hypothermia in clinical and experimental studies48,49. Furthermore, at perfusion rates as low as 0.5 liter per minute per square meter, the latency and amplitude of somatosensory cortical evoked potentials are maintained in the normal range for temperatures of 21 to 25 °C5,50. However, at flow rates of 0.5 liter per minute per square meter in studies in animals, cerebral ATP levels were significantly decreased.
Transient clinical seizures have been reported in 4 to 10 percent of infants in the period immediately after surgery using circulatory arrest51,52,53. In the present study, there were clinical seizures in 11 percent of the infants randomly assigned to circulatory arrest, whereas continuous EEG monitoring detected seizure activity in 26 percent of these infants. Thus, it seems likely that the incidence of seizures has been underestimated in the past, perhaps because the frequent use of paralytic or sedative agents in the early postoperative period obscures the clinical manifestations of seizure activity. The occurrence of seizures after cerebral ischemia is consistent with the concept of excitotoxic mechanisms of neuronal injury after ischemia. High levels of amino acid neurotransmitters have been measured after cerebral ischemia in laboratory animals and result in the excitation and subsequent death of neurons with high concentrations of receptors such as N-methyl-D-aspartate receptors54,55. Excitatory amino acid receptors are present at birth, but their concentrations increase dramatically during the first few weeks of life,56,57,58 offering a possible explanation of the association between age and seizures in our study. Furthermore, experimental data have demonstrated that the seizure threshold decreases with increasing age59. Clinical60,61 or EEG62,63 seizures in the neonatal period are associated with an increased risk of unfavorable outcomes in other infants -- e.g., infants with hypoxic ischemic encephalopathy or hypoglycemia.
In many reparative and palliative operations other than the arterial-switch operation, the accuracy of repair in infants may be facilitated by the circulatory-arrest technique. Thus, cardiovascular surgeons need to balance the technical advantages of circulatory arrest with its potential risks on an individual basis. Furthermore, all infants in this study had some period of circulatory arrest; thus, we could not assess the benefit of a support strategy in which total circulatory arrest is completely circumvented. All the infants in this study were under the age of three months, and most were under the age of one month. The study conclusions should be broadly generalizable to neonatal repair of other forms of congenital heart disease, but not necessarily to the effects of deep hypothermic circulatory arrest in older infants, children, and adults.
These data indicate that a longer period of total circulatory arrest used with deep hypothermia to support vital organs during open-heart surgery in infancy is associated with greater neurologic perturbation in the early postoperative period. Although some neurologic or developmental sequelae of these findings are likely, their incidence, nature, and severity are uncertain. Assessment of the effect of total circulatory arrest on later neurologic and developmental function thus awaits the follow-up of this cohort.
Supported by grants (HL 41786 and RR 02172) from the National Institutes of Health.
We are indebted to the members of the Safety and Data Monitoring Committee, Julien I.E. Hoffman, M.D. (chairman), John W. Kirklin, M.D., Barry M. Lester, Ph.D., Robert J. Levine, M.D., Eli M. Mizrahi, M.D., Joseph G. Reves, M.D., George W. Williams, Ph.D., and Joel I. Verter, Ph.D.; to the perfusionists, Willis G. Gieser, C.C.P., Robert A. LaPierre, B.S., C.C.P., Robert J. Howe, B.S., C.C.P., and Bettina Archilla, B.S.; to the laboratory technician, Nick Morana; to the electroencephalographers, Lewis Kull, Sheila A. McPeck, Susan M. Hegarty, and Wayne A. Cote; to Harvey Meyerson, Ludmila Kyn, and Kunihiko Hayashi for data-base and statistical programming; to the nursing staff for assistance with adherence to protocol; to Donna M. Donati, Donna M. Duva, and Lisa-Jean Buckley for data management; and to Kathleen M. O'Brien for project coordination.
Source Information
From the Departments of Cardiology (J.W.N., G.W., A.Z.W., K.C.L., D.L.W.), Cardiovascular Surgery (R.A.J., D.M.F., F.L.H., J.E.M., A.R.C.), Anesthesia (P.R.H.), Neurology (K.C.K.K., G.L.H., S.L.H., J.C., E.C., J.K.B.), and Radiology (J.C.S.), Children's Hospital; the Departments of Pediatrics (J.W.N., G.W., D.L.W.), Surgery (R.A.J., F.L.H., J.E.M., A.R.C.), Neurology (K.C.K.K., G.L.H., S.L.H., J.C., E.C., J.K.B.), and Radiology (J.C.S.), Harvard Medical School; and the Department of Biostatistics, Harvard School of Public Health (D.W., J.H.W.) -- all in Boston. John K. Barlow, M.D., is deceased.
Address reprint requests to Dr. Jonas at the Department of Cardiovascular Surgery, Children's Hospital, 300 Longwood Ave., Boston, MA 02115.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Neurologic Complications of Heart Surgery in Infants
Larson C. P., Newburger J. W., Jonas R. A., Hickey P. R.
Extract |
Full Text
N Engl J Med 1994;
330:716-717, Mar 10, 1994.
Correspondence
Developmental and Neurologic Status of Children after Heart Surgery
Karl T.R., Thurston J. H., Hauhart R. E., Bellinger D. C., Jonas R. A., Hickey P. R.
Extract |
Full Text
N Engl J Med 1995;
333:391-392, Aug 10, 1995.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
