Understanding how anesthetics cause memory loss is of considerable interest, as some patients experience insufficient amnesia and unexpected recall of adverse surgical events. The incidence of explicit recollection of surgical events that occur during general anesthesia, a disorder referred to as “intraoperative awareness,” is approximately 1.3 per 1,000 patients [2, 3]. The incidence of possible awareness, whereby patients recall events during anesthesia and surgery that cannot be substantiated by operating room personnel, may be as high as 2.4 per 1,000 patients [2]. Given that over 40 million patients undergo anesthesia in North America each year (American Society of Anesthesiologists), thousands of patients may be at risk for serious psychological harm.

At the opposite end of the spectrum of memory disorders associated with anesthesia are the persistent memory deficits that some patients experience. A decline in cognitive performance after general anesthesia is commonly referred to as postoperative cognitive dysfunction (POCD) [4–6]. In the International Study on POCD, approximately 37 % of young adults and 41 % of older patients exhibited cognitive deficits in the days immediately following surgery, with the deficits persisting for up to 3 months in 6 to 13% of patients [4]. Loss of explicit memory and impairment of executive function (which involves skills such as decision-making or planning follow-up appointments) are common features of POCD [7, 8].

The duration of general anesthesia is an independent predictor of cognitive dysfunction, and anesthetics are thereby implicated as a potential causative factor in POCD [6]. However, POCD also occurs in patients undergoing regional anesthesia and conscious sedation, which suggests that inflammation and sedative drugs are contributing factors [8, 9]. Other factors that may contribute to poor performance on psychometric tests include the patient’s underlying illness, disrupted sleep during the hospital stay, the stress and novelty of the hospital environment, and postoperative pain [4, 7].

Cognitive deficits after surgery represent a serious concern as they are associated with poor long-term outcome, premature retirement, loss of independence, and increased long-term mortality [7]. Unfortunately, no treatments or preventive measures are available for POCD. Thus, studies of laboratory animals are required to disentangle the role of anesthesia from the myriad of other potential contributing factors. As described below, research to elucidate the molecular basis of POCD and to identify potential therapeutic strategies is in progress.

Studies using rodent models have shown that a clinically relevant dose of the inhaled anesthetic isoflurane or sevoflurane can impair anterograde spatial memory, as well as retrograde spatial and recognition memory, and that such deficits persist for days to weeks after exposure [10–13]. Neuronal damage and cognitive deficits in rodents are positively correlated with the duration of anesthesia and the number of exposures [14]. Readers interested in a more detailed summary of recent studies of anesthesia-induced neurotoxicity are referred to a commentary [15] and a special issue on anesthetic neurotoxicity and neuroplasticity of the British Journal of Anaesthesia [16].

Fetal and newborn brains are particularly vulnerable to the neurotoxic properties of anesthetics, which cause apoptotic neuronal death by increasing proapoptotic proteins such as caspase-3, cyclin D1, and Bcl-2-interacting protein [17–20]. Inhalational anesthetics also cause mitochondrial dysfunction and increase the levels of reactive oxygen species in neurons, which further contribute to apoptotic cell death [21–23]. At peak periods of brain development, anesthetics can reduce synaptogenesis, thereby causing a sustained reduction in the number of synapses [24].

In the adult brain, inhaled and intravenous anesthetics have been shown to promote the development of Alzheimer’s disease-related cytopathology [25–31]. Transgenic mouse models of Alzheimer’s disease are more susceptible to anesthetic-induced increases in cytokine levels and cognitive deficits than wild-type mice [32]. Exposure to anesthetics increases amyloid-beta aggregation and tau hyperphosphorylation [28–31, 33]. However, the mechanisms underlying anesthetic-induced memory loss in adult brains without any preexisting neuropathology are unknown.

Our recent laboratory studies have focused on the mechanisms underlying anesthesia-induced memory loss. Most general anesthetics cause their primary end points by acting on specific ion channels in the brain and spinal cord [34]. In particular, γ-aminobutyric acid type A (GABA_A) receptors are inhibitory receptors in the brain that are targets for most general anesthetics and intravenous sedative drugs [34, 35], including the volatile anesthetics sevoflurane, desflurane, and isoflurane and the intravenous anesthetics etomidate, propofol, barbiturates, and benzodiazepines (Fig. 54.2).

GABA_A receptors are chloride-permeable ion channels that assemble from 5 subunits encoded by 19 different genes. The particular subunit composition of the receptors influences their pharmacological, physiological, and biophysical properties, as well their regional and cellular patterns of expression. GABA_A receptors are categorized into two main groups, synaptic and extrasynaptic, according to their proximity to the synapse [36]. Synaptic receptors are transiently stimulated by high concentrations of GABA to generate fast inhibitory postsynaptic currents [36]. Extrasynaptic receptors are composed of α5, β3, and γ2 subunits or, alternatively, α4/α6, β2/β3, and δ subunits. These receptors are activated by low, ambient concentrations of GABA to generate a tonic current [36–39] that reduces neuronal excitability [40, 41].

Our recent studies have focused on extrasynaptic α5-subunit-containing GABA_A (α5GABA_A) receptors, which are abundantly expressed in the hippocampus, a brain structure that is essential for learning and memory recall [42]. These receptors are highly sensitive to many classes of neurodepressive drugs, including the general anesthetics isoflurane, midazolam, and propofol [43–45]. These drugs bind directly to the receptors or interact with water-filled binding cavities and act as positive allosteric modulators that markedly increase opening of the integral ion channel [34, 46]. Anesthetics such as propofol also reduce receptor desensitization (closure of the ion channel while the agonist remains bound to the receptor) [47]. Together, the increase in receptor activation and the reduction in desensitization allow channels to remain in the open conducting state for prolonged periods [45].

