Anesthesia and the Neurobiology of Fear and Post-Traumatic Stress Disorder

Abstract

Purpose of Review.

Dysfunction of fear memory systems underlie a cluster of clinically important and highly prevalent psychological morbidities seen in perioperative and critical care patients, most archetypally post-traumatic stress disorder (PTSD). Several sedative-hypnotics and analgesics are known to modulate fear systems, and it is theoretically plausible that clinical decisions of the anesthesiologist could impact psychological outcomes. This review aims to provide a focused synthesis of relevant literature from multiple fields of research.

Recent Findings.

There is evidence in some contexts that unconscious fear memory systems are less sensitive to anesthetics than are conscious memory systems. Opiates may suppress the activation of fear systems and have benefit in the prevention of PTSD following trauma. There is inconsistent evidence that the use of propofol and benzodiazepines for sedation following trauma may potentiate the development of PTSD relative to other drugs. The benefits of ketamine seen in the treatment of major depression are not clearly replicated in PTSD-cluster psychopathologies, and its effects on fear processes are complex.

Summary.

There are multiple theoretical mechanisms by which anesthetic drugs can modulate fear systems and clinically important fear-based psychopathologies. The current state of research provides some evidence to support further hypothesis investigation. However, the absence of effectiveness studies and the inconsistent signals from smaller studies provide insufficient evidence to currently offer firm clinical guidance.

Introduction

Fear memory systems are part of a phylogenetically ancient and highly adaptive evolutionary strategy through which organisms are able to detect and avoid threat, and thus survive. Fear systems initially evolved in neural circuits far too simple to support anything resembling human consciousness, but in humans and other advanced species have developed complex communication and interaction with the conscious brain to influence subjective experience — notably that involving feelings of threat, anxiety, and worry [1].

Dysfunction in fear systems has been linked to several important clinical psychopathologies, including disorders of anxiety and depression, but is most archetypally associated with post-traumatic stress disorder (PTSD) [2]. These fear-based psychopathologies have a high prevalence and incidence in perioperative and intensive care and are thus of substantial clinical relevance to the anesthesiologist. Post-ICU PTSD and depression [3], PTSD-like symptomatology in post-surgical patients [4], PTSD after intraoperative awareness [5], and psychopathologies associated with pain syndromes [6] are all well-described clinical scenarios in which the anesthesiologist encounters the neurobiology of fear systems; indeed, ICU stay and intraoperative awareness are the medical events associated with the highest incidence of PTSD [7]. The neurohumoral stress response to surgery and critical illness includes the release of mediators known to potentiate fear learning, while many of the drugs used by anesthesiologists — both as sedative-hypnotics and as autonomic modifiers — act on targets that can plausibly modulate fear memory processes. Therefore, although the abundance of experimental neuroscience research in this field has mostly focused on understanding mechanisms and has not yet translated into large effectiveness studies or practice guidelines, the anesthesiologist should recognize the relevance of fear systems to perioperative psychological outcomes and understand current thinking on clinical parameters and decisions that may modify those outcomes.

THE ANATOMY AND NEUROBIOLOGY OF FEAR MEMORY SYSTEMS

The amygdala sits at the center of the brain’s fear memory system, which detects and responds to threats through associative learning [8]. Though often semantically confused with the conscious experience of feeling fearful, there is a distinct component of fear memories that do not require conscious awareness [1]. Although explicit (conscious) memory systems are also engaged during fear conditioning, amygdalar responses appear to be automatic [9], though subsequent processing may be modulated through interaction with prefrontal cortex [10]. The amygdala is organized into many nuclei (and further subnuclei) which can be delineated histologically [8], and more recently with high-resolution MRI [11]. In the prevailing (simplified) model of processing within the amygdala, the basolateral amygdala is important in fear acquisition, receiving sensory inputs and having output connections to the hippocampus and prefrontal cortex [12]. The central nucleus of the amygdala is responsible for amygdala-mediated behaviors [12]. Multiple neurotransmitters are involved within the amygdala and in its projections, including: GABA_A and glutamate [12], α₂-adrenergic [13], opioid [14], and serotonin [15]. This creates a complex milieu for manipulation by drugs, and anesthetic and analgesic drugs have obvious overlap in receptor pharmacology. In addition, functional dissection of each of these circuits using modern neuroscience techniques is an active area of investigation.

