A23: Antidotes in Depth

Benzodiazepines are used as first-line anticonvulsants for virtually all xenobiotic induced seizures; as sedatives of choice for most forms of xenobiotic induced agitation; as muscle relaxants for diverse disorders such as serotonin toxicity, neuroleptic malignant syndrome, and poisoning from strychnine or black widow spider envenomation; and as ­sedatives for withdrawal from ethanol, γ-hydroxybutyric acid, and other ­sedatives. Additional distinct indications for benzodiazepines can be found in overdose from chloroquine and possibly other quinine derived antimalarials, and in patients with cocaine associated myocardial ischemia and infarction. This Antidotes in Depth provides a summary of the clinical pharmacology of benzodiazepines in order to provide the reader with the background necessary to administer these drugs as safely and effectively as possible.

Most ancient texts make reference to the use of some plant or natural substance for sedation. By the mid to late 1800s, many of the natural sedatives were replaced by alcohol and chloral hydrate.¹⁷ The search for better sedatives with fewer side effects led to the development of barbiturates in the early 1900s followed by numerous others such as meprobamate, glutethimide, and ethchlorvynol (Chap. 74). Since the introduction of chlordiazepoxide in 1960,⁷⁷ benzodiazepines have gained acceptance as safe and effective drugs for a large variety of clinical indications. Over the intervening years, the number of benzodiazepines has increased, with each new drug demonstrating unique and complex pharmacology.

All benzodiazepines share a common chemical structure, shown in Fig. A23–1. This structure links a benzene ring with a diazepine ring and gives rise to the name used to describe the drug class. The additional phenyl ring is present in all clinically important benzodiazepines and serves as a site of substitution that modulates certain pharmacologic characteristics. The pyrazolopyrimidines (zolpidem, zopiclone, and zaleplon) lack the typical benzodiazepine structure but have similar pharmacologic effects.²⁸ Since these pharmaceuticals are largely unstudied for the antidotal indications described above, they are not discussed in depth here. A discussion of the manifestations and treatment of overdose of benzodiazepines and similar xenobiotics can be found in Chap. 74.

Benzodiazepines target the γ-aminobutyric acid type A (GABA_A) receptor, which is a ligand-gated chloride channel, but have no appreciable binding to GABA_B. When GABA is present on its GABA_A receptor, benzodiazepines increase the frequency of channel opening, resulting in enhanced flow of negatively charged chloride ions into the cell with resultant hyperpolarization.⁸³ However, in the absence of GABA, benzodiazepines have no effect on chloride conductance. The GABA_A receptor is assembled from five subunits that span the cell membrane in a circular fashion to create the chloride channel.⁵⁸ These subunits are coded as α, β, γ, δ, ε, θ, λ, or ρ, and at least 19 isoforms of these subunits (such as α_1-5) are identified.¹³ If the isoforms of the known subunits could assemble randomly, several hundred thousand possible configurations of the GABA_A receptor would be possible.⁸⁵ However, it appears that a minimum of at least one α, one β, and either one γ or one δ subunit are required to form a functional chloride channel,⁸⁹ and as a result the actual number of configurations is limited. The most common configuration consists of one α, two β, and two γ isoforms.^85,89

Rapidly evolving neuroscience has resulted in an exponential expansion in the understanding of benzodiazepine receptors. As a result multiple, potentially confusing nomenclatures have developed.

