Noradrenergic Mechanisms in Fentanyl-Mediated Rapid Death Explain Failure of Naloxone in the Opioid Crisis

Incidence of Fentanyl-Induced Rigidity in the Controlled Surgical Setting.

Current anesthesia literature suggests that WCS is an uncommon event (Çoruh et al., 2013). However, earlier clinical experiences with F/FAs in anesthesia were quite the contrary (Grell et al., 1970). When fentanyl was first introduced to the United States market, as previously mentioned, it revolutionized surgical anesthesia, but only after anesthesiologists learned the nuances of its administration. Early, large-scale clinical studies were the first to describe chest wall rigidity and airway compromise/laryngospasm induced by F/FAs and became a guiding principle in the training of anesthesiologists. The end result is that the incidence of WCS today is quite rare or taken for granted because anesthesiologists have become skilled at compensating for this risk, not unlike early experiences with morphine-induced histamine release and hypotension (Grell et al., 1970).

In 1970, Grell et al. (1970) published a landmark study that looked at the use of fentanyl as an anesthetic agent. Fentanyl was administered to a wide range of 500 surgical subjects (age, procedure type, and comorbidity classification) via intravenous infusion at 30 µg/min or a single bolus dose of 3.3 µg/kg. Of 500 subjects, 499 developed some degree of muscle rigidity, which presented within 60–90 seconds at a minimum dose of 250 µg or ∼3 µg/kg and persisted for 8–15 minutes. The conclusion from the Grell et al. (1970) study was that: “The onset and degree of muscle rigidity was directly related to the size of the dose and rapidity of injection.” A significant limitation of the study was that the degree of rigidity was not quantified, since later studies by Streisand et al. (1993) and Bennett et al. (1997) demonstrated that muscle rigidity occurs with significantly higher doses of F/FAs (e.g., 15 µg/kg for fentanyl and the equivalent of 30 µg/kg of fentanyl with sufentanil 3 µg/kg). In fact, Streisand et al. (1993) only demonstrated rigidity and upper airway symptoms (e.g., difficulty with mask-assisted ventilation) in ∼50% of study subjects with a total dose of 15 µg/kg, and the latter study by Bennett et al. (1997) visually demonstrated vocal cord closure in 28 of 30 subjects with a sufentanil 3 µg/kg bolus over 2 minutes. While these doses of F/FAs would be considered high for routine cases in healthy individuals having elective surgeries such as hernia repairs and cholecystectomies, they have been commonly and safely used in patients with severe cardiovascular disease for procedures such as heart valve replacement and coronary artery bypass grafting (Streisand et al., 1993; Bennett et al., 1997). However, what should remain clear is that the size of the dose becomes relative to the speed of injection, which remains a significant factor in the incidence of complications from F/FA overdose in the current opioid crisis. In the controlled setting of clinical practice, low doses of F/FAs rarely cause WCS.

Fentanyl-Induced Rigidity and Loss of Consciousness.

Grell et al. (1970) demonstrated that muscle rigidity induced by fentanyl has an exceedingly frequent incidence following rapid intravenous injection and occurs at a minimum threshold dose of approximately 250 mcg. However, the study did not measure the concurrent incidence of loss of consciousness with muscle rigidity. In an effort to understand the CNS mechanism underlying WCS, Streisand et al. (1993) looked at the incidence and duration of rigidity and loss of consciousness associated with plasma concentrations. Fentanyl was administered to 12 healthy volunteer subjects at an infusion rate of 150 mcg/min to a total dose of 15 mcg/kg (e.g., ∼1 mg/70 kg). The result was muscle rigidity that occurred in six out of 12 subjects (50%), starting at a plasma fentanyl concentration of 21.5 ± 4.4 ng/ml (range 16–28 ng/ml), which resolved at 6.9 ± 1.5 ng/ml (range 5.2–8.7 ng/ml). Rigidity started at 3 ± 0.9 minutes (range 1–4 minutes) after the peak plasma fentanyl concentration, and lasted for 11.5 ± 5.8 minutes (range 7–23 minutes). All subjects who developed rigidity were apneic, unresponsive, and had no recall of commands to breathe or positive pressure ventilation. Essentially, WCS coincided with unconsciousness of subjects in the experiment. The study concluded, “These findings support the hypothesis that unconsciousness occurs in the unstimulated subject during fentanyl-induced apnea and rigidity” (Streisand et al., 1993).

The results of this study are significant and should be emphasized for several reasons: the study correlates the occurrence of WCS with a loss of consciousness and demonstrates that the plasma concentrations of fentanyl are not significantly different between those subjects with rigidity and those subjects without rigidity. Interestingly, there were no significant changes in recorded hemodynamics among the subjects, indicating that the alteration in level of consciousness in these subjects may have been more of a centrally mediated effect rather than a cerebral perfusion defect or cardiovascular event. A rapid loss of consciousness associated with the FIRMR in this study could have several physiologic explanations. A sudden hemodynamic shift such as a fall in cardiac output or cerebral perfusion pressure can easily cause alterations or loss of consciousness. However, Streisand et al. (1993) found no fundamental differences in hemodynamics between those subjects who became unconscious and those that did not. This suggests a centrally mediated effect resulting from rapid disruption of thalamo-cortical connectivity, as has been previously demonstrated by electroencephalogram monitoring studies in human subjects who lost consciousness after fentanyl administration (Uhrig et al., 2014). The rapid loss of consciousness in the aforementioned studies is consistent with eyewitness and survivor accounts of F/FA overdose and WCS (Somerville et al., 2017).

Respiratory Depression or WCS.

Large-scale human trials with fentanyl have noted that while it may take 7–9 minutes for pulmonary/respiratory mechanics to decline by 50% (e.g., respiratory rate, tidal volume, and minute volume), muscle rigidity routinely occurs 60–90 seconds at a minimum dose of 250 mcg in adults (Grell et al., 1970). Other smaller scale studies have clearly demonstrated that severe rigidity occurs prior to and more rapidly than significant respiratory depression in study subjects receiving high-dose fentanyl infusions of 150 mcg/min up to 15 mcg/kg (Grell et al., 1970). Morphine and heroin by comparison are not known to cause airway compromise due to rigidity (Hamilton and Cullen, 1953; Jenkins et al., 1994) and have a significantly greater time to CNS entry and onset (e.g., 19 minutes for morphine to reach 80% of peak effect) (Owen et al., 1983).

In summary, neither heroin nor morphine induces significant airway compromise in humans due to rigidity as a single agent (Sokoll et al., 1972; Freund et al., 1973). By comparison, F/FAs as single agents produce muscle rigidity and airway compromise rapidly, and these effects are consistently correlated with dose and rate of administration.

