G: Psychotropics

HISTORY AND EPIDEMIOLOGY

The development of antipsychotic drugs changed the practice of psychiatry. Prior to the introduction of chlorpromazine in 1950, patients with schizophrenia were treated with nonspecific sedatives such as barbiturates and chloral hydrate. Highly agitated patients were housed in large mental institutions and often placed in physical restraints, and thousands underwent surgical disruption of the connections between the frontal cortices and other areas of the brain (leucotomy). By 1955, approximately 500,000 patients with psychotic disorders were hospitalized in the United States. The advent of antipsychotics in the 1950s revolutionized the care of these patients. These drugs, originally termed major tranquilizers and subsequently neuroleptics, dramatically reduced the characteristic hallucinations, delusions, thought disorders, and paranoia—the “positive” symptoms of schizophrenia.

Shortly after the introduction of these drugs, it became apparent that they caused significant toxicity following overdose, a common occurrence in patients with mental illness. Moreover, they were also associated with a host of adverse effects, principally involving the endocrine and nervous systems. The latter includes the extrapyramidal syndromes (EPS), a constellation of disorders that are relatively common, sometimes irreversible and occasionally life threatening.

The search for new drugs led to the development of multiple antipsychotics of several chemical classes. These drugs exhibited varying potencies and markedly different adverse effect profiles. The novel antipsychotic clozapine was first synthesized in 1959, but it did not enter widespread clinical use until the early 1970s. Clozapine was unusual because it conferred a relatively low risk of EPS, but also because it was often effective in patients who had not responded well to other xenobiotics. Moreover, unlike the other available xenobiotics, it often improved the “negative” symptoms of schizophrenia such as avolition, alogia, and social withdrawal—symptoms that, while often less outwardly apparent than the positive symptoms, result in significant disability. Reports of life-threatening agranulocytosis led to the withdrawal of clozapine from the market in 1974, although it was reintroduced in 1990.^9,49 However, the unique therapeutic and pharmacologic properties of clozapine led to its characterization as an atypical antipsychotic, the forerunner and prototype of many other second-generation antipsychotics that have now largely supplanted the earlier xenobiotics.

Most antipsychotic toxicity occurs by one of two mechanisms. Following overdose, antipsychotic toxicity is dose dependent and reflects an extension of the effects of the drug on neurotransmitter systems and other biologic processes. The features of antipsychotic drug overdose are therefore generally predictable based upon an understanding of the pharmacology of the drug. Unpredictable (idiosyncratic) adverse reactions also occur in the context of routine therapeutic use. These toxicities result from individual susceptibility, are sometimes pharmacogenetic in nature, and are less reliably correlated with the dose. In both types of toxicity, the severity of illness can range from minor to life threatening, depending on a number of other factors, including concomitant drug exposures, comorbidity, and access to medical care.

The true incidence of antipsychotic overdose and adverse reactions is not known with certainty. Some patients may not seek medical attention, whereas others may be misdiagnosed. Even among those who seek medical attention and are correctly diagnosed, notification of poison centers or other adverse event reporting systems is discretionary and incomplete (Chap. 136). With these limitations in mind, a few observations can be made.

In 2011, poison control centers in the United States were con­tacted about more than 3.6 million potential toxic exposures.¹⁶ Antipsychotic exposures are reported together with sedative-hypnotics, but these collectively represented 168,416 exposures (6.13% of all exposures). The vast majority of poison center calls involving antipsychotics pertain to intentional overdoses in patients 20 years or older, most of whom have a good outcome. However, antipsychotics were associated with more fatalities than any other group (n = 401 deaths), and most exposures involved atypical antipsychotics. Importantly, poison center data underestimate the annual incidence of poisoning and mortality associated with antipsychotics and likely identify only a small minority of adverse drug reactions involving these drugs.

Although all antipsychotics can exhibit significant toxicity in overdose, a substantial body of clinical experience and some observational data suggest that the low potency, first-generation antipsychotics such as thioridazine, chlorpromazine, and mesoridazine are associated with greater toxicity than other antipsychotics.^17,19 Inferences regarding the relative toxicity of the antipsychotics derived from administrative data should be extrapolated to individual patients with caution,^17,40 but at least one well done retrospective cohort study supports the notion that thioridazine is associated with greater cardiovascular toxicity than other antipsychotics.¹⁹

PHARMACOLOGY

Classification

Antipsychotics can be classified in several ways, according to their chemical structure, their receptor binding profiles, or as “typical” or “atypical” antipsychotics. Table 70–1 outlines the taxonomy of some of the more commonly used antipsychotics. Classification by chemical structure was most useful prior to the 1970s, when phenothiazines and butyrophenones constituted most of the antipsychotics in clinical use. Currently, the number of different compounds and their structural heterogeneity renders this scheme of little utility to clinicians. It is worth noting, however, that the phenothiazine antipsychotics bear a high degree of structural similarity to the tricyclic antidepressants (TCAs) (Fig. 70–1) and share many of their manifestations in overdose. The phenothiazines can be further classified according to the nature of the substituent on the nitrogen atom at position 10 of the center ring as either aliphatic, piperazine, or piperidine compounds.

Of greater clinical utility is the classification of antipsychotics according to their binding affinities for various receptors (Table 70–2). However, by far the most widely used classification system categorizes antipsychotics as either typical or atypical. Typical (also called traditional, conventional or, increasingly, first-generation) antipsychotics dominated the first 40 years of antipsychotic therapy. They were subcategorized according to their affinity for the D₂ receptor as either low potency (exemplified by thioridazine and chlorpromazine) or high potency (exemplified by haloperidol). They ameliorated the positive symptoms of schizophrenia, such as hallucinations, delusions, paranoia, and disorganization of thought, but were of little benefit for the sometimes disabling negative symptoms including avolition, alogia, flattening of affect, and social withdrawal. Moreover, they were associated with acute, subacute, and long-term motor disturbances collectively referred to as EPS.

The concept of antipsychotic atypicality has evolved over time with the introduction of new compounds^94,113 and connotes different properties to pharmacologists and clinicians. From a clinical perspective, atypical ­antipsychotics (sometimes termed second-generation antipsychotics) treat both the positive and negative symptoms of schizophrenia, are less likely than traditional drugs to produce EPS at clinically effective doses, and cause little or no elevation of the serum prolactin concentration.⁵⁷ From a pharmacologic perspective, most atypical antipsychotics also inhibit the action of serotonin at the 5-HT_2A receptor. More than a dozen atypical antipsychotics are now in clinical use or under development. Despite their considerably higher cost, these drugs have largely supplanted traditional antipsychotics because of their effectiveness in treating the negative symptoms of schizophrenia and their somewhat more favorable adverse effect profile, in addition to the perception that they cause fewer long-term adverse effects than conventional antipsychotics—a belief that may result, in part, from the use of higher doses of older drugs in studies comparing the tolerability of typical and atypical antipsychotics.⁴⁸ Considerable controversy exists regarding the superiority of these drugs over those of the first-generation antipsychotics, and it is worth noting that the use of the newer antipsychotics for other indications is extremely common, including their use as adjunctive treatment for major depression, eating disorders, attention-deficit hyperactivity disorder, insomnia, posttraumatic stress disorder, personality disorders, and Tourette syndrome.⁶⁸ However, the most extensive off-label use of these drugs is for the management of agitation associated with cognitive impairment in the elderly.

Mechanisms of Antipsychotic Action

Of the many contemporary theories of schizophrenia, the most enduring has been the dopamine hypothesis.¹⁰⁷ First advanced in 1967 and supported by in vivo data,¹ this theory holds that the “positive symptoms” of schizophrenia (hallucinations, delusions, paranoia, and disorganization of thought) result from excessive dopaminergic signaling in the mesolimbic and mesocortical pathways.⁷³ This hypothesis was borne in part from the observation that hallucinations and delusions could be produced in otherwise normal individuals by drugs that increase dopaminergic transmission, such as cocaine and amphetamine, and that these effects could be blunted by dopamine antagonists.

There are at least five subtypes of dopamine receptors (D₁ through D₅), but schizophrenia principally involves excess signaling at the D₂ subtype,¹⁰⁷ and antagonism of D₂ neurotransmission is the sine qua non of antipsychotic activity. Antipsychotics have different potencies at this receptor, reflected by the dissociation constant (K_d), which in turn reflects release of the drug from the D₂ receptor. For example, the receptor releases clozapine and quetiapine more rapidly than it does any other drugs.^105,107

Dopamine receptors are present in many other areas of the central nervous system (CNS), including the nigrostriatal pathway (substantia nigra, caudate, and putamen, which collectively govern the coordination of movement), tuberoinfundibular pathway, hypothalamus and pituitary, and area postrema of the brainstem, which contains the chemoreceptor trigger zone (CTZ). Antipsychotic-related blockade of D₂ neurotransmission in these areas is associated with many of the beneficial and adverse effects of these drugs. For example, D₂ antagonism in the CTZ alleviates nausea and vomiting, whereas blockade of hypothalamic D₂ receptors increases pituitary prolactin release, resulting in gynecomastia and galactorrhea. Blockade of nigrostriatal D₂ receptors underlies many of the movement disorders associated with antipsychotic therapy.^119,132

Antipsychotics interfere with signaling at other receptors to varying degrees, including muscarinic receptors, H₁ histamine receptors, and α-adrenergic receptors. The extent to which these receptors are blocked at therapeutic doses can be used to predict the adverse effect profile of each antipsychotic.²¹ For example, drugs that antagonize muscarinic receptors at clinically effective doses (most notably the aliphatic and piperidine phenothiazines as well as clozapine, loxapine, olanzapine, and quetiapine) often produce anticholinergic adverse effects during routine use and can produce pronounced anticholinergic manifestations following overdose (Table 70–2). Similarly, blockade of peripheral α₁-adrenergic receptors by the aliphatic and piperidine phenothiazines, clozapine, risperidone, and others renders them more likely to cause postural hypotension during therapy and clinically important hypotension following overdose. In contrast, haloperidol overdose is not characterized by marked antimuscarinic effects or hypotension.

Several antipsychotics also block voltage-gated fast sodium channels (I_Na). Although this effect is of little consequence during therapy, in the setting of overdose this can slow cardiac conduction (phase 0 depolarization) and impair myocardial contractility. This effect, most notable with the phenothiazines, is both rate- and voltage-dependent, and is therefore more pronounced at faster heart rates and less negative transmembrane potentials.¹⁸ Blockade of the delayed rectifier potassium current (I_Kr) can produce prolongation of the QT interval, creating a substrate for development of torsade de pointes.⁷⁸ QT interval prolongation is sometimes evident during maintenance therapy, particularly in patients with previously unrecognized repolarization abnormalities or additional risk factors for QT prolongation. This effect may partially explain the dose-dependent increase in risk of sudden cardiac death among patients treated with typical and atypical antipsychotics.^92,93

Several antipsychotics exhibit a relatively high degree of antagonism at the 5-HT_2A receptor, which conveys two important therapeutic properties: (1) greater effectiveness for the treatment of the negative symptoms of schizophrenia and (2) a significantly lower incidence of extrapyramidal side effects. Others antipsychotics are distinguished by unique effects at different receptors. For example, loxapine and clozapine inhibit the presynaptic reuptake of catecholamines and antagonize γ-aminobutyric acid (GABA)_A receptors,¹¹² which may explain the apparent increase in seizure activity with these antipsychotics.⁸⁹ A more detailed description of the pharmacology of the most commonly used second-generation antipsychotics is warranted in light of their increasing role in therapy.

Clozapine, a dibenzodiazepine compound, binds to dopamine receptors (D₁–D₅) and serotonin receptors (5-HT_1A/1C, 5-HT_2A/2C, 5-HT₃, and 5-HT₆) with moderate to high affinity.^9,90,98 It also antagonizes α₁-adrenergic, α₂-adrenergic, and H₁ histamine receptors. It has the highest binding affinity of any atypical antipsychotic at M₁ muscarinic receptors.⁹⁷ Despite this feature, clozapine paradoxically activates the M₄ subtype of the muscarinic receptor and frequently produces sialorrhea during therapy.⁹⁶

Olanzapine, a thienobenzodiazepine, binds avidly to serotonin (5-HT_2A/2C, 5-HT₃, and 5-HT₆) and dopamine receptors (D₁, D₂, and D₄), although its potency at D₂ receptors is lower than that of most traditional antipsychotics.^60,98 It is an exceptionally potent H₁ antagonist, binding more avidly than pyrilamine, which is a widely used antihistamine. It is also has a high affinity for M₁ receptors and is a relatively weak α₁ antagonist.