The amygdala can also significantly affect other brain systems and behaviors. Most notable is the amygdalar influence on the hippocampus [16], which acts to enhance explicit memory for items or events with high emotional valence (the affective quality of an emotion, which ranges from highly positive, though neutral, to highly negative) [17]. Complimentary systems in the amygdala, through interactions with other brain areas, also play a role in reward processing, which has more general influence on decision-making in goal-directed behaviors, through mechanisms of reinforcement learning [18, 19].

Anxiety-spectrum disorders, including panic and PTSD, are among the psychopathologies known to be associated with functional differences in the amygdala, in concert with changes in the anterior cingulate and insula [20]. Patients with anxiety disorder have overactivity of their amygdala, compared to healthy controls, when exposed to the same provocative stimuli [20], and decreased amygdala responsiveness is correlated to successful treatment of anxiety [21]. A hallmark of anxiety and related psychiatric diagnoses, such as PTSD, is more pronounced generalization of fear responses to distinct but related stimuli [22–24], which maps onto some of the clinical symptoms seen in these disorders, despite the absence of real danger.

KEY UNANSWERED QUESTIONS RELATED TO ANESTHETIC PRACTICE

The fundamental commonalities between the receptor systems and signaling mediators involved in fear memory and those targeted by anesthetic drugs give rise to several key questions of potential clinical significance:

• How sensitive are fear systems to the effects of anesthetic drugs? While fear memory shares certain mechanistic processes with explicit memory, there are sufficient differences to reasonably postulate that the two forms of memory may have different sensitivity to the effects of anesthetic drugs.

• Can fear systems function during anesthetic-induced unconsciousness? In distinction to explicit memory, components of fear memory do not require conscious experience. This raises the intriguing possibility that the removal of consciousness by anesthetic drugs may not be sufficient to suppress all fear processes.

• Can anesthetic drugs modulate the development of fear-based psychopathologies? A significant number of studies in the psychiatry and neuroscience literature have investigated the possibility that specifically-timed pharmacologic interventions may either prevent or accentuate the development of PTSD [25, 26]. Several anesthetic drugs act on plausible targets for these effects, raising the possibility that clinical decisions under the control of anesthesiologist may be able to influence the incidence of perioperative psychologic morbidity.

ANESTHETIC MODULATION OF FEAR ACQUISIITION

Most sedative-hypnotic agents have the potential to act on the brain’s fear memory system, as NMDA and GABA_A receptors are heavily involved, and central noradrenergic levels (through ascending arousal pathways via the locus coeruleus) also modulate the process [27]. Additionally, opioid receptors are prevalent in the amygdala [28]. The large body of behavioral work on acquisition of aversive fear conditioning in rodent models has shown that most commonly-used inhalational anesthetics and intravenous sedative-hypnotic agents (particularly propofol, midazolam, and ketamine) inhibit fear learning at higher doses [29]. Perhaps not surprisingly, dexmedetomidine appears to be an outlier, with no significant inhibition of fear learning [29], but it could enhance extinction of acquired fear [30]. Opioids likely also play a role, as fear conditioning is inhibited by morphine [31] and enhanced by opioid antagonists in rodents [32, 33]. In studies done with invasive neuronal recordings, signal transduction generally still occurs during implicit memory experimental paradigms. In one recent example with non-human primates, using a paradigm with aversive tones conditioned to odors, higher doses of midazolam and ketamine did not ablate aversive memory formation, with persistent neuronal activity in the amygdala and dorsal anterior cingulate [34]. This animal work generally supports the notion that hippocampal pathways for explicit memory formation (for neutral stimuli) are more sensitive to anesthetic inhibition, compared to amygdala-based learning circuits that are engaged in aversive stimulation paradigms [29]. Importantly, dose-response studies generally indicate that sub-hypnotic doses of anesthetics could enhance fear learning, a phenomenon with important potential clinical implications and the subject of ongoing research. How observed behavioral changes in lower mammals would translate to conscious experiences or subconscious psychological effects in patients receiving anesthesia or sedation is complex. For the remainder of this section, we will focus on research done in human subjects.