GABA Interneurons.

GABA interneurons are part of an inhibitory network throughout the CNS and are particularly abundant in the LC, where norepinephrine release from the presynaptic terminal is inhibited by GABA efferent signaling (Aston-Jones et al., 2004; Jin et al., 2016; Breton-Provencher and Sur, 2019). Importantly, the LC maintains basal skeletal muscle tone in the axial skeleton and torso via noradrenergic activation of spinal motor neurons. Inhibition of GABA interneurons allows for increases in skeletal muscle tone through increased noradrenergic activity in the LC (Nakamura et al., 2002). Conversely, activation of GABA interneurons can cause a decrease in noradrenergic tone in the adjacent subcoeruleus, as occurs in rapid eye movement sleep or cataplexy, demonstrating that the loss of noradrenergic tone inhibits contractility of all major muscle groups except muscles of respiration and those involved in rapid eye movement (Fraigne et al., 2015). At the level of the GABA interneuron, fentanyl and morphine stimulate μ-opioid receptors and inhibit GABAergic signaling and neurotransmission in several places in the brain, including the periaqueductal gray (Vaughan and Christie, 1997), hippocampus (Alreja et al., 2000), and dorsal raphe nucleus (Jolas and Aghajanian, 1997). Although, there is no description of fentanyl-induced GABA interneuron inhibition in the LC itself, reports demonstrate fentanyl binding to GABA interneuron μ-opioid receptors and inhibition of neurotransmitter output (Nakamura et al., 2002; Griffioen et al., 2004). Recent evidence also indicates the presence of these GABA interneurons in the LC (Breton-Provencher and Sur, 2019). The binding of fentanyl or morphine to μ-opioid receptors located on GABA interneurons in the LC should inhibit interneuron activation and paradoxically increase the release of norepinephrine at the presynaptic terminal, activating spinal motor neurons in the ventral horn of the spinal cord, and followed by postganglionic fibers terminating in skeletal muscle (Lui et al., 1989). What has not been clearly established is the degree to which μ-opioid receptor binding of GABA interneurons supercede or affect the hyperpolarization of LC neuronal cells by μ-opioid receptor agonists (Chiu et al., 1993). However, what is known from electron microscopy and immunocytochemical studies of the LC is that a significant percentage of cells within the dense body of the LC stain positive for GABA, indicating the presence of parasynaptic GABA interneurons and the potential for integrating LC function (Aston-Jones et al., 2004). In addition, electron microscopy studies suggest that μ-opioid ligands act mainly parasynaptically on LC neurons, with approximately 5% of labeled μ receptors associated with axoglial interfaces, indicating that a minor action of μ opioids in the locus may be presynaptic (Moyse et al., 1997).

In a series of experiments using a spinal cord ischemia model in rats, Nakamura et al. (2002) used intrathecal morphine to cause spastic paraparesis. For a reversible injury model, anesthetized rats had intrathecal catheters placed along with an arterial occlusion catheter used to generate spinal cord ischemia on inflation. Motor function was quantified by ambulation after reperfusion with a motor deficit index scale. The presence of spasticity or flaccidity was determined by the presence of an exaggerated flexion response to a pinch of the hind paw. The study used the GABA_A agonist muscimol (intrathecal dose of 0.3, 1, or 3 µg per 10 µl) and the GABA_B agonist baclofen (intrathecal dose of 1, 3, or 10 µg per 10 µl) to demonstrate that GABA agonists attenuate the spastic paraparesis induced by intrathecal morphine (intrathecal dose of 30 mcg) after a noninjurious interval of spinal cord ischemia. The agonists proved ineffective. However, the GABA antagonists bicuculline (intrathecal dose of 4 mcg) and 5-AA (intrathecal dose of 70 mcg) produced a dose-dependent increase in the motor deficit index similar to morphine (intrathecal dose of 10 mcg), and the combination of morphine and GABA antagonists (bicuculline or 5-AA) also produced a further increase in the motor deficit index. In addition, morphine and GABA antagonists had a synergistic, not an additive, interaction in producing spastic paraparesis. Thus, there are μ-opioid receptors on GABA interneurons in the spinal cord; and their stimulation inhibits GABA interneuron signaling, resulting in increased muscle contractility (Nakamura et al., 2002).

Similarly, fentanyl binds μ-opioid receptors on GABA interneurons and inhibits GABA interneuron afferents (Griffioen et al., 2004). In an experiment using fluorescent tracers, infrared cameras, and whole-cell patch-clamp electrophysiology, parasympathetic cardiac vagal neurons were identified and stimulated with fentanyl. This demonstrated the inhibition of GABA neurotransmission to cardiac vagal neurons, resulting in increased parasympathetic output to the heart via the vagus nerve, which may be the underlying mechanism of opioid-induced bradycardia (Griffioen et al., 2004). Fentanyl was tested at concentrations of 0.1, 0.5, 1 and 5 μM, and the amplitude was significantly reduced by 25%–48% with maximal inhibition at 0.1 μM or ∼50 mcg/kg. The results indicate that fentanyl reduces the frequency and amplitude of GABAergic inhibitory postsynaptic currents at pre- and postsynaptic sites in cardiac vagal neurons and the effects are abolished by the μ-opioid receptor antagonist D-Phe-Cys-Tyr-D-Trp-Orn-Thr-Pen-Thr-NH2 (Chieng et al., 1996; Griffioen et al., 2004). Thus, fentanyl inhibits GABA interneurons in the nucleus ambiguus by binding μ-opioid receptors on GABA interneurons innervating cardiac vagal neurons (Griffioen et al., 2004).

Fentanyl Effects on Catecholamines In Vitro and In Vivo.

As previously described, fentanyl differs from morphine since the former has considerably increased lipid solubility, greater selectivity for μ-opioid receptors over κ and δ, but has similar binding affinity as morphine on μ-opioid receptors. In addition, fentanyl, but not morphine, is capable of acting as a norepinephrine reuptake inhibitor (Mather, 1983; Atcheson et al., 1993). High-dose fentanyl (e.g., 1–100 mM) blocks reuptake of [³H]norepinephrine into human neuroblastoma and rat pheochromocytoma cells (Atcheson et al., 1993). Morphine exhibited no such effect at any concentration tested. It is also important to point out that the dose range of fentanyl used in these experiments may have physiologic relevance in humans based on calculated brain lipid opioid concentrations (fentanyl) calculated from plasma opiate concentrations, where 6.9 ng/ml of plasma fentanyl corresponds to a concentration of 2.96 × 10⁻³ M in brain lipid (Freund et al., 1973; Stone and DiFazio, 1988; Hustveit and Setekleiv, 1993; Yamanoue et al., 1993; Sohn et al., 2005).