Risperidone, a benzisoxazole derivative, has high affinity for several receptors, including serotonin receptors (5-HT_2A/2C), D₂ dopamine receptors, and α₁ and H₁ receptors.^60,96,98 It has no appreciable activity at M₁ receptors. Its primary metabolite (9-hydroxyrisperidone) is nearly equipotent as the parent compound at D₂ and 5-HT_2A receptors.⁶⁰ Available orally and as a long-acting parenteral preparation, paliperidone is the major active metabolite of risperidone and exhibits a similar receptor-binding profile.

Quetiapine, a dibenzothiazepine, is a weak antagonist at D₂, M₁, and 5-HT_1A receptors, but it is a potent antagonist of α₁-adrenergic and H₁ receptors.⁶⁰ Of its 11 metabolites, at least 2 are pharmacologically active, but they circulate at low concentrations and likely contribute little to its clinical effect. A considerable proportion of fatalities involving antipsychotic drugs reported to North American Poison Control Centers involve quetiapine, usually in combination with other drugs.¹⁶

Ziprasidone, a benzothiazole derivative, is an antagonist at dopaminergic D₂ and several serotonin (5-HT_2A/2C, 5-HT_1D, and 5-HT₇) receptors, but it also displays agonist activity at 5-HT_1A receptors.^60,61,98 Its α₁-antagonist activity is particularly strong, with a binding affinity approximately one tenth that of prazosin. In addition, it is a strong inhibitor of the delayed rectifier channel (I_Kr) and can significantly prolong repolarization.^61,70

Aripiprazole, a quinolinone derivative, is a novel compound that binds avidly to dopamine D₂ and D₃ receptors and serotonin 5-HT_1A, 5-HT_2A, and 5-HT_2B receptors.^77,98 Some evidence suggests that its efficacy in the treatment of schizophrenia and its lower propensity for EPS may relate to partial agonist activity at dopamine D₂ receptors.⁷⁶ Aripiprazole acts as a partial agonist at serotonin 5-HT_1A receptors but is an antagonist at serotonin 5-HT_2A receptors. Its principal active metabolite, dehydroaripiprazole, has affinity for dopamine D₂ receptors and thus has pharmacologic activity similar to that of the parent compound.⁷⁷

Like aripiprazole, bifeprunox is a partial agonist at D₂ and 5-HT_1A receptors. It has been characterized as a third-generation antipsychotic and has no appreciable affinity for serotonin 5-HT_2A and 5-HT_2C histaminergic or muscarinic receptors.^28,80,106

Amisulpride is a substituted benzamide derivative that preferentially blocks dopamine receptors in limbic rather than striatal structures. At low doses, it blocks presynaptic D₂ and D₃ receptors, thereby accentuating dopamine release, while at high doses it blocks postsynaptic D₂ and D₃ receptors. It has relatively low affinity for serotonergic, histaminergic, adrenergic, and cholinergic receptors.

Sertindole is a second-generation antipsychotic drug that was recently reintroduced into the market after being voluntarily withdrawn in 1998 over concerns about its effects on the QT interval. It binds to striatal D₂ receptors, although less avidly than olanzapine, and also exhibits antagonism at 5-HT2A and α₁-adrenergic receptors.^59,86,111 It is estimated that 3.1% to 7.8% of patients receiving sertindole develop QT intervals greater than 500 milliseconds.¹³⁰

PHARMACOKINETICS AND TOXICOKINETICS

With a few exceptions, the antipsychotics have similar pharmacokinetic characteristics regardless of their chemical classification. Most are ­lipophilic, have a large volume of distribution, and are generally well absorbed, although anticholinergic effects may delay absorption of some antipsychotics. Serum concentrations generally peak within 2 to 3 hours after a therapeutic dose, but this can be prolonged following overdose.

Most antipsychotics are substrates for the various isozymes of the hepatic cytochrome P450 (CYP) enzyme system. For example, haloperidol, perphenazine, thioridazine, sertindole, and risperidone are extensively metabolized by the CYP2D6 system, which is functionally absent in approximately 7% of white patients and overexpressed in 1% to 25% of patients, depending on ethnicity.⁵³ These polymorphisms appear to influence the tolerability and efficacy of treatment with these antipsychotics during therapeutic use^{15,29,30,56,124} but are unlikely to significantly alter the severity of acute antipsychotic overdose.

Drugs that inhibit CYP2D6 (such as paroxetine, fluoxetine, and bupropion) can increase the concentrations of these antipsychotics and increase the risk of adverse effects. In contrast, metabolism of clozapine is primarily mediated by CYP1A2, and increased clozapine concentrations can result following exposure to CYP1A2 inhibitors such as fluvoxamine, macrolide and fluoroquinolone antibiotics, as well as after smoking cessation, because smoking induces CYP1A2.³⁵ The kidney plays a relatively small role in the elimination of antipsychotics, and dose adjustment is generally not necessary for patients with kidney disease.

PATHOPHYSIOLOGY AND CLINICAL MANIFESTATIONS

Table 70–3 lists the adverse effects of antipsychotics. Some of these effects develop primarily following overdose, but others can occur during the course of therapeutic use.

Adverse Effects During Therapeutic Use

The Extrapyramidal Syndromes.

The EPS (Table 70–4) are a heterogeneous group of disorders that share the common feature of abnormal muscular activity. Among the typical antipsychotics, the incidence of EPS appears to be highest with the more potent antipsychotics such as haloperidol and flupentixol, and lower with less potent antipsychotics such as chlorpromazine and thioridazine. Atypical antipsychotics are associated with an even lower incidence of EPS. Although the physiologic mechanisms for this observation are not fully understood, several hypotheses have been put forth. In addition to the aforementioned antagonism of 5-HT_2A receptors, some atypical antipsychotics dissociate more rapidly from the D₂ receptor and incite a lower degree of nigrostriatal dopaminergic hypersensitivity during chronic use.^57,58,71 However, it is important to note that EPS can occur during treatment with any antipsychotic drug, regardless of typicality or potency.

Acute dystonia.

Acute dystonia is a movement disorder characterized by sustained involuntary muscle contractions, often involving the muscles of the head and neck, including the extraocular muscles and the tongue, but occasionally involving the extremities. These contractions are sometimes referred to as limited reactions, reflecting their transient nature rather than their severity. All the currently available antipsychotics are associated with the development of acute dystonic reactions.¹¹⁹ Spasmodic torticollis, facial grimacing, protrusion of the tongue, and oculogyric crisis are among the more common manifestations. Laryngeal dystonia is a rare but potentially life-threatening variant that is easily misdiagnosed because it can present with throat pain, dyspnea, stridor, and dysphonia rather than the more characteristic features of dystonia.³⁸

Acute dystonia typically develops within a few hours of starting of treatment but may be delayed for up to a few days. Left untreated, dystonia resolves slowly over several days once the offending antipsychotic is withdrawn. Risk factors for acute dystonia include male gender, young age (children are particularly susceptible), a previous episode of acute dystonia, and recent cocaine use.^120,132 Although the reaction may appear dramatic and sometimes is mistaken for seizure activity, it is rarely life threatening. Of note, drugs other than antipsychotics can sometimes cause acute dystonia, particularly metoclopramide, antidepressants, some antimalarials, histamine H₂-receptor antagonists, anticonvulsants, and cocaine.¹²⁰

Treatment of acute dystonia.

Acute dystonia is generally more distressing than serious, but rare cases compromise respiration, necessitating supplemental oxygen and, occasionally, assisted ventilation.^38,120 The response to parenteral anticholinergics often is rapid and dramatic, and every effort should be made to administer benztropine as the first-line agent (2 mg intravenously {IV} or intramuscularly {IM} in adults, or 0.05 mg/kg in children). Often, diphenhydramine is more readily available and can be used instead (50 mg IV or IM in adults, or 1 mg/kg in children). Parenteral benzodiazepines such as lorazepam (0.05–0.10 mg/kg IV or IM) or diazepam (0.1 mg/kg IV) should be considered if patients do not respond to anticholinergics, but they may also be effective as initial therapy. It is important to recognize that because the elimination half-life of most anticholinergics is shorter than that of most antipsychotics, dystonia can recur, and administering additional doses of an anticholinergic may be necessary over the next 48 to 72 hours.²⁷ Patients in whom acute dystonia jeopardizes respiration should be observed for at least 12 to 24 hours after initial resolution.

Akathisia.

Akathisia (from the Greek phrase “not to sit”) is characterized by a feeling of inner restlessness, anxiety, or sense of unease, often in conjunction with the objective finding of an inability to sit still. Patients with akathisia frequently appear uncomfortable or fidgety. They may rock back and forth while standing, or may repeatedly cross and uncross their legs while seated. Akathisia can be difficult to diagnose and is easily misinterpreted as a manifestation of the underlying psychiatric disorder rather than an adverse effect of therapy.

Akathisia is common and may be an important determinant of adherence to therapy. Like acute dystonia, akathisia tends to occur relatively early in the course of treatment and coincides with peak antipsychotic concentrations in serum.¹³² The incidence appears highest with typical, high-potency antipsychotics and lowest with atypical antipsychotics. Although most cases develop within days to weeks after initiation of treatment or an increase in dose, a delayed-onset (tardive) variant is also recognized.

The pathophysiology of akathisia is incompletely understood but appears to involve antagonism of postsynaptic D₂ receptors in the mesocortical pathways.^71,119 Interestingly, a similar phenomenon is described in patients following the initiation of treatment with antidepressants, particularly the selective serotonin reuptake inhibitors (Chap. 75).^8,67

Treatment of akathisia.

Akathisia can be difficult to treat. A reduction in the antipsychotic dose is sometimes helpful, as is substitution of another (generally atypical) antipsychotic. Treatment with lipophilic β-adrenergic antagonists such as propranolol may reduce the symptoms of akathisia, but little evidence supports their use.^65,88 Benzodiazepines produce short-term relief, and anticholinergics such as benztropine or procyclidine may reduce manifestations of akathisia, but they are more likely to be effective for akathisia induced by antipsychotics with little or no intrinsic anticholinergic activity.^20,66

Parkinsonism.

Antipsychotics can produce a parkinsonian syndrome characterized by rigidity, akinesia or bradykinesia, and postural instability. It is similar to the idiopathic Parkinson disease, although the classic “pill-rolling” tremor is often less pronounced.⁸⁸ The syndrome typically develops during the first few months of therapy, particularly with high-potency antipsychotics. It is more common among older women, and in some patients it may represent iatrogenic unmasking of latent Parkinson disease. Parkinsonism is thought to result from antagonism of postsynaptic D₂ receptors in the striatum.¹¹⁹

Treatment of drug-induced parkinsonism.

The risk of drug-induced parkinsonism can be minimized by using the lowest effective dose of antipsychotic. The addition of an anticholinergic often attenuates symptoms, at the expense of additional side effects. This strategy often is effective in younger patients, although the routine use of prophylactic anticholinergics is not recommended. Addition of a dopamine agonist such as amantadine is sometimes used, particularly in older patients who may be less tolerant of anticholinergics, but this may aggravate the underlying psychiatric disturbance.⁶⁹

Tardive dyskinesias.

The term tardive dyskinesia was coined in 1952 to describe the delayed onset of persistent orobuccal masticatory movements occurring in three women after several months of antipsychotic therapy.¹¹⁹ The adjective tardive, meaning delayed, was used to distinguish these movement disorders from the Parkinsonian movements described above. The incidence of tardive dyskinesia in younger patients is approximately 3% to 5% per year but rises considerably with age. A prospective study of older patients treated with high potency typical antipsychotics identified a 60% cumulative incidence of tardive dyskinesia after 3 years of treatment.⁵³ Potential risk factors for tardive dyskinesia include alcohol use, affective disorder, prior electroconvulsive therapy, diabetes mellitus, and various genetic factors.¹¹⁹

Several distinct tardive syndromes are recognized, including the classic orobuccal lingual masticatory stereotypy, chorea, dystonia, myoclonus, blepharospasm, and tics. It is generally accepted that the atypical antipsychotics are associated with a lower incidence of tardive dyskinesia and other drug-related movement disorders. However, whether this is true of all atypical antipsychotics is unclear. Among the atypical antipsychotics, clozapine is associated with the lowest incidence of tardive dyskinesia and risperidone with the highest incidence (when higher doses are used), but the reasons for this observation are uncertain.^115,116,119

Treatment of tardive dyskinesia.