Central

The term “central benzodiazepine receptors” is used to refer to benzodiazepine binding sites on GABAergic neurons of the nervous system. Although benzodiazepine binding to the GABA_A receptor cannot occur unless at least one γ subunit is present,¹⁸ the actual site of benzodiazepine binding is located at the interface of an α and a γ subunit; most commonly an α₁ and a γ₂ subunit.^13,85 Since most receptors only contain a single α subunit, only one benzodiazepine binding site usually exists on each GABA_A receptor. Anatomical variations in the α isoforms produce two common patterns of expression that account for some of the clinical variations between benzodiazepines and the pyrazolopyrimidines mentioned earlier. In older nomenclature, benzodiazepine type 1 (BZ₁) receptors were also called ω₁ receptors. They have a predominance of the α₁ isoform, are located in the sensory and motor areas of the brain, and mediate sedative and hypnotic effects. Benzodiazepine type 2 (BZ₂) receptors were also called ω₂ receptors. They have a predominance of the α₂, α₃, or α₅ isoforms, are located in the subcortical and limbic areas of the brain, and mediate anxiolytic and anticonvulsant effects.⁵⁷ Most typical benzodiazepines have substantial affinity for the α₁, α₂, α₃, and α₅ isoforms, which explains their combined sedative-hypnotic, anxiolytic, and anticonvulsant effects. In contrast, the pyrazolopyrimidines (such as zolpidem) have high affinity for α₁, intermediate affinity for α₂ and α₃, and low affinity for α₅ isoforms, which explains their lack of anticonvulsant effect.^28,59 The α₄ confers resistance to benzodiazepines. This older nomenclature (BZ and ω receptors) is now simplified with GABA_A receptors classified as having high, low, or intermediate affinity benzodiazepine binding sites.⁷⁴

On the opposite side of the α isoform where the benzodiazepine receptor is located is an α–β interface. This interface holds the location of a binding site for neurosteroids (Fig. A23–2).³⁸ These neurosteroids are potent modulators of GABA_A receptor function and are important products of the tryptophan-rich sensory protein receptor (peripheral benzodiazepine receptor) stimulation (see below).

Peripheral Tryptophan-Rich Sensory Protein

The term “peripheral benzodiazepine receptor” (PBR) was originally used in the 1970s to define any benzodiazepine binding sites outside of the nervous system.¹¹ Since these PBRs were subsequently identified in most tissues, including the central nervous system, the term is best applied to binding sites not located on GABAergic neurons. Other names for PBRs included BZ₃ or ω₃ receptors to distinguish them from the “central receptors” described above.⁴⁷ However, numerous nonbenzodiazepines have high-affinity binding for these receptors, and their structure and function are so dissimilar from GABA_A-associated benzodiazepine binding sites that other names such as translocator protein (18 kDa), mitochondrial translocator protein (18 kDa), nuclear translocator protein (18 kDa), were appropriate.^68,84,90 The most recent conventions have adopted the term tryptophan-rich sensory protein (TSPO) as this refers to its gene and denotes it as a tryptophan-rich sensory protein.²⁴ Although the term PBR remains attractive, for simplicity and consistency, the current term TSPO will be used in remainder of this discussion.

The TSPO participates in a heterotrimer structure that is composed of an isoquinoline binding protein, which is the actual receptor (TSPO); a voltage-dependent anion channel (VDAC); and an adenine nucleotide transporter (ANT).⁶⁸ Although the actual function of each subunit is not well appreciated, the minimal functional unit is the TSPO (18 kDa protein).⁴⁵ Sequencing of TSPOs demonstrates that they are highly conserved in nature, with DNA from bacteria and fungi having a nearly 50% homology of the isoquinoline binding domain with human DNA.^24,27 Homology among mammals exceeds 75%.²⁷ These findings suggest that the PBRs perform a “housekeeping function,” that is, they are involved in a process or processes that are essential for life. In higher life forms, TSPOs can be found primarily in the brain, adrenal glands, heart, and kidney. The TSPO protein and the VDAC span the outer mitochondrial membrane, while the ANT bridges the outer and inner membranes (Fig. A23–3). TSPOs are implicated in cholesterol and protoporphyrin transport required for the synthesis of neurosteroids, heme, and bile salts; ischemia and reperfusion; regulation of calcium channels; mitochondrial respiration; apoptosis; microglial activation; and the immune response.^27,68,84 More recently, TSPO ligands have been investigated as potential targets for a variety of disorders including anxiety, cancer, ischemia, and others.^{15,19,26,41,42,63,64,67,69,75,80,91} It is hypothesized that TSPO has two major roles: opening of the mitochondrial permeability transition pore (MPTP) leading to calcium influx and apoptosis,^7,23,26,52 and, as noted above, in synthesis of neurosteroids that modulate GABA_A function.^67,75