Notably, this effect on human neuroblastoma and rat pheochromocytoma cells was not antagonized by naloxone, suggesting the site of action for fentanyl’s effect does not involve μ receptors (Atcheson et al., 1993). Subsequent studies with similar models of human SH-SY5Y neuroblastoma cells have demonstrated the presence of endogenous muscarinic (M) 3 receptor sites on these cells, where fentanyl might act as a potential antagonist (Yamanoue et al., 1993; Bundey and Nahorski, 2001). Animal studies have also demonstrated evidence of both α₂- and α₁-adrenergic receptors (specifically, the α_1D subtype) on neuroblastoma cells where fentanyl has α₁-subtype antagonist effects (Baron and Siegel, 1989; Wenner et al., 2016). This study was performed to better understand the relationship between neuronal amine uptake and analgesia and was not analyzed or performed in the context of FIMR or WCS. However, the study demonstrates a key difference between fentanyl and morphine, implicating a mechanism for fentanyl-enhanced noradrenergic activity outside of μ-opioid receptor binding. This appears to be consistent with fentanyl-induced noradrenergic activity in the LC leading to FIMR in the animal model and has been an ongoing unexplained phenomenon, which contradicts the well-established concept from in vitro rat brain slice voltammetry studies that opiates directly inhibit norepinephrine release from LC neurons. In this in vitro voltammetry model, opioids bind μ receptors on LC neurons, hyperpolarize neuronal cell membranes, and result in a subsequent decrease in cell firing and norepinephrine release (Aghajanian, 1982; North and Williams, 1985; Lui et al., 1989; Christie, 1991; Chiu et al., 1993).

Conversely, in vivo animal (sheep) and human studies using intravenous morphine (1 mg/kg), fentanyl (10 mcg/kg), and indwelling intrathecal spinal catheters for serial cerebrospinal fluid (CSF) sampling, have shown increases in CSF levels of norepinephrine and acetylcholine three to four times above baseline levels within 10 minutes of injection with morphine (1 mg/kg), suggesting that fentanyl (e.g., ∼680 times more lipophilic than morphine) or a more lipophilic analog would have a similar centralized effect in a fraction of this time (Bouaziz et al., 1996). Additionally, the rate of infusion in this series of experiments, particularly for fentanyl (e.g., 10 mcg/kg, i.v., infused over 10 minutes), could explain the slow onset of catecholamine rise and is juxtaposed to real-world rapid injection practices commonly seen in F/FA-related overdose. Also noteworthy in this study (Fig. 1) is that the first measurement occurred only 10 minutes after injection and showed a >200% increase in norepinephrine levels, suggesting the possibility that a significant increase may have gone unmeasured prior to the first lumbar CSF sample obtained, and where travel time from brain to distal spinal cord and CSF production must be considered to accurately assess neurotransmitter concentration (Bouaziz et al., 1996). Similar considerations must be made regarding lipophilicity and CSF travel time for the single human subject in the study where drug enters the ventricles of the brain from the vascular space and must travel distally to the lumbar region (e.g., CSF volume in adult human of 1.5–2 ml/kg, production rate of 25 ml/h, and flow rate of 0.3 ml/min) (Bouaziz et al., 1996; Wright et al., 2012; Brinker et al., 2014). In humans, the clinical onset time for F/FA CNS effects is well documented for occurring in a single circulation time with intravenous injection (e.g., 60–90 seconds), suggesting rapid distribution in CSF with CNS entry and transport beginning in ventricles after systemic arterial delivery (McClain and Hug, 1980; Mather, 1983; Bouaziz et al., 1996). More specifically and related to the significance of increased norepinephrine and acetylcholine with opioids, iontophoretic studies of the LC demonstrate that both muscarinic and nicotinic receptors are present in rat and human LC, acetylcholine (ACh) muscarinic receptor (M2 subtype) binding increases norepinephrine release in LC, and the LC itself is rich in acetylcholinesterase (Albanese and Butcher, 1980; Egan et al., 1983; Egan and North, 1985, 1986). The LC receives cholinergic input from cells in the neighboring pontine reticular formation where ACh release may be stimulated by opioids and may involve GABA interneurons or M3 subtype binding in the case of fentanyl (McNaughton and Mason, 1980; Yamanoue et al., 1993).

The contradictory results between the in vitro voltammetry studies and in vivo studies are explained to some degree by this existing literature and help to fulfill some of the essential prerequisites for the activation of skeletal muscle (e.g., chest wall and laryngeal muscles) in WCS via the release of ACh and subsequent release of norepinephrine in the LC by F/FAs (Rasmussen and Jacobs, 1985; Gaumann et al., 1988; Bouaziz et al., 1996). What should be evident at this point is that F/FAs have off-target binding sites (muscarinic and α₁-adrenergic receptors) that may function independently of μ-opioid receptor activation, or alternatively are minimally affected by μ-opioid receptor antagonism, once these muscarinic or α-adrenergic receptors are activated in a secondary capacity from μ-opioid receptor agonism. In the former case, we would expect there to be no effects of naloxone since it has no direct binding capacity at muscarinic or α-adrenergic receptors (Jarrott et al., 1979; Huchet et al., 1986; Hustveit and Setekleiv, 1993); in the latter case, only a partial effect with high-dose naloxone would be expected, as demonstrated by Willette et al. (1982, 1987). In the latter case, where naloxone has a partial effect, the efficacy of naloxone would depend on the balance of these offsite receptors with respect to their dominance, or lack thereof, for μ-opioid receptor effects.

To summarize the case of catecholamine release with F/FAs, there are several mechanisms that could have synergistic effects and increase the plausibility of both cholinergic and noradrenergic mechanisms underlying WCS. Although fentanyl and F/FAs are clearly not agonists at sympathetic receptors, they may increase the release of norepinephrine at the LC via GABA interneuron inhibition, increase the release of ACh for LC cholinergic receptor stimulation, and act as norepinephrine reuptake inhibitors in the LC at the presynaptic terminal. These combined mechanisms would allow for significant sympathetic effects for F/FAs without being direct agonists of α-adrenergic receptors or directly increasing the firing of LC neurons (Sriraman et al., 2007). Additionally, as mentioned previously, Sohn et al. (2005) demonstrated that fentanyl binds α₁-adrenergic receptor subtypes with a rank order of B > A >>> D, which may facilitate selective subtype binding of α₁-adrenergic receptors by norepinephrine and serve to concentrate its effects on the α_1D-receptor subtype. F/FA facilitation of subtype binding and its potential for significant effects on cardiovascular and hepatic systems will be discussed in Human Physiology.

Fentanyl Effects on Catecholaminergic Activity in Human Subjects.