Tardive dyskinesia is highly resistant to the usual pharmacologic treatments for movement disorders. Anticholinergics do not alleviate tardive dyskinesia and indeed may worsen it. Calcium channel blockers, β-adrenergic antagonists, benzodiazepines, and vitamin E have all been used with limited success.³⁶ Clozapine appears to suppress tardive dyskinesia temporarily. Although discontinuation of the causative antipsychotic may not produce total relief of symptoms, when possible, the antipsychotic should be discontinued as soon as signs or symptoms begin.

Neuroleptic malignant syndrome.

Neuroleptic malignant syndrome (NMS) is a potentially life-threatening drug-induced emergency. First described in 1960 in patients treated with haloperidol, this syndrome has been associated with virtually every antipsychotic.³² The reported incidence of NMS ranges from 0.2% to 1.4% of patients receiving antipsychotics,^2,23,114 but less severe episodes may go undiagnosed or unreported. As a result, much of what is known about the epidemiology and treatment of NMS is speculative and based upon case reports and case series.

The pathophysiology of NMS is incompletely understood but appears to involve abrupt reductions in central dopaminergic neurotransmission in the striatum and hypothalamus, altering the core temperature setpoint,⁴³ and leading to impaired thermoregulation and other manifestations of autonomic dysfunction. Blockade of striatal D₂ receptors contributes to muscle rigidity and tremor.^13,25,121 In some cases, a direct effect on skeletal muscle may play a role in the pathogenesis of hyperthermia.⁴³ Altered mental status is multifactorial and may reflect hypothalamic and spinal dopamine receptor antagonism, a genetic predisposition, or the direct effects of hyperthermia and other drugs.⁴⁴ Serotonin also appears to play a role in the pathogenesis of NMS, because antipsychotics that antagonize 5-HT_2A receptors seem to be associated with a lower incidence of NMS.⁴

Although NMS most often occurs during treatment with a D₂ receptor antagonist, withdrawal of dopamine agonists can produce an indistinguishable syndrome. The latter typically occurs in patients with long-standing Parkinson disease who abruptly change or discontinue treatment with dopamine agonists such as levodopa/carbidopa, amantadine, or bromocriptine.¹³ The resulting disorder is sometimes referred to as the parkinsonian-hyperpyrexia syndrome, and mortality rates of up to 4% are reported.⁷⁹ Hospitalization for aspiration pneumonia, a common occurrence in older patients with Parkinson disease, is a particularly high-risk setting for this complication, and is particularly dangerous because the cardinal mani­festations of NMS are easily misattributed to the combined effects of pneumonia and the underlying movement disorder.

The vast majority of NMS cases occur in the context of therapeutic use of antipsychotics rather than following overdose. Postulated risk ­factors for the development of NMS include young age, male gender, extracellular fluid volume contraction, use of high-potency antipsychotics, depot preparations, cotreatment with lithium, multiple drugs in combination, and rapid dose escalation.^2,24,63,81 One large observational study⁸¹ suggests that treatment with high-potency first-generation antipsychotics is associated with a more than 20-fold increase in the risk of NMS, although this may partly reflect heightened suspicion of the disorder in patients receiving those antipsychotics. The mortality rate of NMS associated with first-generation antipsychotics is estimated at approximately 16%, whereas the rate associated with second-generation antipsychotics is estimated at 3%.¹¹⁸

The manifestations of NMS include the tetrad of altered mental status, muscular rigidity (classically described as “lead pipe”), hyperthermia, and autonomic dysfunction. These symptoms can appear in any sequence, although a review of 340 NMS cases found that mental status changes and rigidity usually preceded the development of hyperthermia and autonomic instability.¹²² Occasionally, rigidity is not present when creatine kinase concentrations are elevated but emerges thereafter.⁸² Signs typically evolve over a period of several days, with the majority occurring within 2 weeks of initiation. However, it is important to recognize that NMS can occur even after prolonged use of an antipsychotic, particularly following a dose increase, the addition of another agent, or the development of intercurrent illness. It is also worth noting that the clinical course of NMS can fluctuate rapidly, sometimes waxing and waning dramatically over a few hours.

There is no universally accepted set of criteria for the diagnosis of NMS, and more than a dozen sets of criteria have been proposed.^3,23,34,63 The operating characteristics of these criteria have not been formally evaluated, in part because of the absence of a gold standard. An international group has published the results of a Delphi consensus panel regarding the diagnosis of NMS.⁴² While these too have not yet been validated, the criteria and their relative importance are shown in Table 70–5.

It may be difficult to distinguish NMS from other toxin-induced hyperthermia syndromes, such as the anticholinergic (antimuscarinic) syndrome (Chap. 49) and serotonin toxicity (Chap. 75), all of which share the common features of elevated temperature, altered mental status, and neuromuscular abnormalities. The most important differentiating feature is the medication history, with dopamine antagonists, antimuscarinics, and direct or indirect serotonin agonists (often in combination) as the most likely causal agents, respectively. The time course of the illness may also help differentiate among the disorders. Serotonin toxicity and the antimuscarinic syndrome tend to develop rapidly after exposure to causative substances, while NMS typically develops more gradually, often waxing and waning over several days or more. Occasionally, clinicians must attempt to differentiate NMS from these disorders in the absence of a reliable medication history. The physical examination can be of some utility in this regard.⁸⁷ While NMS is classically characterized by rigidity, the presence of ocular or generalized clonus is more suggestive of serotonin toxicity, particularly when accompanied by shivering and hyperreflexia, features that are not typical of NMS. Because skeletal muscle contraction is effected by nicotinic rather than muscarinic transmission, patients with antimuscarinic syndrome have few muscular abnormalities. However, such patients can sometimes be resistant to physical restraint, giving the appearance of increased muscle tone.

Treatment of neuroleptic malignant syndrome: General measures.

Treatment recommendations are largely based on general physiologic principles, case reports, and case series. Therapy should be individualized according to the severity and duration of illness and the modifying influences of comorbidity.^13,95,123 The provision of good supportive care is the cornerstone for treatment of NMS. It is essential to recognize the condition as an emergency and to withdraw the offending antipsychotic immediately. When NMS ensues after abrupt discontinuation of a dopamine agonist such as levodopa, the drug should be reinstituted promptly. Most patients with NMS should be admitted to an intensive care unit. Supplemental oxygen should be administered, and assisted ventilation may be necessary in cases of respiratory failure, which can result from central hypoventilation, loss of protective airway reflexes, or rigidity of the chest wall muscles.

The hyperthermia associated with NMS is multifactorial in origin and, when present, warrants aggressive treatment. Submersion in an ice-water bath is rapidly effective, although this may be impractical in some settings (Chap. 30). Other strategies include the use of active cooling blankets, the placement of ice packs in the groin and axillae, or evaporative cooling, which can be accomplished by removing the patient’s clothing and exposing the patient to cooled water or towels immersed in ice water, while maintaining constant air circulation with the use of fans.¹²⁷

Hypotension should be treated initially with generous volumes of isotonic crystalloid, followed by vasopressors if necessary. Alkalinization of the urine with sodium bicarbonate may reduce the incidence of myoglobinuric acute kidney injury (AKI) in patients with high creatine kinase concentrations, but maintenance of intravascular volume and adequate renal perfusion are of far greater importance. Tachycardia does not require specific treatment, but bradycardia may necessitate the use of transcutaneous or transvenous electrical pacing. Venous thromboembolism is a major cause of morbidity and mortality in patients with NMS, and prophylactic doses of low-molecular-weight heparin should be considered in patients who likely will be immobilized for more than 12 to 24 hours.

Treatment of neuroleptic malignant syndrome: Pharmacologic measures.

Benzodiazepines are the most widely used pharmacologic adjuncts for treatment of NMS and are considered first-line therapy. Dantrolene and bromocriptine are not well studied, and their incremental benefit over good supportive care is debated.^95,100 However, these drugs are associated with relatively little toxicity, and their use is therefore easily justified, particularly in patients with moderate or severe NMS.

Benzodiazepines are frequently used in the management of NMS because of their rapid onset of action, which is particularly important when patients are agitated or restless. They attenuate the sympathetic hyperactivity that characterizes NMS by facilitating GABA-mediated chloride transport and producing neuronal hyperpolarization, in a fashion analogous to their beneficial effects in cocaine toxicity.⁴⁴ The primary disadvantage of benzodiazepines is that they may cloud the assessment of mental status.

Dantrolene reduces skeletal muscle activity by inhibiting ryanodine receptor calcium release channels, interfering with calcium release from the sarcoplasmic reticulum. In theory, this process should reduce body temperature and total oxygen consumption. It also should lessen the risk of myoglobinuric AKI. Dantrolene has been suggested to be particularly useful when muscular rigidity is a prominent feature of NMS.¹³ It can be given by mouth (50–100 mg/d) or by IV infusion (2–3 mg/kg/d, or up to 10 mg/kg/d in severe cases), although the latter requires laborious reconstitution. A review of 271 published cases of NMS that included information regarding drug treatment found that combination therapy including dantrolene was associated with a prolonged clinical recovery, but also that dantrolene monotherapy was associated with higher mortality than other treatment modalities including supportive care.⁹⁵ The authors concluded that dantrolene was not an evidence-based therapy. However, it is also a relatively nontoxic drug that remains a reasonable therapeutic option in patients with NMS, particularly those with prominent and refractory rigidity.

Bromocriptine is a centrally acting dopamine agonist that is given orally or by nasogastric tube at doses of 2.5 to 10 mg, 3 to 4 times daily. The rationale for its use rests in the belief that reversal of antipsychotic-related striatal D₂ antagonism will ameliorate the manifestations of NMS. Other dopamine agonists anecdotally associated with success include ropinirole, levodopa,^84,110 and amantadine.^41,52,114 When these drugs are used, they should be tapered slowly after the patient improves to minimize the likelihood of recrudescent NMS. In severe cases, dantrolene and a dopamine agonist can be used in combination. Of note, dopaminergic agents may be associated with exacerbation of underlying psychiatric illness.

Electroconvulsive therapy.

Electroconvulsive therapy (ECT) has been reported to dramatically improve the manifestations of NMS, presumably by enhancing central dopaminergic transmission. In one report, five patients received an average of 10 ECT treatments, and resolution generally occurred after the third or fourth session.⁸³ Whether this result represents a true effect of ECT or simply the natural course of NMS with good supportive care alone is not clear. As with drug therapies for NMS, the efficacy of ECT remains unclear and its indications speculative, but its use seems reasonable in patients with severe, persistent, or treatment-resistant NMS and for those with residual catatonia or psychosis following resolution of other manifestations.^13,84

Adverse Effects on Other Organ Systems.

Sedation, dry mouth, and urinary retention occur commonly with antipsychotics, particularly during the initial period of therapy. These symptoms occur most commonly with drugs having potent antihistaminic and antimuscarinic activity. All antipsychotics can lower the seizure threshold, but seizures are uncommon during therapeutic use. Because hypothalamic dopamine inhibits pituitary prolactin release, hyperprolactinemia and galactorrhea can occur. All antipsychotics are associated with metabolic derangements, including weight gain, dyslipidemia, and steatohepatitis. The metabolic syndrome appears most commonly in association with clozapine, olanzapine, and chlorpromazine.³¹ Rare but dramatic instances of glucose intolerance, including fatal cases of diabetic ketoacidosis, are also described.^6,47,91,117 The mechanism of this is incompletely understood, but it is not adequately explained by the weight gain associated with antipsychotic therapy, because glucose disturbances often develop shortly after therapy is instituted. Other idiosyncratic reactions reported with use of antipsychotics include photosensitivity, skin pigmentation, and cholestatic hepatitis (particularly with the phenothiazines), myocarditis, and agranulocytosis, which occurs with many antipsychotic drugs, most notably clozapine (in up to 2% of patients).⁷⁵ Most of these conditions result from an immunologically based hypersensitivity reaction and develop during the first month of therapy. Finally, an increasing number of reports associate antipsychotics with venous thromboembolism.^46,55 This may partially explain the high incidence of thromboembolic disease found in patients with NMS (see above).

Acute Overdose

Antipsychotic overdose can produce a spectrum of toxic manifestations affecting multiple organ systems, but most serious toxicity involves the CNS and cardiovascular systems. Some of these manifestations are present to a minor degree during therapeutic use, although they tend to be most pronounced during the early period of therapy and dissipate with continued use.