Benzodiazepines are usually administered to treat seizures, psychomotor agitation, or sedative-hypnotic withdrawal. Since these disorders often represent life-threatening emergencies, the initial approach to treatment involves the use of parenteral drugs. Parenteral therapy provides guaranteed absorption with a relatively rapid onset of action. As such, this Antidotes in Depth focuses on the most commonly used parenteral benzodiazepines: diazepam, lorazepam, and midazolam. Clinically important pharmacologic parameters of these three drugs are listed in Table A23–1. Of note, since the existing literature is insufficient to absolutely differentiate individual roles for specific benzodiazepines in particular disorders, these three drugs are often used interchangeably for a variety of indications based on availability within the hospital and regional historic preferences. Only through an understanding of the pharmacology can the clinician select the optimal drug, dose, route, and interval to ensure adequate response and limit adverse reactions. As discussed below, it is important to note that the onset and peak effects of sedative and anticonvulsant activity are distinctly different for each drug and among drugs, so pharmacodynamic data are an important adjunct to the interpretation of pharmacokinetic data.

Intravenous administration guarantees complete and immediate absorption. When intravenous access is unavailable, intramuscular lorazepam or midazolam should be considered as they both have good absorptive profiles.^{33,39,40,43,56,78,88} Midazolam may offer added benefits as its role in synthesis of neurosteroids is experimentally defined.⁸¹ In contrast, the absorption following intramuscular diazepam is best described as slow, incomplete, erratic, and dependent on the site of administration and the skill of the person administering it.^{20,21,39,44,86} Intramuscular injection of diazepam should therefore not be considered unless no other alternatives exist.

Other routes of administration are used for rapid drug delivery when intravenous access is not available. For example, and especially in children, intranasal and rectal administration can be considered. When rectal diazepam was compared with intravenous diazepam in children with seizures, the peak concentration was more variable and was delayed by about 20 minutes.⁶⁶ Additionally, failure rates are higher when rectal diazepam is compared with intravenous diazepam.^66,71 Similarly, although midazolam is more rapidly absorbed than diazepam following both nasal and rectal administration, the kinetic profile of nasal or rectal midazolam is still inferior to intravenous administration of either diazepam or midazolam.^35,55,71 Recent evidence has supported a role for intramuscular midazolam for seizures and agitation.^43,56,78

Once absorbed, benzodiazepines must distribute into the central nervous system in order to produce their sedative and anticonvulsant effects. The differences among individual drugs can be evaluated in terms of their pharmacokinetics, such as cerebral drug concentrations, or pharmacodynamics, such as changes in either consciousness or electroencephalographic (EEG) findings. A cat model assessed cerebrospinal fluid (CSF) concentrations of benzodiazepines following intravenous administration along with their effects on the EEG.⁵ Diazepam, midazolam, and lorazepam all appeared rapidly in the CSF.⁵ However, the times to peak concentration, onset, and duration of EEG activities were remarkably different among the three drugs (Table A23–2).

In a similar study, human volunteers were given a one minute intravenous infusion of either diazepam (0.15 mg/kg) or midazolam (0.1 mg/kg) and EEG analysis was used as a surrogate for the pharmacodynamic effects of sedation. Peak EEG effects were present immediately at the end of the diazepam infusion, but were delayed for 5 to 10 minutes after the midazolam infusion.³² These EEG effects lasted for 5 hours in diazepam-treated volunteers in comparison to only 2 hours in midazolam-treated volunteers. The same investigators compared the EEG effects of either intravenous lorazepam (low dose, 0.0225 mg/kg; high dose, 0.045 mg/kg) or diazepam (0.15 mg/kg) in human volunteers.³¹ In comparison to diazepam, the EEG effects of lorazepam were slower in onset, delayed in peak (30 minutes), and prolonged in duration of effect. This may be related in part to a relatively long alpha distribution half-life for lorazepam.³⁰ When another group compared the EEG effects of intranasal to intravenous midazolam they demonstrated the onset of EEG effects within 1.2 minutes of a 5 mg intravenous dose.³⁵ The relative effects of diazepam, lorazepam, and midazolam are shown in Table A23–3.