Several clinical studies further support the hypothesis that WCS is mediated by noradrenergic function and that fentanyl increases norepinephrine release in the CNS and periphery. Hicks et al. (1981) studied 10 adult patients with ischemic heart disease who were induced with a range of high-dose fentanyl for cardiac surgery. Plasma samples of catecholamines were assayed radio-enzymatically for norepinephrine, epinephrine, and dopamine. Norepinephrine levels were significantly elevated after 15 mcg/kg of fentanyl, remained elevated after 30 mcg/kg of fentanyl, but declined toward baseline values at the 50 mcg/kg dose. Also notable was a dose-dependent decrease in left ventricular function at doses greater than 30 mcg/kg of fentanyl. The catecholamine response in the study was interpreted as either a dose- or time-related effect of fentanyl or the stress of induction of anesthesia, rather than the alternative that high-dose fentanyl elevates norepinephrine levels. The study did not monitor or correlate FIRMR or laryngospasm, did not control for the stress of induction, and also did not offer an explanation for the seemingly biphasic response of norepinephrine levels to increasing doses of fentanyl (Hicks et al., 1981).

In a study of 50 healthy adult females having total abdominal hysterectomy under general anesthesia, patients received either 2 or 4 mcg/kg of fentanyl 5 minutes before or 5 minutes after surgical incision with serial catecholamines measured 1 minute prior to opioid dose and serially thereafter for 20 minutes. Serum epinephrine levels decreased significantly at several time points and more so with the higher dose; however, plasma norepinephrine was elevated after fentanyl administration irrespective of time to induction or dose levels and was not attenuated at higher doses (Hori and Nagasaka, 2002). Unlike the study described previously (Hicks et al., 1981), this latter study controlled for the stress of anesthesia induction and incision and similarly yielded a dose-dependent increase in norepinephrine levels to fentanyl at lower doses. Although these two studies were not monitoring for WCS, they each demonstrated that norepinephrine levels were elevated by fentanyl, consistent with the results of in vivo experiments demonstrating that fentanyl facilitates Ach-mediated release of norepinephrine and reinforces a plausible underlying mechanism for WCS (Payne et al., 1986; Bouaziz et al., 1996).

Catecholamine Release and Laryngospasm.

Animal studies suggest that FIMR muscle rigidity in the chest wall and limbs is mediated by noradrenergic receptors in the locus coeruleus, while vocal cord closure/laryngospasm in the intrinsic muscles of the larynx associated with high-dose F/FAs is cholinergically driven and involves the balance of sympathetic and parasympathetic innervation to the intrinsic muscles of the larynx (Willette et al., 1982, 1987; Lui et al., 1989; Lalley, 2003). Literature from human studies indirectly supports the possibility that similar mechanisms control vocal cord closure/laryngospasm (Yasuda et al., 1978; Scamman, 1983; Streisand et al., 1993; Bouaziz et al., 1996; Bennett et al., 1997; Lin et al., 2004; Fodale et al., 2005; Varney et al., 2009; Horng et al., 2010; El Baissari et al., 2014). However, to date there has been no direct demonstration in humans or animals that fentanyl facilitates noradrenergic or cholinergic signaling in the laryngospasm component of WCS. The data presented subsequently examine the existing support for that hypothesis.

The most suggestive clinical study to date was performed by Yasuda et al. (1978), in which the authors measured and compared tracheal pressure increases induced by either morphine (0.5 mg/kg) or fentanyl (0.006 mg/kg) in healthy, adult human subjects (Yasuda et al., 1978). The study involved endotracheal intubation of the patients with a fluid-filled pilot balloon at the level of the vocal cords, which could be used to accurately measure tracheal pressure in real time as equianalgesic doses of the opioids were administered. Both opioids increased tracheal pressure significantly from baseline (morphine increased by 27% and fentanyl increased by 44% above control). Interestingly, morphine demonstrated an initial 8% decrease in tracheal pressure that lasted ∼2 minutes, whereas fentanyl had an ∼14% increase in tracheal pressure within 30 seconds of administration. Also of significance, the tracheal constrictive effect of fentanyl was blocked by pretreatment with the neuroleptic agent droperidol (0.25 mg/kg) and contrasted with the tracheal effects of morphine, which were blocked by pretreatment with the anticholinergic muscarinic antagonist atropine (0.5 mg). Droperidol has α₁-adrenergic antagonist effects and significant anticholinergic effects that are less defined (Stoelting et al., 1975; Boros et al., 1984; Nashan et al., 1984). Atropine had no effect on fentanyl action, indicating that each of these opioids has a unique mechanism of action on the intrinsic laryngeal muscles and smooth muscle of the trachea (Yasuda et al., 1978). A possible confounding variable in the study is that 70% N₂O was used as part of the anesthetic, which is known to cause severe muscle rigidity and airway compromise when combined with morphine (Sokoll et al., 1972; Freund et al., 1973). This confounder for morphine may indicate a further distinction from the underlying mechanisms of F/FAs.

Freund et al. (1973) demonstrated in adult human subjects that moderate- to high-dose morphine (e.g., 2 mg/kg, i.v.) can cause some degree of abdominal wall rigidity and a decrease in thoracic wall compliance as measured by EMG and airway pressures, but the addition of 70% N₂O/30% O₂ in spontaneously ventilating patients caused severe board-like rigidity with immediate airway compromise (could not be mask ventilated) (Freund et al., 1973). Six of the nine adult subjects in the study required endotracheal intubation and all subjects had an 8-fold increase in abdominal wall EMG activity within 3 minutes of N₂O administration (Freund et al., 1973). The muscle rigidity was completely reversed in six of the six intubated patients with 3 mg/kg of intravenous thiopental. It is notable that thiopental is a barbiturate sedative hypnotic agent that is well-known for its supraspinal effects, indicating a central origin for the rigidity. Additionally, spinal monosynaptic reflexes in the study subjects showed significant decrease in amplitude, indicating that the abdominal rigidity and airway compromise induced by morphine and N₂O was unrelated to spinal motoneuron excitability (Freund et al., 1973). Regarding N₂O effects, although the authors could offer no explanation at the time for the dramatic increase in muscle rigidity and airway effects, subsequent research on the molecular mechanisms of N₂O in a rat model have shown significant increases in norepinephrine release (>50% from baseline) from the LC triggered by 60%–70% N₂O (Fukuhara et al., 1998). In the rat model of FIMR demonstrated by Lui et al. (1989), the presumed increase in norepinephrine release from the LC induced by fentanyl caused similar rigidity effects as when N₂O was combined with morphine in humans (Freund, 1974). Although norepinephrine levels were not measured directly by Lui et al. (1989), they did not use N₂O as part of the anesthetic for these experiments. The inhibition of the muscle rigidity induced by F/FAs (fentanyl and alfentanil) with the α₁-adrenergic antagonist prazosin and the α₂-adrenergic agonist clonidine strongly suggests the involvement of elevated levels of norepinephrine in the LC as a significant component of muscle rigidity induced by F/FAs.

