We thank Dr. Laudanski for his thoughtful comments and questions. Our study1 was motivated by the need to understand why critically ill patients, such as those in high inflammatory states (e.g., due to sepsis, trauma, burns, autoimmune disease), are often more sensitive to general anesthetic drugs than are healthier patients. Identifying the factors that cause a standard dose of anesthetic to produce deeper levels of sedation in sicker patients is fundamentally important because of the risks associated with a relative drug overdose.2 Indeed, excessively deep anesthesia may increase postoperative mortality.3–5
Clinical trials are underway to determine whether anesthetic dosing can be safely reduced in high-risk patients; however, discerning the specific factors that modify anesthetic sensitivity in patients is often problematic (see www.anzctr.org.au—ACTRN12612000632897 and www.clinicaltrials.gov—NCT00998894). Consequently, to test specific hypotheses, we used preclinical mouse models, which allow good control over experimental variables. We studied whether inflammation increases anesthetic up-regulation of γ-aminobutyric acid type A (GABAA) receptor activity in neurons in vitro. Next, we tested whether such up-regulation correlates with the enhanced sensitivity for anesthetic behavioral endpoints induced by inflammation in vivo.1
Inflammation in vitro was mimicked by treating hippocampal and cortical neurons with a proinflammatory cytokine, interleukin-1β. As Dr. Laudanski correctly points out, inflammation stimulates the production of an array of cytokines, including interleukin-1β, interleukin-6, and tumor necrosis factor α.6 We selected interleukin-1β for several reasons. An increase in interleukin-1β levels occurs in a variety of medical and surgical disorders, such as cancer,7 aging,8 sepsis-associated encephalopathy,9 and surgical trauma.10 Increased levels of interleukin-1β correlate with cognitive deficits in humans9 and laboratory animals10,11 and may potentiate the cognitive depressive properties of anesthetics. Also, we have previously shown that interleukin-1β increases the cell-surface expression of α5 subunit–containing GABAA receptors in the hippocampus,11 whereas interleukin-6 and tumor necrosis factor α do not.11 These α5 subunit–containing GABAA receptors are highly sensitive to up-regulation by many anesthetics.12 Thus, the choice of interleukin-1β for this study was both reasoned and appropriate. We found that pretreating neurons with interleukin-1β markedly increased anesthetic up-regulation of GABAA receptor function.
Similarly, we selected etomidate and isoflurane for these studies for specific reasons. Etomidate, the most GABAA receptor–selective anesthetic,13 causes minimal hemodynamic effects, which limits the impact of pharmacokinetic factors. Isoflurane, on the other hand, is widely used in clinical practice and has been extensively studied in laboratory animals, so comparisons could be drawn between our results and published data.14 Interestingly, we found that other g-aminobutyric acid–mediated (GABAergic) drugs, including propofol and midazolam, also cause greater up-regulation of GABAA receptor function in hippocampal neurons treated with interleukin-1β (unpublished observation). It is of future interest to determine whether inflammation also potentiates the effects of non-GABAergic anesthetics, including ketamine and nitrous oxide.
Next, we adopted a widely used in vivo model of systemic inflammation to probe whether increased levels of endogenous cytokines increase sensitivity to GABAergic anesthetics. Mice were treated with a low dose of lipopolysaccharide (125 μg/kg) that was sufficient to stimulate an increase in cytokines rather than to mimic sepsis. Such a low dose of lipopolysaccharide does not produce gross hemodynamic instability, as confirmed by studies using echocardiography in isoflurane-treated animals.15 We did not feel compelled to replicate these studies. Our results showed that lipopolysaccharide-treated mice exhibited greater sensitivity to etomidate, as evidenced by the increased hypnotic and immobilizing effects. Lipopolysaccharide-treated mice were also more sensitive to isoflurane but only for behavioral endpoints mediated by GABAA receptors, such as hypnosis but not immobility. Thus, only the GABAA receptor–dependent behaviors were modified in this experimental model. These results show that future studies must measure more than just the immobilizing properties of anesthetics, because minimum anesthetic concentration (MAC) values (considered a gold-standard measure of anesthetic potency) may not correlate with the dose required for hypnosis.
It was suggested by Dr. Laudanski that we measure interleukin-1β levels in the lipopolysaccharide-treated mice. Increased levels of interleukin-1β and other cytokines in the plasma and cerebrospinal fluid of rodents treated with lipopolysaccharide have been reported by others.16,17 Also, cytokines are both autocrine and paracrine cell-signaling factors,18 and measuring circulating cytokine levels has limited value in advancing our understanding of drug actions at the level of neurons.18 In fact, interleukin-1β levels in brain tissue under inflammatory conditions are often higher than those in plasma or cerebrospinal fluid.19 Interleukin-1β is produced locally in the brain in response to endotoxins by a variety of interleukin-1β–producing cells, including macrophages, microglia,20 and astrocytes.17 Moreover, levels of interleukin-1β in the serum are maintained at lower levels because interleukin-1β binds to circulating proteins, such as α2-macroglobulin and complement.19,21 It would be of great interest to measure interleukin-1β concentration at the level of neurons; however, this is a challenging task to achieve with the technologies currently available.
The mechanisms by which inflammation causes anesthetic sensitivity in patients are complex and difficult to resolve in a single study. Therefore, in the Discussion, we stated that “These findings from animal studies cannot be directly extrapolated to patients.”1 Nevertheless, we anticipate that our data will further clinical efforts to identify the biologic underpinning of anesthetic hypersensitivity, in order to optimize anesthetic dosing in critically ill patients.
Supported by operating grants from the Canadian Institutes of Health Research (CIHR), Ottawa, Ontario, Canada, and a Canadian Research Chair (grant numbers: 416838 and 480143; to Dr. Orser); CIHR Fellowship, a Clinician Scientist Transition Award from the Department of Anesthesia, University of Toronto, Toronto, Ontario, Canada, and a Mentored Research Award from the International Anesthesia Research Society, San Francisco, California (to Dr. Avramescu); and Tier I Canada Research Chair in Sleep and Respiratory Neurobiology (to Dr. Horner).
The authors declare no competing interests.