Many chapters in this text discuss the use of benzodiazepines as GABA_A agonists in the management of poisoned patients (Chaps. 14, 15, 75, 76, 78, 81, 82, and 86). In contrast to the highly evidence-based support for GABA_A agonists in sympathomimetic overdose, xenobiotic-induced seizures, and sedative-hypnotic withdrawal, the use of benzodiazepines as modulators of TSPOs for the treatment of poisoned patients is highly speculative. However, as this field is rapidly evolving, discussions of two xenobiotics, chloroquine (Chap. 59) and cocaine (Chap. 78) are warranted.

Chloroquine

Although chloroquine overdose is uncommon, case fatality rates are extremely high, and ingestion of 5 g or more was once considered universally fatal. However, in 1988 a case series of patients with chloroquine overdose who survived with the use of an aggressive new regimen was described.⁷² The protocol, consisting of early endotracheal intubation, high-dose epinephrine infusion, and intravenous diazepam (2 mg/kg over 30 minutes), resulted in the survival of 10 of 11 patients who ingested at least 5 g of chloroquine. This regimen was based on a controlled trial where diazepam improved the hemodynamic and electrocardiographic manifestations of chloroquine poisoning in pigs.⁷³ Although the study was not designed to determine the mechanism of effect of diazepam, there was no difference in the chloroquine concentrations between the treated and control groups, suggesting that diazepam did not change the distribution of chloroquine.

In isolated perfused hearts, high-dose diazepam has positive inotropic effects that are suppressed by a TSPO antagonist.⁵⁰ Unfortunately, translation of these findings to use in human poisoning is not necessarily valid. Benzodiazepines, chloroquine, and experimental TSPO agonists share some structural elements, suggesting a receptor-based mechanism (Fig. A23–4). Functional similarity is also suggested by evidence that both flurazepam and PK 11195 (a TSPO agonist) have antimalarial activity.²² Thus, one could speculate that it is possible that the beneficial effects of high-dose diazepam result from competitive inhibition between chloroquine and TSPOs.

Cocaine

Patients who use cocaine frequently present to emergency departments with chest pain or signs or symptoms that could represent myocardial ischemia or infarction.¹² Unlike most patients with acute cocaine toxicity, however, these patients often present hours after their last drug use and without the classic sympathomimetic findings of acute cocaine toxicity.³⁶ While the pathophysiology of cocaine-induced myocardial ischemia is complex and multifactorial (Chap. 76), one component of delayed myocardial ischemia may result from the vasoconstrictive actions of benzoylecgonine, the principal metabolite of cocaine.⁵⁴ Benzoylecgonine is distinct from cocaine in that it has a much longer half-life,^3,4 does not produce central nervous system stimulation,^61,62 and directly vasoconstricts through modulation of calcium channels.⁵³

Limited research suggests that chronic cocaine use is associated with an increased number of TSPOs on human platelets.¹⁶ Also in humans, cocaine withdrawal is associated with a decrease in TSPOs on neutrophils. In the myocardium, TSPOs are either present on mitochondria or coupled to calcium channels.⁶⁰ Specifically, TSPO ligands have inhibitory effects on myocardial L-type calcium channels.¹⁴ Also in experimental models of cardiac ischemia and reperfusion injury, TSPO agonists limit myocardial infarction size and improve cardiac function.^48,91 This effect most likely occurs through inhibition of the opening of the mitochondrial permeability transition pore,⁶⁵ the opening of which is a common final mechanism of cell death. Additionally, although the exact mechanism is unclear, TSPO agonists directly antagonize the vasoconstrictive effects of norepinephrine in rat aortic tissue.²⁹