Perhaps the most significant observation to be made from this study in the context of F/FAs is that it helps to distinguish the effects of fentanyl and morphine from each other. The study clearly demonstrates that morphine itself has no significant impact on upper airway function. Specifically, it does not cause either laryngospasm or vocal cord closure on its own, but requires N₂O for the upper airway to become compromised. Conversely, fentanyl and its analogs can cause laryngospasm/vocal cord closure as single agents and seem to be significantly impacted by α-adrenergic receptor activity and norepinephrine levels in the LC (Scamman, 1983; Lui et al., 1989; Bennett et al., 1997). The work by Freund et al. (1973) establishes supportive evidence that the laryngospasm component of WCS appears to be unique to F/FAs, which was later supported by Scamman (1983).

A more recent case study examined the occurrence of severe vocal cord closure/laryngospasm with the use of tramadol, a relatively weak μ-opioid receptor agonist that also acts as a monoamine reuptake inhibitor. The case report (Fodale et al., 2005) described an elderly male with mild age-related compromise of hepatic metabolism, who following an intravenous tramadol dose (e.g., 2 mg/kg bolus followed by 0.4 mg/kg per hour) experienced a severe episode of vocal cord laryngospasm. Fiberoptic bronchoscopy of the glottis documented vocal cord closure without edema, indicating laryngospasm presumably induced by the opioid since he was not taking any other medications known to induce laryngospasm. The episode gradually resolved over ∼25 minutes after 1 mg of intravenous naloxone was administered (Fodale et al., 2005). This study is significant in demonstrating that tramadol, a weak μ agonist (e.g., morphine is 6000 times more potent), produced fentanyl-like effects, possibly indicating that both drugs act as monoamine reuptake inhibitors in the LC (Lui et al., 1989; Atcheson et al., 1993; Halfpenny et al., 1999; Fodale et al., 2005).

Supporting this hypothesis, in rat brain slice real-time voltammetry of norepinephrine efflux and reuptake, the (−) tramadol enantiomer induced significant efflux and blocked norepinephrine reuptake in the LC (Halfpenny et al., 1999). It is also significant that the short half-life of naloxone (e.g., 30–60 minutes vs. tramadol half-life of 360 minutes) and its rapid egress from CNS after entry due to its high lipophilicity (e.g., 15-fold greater than morphine), indicates that it is usually removed from the site of action in the CNS within 10 minutes (Dean et al., 2009). In fact, 1 mg of naloxone (15 mcg/kg dose) in the case subject (∼65 kg) would be expected to produce a ∼50% occupancy of μ receptors, which makes it unlikely that the reversal of the laryngospasm was due to naloxone (Clarke et al., 2005).‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬‬

Vocal cord dysfunction (VCD) is a pulmonary condition that presents with airway compromise similar to asthma, but is unrelated and unresponsive to asthma therapeutics (e.g., bronchodilators, nebulizers, epinephrine, and inhaled steroids). VCD is an uncommon condition characterized by paradoxical vocal fold motion during inspiration, which leads to stridor, tachypnea, chest tightness, and a feeling of suffocation or air hunger (Mikita and Mikita, 2006; Morris and Christopher, 2010; see “It Is Hard to Breathe when Your Vocal Cords Don’t Work!” (https://www.emrap.org/episode/february2015/airwaycorner)]. VCD is most commonly observed among conditioned athletes, and is thought to emerge as a symptom of autonomic dysfunction that imbalances vagal innervation at intrinsic muscles of the larynx during high aerobic stress. VCD has responded inconsistently to various therapeutics including cholinergic, anticholinergic, and noradrenergic agents (Morris et al., 1999). Success with therapeutic strategies for reducing sympathetic or adrenergic innervation to the intrinsic muscles of the larynx for treatment of VCD reinforces the syndrome as a sympathetic phenomenon. Varney et al. (2009) used low-dose amitriptyline (e.g., 20 mg), the sedating tricyclic reuptake inhibitor, to gradually treat VCD over several weeks. As the sedative effects increased over time for the study participants, sleep improved and symptoms of VCD improved in ∼88% of the 62 patients studied (Varney et al., 2009; Ramakrishna and Subhash, 2012). It is quite possible that the treatment mechanism involves amitriptyline’s downregulation of α₁-adrenergic receptors (Ramakrishna and Subhash, 2012).

The relationship between catecholamine release and laryngospasm is readily observed in the routine securing of the airway with an endotracheal tube during the induction of general anesthesia. Laryngospasm, hypertension, and cardiac arrhythmias commonly result from inadequate blockade of the sympathetic nervous system during instrumentation of the airway and point to the influence of adrenergic innervation in addition to the cholinergically mediated innervation of laryngeal musculature. Specifically, noradrenergic nerve fibers directly innervate the intrinsic muscles of the larynx via laryngeal nerves in animals and humans and are contributed to vagal nerves via the superior cervical and middle cervical ganglia (Hisa, 1982; Hisa et al., 1982; Basterra et al., 1989; Yoshida et al., 1992). It is also important to note that prior to intubation or laryngeal mask placement anesthesiologists will administer opioids along with sedative hypnotics (such as propofol) to blunt the sympathetic activation that occurs with instrumentation of the airway. If opioids are not concurrently administered, a severe adrenergic response is likely to occur. The dose of propofol, an alkylphenol derivative that acts at the GABA_A receptor, provides sedation and amnesia but no analgesia. The necessary dose of propofol as a single agent to blunt an adverse sympathetic response is two to three times greater than when used with an opioid and a paralytic, and likely to cause severe hypotension and inadequate blockade of laryngospasm (Koenig et al., 2015). Paradoxically, the blunting of sympathetic airway responses with opioids is balanced against the clinical phenomenon of fentanyl-induced cough, which is thought to result from centrally mediated sympathetic inhibition and vagal predominance in the laryngeal muscle innervation (El Baissari et al., 2014). The effect can be benign or produce life-threatening laryngospasm. It occurs immediately after injection of fentanyl and can occur at low doses, but seems to share a similar underlying mechanism with the severe laryngospasm seen in WCS since both seem to be managed successfully with α-adrenergic and anticholinergic agents (Basterra et al., 1989; Yoshida et al., 1992; Lin et al., 2004; Horng et al., 2010; El Baissari et al., 2014).

