The Tin Man And The Scarecrow: Must we have a normal heart to have a normal brain?
By Scott G. Walker, MD
Riley Hospital for Children
Wise-Faberowski L, Loepke A: Anesthesia during surgical repair for congenital heart disease and the developing brain- neurotoxic or neuroprotective? Pediatr Anesth 2011; 21: 554-559
Guerra GG, Robertson CM, et al.: Neurodevelopmental outcome following exposure to sedative and analgesic drugs for complex cardiac surgery in infancy. Pediatr Anesth 2011; 21: 932-941
DiNardo JA: Should what we know about neurobehavioral development, complex congenital heart disease, and brain maturation affect the timing of corrective cardiac surgery? Pediatr Anesth 2011; 21: 781-786
“They… soon they found themselves in the midst of a great meadow of poppies. Now it is well known that when there are many of these flowers together their odor is so powerful that anyone who breathes it falls asleep, and if the sleeper is not carried away from the scent of the flowers, he sleeps on and on forever.”1 - L. Frank Baum
In L. Frank Baum’s The Wonderful Wizard of Oz, as the Tin Man and the Scarecrow approach the Emerald City in search of a heart and a brain, they encounter a field of poppies. In crossing it, their friends Dorothy (a human subject) and Toto and the Lion (animal models) undergo general anesthesia by inhalation induction. The Tin Man and the Scarecrow, being mechanical and stuffed, are unaffected. Given their cardiac and neurologic “lesions,” would they have benefited from an interval of anesthesia? Or would it have caused more harm? Can we extrapolate the outcomes of Toto and the Lion to that of Dorothy? Would it matter if Dorothy were a newborn?
Three recent studies published in Pediatric Anesthesia seek answers to questions about the effect of congenital heart disease (CHD) surgery and anesthesia on the developing brain. Wise-Faberowski and Loepke provide a concise review of available data. Guerra et al. analyze the association between cumulative anesthetic dose and neurologic outcomes in 95 survivors of neonatal heart surgery. DiNardo considers the question of timing and whether benefit can be derived from allowing an interval of brain maturation prior to surgical intervention.
In the article by Wise-Faberowski and Loepke, the authors note that neurologic outcome following congenital heart disease surgery is quite variable. Many children are totally normal, but up to half are affected by neurologic impairment. They note that while the etiology of neurologic insults in this population has been well studied and is clearly multifactorial, little is known about the contribution of anesthetic exposure despite the recent interest in this topic. Three factors are considered by the authors to be particularly important in terms of neurologic outcome: 1) chronic hypoxia, 2) cardiopulmonary bypass, and 3) prolonged anesthetic exposure.
In discussing the contribution of chronic hypoxia, the authors point out that the developing brain may or may not be affected by chronic hypoxia differently than the mature brain. It is possible that the developing hypoxic brain may be afforded a protective adaptive response, as supported by several rat studies. Implicated in this adaptive response is the predominance of the 2B subunit composition of N-methyl-D-aspartate (NMDA) receptors in the immature brain, versus 2A in the more developed brain. Although volatile anesthetic agents are known to act on NMDA 2B subunit receptors, the influence of these agents on an adaptive response to hypoxia is not known. Neurogenesis may also play a protective role, a process also noted to be influenced by narcotic agents.
Cardiopulmonary bypass is considered as a model of ischemia/reperfusion known to cause brain injury, and the practice of hyperoxic perfusion during bypass is brought into question. However, the authors point out that no data exists on the influence of anesthesia on this process in chronically hypoxic humans or animals.
Anesthesia for CHD surgery has historically included high-dose narcotics, in large part through the influence of the Boston Circulatory Arrest trials that showed evidence of improved outcome with this technique. However, Wise-Faberowski and Loepke point out that while the fetal brain is well equipped to mount an effective stress response and that opioid binding is independent of development, opioid receptor subtype expression does change during maturation. Specifically, the fetus has decreased expression of the mu receptor, which is involved in neuronal and glial proliferation. The authors contend that narcotic inhibition of neuronal myelination early in development, combined with preoperative deficits and the insult of chronic hypoxia and cardiopulmonary bypass, is a reason to question the use of perioperative narcotics in newborns with CHD. The effects of volatile anesthetics are also considered. Conjecture is made that given the known interaction of volatile anesthetics with GABAA receptors, the predominance of the alpha-2 subunit composition of these receptors in early development, and the association between GABAA alpha-2 receptors and promotion of synaptic growth and development, it is possible that use of volatile agents in the developing infant may promote neuronal cell death. To support this, reference is made to a rat study in which neuronal apoptosis was increased after exposure to isoflurane at post-natal day 7.