Benzodiazepines are commonly used to treat the agitation associated with sympathomimetic overdose (Chaps. 76 and 78). While it is assumed that the normalization of vital signs results from a decrease in central nervous system stimulation and psychomotor agitation, effects on TSPOs may be contributory. In the same isolated perfused hearts described above, low-dose diazepam has a negative inotropic effect that results from an interaction with calcium currents.⁴⁹ As noted above, interpreting the implications of varied benzodiazepine doses in an isolated rat model with regard to human beings may not be valid. More importantly, two randomized controlled studies evaluated the use of benzodiazepines in patients with cocaine-associated chest pain.^9,37 In the first study, patients were randomized to receive nitroglycerin, diazepam, or combined therapy.⁹ Both drugs were associated with an improvement in chest pain, and combined therapy appeared to offer no additional benefit. The second study randomized patients to nitroglycerin or combined nitroglycerin with lorazepam therapy.³⁷ In this trial, combined therapy was better than either therapy alone, possibly suggesting that benzodiazepines and nitroglycerin relieve chest pain by different mechanisms.

The most common adverse effect of benzodiazepines is central nervous system depression. While this is unavoidable in some cases, it can generally be limited by selecting the optimal drug and the proper dose and dosing interval for the drug being used. Extra caution is advised in elderly patients as they appear to be more sensitive, particularly to the sedative and respiratory depressant effects of midazolam.¹ In contrast, paradoxical reactions may occur where patients become more agitated following benzodiazepine administration.^76,79,87 These infrequent reactions probably result from disinhibition and may respond to larger doses of benzodiazepines. Although paradoxical agitation also responds to flumazenil (Antidotes in Depth: A22), reversal would generally be inappropriate following the use of benzodiazepines for most toxicological indications.

Intravenous benzodiazepines produce a mild reduction in heart rate and both systolic and diastolic blood pressure. The effects of midazolam may be greater than diazepam,⁷⁰ but this may merely be based on an inability to determine an equivalent dosing regimen. While these reductions result in part from diminished sympathetic tone, direct myocardial effects are rarely severe and are often considered desirable in the overdose setting.⁷⁰

Whereas respiratory depression is generally not a concern with oral benzodiazepines, parenteral administration is documented to impair ventilation. Early investigations demonstrated that intramuscular diazepam (10 mg) blunted the hypoxic ventilatory drive in normal subjects.⁴⁶ Intravenous diazepam impaired the ventilatory response to a rising PCO₂ in normal volunteers.⁸ The impaired response to a rising PCO₂ was evident almost immediately and lasted for at least 25 minutes following injection of 0.4 mg/kg of diazepam in healthy volunteers.³⁴ Studies with midazolam demonstrate similar alterations in respiratory physiology²⁵ that are comparable in magnitude to those reported with diazepam.¹⁰ Apnea is reported following intravenous midazolam and appears to be dose and rate related, with doses greater than or equal to 0.15 mg/kg being of particular concern.⁴⁸ Individuals with preexisting pulmonary disorders, extremes of age, or other central nervous system or respiratory depressants may be more susceptible. When intramuscular midazolam is used for the treatment of agitation, some transient respiratory depression is noted, but the need for external ventilator support is uncommon.^40,43,56

Pregnancy and Lactation

Diazepam, lorazepam, and midazolam are all listed as category D drugs in pregnancy and are likely unsafe in lactation. These classifications are likely based on repetitive use and have little applicability to short-term benzodiazepine administration in emergent situations. However, the risks, benefits, and alternatives should always be considered prior to giving any xenobiotic to a pregnant woman. If a benzodiazepine is given to a lactating mother, “pumping and dumping” would be advised to prevent sedation in the child (Chap. 31).