Finally, when discussing muscle rigidity induced by F/FAs, it is important to note that centrally mediated skeletal muscle rigidity has a number of possible neurotransmitter-based etiologies including endogenous deficiencies, pharmacological induction (e.g., selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, antipsychotics, and anticholinergics), and neurologic conditions (e.g., Parkinson’s disease and stiff person syndrome) (Shanthanna, 2010). In fact, there have been reports in the anesthesia literature of serotonin and neuroleptic malignant syndromes occurring when selective serotonin reuptake inhibitors and neuroleptics/antipsychotics are used with anesthetic drugs such as fentanyl, which has been shown to have some proserotonergic effects (Gollapudy et al., 2012). However, this may be due to multiple concurrent neurotransmitter disruptions resulting from polypharmacy, since fentanyl, as a single agent, has never been documented to cause either serotonin or neuroleptic malignant syndromes (Pilgrim et al., 2011). In addition to proserotonergic effects, fentanyl has demonstrated anticholinergic effects (muscarinic receptor antagonism) at high doses in postoperative anesthesia patients with clinical symptoms similar to central anticholinergic syndrome (confusion, agitation, delirium, vigorous shivering, hot dry skin, purposeless movements, severe tachycardia, hypertension, and changes in vocal cord function) (Hustveit, 1994). Subsequent studies by Hustveit (1994) in rat brain demonstrated that fentanyl, but not morphine, had significant muscarinic receptor binding with K_i values in the micromolar range, in which it was concluded that muscarinic receptor binding of fentanyl is likely to occur during high-dose fentanyl anesthesia (50–150 mcg/kg) (Stone and DiFazio, 1988; Hustveit and Setekleiv, 1993; Yamanoue et al., 1993). In fact, the prerequisite for central anticholinergic syndrome is the antagonism of muscarinic receptors in the CNS (Hustveit, 1994).

As illustrated from the previous review, there is significant evidence that GABAergic, cholinergic, and noradrenergic pathways play a significant role in F/FA-induced WCS and will guide the development of effective therapeutic compounds.

Fentanyl’s Effects on Larynx and Vocal Cords: The Vagus Nerve.

The complexity of the innervation of vocal cords and the innervation of the airway is significant and a full review is outside the focus of this article. The reader is referred to the following studies for detailed reviews: Rex (1970), Basterra et al. (1989), and Jette and Thiebeault (2011).

Among humans, the clinical model of WCS demonstrates the significant involvement of the larynx and vocal cords, in which laryngospasm is the key feature of WCS (Scamman, 1983; Bennett et al., 1997). Laryngospasm is defined as the involuntary closure or occlusion of the glottic opening controlled by intrinsic muscles of the larynx. Although the underlying molecular mechanisms of WCS have not been definitively established, the combined synthesis of the animal and human data in this review may provide a plausible path to explain how fentanyl induces laryngospasm as a component of WCS in humans. The path begins with increased norepinephrine in the LC by fentanyl (indirectly), followed by activation and/or inhibition of noradrenergic and cholinergic signal paths through the CNS, and terminates on laryngeal muscles, resulting in laryngospasm. In this model, the predominant autonomic tone of the system will determine the mechanism of vocal cord closure. Pharmacologic points of therapeutic intervention will be suggested by the mechanism described.

Diaphragm, Abdomen, and Thoracic Wall Innervation.

Along with the glottic muscles, the main muscles of respiration include the diaphragm, intercostals, and abdominal wall muscles. The diaphragm receives its motor nerve impulses from the medullary centers via the phrenic nerve, which carries motor fibers from C3, C4, and C5 and also receives sympathetic nerve fibers from the cervical sympathetic chain. The intercostal muscles are stimulated by way of cervical, thoracic, and lumbar motor nerves and spinal motor neurons that terminate on skeletal muscles of the thorax and abdomen (De Troyer et al., 2005). The control of these muscle groups is complex, receiving multiple control signals from higher brain centers for voluntary control and subcortical regions for autoregulatory and reflexive controls (De Troyer et al., 2005).

Respiration and the Feedback Loop of Respiratory Mechanics with Medulla DRG and VRG.

The regulation and rhythmicity of breathing is based on a feedback loop of acid/base balance and blood chemistry (e.g., blood pH and PaCO₂ levels) that reflexively changes with metabolic and respiratory function. The dynamics of respiration increase or decrease based on afferent signals from the vagus and glossopharyngeal nerves (cranial nerves X and IX, respectively) that include peripheral chemoreceptors and mechanical changes in the lung parenchyma, airway, and joint proprio-receptors (Ikeda et al., 2017). These monitoring signals return to respiratory centers or respiratory-related neurons in the pons (i.e., apneustic and pneumotaxic centers), medulla (i.e., the DRG), central chemoreceptors, and VRG. The DRG is composed mostly of inspiratory neurons and subsequently controls the rhythmic initiation of inspiration impulses by sending signals to the motor nerves governing the diaphragm and external intercostal muscles. The VRG contains inspiratory and expiratory neurons and sends signals via the nucleus ambiguus to laryngeal and pharyngeal muscles and to the diaphragm, abdominal, and intercostal muscles. The DRG and VRG receive input from the pons and LC, which essentially stimulates or inhibits respiratory drive reflexively and can override the rhythmicity of the vagus and medullary feedback loop (Ikeda et al., 2017). The FIRMR component of WCS involves the innervation of the diaphragm, chest, and abdominal wall via the LC and spinal motor neurons, while the vocal cord/laryngeal component involves innervation of all intrinsic muscles of the larynx via the LC, superior cervical ganglia, vagus nerve, and vagal nuclei. In this scenario, LC activation by fentanyl could allow for increased medullary DRG- and VRG-driven muscle contraction, resulting in discoordination of medullary rhythmic respiratory drive signals and feedback loops. The subsequent triggering of diaphragm contraction and external intercostal muscle contraction against closed vocal cords is activated via the aberrant sympathetic signaling effects of fentanyl on the LC, vagal, and medullary nuclei. Additionally, the vagus nerve fibers contain μ-opioid receptors, innervate stretch receptors in lung parenchema, and can send afferent signals to vagal nuclei when activated by opioid binding (Willette and Sapru, 1982). This activity can trigger the Hering-Breur reflex discussed next (Chávez et al., 1998).

Hering-Breuer Reflex.

The scenario described previously is possibly complicated by the activation of another reflex arc involving the pulmonary stretch receptors located in the visceral pleura of the lung and the walls of the bronchi and bronchioles that then activate the Hering-Breur reflex. The Hering-Breur reflex arc was first noted in 1868 when Josef Breur and Ewald Hering observed that a maintained distension of the lungs during mechanical ventilation of anesthetized animals caused a cessation of inspiratory drive (Dutschmann et al., 2014). The reflex acts to prevent overinflation of the lung by sending inhibitory signals to the DRG to cease further inspiration.