Impaired consciousness is a common and dose-dependent feature of antipsychotic overdose, ranging from somnolence to coma. It may be associated with impaired airway reflexes, but significant respiratory depression is uncommon. Many antipsychotics, including several of the atypical drugs, are potent muscarinic antagonists and can produce anticholinergic features in overdose.^11,21,26 Peripheral manifestations include tachycardia, decreased production of sweat and saliva, flushed skin, urinary retention, diminished bowel sounds and mydriasis, although miosis also occurs. These findings may be present in isolation or coexist with central manifestations, including agitation, delirium, psychosis, hallucinations, and coma, some of which may be mistakenly attributed to the underlying psychiatric illness.

Mild elevations in body temperature are common and reflect impaired heat dissipation as a result of impaired sweating, and increased heat production in agitated patients. Hyperthermia should always prompt a search for other features of NMS. Tachycardia is a common finding in patients with antipsychotic overdose and reflects reduced vagal tone and, when present, a compensatory response to hypotension. Bradycardia is distinctly uncommon, and while it may be a preterminal finding, its presence should prompt a search for alternate causes including an ingestion of negative chronotropic drugs such as β-adrenergic antagonists, calcium channel blockers, cardiac glycosides, and opioids. Hypotension is a common feature of antipsychotic overdose and is generally due to peripheral α₁-adrenergic blockade and decreased myocardial contractility.

The electrocardiographic (ECG) manifestations of antipsychotic overdose exhibit similarities to those of cyclic antidepressant toxicity (Chaps. 16 and 71) and include widening of the QRS complex and a rightward deflection of the terminal 40 msec of the QRS complex, typically manifesting as a tall, broad terminal component of the QRS complex in lead aVR. These changes reflect blockade of the inward sodium current (I_Na). Prolongation of the QT interval results from blockade of the delayed rectifier potassium current (I_Kr), creating a substrate for development of torsade de pointes.⁷⁸ This situation is sometimes evident during maintenance therapy and may underlie the apparent increase in sudden cardiac death among users of antipsychotics.^92,93 A published meta-analysis of the operating characteristics of the ECG in patients with cyclic antidepressant toxicity found the ECG was a relatively poor predictor of seizures, dysrhythmia, and death.⁷ However, the ECG is a dynamic instrument, particularly in the initial hours following overdose, and few studies have evaluated longitudinal changes in the ECG.⁶⁴

DIAGNOSTIC TESTS

The diagnosis of antipsychotic poisoning is supported by the clinical history, the physical examination, and a limited number of adjunctive tests. Both the clinical and ECG findings are nonspecific and can occur following overdose of several different drug classes, including CAs, skeletal muscle relaxants, carbamazepine, and first-generation antihistamines such as diphenhy­dramine. Moreover, the absence of typical ECG changes does not exclude a significant antipsychotic ingestion, particularly in the initial phase of an overdose, and at least one additional ECG should be performed in the ­subsequent 2 to 3 hours.

Abdominal radiography may reveal radioopacities in the gastrointestinal tract, as some solid dosage forms of phenothiazines are radiopaque. However, these tests are neither sensitive nor specific, and they are not routinely recommended.

Serum concentrations of antipsychotics are not widely available, do not correlate well with clinical signs and symptoms, and do not help guide therapy. Comprehensive urine drug screens using high-performance liquid chromatography, gas chromatography–mass spectrometry, or tandem mass spectrometry may indicate the presence of antipsychotics, but these tests are available at only a few hospitals and in most instances provide only a qualitative result. Blood and urine immunoassays for CAs may yield a false-positive result in the presence of phenothiazines.^5,99

MANAGEMENT

The care of a patient with an antipsychotic overdose should proceed with the recognition that other drugs, particularly other psychotropics, may have been coingested and can confound both the clinical presentation and management. Regularly encountered coingestants include other psychotropic drugs such as antidepressants, sedative-hypnotics, anticholinergics, valproic acid, and lithium, as well as ethanol and nonprescription analgesics such as acetaminophen and aspirin.

Supportive care is the cornerstone of treatment for patients with antipsychotic overdose. Supplemental oxygen should be administered if hypoxia is present. Patients with altered mental status should receive thiamine and naloxone, along with parenteral dextrose if hypoglycemia present. Intubation and ventilation are rarely required for patients with single drug ingestions but may be necessary for patients with very large overdoses of antipsychotics or ingestion of other CNS depressants. All symptomatic patients should undergo continuous cardiac monitoring. In addition, an ECG should be recorded upon presentation and reliable venous access obtained. Asymptomatic patients with a normal ECG obtained 6 hours after overdose are at exceedingly low risk of complications and no longer require cardiac monitoring. Symptomatic patients and those with an abnormal ECG should have continuous monitoring for a minimum of 24 hours.

Gastrointestinal Decontamination

Gastrointestinal decontamination with activated charcoal (1 g/kg by mouth or nasogastric tube) should be considered for patients who present within a few hours of a large or multidrug overdose. Although this intervention is time sensitive, many antipsychotics exhibit significant antimuscarinic activity and slow gastric emptying, thereby increasing the likelihood that activated charcoal will be beneficial. Although it is unknown whether activated charcoal improves clinically important outcomes, a Bayesian analysis of pharmacokinetic data from a series of quetiapine overdoses concluded that activated charcoal use led to a 35% reduction in the fraction of quetiapine absorbed.⁵⁰ Orogastric lavage and whole-bowel irrigation likely will not improve clinical outcomes and should not be routinely employed in the management of antipsychotic overdose.

Treatment of Cardiovascular Complications

Vital signs should be monitored closely. Hypotension may result from peripheral α-adrenergic blockade and is most likely to occur with older, low potency antipsychotics such as thioridazine.⁷⁶ The hypotension should be treated initially with appropriate titration of 0.9% sodium chloride solution (30–40 mL/kg). If vasopressors are required, direct-acting agonists such as norepinephrine or phenylephrine are preferred over dopamine, which is an indirect agonist and likely will be ineffective. Vasopressin or its analogs may also be used, although direct-acting vasopressors should be used with great caution in patients who have coingested a negative inotropic drug such as a β-adrenergic antagonist or calcium channel blocker. Continuous blood pressure monitoring may be warranted in such cases.

Progressive widening of the QRS complex is uncommon and reflects sodium channel blockade and slowing of phase 0 depolarization in the His-Purkinje system. This may be associated with reduced cardiac output and malignant ventricular dysrhythmias. Much of what is known about the treatment of sodium channel blocker toxicity derives from the cyclic antidepressant literature. Treatment recommendations are extended to sodium-channel-blocking antipsychotic drugs by analogy. Sodium bicarbonate (1–2 mEq/kg) is the first-line therapy for ventricular dysrhythmias and should be considered for patients with dysrhythmias or QRS widening >0.12 msec. The rationale for this strategy is based upon the treatment of cyclic antidepressant overdose (Antidotes in Depth: A5). At least two mechanisms underlie the beneficial effects of sodium bicarbonate. First, the degree of sodium channel blockade is partially overcome by an increase in extracellular sodium. Indeed, hypertonic saline alone may be beneficial. Second, the binding of these drugs to the sodium channel is pH dependent, with less extensive binding at higher pH.

Repeated boluses of bicarbonate can be given to achieve a target blood pH of 7.5, although many toxicologists recommend continuous infusions.¹⁰⁹ If the patient is intubated, hyperventilation also can be used but is not comparably efficacious. If significant conduction abnormalities or ventricular dysrhythmias persist despite the use of sodium bicarbonate, lidocaine (1–2 mg/kg followed by continuous infusion) is a reasonable second-line antidysrhythmic. Although lidocaine is also a sodium channel antagonist, it exhibits rapid on/off sodium channel binding with preferential binding in the inactivated state and may lessen the cardiotoxicity associated with antipsychotic drug overdose.¹⁰⁸ Class IA antidysrhythmics (procainamide, disopyramide, and quinidine), class IC antidysrhythmics (propafenone, encainide, and flecainide), and class III antidysrhythmics (amiodarone, sotalol and bretylium) can aggravate cardiotoxicity and should not be used. When administering sodium bicarbonate to patients with antipsychotic overdose, caution must be taken to avoid hypokalemia, as many of these antipsychotics block cardiac potassium channels thereby prolonging the QT interval. Hypokalemia can exacerbate this blockade, potentially leading to torsade de pointes, particularly in overdoses involving amisulpride or ziprasidone.

Sinus tachycardia related to anticholinergic activity should not be treated unless it is associated with active ischemia, which, although uncommon, may complicate antipsychotic overdose in patients with existing coronary disease. Should symptomatic sinus tachycardia occur, a short-acting β-adrenergic antagonist such as esmolol may be preferable. Prolongation of the QT interval requires no specific treatment other than monitoring and correction of potential contributing causes such as hypokalemia and ­hypomagnesemia. Torsade de pointes should be treated with cardioversion followed by IV magnesium sulfate, taking care to prevent hypotension, which is dose- and rate-dependent. Overdrive pacing with isoproterenol or transcutaneous or transvenous pacing should be considered if the patient does not respond to magnesium sulfate, although in theory this therapy may worsen the rate-dependent sodium channel blockade.

Many antipsychotics, including olanzapine, quetiapine, and sertindole, exhibit a high degree of lipophilicity in addition to significant cardiovascular toxicity. Recently, considerable enthusiasm has emerged for the use of high-dose intravenous fat emulsion therapy for patients with significant overdoses of drugs displaying these characteristics. The rationale for this therapy rests, in part, in the concept that highly lipophilic drugs will selectively partition into the exogenous fat, thereby minimizing toxicity at the biophase. This treatment has been extensively studied in animal models of bupivacaine toxicity,^33,125,126 but published experience with antipsychotics is limited to a handful of case reports.^39,72,133 Dosing for fat “rescue” is not well established, but a popular protocol involves 20% lipid emulsion given as a bolus (1.5 mL/kg) followed by an infusion of 0.25 mL/kg/min for 30 to 60 minutes, with adjustment of therapy according to the individual response (Antidotes in Depth: A20). Extracorporeal circulatory support has been associated with survival in severe quetiapine overdose and may be an option in selected centers.⁶²

Treatment of Seizures

Seizures associated with antipsychotic overdose are generally short lived and often require no pharmacologic treatment. Multiple or refractory seizures should prompt a search for other causes, including hypoglycemia and ingestion of other proconvulsant medications. When treatment is necessary, benzodiazepines such as lorazepam or diazepam generally suffice, although phenobarbital may be necessary. Although phenytoin is part of the standard algorithm for status epilepticus, it is of limited effectiveness for xenobiotic-induced seizures; barbiturates such as phenobarbital or thiopental are preferred. Refractory seizures should respond to propofol infusion or general anesthesia. Seizures complicated by hyperthermia are considerably more ominous and warrant aggressive lowering of body temperature with aggressive rapid cooling measures. Finally, seizures can abruptly lower serum pH and may abruptly increase the cardiotoxicity of these drugs by enhancing binding to the sodium channel; therefore, an ECG should be obtained following resolution of seizure activity.

Treatment of the Central Antimuscarinic Syndrome

Many of the older and newer-generation antipsychotics have pronounced anticholinergic properties. Case reports and observational studies suggest that the cholinesterase inhibitor physostigmine (Antidotes in Depth: A9) can safely and effectively ameliorate the agitated delirium associated with the central anticholinergic syndrome by indirectly increasing synaptic acetylcholine concentraions.^{25,102, 103, and 104} Doses of 1 to 2 mg are usual, administered cautiously after ensuring QRS widening is not present. Although benzodiazepines will control agitation, they will further impair alertness, obfuscating the assessment of mental status and increasing the risk of complications.²²

Physostigmine has been used successfully in patients with antipsychotic overdose,^{22,101,104,128,129} but it should be used with caution. It should not be used in patients with dysrhythmias, any degree of heart block, or widening of the QRS complex. If physostigmine is used, it should be given in 0.5 mg increments slowly every 3 to 5 minutes, with close observation of the patient. If bradycardia, bronchospasm, or bronchorrhea develops, these can be treated with glycopyrrolate 0.2 to 0.4 mg IV. Atropine is often more widely available and could be used, but it crosses the blood–brain barrier and may aggravate any associated delirium. The effects of physostigmine are transient, typically ranging in duration from 30 to 90 minutes, and additional doses are often necessary. Of note, physostigmine does not prevent other complications of antipsychotic overdose, particularly those involving the cardiovascular system.

