In Reply:
We thank Drs. Engelhardt, Blaylock, and Weiss for their comments on our article,1and we are happy to clarify the issues of ketamine dosing. Intrathecal ketamine (30 mg/kg) produced irreversible sedation and respiratory depression in P3 pups, and excitation and convulsions in P21 animals after emergence from anesthesia that necessitated termination. As a result, the same degree of dose escalation was not possible with ketamine as with morphine. Rather than indicating that “sublethal doses” of ketamine are associated with apoptosis, this emphasizes the narrower therapeutic window between analgesia and dose-limiting side effects with ketamine.
The authors incorrectly stated that “no data on analgesic action” were provided for 0.1–0.3 mg/kg intrathecal ketamine. Figure 1B clearly presents dose-response data for antihyperalgesic effects of intrathecal ketamine in both P3 and P21 pups.1Because of ketamine's mode of action, increases in baseline sensory threshold (i.e. , antinociceptive effects) are not seen. The increased primary afferent input after injury (i.e. , carrageenan-induced hind paw inflammation) results in activation of N -methyl d-aspartate-mediated sensitization in the spinal cord, and dose-dependent reversal of hyperalgesia by ketamine can now be demonstrated. This pattern of response is discussed and referenced under “Dose-dependent effects” in our manuscript. Significant reversal of hyperalgesia was seen 30 min after intrathecal ketamine 3 mg/kg in P3 rats and 15 mg/kg in P21 rats. As we had shown that ketamine produced apoptosis within this analgesic dose range in P3 pups, repeating the same experiments with subtherapeutic doses would represent unnecessary use of animals and resources.
Engelhardt et al. refer to doses that “are the comparative and relevant equivalents commonly employed for caudal anesthesia.” We are surprised that the authors expect there to be a direct correlation between the doses uesd in different species, at different ages, and by different routes. Again, these issues were covered in our discussion.1Our rationale for describing results in terms of a therapeutic index was to provide a ratio of toxic to functional doses that could facilitate comparison of different drugs at different developmental stages. Similarly, studies evaluating apoptosis following systemic ketamine have related doses to functionally significant general anesthetic effects in different species at specific developmental ages (i.e. , infusion of ketamine at doses producing surgical anesthesia in primates2or the ED50for loss of righting reflex in rodents3).
Our evaluation of apoptosis was confirmed using multiple methodologies. The low standard errors and statistically significant differences comparing ketamine but not morphine with saline control groups, in the setting of evaluation of appropriately randomly coded sections, support the validity of our method of quantification. Although we agree that the size or dermatomal level may have influenced the total number of cells, the whole lumbosacral spinal cord section and not selected fields of view were examined, and for each animal, counts were performed in at least four randomly selected sections for each different evaluation (activated caspase-3 immunohistochemistry, FluoroJade C, and hematoxylin and eosin staining). The “wandering mean” method of counting would not be appropriate because of the uneven distribution of positive cells (as stated in the article, more apoptotic cells were identified in the dorsal rather than ventral horn; fig. 41). Consistent with previous reports, the level of baseline neuronal apoptosis was higher in the spinal cord in early postnatal life, and consistent reductions in apoptotic counts with increasing age at time points from P4 to P22 were found in naïve and saline control groups using all our methodologies. Results from activated-caspase 3 immunohistochemistry correlated with those obtained by counting apoptotic profiles in hematoxylin and eosin-stained sections,4and although not reported in the manuscript, double staining with activated-caspase 3 and the neuronal marker NeuN was confirmed in preliminary experiments. Numbers of FluoroJade C-positive cells were consistent with but higher than those with activated-caspase 3, as Flouro-Jade C staining is a more sensitive method and will capture dead cells at most stages of cell death, dying by any cell death pathway. Importantly, the authors do not acknowledge that single-dose intrathecal ketamine at P3 also produced long-term functional changes in mechanical sensory threshold and paw placement during gait, or that data from adult models have shown toxicity following intrathecal ketamine, providing further evidence for caution with neuraxial administration of ketamine.
Engelhardt et al. refer to apoptotic effects of local anesthetics. The referenced study confirmed concentration-dependent apoptosis when cultured cells were exposed to local anesthetics for 24 h, and the susceptibility to apoptosis was related to the physicochemical properties of different local anesthetics.5Interestingly, using the same methodology, concentration- and time-dependent apoptosis after exposure to preservative-free racemic ketamine and S -ketamine has recently been reported.6The potential limitations of extrapolating in vitro to in vivo data are noted in the discussion of both studies. In particular, in vitro evaluation of isolated cells in culture does not allow comparison of specific age-dependent effects, particularly if mechanisms are reliant on alterations in synaptic transmission.5,6As noted in our discussion, prolonged exposure to morphine also produces apoptosis in cell culture, but was not seen in our in vivo model,4and further studies of age-dependent local anesthetic effects in vivo are warranted.
We agree with Engelhardt et al. that hypoxia has the potential to cause major morbidity and mortality in children; but this is a different issue and we would hope that the risk is small in children undergoing elective surgery with adequate monitoring and skilled anesthetic care. Similarly, complications after regional analgesia are fortunately rare, although the focus has been on major neurologic complications rather than on the more sensitive evaluations of sensory function. Our concern relates to neuraxial administration of drugs without adequate preclinical safety data, particularly in clinical studies of healthy children undergoing unilateral hernia repair as referenced by Englehardt et al. , where equivalent analgesia can be achieved with less invasive measures7or with neuraxial drugs with a more comprehensive safety record. As stated previously,8we are very aware and supportive of the important role of regional analgesia for perioperative analgesia in children. However, preclinical assessment of the relative toxicity of current and potential new neuraxial drugs, particularly in early life when the developing nervous system may be more vulnerable, has an important role in informing clinical choice. If a range of drugs have similar analgesic efficacy, it would seem prudent to choose a drug with a wide therapeutic index and documented safety record.
* UCL Institute of Child Health and Great Ormond Street Hospital, London, United Kingdom. suellen.walker@ich.ucl.ac.uk.