The reflex may also include VRG-mediated closure of the vocal cords. In anesthesia with human subjects, overdistention of the lungs by manual ventilation can result in severe laryngospasm and may also be evident in the clinical phenomenon of the exercise-induced VCD (Dutschmann et al., 2014). Further complicating matters, and as mentioned previoulsy, μ-opioid receptors are located on the vagal fibers that innervate the actual juxtacapillary stretch receptors surrounding the alveoli. Activation of these fibers sends afferent signals to medullary respiratory centers and retrograde signals via the vagus nerve to the heart and vocal cords in an animal model where morphine and peptide opioids were delivered into the right atrium (Willette and Sapru, 1982). Delivery of these opioids into the right atrium caused significant electrical pulse activity, measured by EMG, to the cardiac vagal fibers and vagal fibers leading to the vocal cords. The peak pulse occurred within 1 to 2 seconds with rapidly decreasing amplitude over ∼10–15 seconds (Willette and Sapru, 1982). Vocal cord closure was not noted, since this had the limitation of being a tracheostomy rat model; however, pulmonary peak pressures were elevated during the initial EMG peak.

Cardiac Function in WCS.

Cardiac output can be significantly inhibited by vagus nerve activation in the nucleus ambiguus. Fentanyl binds μ-opioid receptors in the vagal nuclei and GABA interneurons of the nucleus ambiguus and can cause severe bradycardia and decreased cardiac output, with direct consequences for cerebral and hepatic perfusion pressures. As noted previously, fentanyl also antagonizes α₁-adrenergic receptor subtypes in a rank order of 1B > 1A >> 1D. This may have significant consequences on the function of cardiac myocytes and coronary arteries, which have selective distributions of these receptor subtypes. With regard to the heart, α_1A- and α_1B-adrenergic subtypes are concentrated almost exclusively on cardiac myocytes, while the coronary arteries exclusively use α_1D subtypes as noradrenergic inputs to control myocardial contractility/ventricular function and coronary artery perfusion of the mycocardium, respectively (Jensen et al., 2009). Selective binding and competitive antagonism of the cardiac α_1A- and α_1B-adrenergic subtypes by F/FAs can prevent the binding of catecholamines to the myocardial cells, resulting in depressed contractility and ventricular function (Hicks et al., 1981).

Similarly, fentanyl can antagonize α_1D receptors in coronary vessels, but has significantly weaker binding than at 1A and 1B receptors and competes poorly with norepinephrine at the 1D subtype (Sohn et al., 2005). The competitive binding of F/FAs to the 1A and 1B subtypes may simply help to concentrate norepinephrine and facilitate norepinephrine binding to coronary vessel 1D receptors, resulting in coronary artery vasoconstriction and a rapid decrease in coronary perfusion pressure and myocardial function.

Although the exact mechanisms in F/FA overdose and toxicity will need further attention, significant F/FA effects on the cardiovascular system cannot be discounted. A study by Wang et al. (2006) in chamber-isolated mouse myocardium demonstrated that each respective ventricle has different inotropic responses with the administration of the α_1A-adrenergic receptor agonist phenylephrine. Right ventricle tissue stimulation results in measurable negative inotropic effects in isolated myocardial fibers, while left ventricle stimulation results in positive inotropic effects. Myocardial α₁-adrenergic receptor subtype distribution in the mouse model is relevant in the human cardiovascular system (O’Connell et al., 2013). F/FA-selective antagonism of α receptors in human ventricular myocardium may result in paradoxical, inotropic isolation of the ventricles. The reversal of these demonstrated effects could promote left ventricle negative inotropy, while simultaneously promoting right ventricle positive inotropy in the presence of norepinephrine. The physiologic result would likely be severe pulmonary edema due to high forward pressure (right ventricle positive inotropy) against high back pressure (left ventricle negative inotropy), meeting in the pulmonary capillary beds and resulting in plasma ultrafiltrate or exudate in the alveoli (Wang et al., 2006).

Muscarinic receptor subtypes must also be considered in coronary artery vasoconstriction, since fentanyl has been shown to act as an M3 antagonist and the M3 receptor has been demonstrated to be the most prevalent subtype in porcine coronary arteries. Ach stimulation of M3 receptors appears to act as a coronary vasoconstrictor and could be potentially antagonized by fentanyl. The impact of these experimentally isolated effects as demonstrated by Yamanoue et al. (1993) remains unclear in the face of the α-adrenergic effects described, but would seem worthy of further exploration to determine the dominance of each effect (Yamanoue et al., 1993; Sohn et al., 2005).

These proposed cardiac mechanisms may explain the rapid onset of vascular, hepatic, and CNS effects described in the public health and autopsy data findings as mentioned previously, including: 1) rapid onset of cyanosis, 2) immediate loss of consciousness (central thalamocortical inhibition and/or decreased cerebral perfusion decrease), and 3) decreased hepatic metabolism (Somerville et al., 2017). Additionally, the selective effects of fentanyl on α₁-adrenergic receptors and their specific cardiovascular distribution may offer at least a partial explanation for the vascular component of F/FA-induced WCS.

Hepatic Function in WCS.

By decreasing cardiac output, high-dose fentanyl can secondarily cause decreased hepatic artery perfusion and/or potentially decreased cerebral perfusion, resulting in significant malfunction in fentanyl metabolism and rapid loss of consciousness, respectively (Griffioen et al., 2004). Hepatic perfusion normally constitutes approximately 25% of cardiac output, and hepatic metabolic function can be significantly compromised with decreases in blood flow of as little as 10% (Tuñón et al., 1999). The liver has a dual blood supply: 70% of hepatic blood flow is supplied by the portal vein and the rest by the hepatic artery. Portal vein flow is not tightly regulated, but is directly affected by systemic hypotension or decreases in cardiac output. In hepatic perfusion, studies up to a 50% decline in portal flow can occur if hepatic artery tone is compromised (Lautt, 1996; Møller and Henriksen, 2008; Schuppan and Afdhal, 2008).

Burns et al. (2016) noted that F/FA overdose decedents showed no detectable fentanyl metabolites in 20 of 48 cases (42%) and low levels (<1 ng/ml) in 25 of 48 cases (52%). Given the rapidity of normal fentanyl hepatic metabolism, this study indicates that hepatic metabolism becomes compromised within the first 2 minutes of F/FA overdose (McClain and Hug, 1980). Hypothetically, if a baseline of cardiac output continued, a compromise in hepatic metabolism could result in recirculation of high plasma/serum levels of F/FA back to the CNS with little to no decrease from the first-pass redistribution into the brain from the initial intravenous injection. If hepatic metabolism did not return, the effects of WCS would continue inevitably to death.