Other commonly used cholinesterase inhibitors, such as edrophonium, neostigmine, or pyridostigmine, should not be used to treat anticholinergic delirium because they do not cross the blood–brain barrier. Case reports involving other anticholinergics suggest that cholinesterase inhibitors used for treatment of dementia (eg, tacrine, donepezil, and galantamine) may be alternatives to physostigmine for patients able to take medications orally.^51,74,85

Enhanced Elimination

No pharmacologic rationale supports the use of multiple-dose activated charcoal or manipulation of urinary pH to increase the clearance of antipsychotics. One volunteer study found that urinary acidification may increase remoxipride elimination,¹³¹ but this practice is impractical and possibly dangerous. Because most antipsychotics exhibit large volumes of distribution and extensive protein binding (Table 70–1), extracorporeal removal is unwarranted and should be considered only if the patient has coingested other xenobiotics amenable to extracorporeal removal.

SUMMARY

• Over the past decade, atypical antipsychotics have supplanted traditional antipsychotic drugs, which were associated with greater toxicity in overdose and a higher incidence of extrapyramidal symptoms. Consequently, atypical antipsychotics are now implicated in the majority of overdoses.

• With all antipsychotics, significant toxicity can occur either during the course of therapy or following overdose.

• Of the various toxicities that arise during therapeutic use, NMS is the most dangerous. It may be difficult to recognize, but should be considered in any unwell patient treated with antipsychotics, particularly in the 2 weeks following a change in therapy or in a patient with severe intercurrent illness.

• The principal manifestations of antipsychotic overdose involve the CNS and cardiovascular system. Depressed mental status, hypotension, and anticholinergic signs are nonspecific features that support the diagnosis of antipsychotic overdose, particularly in conjunction with typical ECG findings of sodium channel blockade and QT interval prolongation. Most fatalities following antipsychotic overdose occur in cases involving coingestion of other CNS depressants or cardiotoxic medications.

• Supportive care is the mainstay of therapy for patients with antipsychotic overdose, although selective use of nonspecific antidotes, such as activated charcoal, sodium bicarbonate, or physostigmine may improve outcomes in selected patients. Particularly severe or refractory cardiovascular toxicity may warrant a trial of lidocaine or intravenous lipid emulsion, although these interventions are not well studied in this setting.

Acknowledgments

Frank LoVecchio, MD, and Neal A. Lewin, MD, contributed to this chapter in previous editions.

References

HISTORY AND EPIDEMIOLOGY

The term cyclic antidepressant (CA) refers to a group of pharmacologically related xenobiotics used for treatment of depression, neuralgic pain, migraines, enuresis, and attention deficit hyperactivity disorder. Most CAs have at least three rings in their chemical structure. They include the traditional tricyclic antidepressants (TCAs) imipramine, desipramine, amitriptyline, nortriptyline, doxepin, trimipramine, protriptyline, and clomipramine, as well as other cyclic compounds such as maprotiline and amoxapine.

Imipramine was the first TCA used for treatment of depression in the late 1950s. However, the synthesis of iminodibenzyl, the “tricyclic” core of imipramine, and the description of its chemical characteristics date back to 1889. Structurally related to the phenothiazines, imipramine was originally developed as a hypnotic for agitated or psychotic patients and was serendipitously found to alleviate depression. From the 1960s until the late 1980s, the TCAs were the major pharmacologic treatment for depression in the United States. However, by the early 1960s, cardiovascular and central nervous system (CNS) toxicities were recognized as major complications of TCA overdoses. The newer CAs developed in the 1980s and 1990s were designed to decrease some of the adverse effects of older TCAs, improve the therapeutic index, and reduce the incidence of serious toxicity. Other CAs include the tetracyclic drug maprotiline and the dibenzoxapine drug amoxapine.

The epidemiology of CA poisoning has evolved significantly in the past 30 years, resulting in great part from the introduction of the selective serotonin reuptake inhibitors (SSRIs) and other newer antidepressants for the treatment of depression. Although the use of CAs for depression has decreased over the past 20 years, other medical indications, including chronic pain, obsessive-compulsive disorder, and, particularly in children, enuresis and attention deficit hyperactivity disorder have emerged, resulting in their continued use. The antidepressants are a leading cause of drug-related self-poisonings in the developed world, primarily because of their ready availability to people with depression or chronic pain who by virtue of their diseases are at high risk for overdose. However, despite the increase in SSRI use and overdose, patients with TCA overdoses continue to have higher rates of hospitalization and fatality than do those with SSRI overdose.

Children younger than 6 years have consistently accounted for approximately 12% to 13% of all CA exposures reported to poison centers during each of the last 15 years (Chap. 136). Despite the emergence of the SSRIs in the early 1990s, TCAs are still frequently prescribed by pediatric office based practices for many of the conditions noted above. Following the October 2004 US Food and Drug Administration Black Box Warning about the increased risk of suicidal behavior associated with antidepressant use, several reports have described significant declines in antidepressant dispensing in children compared to historical trends.²⁵ Nevertheless, CA poisoning likely will continue to be among the most lethal unintentional drug ingestions in younger children because only one or two adult-strength pills can produce serious clinical effects in young children.

PHARMACOLOGY

In general, the TCAs can be classified into tertiary and secondary amines based on the presence of a methyl group on the propylamine side chain (Table 71–1). The tertiary amines amitriptyline and imipramine are metabolized to the secondary amines nortriptyline and desipramine, respectively, which themselves are marketed as antidepressants. In therapeutic doses, the CAs produce similar pharmacologic effects on the autonomic system, CNS, and cardiovascular system. However, they can be distinguished from each other by their relative potencies.¹¹²

At therapeutic doses, CAs inhibit presynaptic reuptake of norepinephrine and/or serotonin, thus functionally increasing the amount of these neurotransmitters at CNS receptors. The tertiary amines, especially clomipramine, are more potent inhibitors of serotonin reuptake, whereas the secondary amines are more potent inhibitors of norepinephrine reuptake. Although these pharmacologic actions formed the basis of the monoamine hypothesis of depression in the 1960s, antidepressant actions of these drugs appear to be much more complex.

Extensive research has led to the “receptor sensitivity hypothesis of antidepressant drug action,” which postulates that following chronic CA administration, alterations in the sensitivity of various receptors are responsible for antidepressant effects. Chronic TCA administration alters the number and/or function of central β-adrenergic and serotonin receptors. In addition, TCAs modulate glucocorticoid receptor gene expression and cause alterations at the genomic level of other receptors.⁸ All of these actions likely play a role in the antidepressant effects of TCAs.

Additional pharmacologic mechanisms of CAs are responsible for their side effects with therapeutic dosing and clinical effects following overdose. All of the CAs are competitive antagonists of the muscarinic acetylcholine receptors, although they have different affinities. The CAs also antagonize peripheral α₁-adrenergic receptors. The most prominent effects of CA overdose result from binding to the cardiac sodium channels, which is also described as a membrane-stabilizing effect (Fig. 71–1) (Chap. 16). The tricyclic antidepressants are potent inhibitors of both peripheral and central postsynaptic histamine receptors. Finally, animal research demonstrates that the CAs interfere with chloride conductance by binding to the picrotoxin site on the γ-aminobutyric acid (GABA)–chloride complex.¹⁰²

Amoxapine is a dibenzoxapine CA derived from the active antipsychotic loxapine. Although it has a three-ringed structure, this drug has little similarity to the other tricyclics. It is a potent norepinephrine reuptake inhibitor, has no effect on serotonin reuptake, and blocks dopamine receptors. Maprotiline is a tetracyclic antidepressant that predominantly blocks norepinephrine reuptake. Both of these CAs have a slightly different toxic profile than the traditional TCAs.^55,56,112

PHARMACOKINETICS AND TOXICOKINETICS

The CAs are rapidly and almost completely absorbed from the gastrointestinal (GI) tract, with peak concentrations 2 to 8 hours after administration of a therapeutic dose. They are weak bases (high pK_a). In overdose, the decreased GI motility caused by anticholinergic effects and ionization in gastric acid delay CA absorption. Because of extensive first-pass metabolism by the liver, the oral bioavailability of CAs is low and variable, although metabolism may become saturated in overdose, increasing bioavailability.

The CAs are highly lipophilic and possess large and variable volumes of distribution (15–40 L/kg). They are rapidly distributed to the heart, brain, liver, and kidney, where the tissue to plasma ratio generally exceeds 10:1. The octanol/water partition coefficient (Log P) is an often cited measure of lipid-solubility with the Log D representing Log P at physiological pH—a more representative measure. The latter pharmacologic property becomes important when evaluating the potential effectiveness of lipid emulsion therapy for CA toxicity. Some examples of Log D values for CAs are amitriptyline, 3.96; nortriptyline, 2.86; imipramine, 2.06; desipramine, 1.05; and doxepin, 2.93.

Less than 2% of the ingested dose is present in blood several hours after overdose, and serum CA concentrations decline biexponentially. The CAs are extensively bound to α₁-acid glycoprotein (AAG) in the plasma, although differential binding among the specific CAs is observed.² Changes in AAG concentration or pH can alter binding and the percentage of free or unbound drug.^87,95 Specifically, a low blood pH (which often occurs in a severely poisoned patient) may increase the amount of free drug, making it more available to exert its effects. This property serves as one basis for alkalinization therapy (see below).

The CAs undergo demethylation, aromatic hydroxylation, and glucuronide conjugation of the hydroxy metabolites. The tertiary amines imipramine and amitriptyline are demethylated to desipramine and nortriptyline, respectively. The hydroxy metabolites of both tertiary and secondary amines are pharmacologically active and may contribute to toxicity. The glucuronide metabolites are inactive.

Genetically based differences in the activity of the CYP2D6 enzymes, which are responsible for hydroxylation of imipramine and desipramine, account for wide interindividual variability in metabolism and steady-state serum concentrations.¹⁹“Poor metabolizers” may recover more slowly from an overdose or demonstrate toxicity with therapeutic dosing.¹⁰⁶ The metabolism of CAs also may be influenced by concomitant ingestion of ethanol and other medications that induce or inhibit the CYP2D6 isoenzyme (Chap. 13, Appendix). Patient variables such as age and ethnicity also affect CA metabolism.

Elimination half-lives for therapeutic doses of CAs vary from 7 to 58 hours (54–92 hours for protriptyline), with even longer half-lives in the elderly. The half-lives may also be prolonged following overdose as a result of saturable metabolism. A small fraction (15%–30%) of CA elimination occurs through biliary and gastric secretion. The metabolites are then reabsorbed in the systemic circulation, resulting in enterohepatic and enterogastric recirculation and reducing their fecal excretion. Finally, less than 5% of CAs are excreted unchanged by the kidney.

PATHOPHYSIOLOGY

The CAs slow the recovery from inactivation of the fast sodium channel, slowing phase 0 depolarization of the action potential in the distal His-Purkinje system and the ventricular myocardium (Fig. 71–1 and Fig. 22–2; Chap. 16). Impaired depolarization within the ventricular conduction system slows the propagation of ventricular depolarization, which manifests as prolongation of the QRS interval on the electrocardiogram (ECG) (Fig. 71–1). The right bundle branch has a relatively longer refractory period, and it is affected disproportionately by xenobiotics that slow intraventricular conduction. This slowing of depolarization results in a rightward shift of the terminal 40 millisecond (T40-msec) of the QRS axis and the right bundle branch block pattern that is noted on the ECG of patients who are exposed to, or overdose with, a CA.¹¹⁴

Because CAs are weakly basic, they are increasingly ionized as the ambient pH falls, and less ionized as the pH rises. Changing the ambient pH therefore likely alters their binding to the sodium channel. That is, since it is probable that 90% of the binding of CA to the sodium channel occurs in the ionized state, alkalinizing the blood facilitates the movement of the CA away from the hydrophilic sodium channel and into the lipid membrane.

Sinus tachycardia is due to the antimuscarinic, vasodilatory (reflex tachycardia), and sympathomimetic effects of the CAs. Wide-complex tachycardia most commonly represents aberrantly conducted sinus tachycardia rather than ventricular tachycardia. However, by prolonging anterograde conduction, nonuniform ventricular conduction may result, leading to reentrant ventricular dysrhythmias.

Electrophysiologic studies in a canine model demonstrate that QRS prolongation is rate dependent, a characteristic effect of the class I antidysrhythmics (Chap. 64). In these studies, when the heart rate could not accelerate because of a crushed sinus node, the dogs never developed QRS prolongation. Furthermore, pharmacologic induction of bradycardia prevents or narrows wide-complex tachycardia by allowing time for recovery of the sodium channel from inactivation.^4,94 However, since bradycardia adversely affects cardiac output, induction of bradycardia is not recommended.

