Of Mice and Men: Should We Extrapolate Rodent Experimental Data to the Care of Human Neonates?

Over the years, Dr. Olney et al.  1have enlightened the scientific community with their research on the N -methyl-d-aspartate receptor and its role in human disease. Recently, Dr. Olney’s group has examined the effects of commonly used anesthetics and anticonvulsants on the developing brain. In several high-profile scientific journals, they reported that the administration of these drugs, including ketamine, ethanol, phencyclidine, nitrous oxide, isoflurane, propofol, barbiturates, diazepam, and other anticonvulsants all increase apoptotic neurodegeneration in developing rat brain.2–5We and others have replicated some of the studies with ketamine, and there is no doubt regarding the scientific validity of their findings.6,7 

The direct applicability of these experimental findings to the clinical practice of pediatric anesthesia and critical care, however, should be questioned. Likewise, the implication that all types of anesthetic and sedative agents may have similar potentially deleterious effects on neuronal development in neonates is premature and inappropriate. Multiple lines of evidence cast doubt on the clinical relevance of these experimental paradigms, as recently reviewed in our Special Article published in the August 2004 issue of Anesthesiology.8Dr. Olney et al.  9eloquently provided a counterpoint editorial in the same issue of Anesthesiology, which mainly questioned our assertion that repeated large doses of ketamine were responsible for mediating the neurodegenerative changes noted in neonatal rat brains. Similar doses and durations of administration, particularly in the absence of surgical stimulation, are never used in pediatric anesthesia.

Recent data from the Neurotoxicology division at the National Center for Toxicological Research (Jefferson, Arkansas) further confirm our findings that the single- or smaller-dose regimens of ketamine do not increase neuronal apoptosis in the brains of 7-day-old (P7) rat pups.7Scallet et al.  7examined the effect of various dosing regimens on the different assays for neuroapoptosis in P7 rat pup brains. They reported no differences in the appearance of apoptotic neurons in rat pups receiving saline, seven doses of ketamine (10 mg/kg) over 9 h, and a single dose of ketamine (20 mg/kg), whereas rat pups receiving higher doses, seven doses of ketamine (20 mg/kg) over 9 h, had significant increases in neuronal apoptosis. Scallet et al.  found that the magnitude and pattern of neuronal changes were similar to those reported earlier.2,6Unlike Olney et al. , these investigators also measured blood concentrations of ketamine to show that the repeated low-dose (10 mg/kg) and single-dose regimens resulted in the blood concentrations achieved in clinical practice, whereas the repeated high-dose regimen (20 mg/kg) resulted in sevenfold higher ketamine concentrations. Furthermore, Scallet et al.  verified the presence of neurodegeneration with several histochemical methods, including cupric silver and Fluoro-Jade-B staining for nonspecific neurodegeneration, 4′,6-diamido-2-phenylindole for nuclear DNA, and caspase-3 immunohistochemistry as a marker for neuronal apoptosis. These findings confirmed that ketamine doses and duration both play important roles in ketamine-induced neurodegeneration.2,6,7 

In their editorial, Olney et al.  9disagree with our concern that neurodegenerative changes after repeated high doses of ketamine may result from hypoxic/ischemic damage because of the hemodynamic effects of high-dose ketamine. “It is excitotoxic, and ultrastructurally does not resemble apoptosis,” they state. The neuronal cell death that occurs soon after exposure to hypoxia/ischemia does result from excitotoxicity, but rat brains were not sampled at this time period in the studies by Olney et al.  2,4A voluminous literature shows a delayed neuronal cell death after hypoxia/ischemia that results from neuronal apoptosis in the immature rat brain and also occurs at the same time periods (4–24 h) as those sampled by Olney et al.  10–13Excitotoxicity cannot account for hypoxia-induced injury in immature neurons but may explain the effects of hypoxia on more mature neurons,14and even the excitotoxic damage of immature neurons leads to apoptosis at 6–24 h,15at the time periods studied by Olney et al.  2,4 

In their editorial, Olney et al.  9state that arterial blood gases showed oxygen saturations of 97–99% (contained in the legend for their fig. 2). To our knowledge, the complete data from arterial blood gases obtained at 0, 15, 30, 60, 120, 180, or 240 min in neonatal rats given high-dose ketamine have not been reported in the literature and cannot guarantee the absence of significant hypoxia occurring between these sampling times. In contrast, pediatric anesthesiologists routinely and continuously monitor oxygenation (oximetry), ventilation (capnography), and other hemodynamic variables during neonatal anesthesia and respond to these monitors within seconds after noting a clinically important change.

Similarly, our concern for nutritional deprivation was also dismissed by Olney et al.  9because of the lack of “a robust pattern of neuroapoptosis” in the controls from their experiments, stating that “all animals are sacrificed 4 to 8 h later for histologic evaluation of the brains.” Their experiments, however, showed that rat pups were killed at 4, 8, 12, 16, 24, or 48 h after ketamine and that apoptosis was most prominent in the brains sampled at 12, 16, and 24 h.2If newborn rat pups were exposed to nutritional deprivation for these periods of time, increasing patterns of neuronal apoptosis would be expected.16,17 

Last, we must restate our concern that neonatal rats or humans exposed to anesthetic agents in the presence versus  the absence of surgically induced pain or stress would manifest drastically different effects and that the observations that the neuronal effects of all anesthetic agents and ketamine in particular are dose dependent.8At the Society for Neuroscience annual meeting on November 8–12, 2003 (New Orleans, Louisiana), Olney et al.  reported that a single sedating dose of ketamine increases neurodegeneration, but this finding was noted from neonatal mice, highlighting the importance of species-dependent variations in the response to pharmacologic agents.18,19 

We agree with Drs. Olney and Todd that human clinical studies are needed to truly examine the potential neurotoxic effect of anesthetic drugs in the developing human brain.9,19,20Given the ethical and societal constraints, randomized controlled trials mirroring the experimental designs used in rodent models are not possible. We examined children at 4 and 8 yr after the surgical repair of congenital heart defects as neonates using a standardized anesthetic regimen including high-dose barbiturates and opiates.21–23Despite the known neurologic sequelae of cardiopulmonary bypass and deep hypothermic circulatory arrest,24their performance IQ scores were not significantly different from the population norms (fig. 1). Certainly their neurologic outcome was no worse than that of peer groups with similar congenital heart defects repaired later in life and exposed to anesthetic agents after the “vulnerable” neonatal and infant periods of brain development at issue in the study of Newburger et al.  25The longer operative repair (and anesthetic exposure) was postponed after infancy in this study population, the greater the cognitive impairment was. These findings suggest that even under extreme conditions encountered in the conduct of pediatric anesthesia, neonatal exposure to anesthetic agents does not necessarily confer substantial neurologic deficits and is, at worst, only one factor among many potentially influencing brain development.

Regardless of this interpretation of such data, the controversy initiated by the studies of Olney and others should provoke clinical investigators to examine the long-term neurocognitive effects of prolonged exposure to anesthetic drugs in pediatric patients, to help sort out the differences between mice and men. Further knowledge in these areas is required before changes in clinical practices in neonatal and pediatric anesthesia can be recommended; such changes may not be necessary or may even be deleterious to long-term neurologic outcome. We are all indebted to Olney et al.  for their contribution in raising these issues.

* Children’s Hospital Boston and Harvard Medical School, Boston, Massachusetts. sulpicio.soriano@childrens.harvard.edu