In summary, WCS is a syndrome that includes laryngospasm, respiratory muscle rigidity/contraction, cardiovascular compromise, and a concurrent decline in hepatic metabolism. Each component can be explained as a consequence of α₁-adrenergic, noradrenergic, and cholinergic receptor activity in addition to μ-opioid receptor activation.

Drug Development.

Based on the data presented here, there should be little doubt that laryngospasm, as part of WCS, plays a critical role in overdose death from F/FAs and that noradrenergic systems, cholinergic systems, and μ-opioid receptors are significant and interdependent components of this syndrome. This suggests that the most effective antagonist (immediate reversal therapy) and prophylaxis treatment strategies would consist of agents that can concurrently treat respiratory depression (μ-opioid receptor mediated) and the rapid respiratory failure of WCS (α-adrenergic, μ-opioid receptor, and muscarinic receptor mediated). Several points are worth clarifying here. The significant lethality that is occurring with F/FAs is partly due to their combination with other opiates such as heroin, which can have synergistic effects on respiratory drive when combined with F/FAs (Kiyatkin, 2019). All opioids cause some degree of dose-dependent respiratory depression and it is likely to emerge in a subject that has taken a large enough dose of F/FAs to cause WCS, that is, provided that the subject survives the acute mechanical failure of respiration. Therefore, naloxone and longer-acting μ-opioid antagonists such as nalmafene or naltrexone should be included in all opioid overdose treatment regimens. However, based on the mechanisms described, naloxone is clearly not effective as a sole therapeutic agent in F/FA overdose and appears to have a particularly minimal effect on the laryngospasm component of WCS as demonstrated in the animal model described by Willette et al. (1982, 1987).

Effective therapies in the future for F/FA overdose or toxic exposure prophylaxis will consist of α₁-adrenergic antagonists, possibly α₂ agonists, anticholinergic agents, and μ-opioid receptor antagonists. In the case of the α antagonists, it should be pointed out that although prazosin demonstrated effective inhibition of FIMR in the animal experiments noted, it is a nonselective agent that can cause profound hypotension in humans as it was used in these experiments. However, the effect of hypotension may be ameliorated to some degree by either using selective agents that have high affinity for α_1D subtypes (tamsulosin and terazosin) or combining smaller doses of prazosin with selective agents (Jerussi et al., 1987; Lui et al., 1990, 1993, 1995; Fu et al., 1997; Burns et al., 2016). It is also important to note that while both fentanyl and prazosin antagonize α₁-adrenergic receptors, fentanyl is a very weak antagonist (K_i = ∼1000 nM) at these receptors compared with prazosin (K_i = ∼0.1 nM) (Sohn et al., 2005). This is a critical distinction, and as a result prazosin can completely block norepinephrine activity at all α₁-adrenergic subtypes, while fentanyl binds weakly at α_1D subtypes (rank order of 1B > 1A >> 1D) compared with norepinephrine, which has its greatest affinity for the 1D subtype (Morrow and Creese, 1986; Sohn et al., 2005) Although the significance of this fact in the context of WCS was not pointed out by Sohn et al. (2005), this suggests that any available norepinephrine activity may be focused on interaction at the α_1D subtype (Yamanoue et al., 1993).

It should also be pointed out that α₂ agonists (clonidine) have been noted to be effective for treatment of FIMR in animal models and may be effective if used in combination with the other agents listed here, with the added benefit of decreasing the norepinephrine release seen in acute withdrawal and rapid reversal of opioid overdose in opioid-tolerant individuals (Jerussi et al., 1987; Weinger et al., 1989). Finally, the effects of selective anticholinergic agents (M1 and M2) or nonselective anticholinergic agents, with greater affinity for M1–M5 receptors than fentanyl, may also be effective agents to use in combination with α-adrenergic and μ-opioid receptor antagonist formulations. In this case, droperidol may be an effective agent, since it combines both α₁-adrenergic antagonist properties with anticholinergic effects that appear to inhibit the upper airway effects and decreased thoracic wall compliance that has been demonstrated with fentanyl (Stoelting et al., 1975; Yasuda et al., 1978; Parmentier and Dagnelie, 1979).

Education and Treatment Protocol Development for Medical Practitioners.

Clear guidelines are needed regarding administration of antagonist therapies. Benefits of using non-μ-opioid receptor antagonist therapy prophylaxis agents to prevent death from WCS in active heroin/fentanyl users should be explored as part of education programs. EMS and paramedic protocols for managing F/FA overdose should promote intubation and use of rapid-acting muscle paralytics such as succinylcholine as opposed to administering excessive doses of naloxone. Emergency rooms may establish notification protocols to request anesthesia call teams once they are alerted that an overdose involving fentanyl or synthetic opioids is en route.

Education/Protocol Development: Vulnerable Populations (Drug Users).

Recommendations should be made to active opioid users to modify injection practices with drug testing kits, test shots (small test dose preceding a larger dose), safe injection sites, and the use of chaperones or trained observer support. Prophylaxis agents that can be adapted to the population at risk should be made available for vulnerable populations that are still actively using opiates. These populations would be well suited for prophylactic muscle rigidity antagonist therapy that would strictly isolate prophylaxis against chest wall rigidity and laryngospasm and would exclude opioid antagonist therapy. Here, the rational is that if patients are exposed to fentanyl while using heroin, they would be less likely to suffer the rapid failure of respiratory mechanics associated with WCS. This would create the temporal window to use a reversal drug to treat respiratory depression and/or get expert medical attention.

Environmental and First-Responder Protective Strategies.

First responders including paramedics, emergency medical providers, firefighters, law enforcement, military personnel, and service animals are at risk of environmental exposure if F/FAs are an environmental contaminant. The task of managing and maintaining safety can be compromised if prophylaxis agents are unavailable. Prophylaxis agents can consist of antagonists for both muscle rigidity (chest wall and laryngeal muscles) and respiratory depression, treating both side effects concurrently in the case of large-scale exposure to F/FAs. This can allow for a therapeutic window for the administration of more concentrated reversal agents and treatments as needed. F/FAs have been weaponized previously on the battlefield and in covert military operations around the world, creating an imminent danger that needs addressing (Armenian et al., 2018). The threat of aerosolized forms of the potent fentanyl analog carfentanil was actualized in the Moscow theater incident of 2003, in which ∼1800 civilians were held hostage. Military responders piped carfentanil into the ventilation system of the theater to end the standoff. Civilian casualties exceeded 100, despite the reported availability and administration of naloxone for treatment (Pilch and Dolnik, 2003). F/FAs are readily transported through mail systems and traditional ports of entry (Armenian et al., 2018), which should instigate the concentrated development of effective strategies for protection.