When combined with midazolam and N2O these rats also exhibited learning and memory deficits remote to the exposure. These findings are balanced against the observation that this study (and others similar to it) lacked physiologic control during anesthetic exposure. For example, in another study by Loepke and his group, about 1 in 5 rats did not survive the anesthetic exposure, compared to a 100% survival rate in the non-exposed group2. Also, considering the issue in the context of congenital heart disease, preliminary evidence suggests that hypothermia, a component of cardiopulmonary bypass, prevents the neurotoxic effects of anesthetics. Furthermore, other animal studies actually support a neuroprotective effect of volatile anesthetics. The authors refer to two studies, one in adult rats and one in piglets, that provide evidence of neuroprotection conferred by volatile anesthetics during ischemia and deep hypothermic circulatory arrest (DHCA), respectively. No evaluation of this effect has been made in a model of chronic hypoxemia, however.
Wise-Faberowski and Loepke appropriately leave us with more questions than answers. We do not yet know whether anesthetic neurotoxicity is different in the developing normoxic vs. hypoxic brain, or whether hyperoxia, as is frequently present during CPB, may be damaging to a brain that is apparently adapted to tolerate hypoxic conditions (e.g. in utero), or whether these adaptive mechanisms are disrupted by volatile agents and hyperoxia, and what the neurodevelopmental outcome of this might be.
Guerra et al. sought to take a very direct look at neurodevelopmental outcome associated with anesthetics and CHD surgery. Ninety-five chromosomally normal survivors of CHD surgery who were operated on at ≤6 weeks of age over a recent 3 year period were analyzed in this study from the Western Canadian Complex Pediatric Therapies Follow-up Group. In addition to prospective demographic and perioperative variables, the cumulative dose from each class of anesthetics used was calculated from retrospective data. Assessments were then made at 18- and 24-months of age using the Bayley Scales of Infant (or Infant and Toddler) Development, as well as a parental questionnaire, the Adaptive Behavior Assessment System (ABAS).
From these, four outcomes of interest were identified and analyzed: 1) adaptive skills, 2) adaptive motor skills, 3) significant mental and motor delay, and 4) significant vocabulary delay. Anesthetic exposure was measured over the entire interval between birth and 6 weeks of age, including exposure outside that of the admission for cardiac surgery (e.g., in the cardiac catheterization laboratory or during other procedures). Volatile agents were measured in terms of MAC hours, benzodiazepines in terms of midazolam equivalents, and opioids in terms of morphine equivalents. Ketamine and chloral hydrate were also included in the analysis, and each intravenous drug class was reported in terms of total mg/kg, average mg/kg/day on days the drugs were given, and total number of days drugs were given. Demographics were similar, and of note, 85% of the patients had cyanotic heart disease.
To very briefly summarize the findings of this study, no correlation was found between sedation variables and any of the four neurodevelopmental outcomes of interest when multivariate analysis was performed. Instead, correlation with poorer outcome was found with several non-sedation variables, such as duration of ventilation, years of mother’s schooling, and use of DHCA, among others. Limitations of the study, as noted by the authors, include the fact the the Bayley Scales changed during the course of the study, with about half analyzed with the second edition and about half with the third edition. Also noted were a small sample size, uncertain validity of extrapolation to patients without CHD, inconsistent use of ketamine, and the question of whether a preconditioning effect of intraoperative anesthesia might have prevented neurotoxicity related to postoperative sedation.
In support of their findings, the authors refer to two studies similarly constructed that looked at preterm neonates and also found no correlation between sedation variables and neurobehavioral outcomes. They conclude that whereas no evidence in their study suggests an association between anesthetics and neurologic outcome, this is preliminary data and further studies would be needed to draw definitive conclusions.
Finally, DiNardo presents us with a review article that carefully considers the existing data on CHD and brain injury (placed in the context of what is known about brain development) and asks whether we can expect better neurologic results if we wait awhile before we expose these patients to surgery and anesthesia. He starts by reiterating the clear evidence that patients with CHD, particularly complex CHD, have a high incidence of neurobehavioral problems in both short and long term follow up, then reminds us that this may be related to either fixed or modifiable factors. Modifiable factors include those things we do either pre-, intra-, or post-operatively in the management of these patients, including timing of surgery, conduct of CPB and anesthesia, and ICU treatment strategies. For example, patients with D-transposition of the great vessels often require balloon atrial septostomy (BAS) and this combination of lesion and treatment logically makes the patient vulnerable to paradoxical emboli and stroke. However, existing data conflict over whether there is, in fact, a causal correlation between BAS and stroke, illustrating the complex nature of preoperative brain injury in neonates. From the standpoint of intraoperative management, numerous modifications have been employed to reduce brain injury, including changes in hematocrit and acid-base management, as well as the use of antegrade cerebral perfusion (ACP) to reduce the duration and impact of DHCA.
