THE QUESTION OF WHAT HAPPENS TO TIME IN THE ANESTHETIZED PATIENT IS PHILOSOPHICALLY FASCINATING, CLINICALLY SIGNIFICANT, AND UNDER-studied. Unlike the case of natural sleep in which the sleeper retains a rudimentary awareness of a long night's rest or of a brief and insufficient slumber, for the patient undergoing general anesthesia, the cognitive sense of time elapsed fully disintegrates. Whether this remains true globally across the CNS or whether the underlying neural circuits preserve some subconscious measure of elapsed time is an open question.
Seminal studies conducted over the past two decades have suggested that general anesthetics interact with the neural systems regulating endogenous arousal.1,2 As the concentration of an inhaled anesthetic drug is increased to the point of loss of consciousness, predictable EEG changes arise including anteriorization of EEG power, an initial increase of β power followed by subsequent slowing into spindle-like waves in the α frequency range, with further slowing into the θ range and eventual migration of power in δ frequencies.3 Functional imaging studies reveal similar changes in cortical and thalamic activity as subjects transition into NREM sleep or anesthetic hypnosis.4 Finally, both states are accompanied by an effective spatial and temporal breakdown in cortical communication.5,6 Complementarity between the states of sleep and general anesthesia is further highlighted by the findings that both sleep deprivation and the endogenous somnogen adenosine potentiate the hypnotic properties of several anesthetics.7,8 Hence, it is clear that endogenous regulation of arousal state has mechanistic significance for anesthesiology. However, might general anesthesia also contribute to the understanding of sleep neurobiology?
Clearly the states of general anesthesia and NREM sleep are not identical. Among the many obvious distinctions, sleep is a homeostatically regulated process. Despite efforts to subvert our need for sleep, and the hope that the homeostat might accommodate over time, a recent study by Tononi, Cirelli, and colleagues demonstrates that homeostat is not easily reset. Increasing sleep pressure intrudes into wakefulness9 and impairs performance.10
Enter general anesthesia. Years ago, insightful studies by Tung and colleagues demonstrated that during a prolonged 12-hour infusion of general anesthetic propofol, administered during the rest phase, rats placed into a state of anesthetic-induced unconsciousness did not accrue any sleep debt.11 Moreover, preexisting sleep debts amassed during 24 hours of sleep deprivation fully dissipated during a subsequent propofol general anesthetic.12 Together these studies suggested the possibility that general anesthesia produced by propofol at least, might directly satisfy the homoeostatic drive or else erase the homeostat's record of rising sleep pressure. Several crucial questions remained unaddressed leading up to the work by Nelson and colleagues in this issue.13 First, might all general anesthetics be capable of appeasing the homeostatic clock? Work by Mashour and colleagues suggests not.14 Second, would the slow delta waves common to both NREM sleep and states of volatile anesthetic-induced unconsciousness be sufficient to mimic recovery sleep?
Nelson and colleagues address these questions with their work presented in this issue.13 Following undisturbed baseline EEG recordings, rats were sleep deprived for four hours using gentle handling and novel objects during their rest period. Animals were next exposed to a mock anesthetic or to one hour of isoflurane or desflurane at hypnotic concentrations that produce continuous slow waves. While the anesthetized groups did not show any changes in gross sleep architecture or continuity measures when compared to controls, they did exhibit significantly reduced levels of 0.5–4.0Hz slow wave activity (SWA) during the rebound sleep, when compared to baseline and mock-anesthetized controls. As with propofol, this suggests that volatile anesthetics relieve a component of homeostatic pressure that accumulates during sleep deprivation and that the slow waves of volatile-anesthetic-induced hypnosis may substitute for those of endogenous NREM sleep. Surprisingly however, extremely deep supra-hypnotic doses of desflurane that cause seconds of isoelectric EEG with interspersed bursting (burst suppression), but that lack slow wave activity, were also shown to discharge the SWA in the rebound following 4 hours of sleep deprivation. While such extremely deep levels of desflurane may have been associated with hypoxia, hypotension, and irreversible tissue ischemia—hypothetically including injury to the homeostatic time keeper in these spontaneously breathing rats—the normal appearance of cortical EEG and cumulative slow wave energy (SWE) in the deeply anesthetized desflurane group that resemble uncorrected SWE in the lower concentration anesthetic exposures make this possibility less likely. Alternatively, results with high-dose desflurane exposed rats suggest that anesthetic-induced slow waves are not necessary to discharge sleep pressure.
What remains unclear is whether anesthetic dissipation of sleep debt employs the same neurochemical mechanisms as natural sleep. An alternate possibility is that anesthetics may disrupt the networks responsible for the rebound slow wave activity, thus reducing SWA without achieving the restorative benefits associated with natural sleep. By varying either the amount of sleep deprivation or length of anesthetic, it would be possible to determine whether homeostatic drive is being gradually satisfied or merely disrupted at a certain threshold. The finding that REM sleep rebound after 24 hours of selective REM sleep deprivation is not attenuated by six hours of isoflurane exposure confirms that the homeostatic clock is not fully disabled during isoflurane anesthesia.14 Exploiting the distinct effects of individual anesthetic drugs upon the sleep homeostat11,13,14 may offer an opportunity to elucidate the homeostat's neurochemical underpinnings, as closely related anesthetic drugs are known to exert different effects upon common neuronal targets.15,16
Sleep deprivation has negative consequences on multiple behavioral measures, including learning, memory, and task performance.10,17 While the effects of anesthetics upon the sleep homeostat are provocative, future studies must also examine higher cognitive functions. With respect to these, the anesthetic seems unlikely to be as restorative as recovery sleep.14 Although finding a “magic bullet” to address the sleep deprivation common to our modern society without actually paying the sleep debt will likely remain elusive, exploring a possible dichotomy between a reduction in slow wave pressure and task performance may nonetheless yield new insights into the neural and molecular source of the restoration provided by sleep.