Behavioral studies of humans and laboratory animals have shown that the expression and function of α5GABA_A receptors regulate memory behaviors. Reducing α5GABA_A receptor activity through pharmacological inhibition (by the selective inverse agonist L-655,708) or through genetic knockdown of the gene encoding the α5 subunit (to generate Gabra5–/– mice) improves performance for certain memory tasks. L-655,708 also enhances long-term potentiation of excitatory synaptic plasticity in the hippocampus, a molecular substrate of memory [48]. Conversely, drugs that increase the activity of α5GABA_A receptors, such as the intravenous anesthetic etomidate, cause memory deficits and prevent the induction of long-term synaptic plasticity [49, 50].

Studies with genetically modified mice have further implicated α5GABA_A receptors in drug-induced memory loss. Wild-type mice treated with a low dose of etomidate were unable to learn a fear-conditioning task during treatment [49]. In contrast, Gabra5−/− mice treated with etomidate and wild-type mice co-treated with etomidate and L-655,708 formed new memories during training on the fear-conditioning task [49]. Collectively, these results show that α5GABA_ARs are required for the acute memory blockade induced by etomidate (Fig. 54.3).

In recent studies, we have examined whether α5GABA_A receptors also contribute to long-term memory deficits that persist after elimination of the anesthetic. In a preclinical mouse model, we showed that a single, 1-h treatment with isoflurane or sevoflurane caused learning and memory deficits for contextual fear memory and object recognition memory for at least 48 h [51, 52]. Such memory loss occurred at a time when concentrations of anesthetic in the brain were undetectable or at the limits of detection [51].

More importantly, pharmacological or genetic inhibition of α5GABA_A receptors prevented these memory deficits [51, 52]. This result suggests that α5GABA_A receptors are required for triggering postanesthetic memory deficits and contribute to the memory deficits observed after anesthesia. Ongoing studies are evaluating the effectiveness of α5GABA_A receptor inhibitors for treating memory deficits after intravenous anesthesia in rodents and nonhuman primates.

Systemic inflammation after major surgery may also contribute to postoperative memory loss. The levels of many proinflammatory cytokines, including tumor necrosis factor-α, interleukin 1β (IL-1β), and interleukin 6, are markedly increased in the circulation and brain after surgery [53–55]. These cytokines, which are released from macrophages at the site of injury, stimulate innate immune responses in the brain and activate resident microglia to release additional cytokines.

Rodent models of various types of surgery, including hepatectomy, splenectomy, and orthopedic tibial osteotomy, have confirmed that inflammation is associated with postoperative memory deficits [53, 56–58]. In particular, increased levels of cytokines were associated with memory loss for certain hippocampus-dependent tasks, such as contextual fear-conditioning and spatial navigational memory [53, 56].

Innovative strategies aimed at reducing memory deficits after surgery and anesthesia have included treatment with anti-inflammatory drugs or selective antagonists for certain proinflammatory cytokine receptors [53, 59]. For example, the anti-inflammatory antibiotic minocycline [53] and the COX-2 inhibitor parecoxib [60] reduce cognitive deficits associated with postoperative inflammation in a mouse model. A major limitation of this approach is that reductions in the immune response could delay wound healing or increase the risk of postoperative infection. Therefore, we have studied an alternative approach to identify and target specific receptors in the hippocampus that mediate the memory-blocking effects of proinflammatory cytokines.

We recently investigated whether systemic inflammation increases the expression of α5GABA_A receptors in hippocampal neurons. We found that IL-1β enhances the activity and surface expression of α5GABA_A receptors, an effect that contributes to memory deficits [61]. The increased expression of α5GABA_A receptors depended on binding of IL-1β to the interleukin-1 receptor and involved activation of p38 mitogen-activated protein kinase-dependent signaling pathways [61]. Importantly, memory performance was impaired in wild-type mice in which systemic inflammation was induced by injection of the bacterial endotoxin lipopolysaccharide or IL-1β, but Gabra5−/− mice did not exhibit such memory deficits. Pharmacological inhibition of α5GABA_A receptors also reduced memory deficits induced by systemic inflammation (Fig. 54.4). Collectively, these results suggest that proinflammatory cytokines contribute to memory deficits by enhancing the number of functional inhibitory α5GABA_A receptors in the brain. Animal models have shown that anesthetics further stimulate the release of proinflammatory cytokines [62], which may exacerbate memory deficits.

The results summarized above raise many questions. For example, does a reduction in the expression levels of α5GABA_A receptors contribute to intraoperative awareness? Polymorphisms exist in the human gene that encodes the α5 subunit [63], but whether such polymorphisms increase the risk of awareness in patients is unknown. Additionally, it will be of great interest to determine whether inverse agonists that preferentially inhibit α5GABA_A receptors can either prevent or treat memory loss associated with anesthesia in patients.

It will also be important to determine whether different classes of anesthetics differ in their ability to cause postoperative cognitive deficits. For example, preliminary studies in rodent models have shown that sevoflurane but not desflurane causes an increase in reactive oxygen species and subsequent cell death in the young brain [64, 65]. The sedative dexmedetomidine may also be neuroprotective rather than neurotoxic [66–68]. Lastly, studies in both animals and humans should help in determining whether the cognitive deficits associated with inflammation can be reduced without impairing or delaying wound healing.

From a practical perspective, high-risk patients should be informed about the risks of cognitive deficits after anesthesia and surgery, and appropriate precautions should be taken, such as writing down critical information about drugs and treatment recommendations during the postoperative period. Family members or other care providers may wish to accompany patients to ensure that postoperative care instructions are received and recorded correctly. Also, the exposure of health care personnel to low levels of anesthetics should be minimized until we develop a better understanding of the long-term consequences of anesthetic-induced neurotoxicity.