In theory, an analysis of the pharmacokinetic and pharmacodynamic parameters presented in Tables A23–1 and A23–3 should allow clinicians to choose among the benzodiazepines based on the particular clinical situation. In reality, however, these choices may be difficult in patients with complex clinical presentations of uncertain etiologies. In addition, clinical studies seem to suggest less variability than might be predicted based on pharmacokinetic and pharmacodynamic modeling. For example, in an analysis of 28 studies of conscious sedation, 19 failed to demonstrate a significant difference in the times to recovery between diazepam and midazolam.⁶ Likewise, although multiple studies demonstrate that lorazepam is at least as effective and likely superior to diazepam in terminating status epilepticus, its onset of action does not appear to be delayed.^2,51,82 In a double-blind trial of intravenous lorazepam versus intramuscular midazolam for patients with prehospital status epilepticus, patients who received intramuscular midazolam stopped seizing sooner and in greater numbers.⁷⁸

Some of the observed inability to translate controlled pharmacologic analyses into clinical practice may result from imprecision in determining equivalent doses of these drugs. We offer the following guidance. Each of these three benzodiazepines discussed will likely have some efficacy in all clinical scenarios for which a benzodiazepine is indicated. It may therefore be preferable for many clinicians to understand and master the pharmacology and limitations of a single drug rather than attempt to use all three drugs selectively and run the risk of an improper dose or suboptimal therapeutic interval leading to an adverse effect. For those who wish to use these drugs selectively, three variables should be considered: onset of effect, peak effect, and duration of effect. For example, in patients with extreme psychomotor agitation intravenous diazepam or midazolam would be preferred both for their rapid onset and rapid peak effects. The relatively slow onset and delayed peak of lorazepam might lead to administration of multiple doses before the full effect of the first dose is appreciated, with resultant oversedation. When a short duration of sedation is anticipated, such as when treating a patient with toxicity following intravenous or inhaled use of cocaine, midazolam might be preferable over diazepam or lorazepam in that the duration of the effects of midazolam better matches the duration of effects of cocaine, thereby limiting oversedation after cocaine has been rapidly metabolized. For disorders where very long periods of agitation are expected, such as alcohol withdrawal, the choice of diazepam with its active metabolites (desmethyldiazepam and oxazepam) may be preferable over lorazepam and midazolam. With diazepam, less frequent dosing may be required and an auto-tapering effect may result at the end of therapy, because these metabolites persist longer than diazepam alone. When status epilepticus is present in the absence of substance use or withdrawal, midazolam in large doses seems to offer a clear advantage. It is important to note that switching from one benzodiazepine to another is rarely indicated and increases the risk of an adverse drug event from unpredictable peak effects or improper therapeutic doses or dosing intervals. Specific recommendations can be found in Chaps. 15, 76, 78, and 81.

Parenteral benzodiazepines are available from multiple manufacturers in varying concentrations. It is essential to recognize that some formulations contain several significant and varied excipients (Table A23–1). Large doses or prolonged continuous infusions of benzodiazepines may result in toxicity from these excipients (Chap. 55).

• Although benzodiazepines are commonly used in a variety of medical settings, subtle differences exist in their pharmacokinetics and pharmacodynamics. Optimal use of these drugs requires a thorough understanding of these differences.

• Clinicians should consider the desired onset of action, peak effect, and duration of action when choosing among different benzodiazepines.

• In general, intravenous administration is preferred in an emergency because of rapid and reliable absorption. Because of a rapid onset of action and short time to peak effect, sedation is best achieved with midazolam when a short duration of action is desirable, or diazepam for a longer duration of action.

• When intravenous access is not available, pharmacokinetic and pharmacodynamic parameters favor the use of midazolam for sedation.

• When used for sedation, the delayed peak effect of lorazepam (regardless of route of administration) is undesirable.

• Intravenous lorazepam is highly effective for seizures, but a continuous infusion of midazolam is preferred for status epilepticus.

• Although the use of benzodiazepines as TSPO (peripheral benzodiazepine receptor) agonists may explain some unique therapeutic effects of these drugs, existing research is too limited to offer guidance with regard to choice of drug, dose, or dosing interval required to optimize peripheral benzodiazepine receptor response.

References