Though not a study of fear conditioning per se, seminal work on anesthetic modulation of memory in humans [35] demonstrated that sevoflurane (0.25% concentration) was able to inhibit explicit memory for emotionally-charged images, and this with associated with decreased amygdala to hippocampal functional connectivity (measured with PET imaging). Studies employing sedation during functional MRI have shown that bilateral amygdalar activation to emotional images persisted under dexmedetomidine [36] and propofol [37], with no drug-based differences in the subsequent memory performance advantage seen with emotionally-charged (vs. neutral) pictures. This, coupled with significant reductions in hippocampal activity under propofol [37], suggests that amygdala-based memory systems are more resistant to inhibition with at least some anesthetics in humans. However, in contrast, ketamine seems to inhibit amygdala activity while viewing emotional images [38] or experiencing painful stimulation [39]. Similarly, pain-related activity in the amygdala was not demonstrated under sedation with midazolam in the context of noxious stimulation [39]. In sum, this work suggests that mechanistically-distinct sedative-hypnotic agents differentially affect localized brain activity in fear learning centers with a range of stimuli.

In addition to localized inhibition of fear centers, disruption of functional connectivity between the amygdala and higher brain areas may reflect a possible means by which cortical integration for fear learning may be disrupted. However, much anesthetic connectivity research is more focused on broader network-level brain changes and has employed methodology that makes it challenging to resolve amygdalar connectivity changes. Thus, a relatively small number of studies are reviewed here. Ketamine has been shown to reduce amygdalar connectivity in the resting-state [40], as well during periodic painful stimulation [39]. Morphine decreases functional connectivity between the amygdala and other areas important in fear learning, such as the dorsal anterior cingulate cortex [41]. Complimentary analyses for low-dose midazolam during pain have shown increased functional connectivity between the amygdala and prefrontal cortex [39], from which one might infer a modulation of fear learning capacity that is specific to both anesthetic agent and dose.

Classical fear-conditioning paradigms have not been studied using patients in the clinical setting, likely because of ethical and technical challenges to employing such a paradigm [42]. Based on work employing lexical tasks and looking for perceptual priming [43], implicit memory of this type may occur at lighter planes of sedation, but are ablated at higher doses [43–45]. However, there are replication failures in this space [46], and the clinical or psychological relevance of these findings are unclear. One such study [47] employed recordings of words with different emotional valences played during laparoscopic cholecystectomies under isoflurane anesthesia. The investigators found a small subsequent participant response time advantage for emotionally negative words, and this correlated to EEG spectral changes at the time of exposure [47].

ANESTHETIC MODULATION OF FEAR-BASED PATHOLOGIES

How anesthetic and analgesic agents might be best employed clinically to prevent PTSD and other psychiatric sequalae during surgery or following traumatic experiences is an important area of investigation, and not currently fully understood. Animal research supportive of this goal often focuses on fear consolidation or extinction, looking for experimental methods to reduce fear-related behaviors after fear-conditioning. As one example, administration of morphine after a conditioning paradigm impairs fear consolidation [33] and also fear generalization [48]. Replications of this effect have elucidated a complex interaction with opioid signaling and the adrenal axis [49]. Consistent with an important role of the adrenergic system, dexmedetomidine may protect against the development of PTSD-like behaviors after aversive conditioning in rodents [50]. However, faciliatory effects for PTSD behaviors have been demonstrated for propofol and ketamine in pre-clinical work [50].

Clinical studies examining the relationship between anesthetics and analgesics and the development of PTSD are generally retrospective correlation studies. PTSD is relatively common in survivors of critical illness, and use of several agents have been associated with this neuropsychiatric sequala, including benzodiazepines [3, 51], and propofol [52]. One randomized study in critically-ill patients requiring mechanical ventilation found no difference in PTSD between groups receiving no sedation versus sedation, using a regimen of propofol followed by midazolam after 48 hours [53].