A Brugada ECG pattern, specifically type 1 or “coved” pattern, is rarely associated with CA overdose. The Brugada syndrome originates from a structural change in the myocardial sodium channel that results in functional sodium channel alterations similar to those caused by the CAs.^11,77 It is possible that this small cohort of patients may have had subclinical Brugada syndrome that was uncovered by the CA (Chap. 16).

QT interval prolongation can occur in the setting of both therapeutic use and overdose of CAs. This apparent prolongation of repolarization results primarily from slowed depolarization (ie, QRS prolongation) rather than altered repolarization.⁹⁰ Although QT prolongation predisposes to the development of torsade de pointes, this dysrhythmia is uncommon in patients with CA poisoning due to the prominent tachycardia.

Hypotension is caused by direct myocardial depression secondary to altered sodium channel function, which disrupts the subsequent excitation-contraction coupling of myocytes and impairs cardiac contractility. Peripheral vasodilation from α-adrenergic blockade by CAs also contributes prominently to postural hypotension. In addition, downregulation of adrenergic receptors may cause a blunted physiologic response to catecholamines.⁷⁶

Agitation, delirium, and depressed sensorium are primarily caused by the central anticholinergic and antihistaminic effects. Hemodynamic effects are likely to contribute in only the most severely poisoned patients. Details regarding the exact mechanism of CA-induced seizures remain elusive. CA-induced seizures may result from a combination of an increased concentration of monoamines (particularly norepinephrine), muscarinic antagonism, neuronal sodium channel alteration, and GABA inhibition.⁷⁸

Acute respiratory distress syndrome (ARDS) may occur in the setting of CA overdose. In one study, amitriptyline exposure caused dose-related vasoconstriction and bronchoconstriction in isolated rat lungs.¹⁰⁵ Many substances implicated in ARDS, such as platelet-activating factor and protein kinase activation, were important in mediating amitriptyline-induced impairment of lung function in this experimental model. Another animal model demonstrated that acute amitriptyline poisoning causes dose-dependent rises in pulmonary artery pressure, pulmonary edema, and sustained vasoconstriction that could be attenuated by either calcium channel inhibition or a nitric oxide donor.⁶⁶

CLINICAL MANIFESTATIONS

The toxic profile is qualitatively the same for all of the first-generation TCAs but is slightly different for some of the other CAs.¹¹² The progression of clinical toxicity is unpredictable and may be rapid. Patients commonly present to the emergency department (ED) with minimal apparent clinical abnormalities, only to develop life-threatening cardiovascular and CNS toxicity within hours.

The CAs have a low therapeutic index, meaning that a small increase in serum concentration over the therapeutic range may result in toxicity. Acute ingestion of 10 to 20 mg/kg of most CAs causes significant cardiovascular and CNS manifestations (therapeutic dose, 2–4 mg/kg/d). Thus, in adults, ingestions of more than 1 g of a CA is usually associated with life-threatening effects. As few as two 50 mg imipramine tablets may cause significant toxicity in a 10 kg toddler (ie, 10 mg/kg). In a series of children with unintentional TCA exposure, all patients with reported ingestions of more than 5 mg/kg manifested clinical toxicity.⁷²

Acute Toxicity

Most of the reported toxicity from CAs derives from patients with acute ingestions, especially in patients who are chronically taking the medication. Clinical manifestations of these two cohorts do not appear to be different, and most studies do not distinguish between them.

Acute Cardiovascular Toxicity.

Cardiovascular toxicity is primarily responsible for the morbidity and mortality attributed to CAs. Refractory hypotension due to myocardial depression probably is the most common cause of death from CA overdose.^20,104 Hypoxia, acidosis, volume depletion, seizures, or concomitant ingestion of other cardiodepressant or vasodilating drugs can exacerbate hypotension.

The most common dysrhythmia observed following CA overdose is sinus tachycardia (rate, 120–160 beats/min in an adult), and this finding is present in most patients with clinically significant TCA poisoning. The ECG typically demonstrates intraventricular conduction delay that manifests as a rightward shift of the T40-msec QRS axis and a prolongation of the QRS complex duration. These findings can be used to identify and risk stratify, respectively, patients with CA poisoning (see Diagnostic Testing). PR, QRS, and QT interval prolongation can occur in the setting of both therapeutic and toxic amounts of TCAs.⁶⁸

Wide-complex tachycardia is the characteristic potentially life-threatening dysrhythmia observed in patients with severe toxicity (Fig. 71–2A-C). Ventricular tachycardia may be difficult to distinguish from aberrantly conducted sinus tachycardia which occurs more commonly. In the former cases, the preceding P wave may not be apparent because of prolonged atrioventricular conduction, widened QRS interval, or both. Ventricular tachycardia occurs most often in patients with prolonged QRS complex duration and/or hypotension.^63,108 Hypoxia, acidosis, hyperthermia, seizures, and β-adrenergic agonists may also predispose to ventricular tachycardia.^63,108 Fatal dysrhythmias are rare, as ventricular tachycardia and fibrillation occur in only approximately 4% of all cases.^41,85 Both the Brugada type I ECG pattern and torsade de pointes are uncommon with acute TCA overdose.

Acute Central Nervous System Toxicity.

Altered mental status and seizures are the primary manifestations of CNS toxicity. Delirium, agitation, and/or psychotic behavior with hallucinations may be present and most likely result from antagonism of muscarinic and histaminergic receptors. These alterations in consciousness usually are followed by lethargy, which is followed by rapid progression to coma. The duration of coma is variable and does not necessarily correlate or occur concomitantly with ECG abnormalities.⁵⁶

Seizures usually are generalized and brief, most often occurring within 1 to 2 hours of ingestion.²⁸ The incidence of seizures is similar to ventricular tachycardia and occurs in an estimated 4% of patients presenting with overdose and 13% of fatal cases.¹¹⁵ Status epilepticus is uncommon. Abrupt deterioration in hemodynamic status, namely hypotension and ventricular dysrhythmias, may develop during or within minutes after a seizure.^28,63,108 This rapid cardiovascular deterioration likely results from seizure-induced acidosis that exacerbates cardiovascular toxicity. The risk of seizures with CA overdoses may be increased in patients undergoing long-term therapy or who have other risk factors such as history of seizures, head trauma, or concomitant drug withdrawal.¹⁰⁰ Myoclonus and extrapyramidal symptoms may also occur in CA-poisoned patients.

Cessation of CAs may produce a drug discontinuation syndrome in some patients, which is typified by GI and somatic distress, sleep disturbances, movement disorders, and mania.³⁷

Other Clinical Effects.

Anticholinergic effects can occur early or late in the course of CA toxicity. Pupils may be dilated and poorly reactive to light. Other anticholinergic effects include dry mouth, dry flushed skin, urinary retention, and ileus. Although prominent, these findings are typically clinically inconsequential.

Reported pulmonary complications include ARDS, aspiration pneumonitis, and multisystem organ failure. ARDS may be the result of aspiration, hypotension, pulmonary infection, and excessive fluid administration, along with the primary toxic effects of CAs.^96,97 Bowel ischemia, pseudoobstruction, and pancreatitis are reported in patients with CA overdose.⁷⁴

Death directly caused by CA toxicity usually occurs in the first several hours after presentation and is secondary to refractory hypotension in patients who reach health care facilities. Late deaths (>1–2 days after presentation) usually are secondary to other factors such as aspiration pneumonitis, adult respiratory distress syndrome from refractory hypotension, and/or infection.²¹

Chronic Toxicity

Chronic CA toxicity usually manifests as exaggerated side effects, such as sedation and sinus tachycardia, or is identified by supratherapeutic drug concentrations in the blood in the absence of an acute overdose.³⁹ Unlike chronic theophylline and aspirin poisoning, chronic CA toxicity does not appear to cause life-threatening toxicity.

A sparse literature describes the clinical course of this cohort. However, a recent case report described chronic amitriptyline overdose in a child (15 mg/kg a day for a month), which resulted in status epilepticus and significant cardiac conduction abnormalities but normal neurological outcome.²⁶ Genetic analysis of this patient’s CYP450 system showed two copies of wild-type alleles for the genes responsible for CYP2D6 activity; thus, concluding the patient was not a “rapid metabolizer.” Several possible protective mechanisms are presented and further illustrate the complexity of this drug, its metabolism, and toxicity. These include a unique pharmacogenomics profile that yields an abnormal receptor profile or metabolic pathway (eg, polymorphism in the gene for myocardial fast sodium channels), another medication causing a beneficial drug-drug interaction, a cardioprotective role for caffeine if the patient’s intake was high causing adenosine receptor anta­gonism and the protective effect from other drugs the patient was taking—guanfacine and clonidine—adrenergic medications that may have offered some protective effect from the α-adrenergic blockade caused by amitriptyline.

Several reports describe sudden death in children taking therapeutic doses of CAs.^88,89,107 QT prolongation with resultant torsade de pointes, advanced atrioventricular conduction delays, blood pressure fluctuations, and ventricular tachycardia are postulated mechanisms, although whether any of these effects contributed to the deaths is unknown. Prospective studies using 12-lead ECG, 24-hour ECG recording, and Doppler echocardiography in children receiving therapeutic doses of CAs have failed to find any significant cardiac abnormalities when compared to children not taking CAs.^14,31 However, authors recommend that CAs not be initiated or continued in any child with a resting QT interval greater than 450 msec or with bundle branch block.³⁴ This is an ongoing area of research as it becomes problematic in making decisions about pharmacotherapy interventions.

Unique Toxicity from “Atypical” Cyclic Antidepressants

Although the incidence of serious cardiovascular toxicity is lower in patients with amoxapine overdoses, the incidence of seizures is significantly greater than with the traditional CAs.^56,65 Moreover, seizures may be more frequent, or status epilepticus may develop.⁷⁷ Similarly, the incidences of seizures, cardiac dysrhythmias, and duration of coma are greater with maprotiline toxicity compared to the CAs.⁵⁵

DIAGNOSTIC TESTING

Diagnostic testing for patients with CA poisoning primarily relies on indirect bedside tests (ECG) and other qualitative laboratory analyses. Quantification of CA concentration provides little help in the acute management of patients with CA overdose but provides adjunctive information to support the diagnosis.

Electrocardiography

The ECG can provide important diagnostic information and may predict clinical toxicity after a CA overdose. CA toxicity results in distinctive and diagnostic ECG changes that may allow early diagnosis and targeted therapy when the clinical history and physical examination are unreliable.

A T40-msec axis between 120° and 270° is associated with CA toxicity and was a sensitive indicator of drug presence in one study.^22,79,114 A terminal QRS vector between 130° and 270° discriminated between 11 patients with positive toxicology screens for CAs and 14 patients with negative toxicology screens.⁷⁹ With further analyses, this report concluded that the positive and negative predictive values of this ECG parameter for CA ingestions were 66% and 100%, respectively, in a population of 299 general overdose patients. A retrospective study reported that a CA-poisoned patient was 8.6 times more likely to have a T40-msec axis greater than 120° than was a non–CA-poisoned patient.¹¹⁴ This parameter was a more sensitive indicator of CA-induced altered mental status but not necessarily of seizure or dysrhythmia. However, the T40-msec axis is not easily measured in the absence of specialized computer-assisted analysis, which limits its practical utility. An abnormal terminal rightward axis can be estimated by observing a negative deflection (terminal S wave) in leads I and aV_L and a positive deflection (terminal R wave) in lead aV_R (Fig. 71–3).

The maximal limb lead QRS complex duration is an easily measured ECG parameter that is a sensitive indicator of toxicity. One investigation reported that 33% of patients with a limb lead QRS interval greater than or equal to 100 msec developed seizures and 14% developed ventricular dysrhythmias.¹⁷ No seizures or dysrhythmias occurred in those patients whose QRS interval remained less than 100 msec. There was a 50% incidence of ventricular dysrhythmias among patients with a QRS duration greater than or equal to 160 msec. No ventricular dysrhythmias occurred in patients with a QRS duration less than 160 msec. Subsequent studies confirmed that a QRS duration greater than 100 msec is associated with an increased incidence of serious toxicity, including coma, need for intubation, hypotension, seizures, and dysrhythmias, making this ECG parameter a useful indicator of toxicity.^22,61

Evaluation of lead aV_R on a routine ECG may also predict toxicity (Figs. 71–2 and 71–3). When prospectively studied, 79 patients with acute CA overdoses demonstrated that the amplitude of the terminal R wave and R/S wave ratio in lead aV_R (RaV_R, R/SaV_R) were significantly greater in patients who developed seizures and ventricular dysrhythmias.⁶¹ The sensitivity of RaV_R= 3 mm and R/SaV_R= 0.7 in predicting seizures and dysrhythmias was comparable to the sensitivity of QRS = 100 msec.