Nonetheless, patients managed with ACP, including those monitored with near infrared spectroscopy (NIRS) and trans-cranial Doppler, still have a significant incidence of ischemic brain injury on postoperative MRI. In the immediate postoperative period, tight management of hemodynamic variables has been shown to be critical in reducing insults such as white matter injury (WMI) and other ischemic brain lesions, although keeping the NIRS values above a threshold value (e.g. 45%) may or may not be protective. In addition to these modifiable factors, DiNardo considers the relationship between CHD and abnormal neuroanatomic development. Note is made of the association between certain genetic syndromes (e.g. DiGeorge) and CHD, and the clear connection between most of these syndromes and abnormal CNS development. An especially strong point is made of the fact that brain maturation in the patient with complex CHD is delayed to the extent that it closely compares to that of the premature neonate, a population in whom there is an indisputable connection between degree of WMI and developmental impairment. An elegant argument is made that abnormal patterns of cerebral blood flow in fetuses with complex CHD, as well as alterations in cerebral vascular resistance and CO2 reactivity, are the most significant contributors to the development of intrauterine findings such as reduced brain volume and head size. DiNardo suggests that given the immaturity of these patients’ brains and their subsequent lack of appropriate cerebral autoregulation, re-evaluation should be made of acceptable lower limits of intraoperative blood pressure and the practice of hyperventilation to reduce pulmonary vascular resistance. He also praises ongoing efforts to ameliorate the perioperative inflammatory response, using the novel application of minocycline or erythropoietin to this problem as examples.
Finally, the hybrid stage I procedure to palliate hypoplastic left heart syndrome (HLHS) is considered as a clear example of an effort to delay open heart surgery, which, among other advantages, allows an interval of brain maturation. Is this strategy definitively associated with improved neurologic outcome? Unfortunately the evidence is not there yet. Nor, DiNardo points out, is it clear that a prolonged interval of hypoxemia, as might accompany a delay in surgical intervention, is not at least as damaging to the brain as surgery itself. Similarly, there is evidence that cardiac transplantation to palliate HLHS, even if it is delayed beyond the apparent vulnerable period of neonatal brain development and is accompanied by a shorter interval of chronic hypoxemia, is no better than Fontan staging at reducing neurobehavioral deficits later in childhood.
Ultimately, this review leaves open the question as to whether or not future strategies in CHD palliation should target methods to delay surgery and CPB until the brain can more fully mature. I would add that there are also compelling reasons not to delay surgery in this population unrelated to the CNS, such as evidence that indices of cardiac function may be improved in single ventricle patients who are volume-unloaded and/or relieved of cyanosis early rather than late3-5.
So, from the articles in question, what do we know, what do we think, and what do we not know? We know that patients born with abnormal hearts are also somewhat likely to have or to develop abnormal brains. We know that the causes of this are multifactorial, but almost certainly relate to, among other things, intrauterine blood flow patterns, postnatal hypoxia, genetic factors, and cardiopulmonary bypass during surgical intervention. We think that neonates have adaptive mechanisms to protect against hypoxia that are vulnerable to disruption. We don’t know if anesthetics play a role in neurologic deficits in this population, either via toxic or protective means, and whether or not it matters which anesthetic is used. We don’t know if the hypoxic human brain develops differently than the normoxic human brain. We don’t know if delaying surgery might improve neurologic outcome in these patients.
“They carried the sleeping girl to a pretty spot beside the river, far enough from the poppy field to prevent her breathing any more of the poison of the flowers, and here they laid her gently on the soft grass and waited for the fresh breeze to waken her.”1 - L. Frank Baum
References
- Baum LF. The Wonderful Wizard of Oz. George M. Hill Co., Chicago IL, 1900
- Istaphanos GK, Howard BS, et al. Comparison of Neuroapoptotic Properties of Equipotent Anesthetic Concentrations of Desflurane, Isoflurane, or Sevoflurane in Neonatal Mice. Anesthesiology 2011; 114: 578-87
- Mahle WT, Wernovsky G, et al. Impact of Early Ventricular Unloading on Exercise Performance in Preadolescents with Single Ventricle Fontan Physiology. J Am Coll Cardiol 1999; 34: 1637-43
- Milanesi O, Stellin G, Colan SD, et al. Systolic and Diastolic Performance Late After the Fontan Procedure for a Single Ventricle and Comparision of Those Undergoing Operation at <12 Months of Age and at > 12 Months of Age. Am J Cardiol 2002; 89: 276-80
- Anderson PAW, Sleeper LA, et al. Contemporary Outcomes After the Fontan Procedure (A Pediatric Heart Network Multicenter Study). J Am Coll Cardiol 2008; 52: 85-98