An important thread of clinical research on PTSD comes from studies of different agents used to manage pain or provide sedation following traumatic injuries. In a small cohort of traumatically injured patients, acute pain predicted PTSD development, and after adjusting for injury severity, there was a correlation between higher doses of morphine and decreased PTSD severity [54]. There is also a lower PTSD rate associated with morphine use after combat trauma [55]. Interestingly, there appears to be a higher PTSD rate when propofol is used for ICU sedation in the first 72 hours following motor vehicle accident [56]. This may seem counter to peri-operative clinical experience, in which propofol can achieve anxiolysis in anxiety-provoking (but non-life threatening) situations [57, 58]; however, one plausible explanation is the phenomenon of retrograde facilitation, in which drugs that inhibit memory consolidation (including GABA_A agonists) will have an augmenting effect for memories formed immediately prior to administration [59]. One randomized study, comparing propofol to sevoflurane for general anesthesia in emergency trauma surgery, showed a two-fold higher incidence of PTSD with propofol [60], suggesting a possible neuropsychiatric advantage for volatile agents following a trauma. Perioperative midazolam use has been studied in burn-injured soldiers, with no association (positive or negative) to subsequent PTSD [61].

Ketamine has become of intense interest in psychiatry for use in treatment-resistant depression [62], though there is less evidence that perioperative use improves mood [63]. Ketamine use in treating PTSD is controversial [64], and may work better in combination with other non-pharmacologic therapies [65]. However, ketamine can reduce PTSD symptom severity [66], including when treating concomitant depression [67]. In contrast, use of ketamine during initial trauma care has shown an association with PTSD [68, 69]. However, perioperative ketamine used for burn surgery [70] and other combat injuries [71] has not been correlated with an increased (nor decreased) incidence of PTSD.

Prevention of adverse psychiatric sequalae is particularly challenging in the setting of unintended awareness with recall memory during general anesthesia, typically during surgery. This is due to the relative rarity and heterogeneity in the conditions under which these events occur, including patient comorbidities, anesthetic technique, and pre-existing psychiatric illness. The importance of the latter is highlighted by the surprisingly high incidence of PTSD in surgical patients who did not have anesthetic awareness [72]. Feelings of immobility and helplessness are risk factors for subsequent psychological sequalae [73]. This suggests that neuromuscular blockers, in addition to being associated with higher incidence of awareness [74], are also more likely to result in PTSD following such an event. It might be surprising that pain is not a common complaint following anesthetic awareness [75], despite consciousness and pain (with subsequent amnesia) occurring more often than previously thought during routine anesthesia [76].

Taken together, anesthetic and analgesic agents have the capacity to modulate fear memory formation during and after severely aversive conditions, and this could have important neuropsychiatric implications. Patients with preexisting psychiatric comorbidities are likely at more risk, and extra precautions may be warranted in these cases. At present, the literature is too disparate to make strong specific recommendations for the prevention of PTSD during emergency trauma care, anesthesia or ICU sedation. However, minimizing the use of neuromuscular blocking drugs and providing adequate analgesia are reasonable recommendations. Further, the use of a multi-modal sedative-hypnotic strategy during painful procedures and critical illness may be more likely to prevent adverse psychiatric sequalae than approaches employing individual drugs.

CONCLUSION

Dysfunction of fear memory systems underlie a cluster of clinically significant and highly prevalent psychological morbidities seen in perioperative and critical care patients. The collective human and non-human literature suggests that both the physiologic response to surgery and critical illness and the drugs used in anesthetic care have definite but complex effects on fear memory systems. While research continues, the complexity of the effects and relative paucity of human studies mean that there is presently insufficient evidence to provide clinical practice recommendations. There is some basis to support the hypothesis that unconscious fear-based memory systems may be more refractory to suppression by some sedative-hypnotic anesthetic drugs than are conscious memory systems; the boundaries and dose-dependence of these differences are not yet defined. There is some suggestion that the actions of opiates are effective at suppressing fear memory processes, and that higher doses of opiates may reduce the incidence and severity of PTSD following trauma. Some studies of sedation used in the post-trauma setting suggest that use of the highly GABAergic-selective sedatives (notably propofol) may be less effective at preventing the development of PTSD than alternatives. Despite the clear benefits of ketamine in treating depression, studies of its effects on fear-based memory processes and related psychological outcomes appear more complex and inconsistent. Despite the current lack of clarity, anesthesiologists should recognize the potential for meaningful clinical benefit to emerge from this sphere of research and maintain an awareness of the ongoing literature.

KEY POINTS.

• Fear memory systems, involved in the development of PTSD, are activated by aversive experiences that include physical trauma, critical illness, and surgery.

• Nearly all anesthetic and analgesic drugs have some effect on fear memory systems, but the interactions are complex and not completely understood.

• How specific agents can be best used to prevent PTSD in clinical anesthesiology practice remains an area for further study.