The type 1 Brugada pattern is similar to a right bundle branch block (rSR′), with downsloping ST elevations (“coved”) in the right precordial leads (V1–V3).^11,77 This pattern is neither highly sensitive nor specific for CA toxicity, and it is reported in patients with cocaine and phenothiazine toxicity as well as those on class IA antidysrhythmic therapy. In one series of more than 400 patients with CA overdose, a significant increase in adverse outcomes (ie, seizures, widened QRS interval, and hypotension) was identified in those patients with a Brugada ECG pattern compared to those who did not have the pattern.¹¹ However, there were no deaths or dysrhythmias in the nine patients with this pattern.

Serial ECGs should be obtained because the ECG changes can be dynamic. ECG parameters should always be interpreted in conjunction with the clinical presentation, history, and course during the first several hours to assist in decision making regarding interventions and disposition.⁶²

Laboratory Tests

Determination of serum CA concentrations has limited utility in the immediate evaluation and management of patients with acute overdoses. In one study, serum drug concentrations failed to accurately predict the development of seizures or ventricular dysrhythmias.¹⁷ The pharmacologic properties of CAs—namely, large volumes of distributions, prolonged absorption phase, long distribution half-lives, pH-dependent protein binding, and the wide interpatient variability of terminal elimination half-lives—explain the limited value of serum concentrations in this situation. Any concentration above the therapeutic range (50–300 ng/mL, including active metabolites) may be associated with adverse effects, and is an indication to decrease or discontinue the medication.

Although CA concentrations greater than 1000 ng/mL usually are associated with significant clinical toxicity such as coma, seizures, and dysrhythmias, life-threatening toxicity may be observed in patients with serum concentrations less than 1000 ng/mL.^17,57 Serious toxicity at lower concentrations probably results from a number of factors, including the timing of the specimen in relation to the ingestion and the limitations of measuring the concentration in blood and not the affected tissue. Quantitative concentrations usually cannot be readily obtained in most hospital laboratories. However, qualitative screens for CAs using an enzyme-multiplied immunoassay test are available at many hospitals. Unfortunately, false-positive results can occur with many drugs such as carbamazepine, cyclobenzaprine, thioridazine, diphenhydramine, quetiapine, and cyproheptadine (Chap. 6). Thus, the presence of a CA on a qualitative assay should not be relied upon to confirm the diagnosis of CA poisoning in the absence of corroborating historical or clinical evidence.

Quantitative concentrations may be helpful in determining the cause of death in suspected overdose patients. CA concentrations reported in lethal overdoses typically range from 1100–21,800 ng/mL. CA concentrations may increase more than fivefold because of postmortem redistribution (Chap. 34).⁵ Measurement of liver CA concentration or the parent-to-metabolite drug ratio may be useful in the postmortem setting.

MANAGEMENT

Any person with a suspected or known ingestion of a CA requires immediate evaluation and treatment (Table 71–2). The patient should be attached to a cardiac monitor, and intravenous access should be secured. Early intubation is advised for patients with CNS depression and or hemodynamic instability because of the potential for rapid clinical deterioration. A 12-lead ECG should be obtained for all patients. Laboratory tests, including concentrations of glucose and electrolytes, should be performed for all patients with altered mental status, as well as blood gas analysis to both assess the degree of acidemia and guide alkalinization therapy. Aggressive interventions for maintenance of blood pressure and peripheral perfusion must be performed early to avoid irreversible damage. Both children and adults receiving cardiopulmonary resuscitation have recovered successfully despite periods of asystole exceeding 90 minutes.^24,27,79,101 The options for GI decontamination discussed in the following section should then be considered.

Gastrointestinal Decontamination

Induction of emesis is contraindicated, given the potential for precipitous neurologic and hemodynamic deterioration. Because of the potential lethality of large quantities of CAs, orogastric lavage should be considered in the symptomatic patient with an overdose. Although the benefits of orogastric lavage for CA toxicity are not substantiated by controlled trials, the potential benefits of removing significant quantities of a highly toxic drug must be weighed against the risks of the procedure (Chap. 8).¹⁸ Because the anticholinergic actions of some CAs may decrease spontaneous gastric emptying, attempts at orogastric lavage up to 12 hours after ingestion may yield unabsorbed drug. Because of the potential for rapid deterioration of mental status and seizures, orogastric lavage should be performed only after endotracheal intubation has ensured airway protection. Orogastric lavage in young children with unintentional ingestions of CAs may be associated with more risk and impracticalities, such as the inadequate hole size of pediatric tubes, and less benefit given the amount of drug usually ingested. Activated charcoal should be administered in nearly all cases. Irrespective of age, an additional dose of activated charcoal several hours later is reasonable in a seriously poisoned patient in whom unabsorbed drug may still be present in the GI tract or in the case of desorption of CAs from activated charcoal. It is important to monitor for the development of an ileus to prevent abdominal complications from additional doses of activated charcoal.⁷⁴

Wide-Complex Dysrhythmias, Conduction Delays, and Hypotension

The mainstay therapy for treating wide-complex dysrhythmias and for reversing conduction delays and hypotension is the combination of serum alkalinization and sodium loading. Increasing the extracellular concentration of sodium, or sodium loading, may overwhelm the effective blockade of sodium channels, presumably through gradient effects (Fig. 71–1). Controlled in vitro and in vivo studies in various animal models demonstrate that hypertonic sodium bicarbonate effectively reduces QRS complex prolongation, increases blood pressure, and reverses or suppresses ventricular dysrhythmias caused by CAs.^{82,92, 93, and 94} These studies showed a clear benefit of hypertonic sodium bicarbonate when compared to hyperventilation, hypertonic sodium chloride, or nonsodium buffer solutions. A systematic review of all animal and human studies published before 2001 revealed that alkalinization therapy was the most beneficial therapy for consequential dysrhythmias and shock¹⁶ (Antidotes in Depth: A5).

The optimal dosing and mode of administration of hypertonic sodium bicarbonate and the indications for initiating and terminating this treatment are unsupported by controlled clinical studies. Instead, the information is extrapolated from animal studies, clinical experience, and an understanding of the pathophysiologic mechanisms of CA toxicity. A bolus, or rapid infusion over several minutes, of hypertonic sodium bicarbonate (1–2 mEq/kg) should be administered initially.^70,98 Additional boluses every 3 to 5 minutes can be administered until the QRS interval narrows and the hypotension improves (Fig. 71–2). Blood pH should be carefully monitored after several bicarbonate boluses, aiming for a target pH of no greater than 7.50 or 7.55. Because CAs may redistribute from the tissues into the blood over several hours, it may be reasonable to begin a continuous sodium bicarbonate infusion to maintain the pH in this range. Differences in outcomes between repetitive boluses versus bicarbonate infusions are not well studied. Although diluting sodium bicarbonate in 5% dextrose in water and infusing it slowly renders it less able to increase the sodium gradient across the cell, the beneficial effects of pH elevation still warrant its use once the patient is stabilized. No evidence supports prophylactic alkalinization in the absence of cardiovascular toxicity (eg, QRS < 100 msec). In addition, alkalization would inevitably cause a decrease in potassium, which may cause QT prolongation and potentially contribute to other dysrhythmias. Hypertonic sodium chloride (3% NaCl) reverses cardiotoxicity in several animal studies,^47,71,82 and numerous reports and extensive clinical experience support its efficacy in humans.^16,48,49,73 However, the dose of hypertonic saline for CA poisoning has never been evaluated in humans for safety or efficacy, and the dose suggested by animal studies (up to 15 mEq/kg) exceeds the amount that most clinicians would consider safe (1–2 mEq/kg). Hypertonic sodium chloride is associated with a hyperchloremic metabolic acidosis, an undesired effect that highlights one benefit of hypertonic sodium bicarbonate. However, hypertonic saline could be considered in situations in which alkalinization with sodium bicarbonate is not possible.

Hyperventilation of an intubated patient is a more rapid and easily titratable method of serum alkalinization but is not as effective as sodium bicarbonate in reversing cardiotoxicity.^51,70 Simultaneous hyperventilation and sodium bicarbonate administration may result in profound alkalemia and should be performed only with extreme caution and careful monitoring of pH. Hyperventilation without bicarbonate administration may be indicated in patients with ARDS or congestive heart failure in whom administration of large quantities of sodium is contraindicated.

Alkalinization and sodium loading with hypertonic sodium bicarbonate and or hypertonic saline along with controlled ventilation (if clinically indicated) should be administered to all CA overdose patients presenting with major cardiovascular toxicity and altered mental status. Indications include conduction delays (QRS > 100 msec) and hypotension. It is imperative to initiate treatment until CA toxicity can be excluded because of the risk of rapid and precipitous deterioration. Although commonly assumed, it is unclear whether the failure of the QRS complex to narrow with sodium bicarbonate treatment excludes CA toxicity.

It is unclear whether alkalinization and sodium loading is effective for reversing the Brugada pattern. The sparse available literature is equivocal.^12,77 It would seem prudent to administer sodium bicarbonate in the presence of a presumed CA-induced Brugada pattern, especially with concomitant signs of other CA toxicity.

Alkalinization may be continued for at least 12 to 24 hours after the ECG has normalized because of the redistribution of the drug from the tissue. However, the time observed for resolution or normalization of conduction abnormalities is extremely variable, ranging from several hours to several days, despite continuous bicarbonate infusion.⁶² We recommend stopping alkalinization when the patient’s mental status improves and there is improvement, but not necessarily normalization, of abnormal ECG findings.

Antidysrhythmic Therapy

Lidocaine is the antidysrhythmic most commonly advocated for treatment of CA-induced dysrhythmias, although no controlled human studies demonstrate its efficacy.⁸⁶ Because lidocaine has sodium channel blocking properties, some investigators argue against its use in CA poisoning.¹ These theoretical concerns are not well supported in the literature, and the class IB antidysrhythmic channel binding kinetics may prove favorable. Although limited data also suggest that the IB antidysrhythmic phenytoin prevents or reverses conduction abnormalities,^43,69 these data were poorly controlled for other confounding factors, such as blood pH and sodium bicarbonate administration; they had very small numbers; and, in some, the cardiotoxicity was not severe. Since phenytoin exacerbates ventricular dysrhythmias in animals²¹ and fails to protect against seizures,¹⁰ its use is no longer recommended.

The use of class IA (quinidine, procainamide, disopyramide, and moricizine) and class IC (flecainide, propafenone) antidysrhythmics is absolutely contraindicated because they have similar pharmacologic actions to CAs and thus may worsen the sodium channel inhibition and exacerbate cardiotoxicity. Class III antidysrhythmics (amiodarone, bretylium, and sotalol) prolong the QT interval and, although unstudied, may be contraindicated as well (Chap. 64).

Because magnesium sulfate has antidysrhythmic properties, it may be beneficial in the treatment of ventricular dysrhythmias. Animal studies of the effects of magnesium on CA-induced dysrhythmias yield conflicting results.^52,53 However, successful use of magnesium sulfate in the treatment of refractory ventricular fibrillation after TCA overdose is reported.^24,27,54,91 A case control study suggested that magnesium sulfate and sodium bicarbonate resulted in lower fatality incidence and shorter intensive care unit stay compared to sodium bicarbonate alone.²⁹ When dysrhythmias fail to reverse after alkalinization, sodium loading, and a trial of lidocaine, or magnesium sulfate may be warranted.

Slowing the heart rate in the presence of CAs may allow more time during diastole for CA unbinding from sodium channels and result in an improvement in ventricular conduction.^3,92 This may abolish the reentry mechanism for dysrhythmias and was one rationale for the past use of physostigmine and propranolol. Thus, decreasing the sinus rate may itself be effective in abolishing ventricular dysrhythmias by eliminating rate-dependent conduction slowing. Propranolol terminated ventricular tachycardia in an animal model but also caused significant hypotension and death.⁹⁴ In one case series, patients developed severe hypotension or had a cardiac arrest shortly after receiving a β-adrenergic antagonist.³⁶ Other animal studies suggest that preventing or abolishing tachycardia by sinus node destruction, or by using bradycardic agents that impede sinus node automaticity without affecting myocardial repolarization or contractility, may successfully prevent CA-induced ventricular dysrhythmias.^3,4 The combined negative inotropic effects of β-adrenergic antagonists and CAs, along with the significant cardiac and CNS effects reported with physostigmine use, do not support their routine use in the management of CA-induced tachydysrhythmias.

Hypotension

Standard initial treatment for hypotension should include volume expansion with isotonic saline or sodium bicarbonate. Hypotension unresponsive to these therapeutic interventions necessitates the use of inotropic or vasopressor support and possibly extracorporeal cardiovascular support.

No controlled human trials are available to guide the use of vasopressor therapy. The pharmacologic properties of CAs complicate the choice of a specific agent. Specifically, CA blockade of neurotransmitter reuptake theoretically could result in depletion of intracellular catecholamines. This could blunt the effect of dopamine, which is dependent on the release of endogenous norepinephrine for its inotropic activity. This suggests that a direct-acting vasopressor such as norepinephrine is more efficacious than an indirect-acting catecholamine such as dopamine.

In fact, limited clinical data suggest that norepinephrine is more efficacious than dopamine.¹⁰⁹ In a retrospective study of 26 adult hypotensive patients receiving nonstandardized therapy, response rates to norepinephrine (5–53 μg/min) were significantly better than response rates to dopamine (5–10 μg/kg/min).¹¹⁰ Patients who did not respond to dopamine at vasopressor doses (10–50 μg/kg/min) responded to norepinephrine (5–74 μg/min). Animal data comparing various treatments are conflicting, and their direct applicability to clinical human poisoning is limited.^32,111 Both norepinephrine and epinephrine increased the survival rate in CA-poisoned rats. In addition, epinephrine was superior to norepinephrine when used both with and without sodium bicarbonate, and the most effective treatment regimen in this study was epinephrine plus sodium bicarbonate; neither precipitated dysrhythmias. The authors propose that epinephrine is more efficacious because it augments myocardial perfusion more than norepinephrine and improves the recovery of CA sodium channel blockade by hyperpolarization of the membrane potential through its stimulation of increased potassium intracellular transport.

Based on the available data, pharmacologic effects, theoretical concerns, and experience, norepinephrine (0.1–0.2 μg/kg/min) is recommended for hypotension that is unresponsive to volume expansion and hypertonic sodium bicarbonate therapy. Central venous pressure and or pulmonary artery catheterization may be necessary to guide the choice of additional vasopressor or inotropic agents, especially in the presence of other cardiodepressant drugs.

If these measures fail to correct hypotension, extracorporeal life support measures should be considered. Extracorporeal membrane oxygenation, extracorporeal circulation, and cardiopulmonary bypass are successful adjuncts for refractory hypotension and life support when maximum therapeutic interventions fail.^42,101,113 These modalities can provide critical perfusion to the heart and brain and maintain metabolic function while giving the body time to metabolize and eliminate the CA by maintaining hepatorenal blood flow.

Emerging Therapies

Vasopressin is increasingly being used in the setting of vasodilatory shock with successful increases in arterial blood pressure based on its vasoconstrictive actions from several mechanisms. Its successful use for intractable hypotension due to CA toxicity, unresponsive to α-receptor agonists and pH manipulation, has been described and warrants further investigation.⁹

Intravenous fat emulsion is reported to be effective in reversing cardiovascular toxicity due to several lipophilic drugs including amitriptyline and clomipramine. Its utilization and effectiveness appears logical given their pharmacological properties—Log D and Log P—octanol/water partition coefficient discussed previously.

Several controlled animal studies have demonstrated improved survival in clomipramine-induced cardiovascular collapse when intravenous lipid emulsion is given either as pretreatment or resuscitation in comparison with saline controls and sodium bicarbonate infusion.⁴⁴ Other animal studies failed to demonstrate any benefit.^7,64 Specifically, one failed to demonstrate a statistically significant benefit in amitriptyline-poisoned rats pretreated with intravenous fat emulsion.⁷ Case series and case reports demonstrate clinical improvement when lipids have been administered for cardiovascular ­collapse or instability refractory to other therapies.^{38,44,50,58,99} The dosing and timing of administration are variable as well as other concomitant therapies, making it difficult to reach any definitive conclusions regarding its effectiveness. In addition, significant adverse reactions and complications have been noted including ARDS and pancreatitis. More data is emerging allowing more evidence-based criteria for its use and dosing. Certainly for patients with refractory hypotension and or ventricular dysrhythmias, fat emulsion therapy should be strongly considered, given the high mortality rate with these medications (Antidotes in Depth: A20).

Central Nervous System Toxicity

Seizures caused by CAs usually are brief and may stop before treatment can be initiated. Recurrent seizures, prolonged seizures (>2 minutes), and status epilepticus require prompt treatment to prevent worsening acidosis, hypoxia, and development of hyperthermia and rhabdomyolysis. Benzodiazepines are effective as first-line therapy for seizures. If this therapy fails, barbiturates or propofol should be administered. Propofol controlled refractory seizures resulting from amoxapine toxicity.⁷⁵ Failure to respond to barbiturates or propofol should lead to consideration of neuromuscular paralysis and general anesthesia with continuous electroencephalographic monitoring. Phenytoin is not recommended for seizures because data not only demonstrate a failure to terminate seizures but also suggest enhanced cardiovascular toxicity.^9,21

Use of flumazenil in a patient with known or suspected CA ingestion is contraindicated. Several case reports of patients with CA overdoses describe seizures following administration of flumazenil⁵⁹ (Antidotes in Depth: A27). Physostigmine was used in the past to reverse the acute CNS toxicity of CAs (Antidotes in Depth: A9). However, physostigmine is not recommended because it may increase the risk of cardiac toxicity, cause bradycardia and asystole, and precipitate seizures in acutely CA-poisoned patients.⁸⁴

Enhanced Elimination

No specific treatment modalities have demonstrated clinical significant efficacy in enhancing the elimination of CAs. Some investigators propose multiple doses of activated charcoal to enhance CA elimination because of their small enterohepatic and enterogastric circulation.⁶⁷ Human volunteer studies and case series of patients with CA overdoses suggest that the half-life of CAs may be decreased by multiple-dose activated charcoal (MDAC).¹⁰⁷ Activated charcoal reduced the apparent half-life of amitriptyline to 4 to 40 hours in overdose patients, compared to previously published values of 30 to more than 60 hours.¹⁰⁷ Changes in the severity or duration of clinical toxicity, however, were not reported. Other investigators showed that in human volunteers MDAC reduced the half-life of therapeutic doses of amitriptyline approximately 20% compared with no activated charcoal administration. However, the methodologic flaws and equivocal findings of these studies and the lack of any positive outcome data for this intervention from additional studies do not provide evidence supporting its use in this setting.^23,40 Pharmacokinetic properties of CAs (large volumes of distribution, high plasma protein binding) weighed against the small increases in clearance and the potential complications of MDAC, such as impaction, intestinal infarction, and perforation, do not warrant its routine use.^23,74 One additional dose of activated charcoal may be given to decrease GI absorption in patients with evidence of significant CNS and cardiovascular toxicity if bowel sounds are present.

Measures to enhance urinary CA excretion have a minimal effect on total clearance. Urinary alkalinization does not enhance, and may reduce, urinary clearance due to passive reabsorption of the unionized CA from an alkaline urine. Hemodialysis is ineffective in enhancing the elimination of CAs because of their large volumes of distribution, high lipid solubility, and extensive protein binding.⁴⁵ Hemoperfusion overcomes some of the limitations of hemodialysis but may not be effective because of the large volumes of distributions of CAs. Although several uncontrolled case reports and a case series described improvement in cardiotoxicity during hemoperfusion, this finding may be coincidental.^13,24,35

Hospital Admission Criteria

All patients who present with known or suspected CA ingestion should undergo continuous cardiac monitoring and serial ECG for a minimum of 6 hours. Recommendations in the older literature for 48 to 72 hours of intensive care unit monitoring even for patients with minor CA ingestions stem from isolated case reports of late-onset dysrhythmias, CNS effects, and sudden deaths.⁸³ However, review of these cases shows inadequate GI decontamination, inadequate therapeutic interventions, and significant ongoing complications of overdose. Several retrospective studies demonstrate that late, unexpected complications in CA overdoses such as seizures, dysrhythmias, and death did not occur in patients who had few or no major signs of toxicity at presentation or a normal level of consciousness and a normal ECG for 24 hours.^20,27,30,85 A disposition algorithm has been proposed based on clinical signs and symptoms.^6,108 If the patient is asymptomatic at presentation, undergoes GI decontamination, has normal ECGs, or has sinus tachycardia (with normal QRS complexes) that resolves, and the patient remains asymptomatic in the health care facility for a minimum of 6 hours without any treatment interventions, the patient may be medically cleared for psychiatric evaluation (if appropriate) or discharged home as appropriate.

A prospective study of 67 patients used the Antidepressant Overdose Risk Assessment (ADORA) criteria to identify patients who were at high risk for developing serious toxicity and proposed criteria for hospitalization.³³ In this study, the presence of QRS interval greater than 100 msec, cardiac dysrhythmias, altered mental status, seizures, respiratory depression, or hypotension on presentation to the ED (or within 6 hours of ingestion, if the time was known) was 100% sensitive in identifying patients with significant toxicity and subsequent complications. Criteria specific for intensive care unit admission (other than patients requiring ventilatory and or blood pressure support), versus an inpatient bed with continuous cardiac monitoring, are less clear and probably are institution dependent.¹⁰³

The disposition of patients with persistent isolated sinus tachycardia, prolonged QT interval with no concomitant altered mental status, or blood pressure changes, is not clearly defined. Previous studies demonstrate that these two parameters alone are not predictive of subsequent clinical toxicity or complications.^33,34 In addition, the sinus tachycardia may persist for up to one week following ingestion. However, a study of isolated CA overdose patients reported that a heart rate greater than 120 beats/min and QT interval greater than 480 msec were associated with an increased likelihood of major toxicity.²² These patients are candidates for observation units with continuous ECG monitoring and serial ECGs for 24 hours.

Inpatient Cardiac Monitoring

The duration of cardiac monitoring in any patient initially exhibiting signs of major clinical toxicity depends on many factors. Certainly the duration of CA cardiotoxicity and neurotoxicity may be prolonged, and using normalization of ECG abnormalities as an end point for therapy and discharge is problematic. Some studies document the variable resolution and normalization of QRS prolongation and T40-msec axis rotation.^81,97 Based on the available literature, it is reasonable to recommend that after the mental status and blood pressure normalize, and the ECG improves, patients who exhibited significant poisoning should be monitored for another 24 hours off of all of therapy, including alkalinization, antidysrhythmics, and inotropics/vasopressors.

SUMMARY

• CA poisoning continues to be a cause of serious morbidity and mortality worldwide.

• The distinctive characteristics of these drugs can cause significant CNS and cardiovascular toxicity, the latter being responsible for mortality as a result of overdose of these drugs. Cardiovascular toxicity ranges from mild conduction abnormalities and sinus tachycardia to wide-complex tachycardia, hypotension, and asystole. CNS toxicity includes delirium, lethargy, seizures, and coma.

• The ECG is a simple, readily available diagnostic test that can predict the development of significant toxicity, particularly seizures and/or dysrhythmias.

• Management strategies are based primarily on the pathophysiology of these drugs, namely, sodium channel blockade in the myocardium. Alkalinization and sodium loading with hypertonic sodium bicarbonate is the principal therapy for cardiovascular toxicity.

• Guidelines for observing or admitting patients to the hospital may be based on initial clinical presentation or development of clinical effects and ECG changes.

References

HISTORY AND EPIDEMIOLOGY

The Swedish chemistry student Arfwedson discovered lithium in 1817.^8,131 Lithium derives its name from the Greek word for stone, lithos, from which it was first isolated. The therapeutic use of lithium has a long history beginning in the mid 19th century, when lithium salts were used to treat individuals with gout and were noted to improve symptoms of mania and depression.¹¹⁶ The soft drink 7-Up was originally formulated with lithium as its “active ingredient.”⁸ During the 1930s and 1940s, lithium was used as a salt substitute (“Westsal”) for patients with congestive heart failure but was discontinued after several cases of acute lithium poisoning were described.⁸⁷ The beneficial effects of lithium on bipolar disorder were “rediscovered” by Cade in 1949, when he noticed the calming effect of lithium carbonate on guinea pigs.⁵² The same year, however, the US Food and Drug Administration (FDA) banned the use of lithium in response to reported poisonings.¹³⁹ The FDA lifted the ban in 1970 and approved the use of lithium for the treatment of mania.

Currently, lithium is the most efficient long-term therapy for treatment and prevention of bipolar affective disorders,^51,81 with a demonstrated antisuicidal effect and an ability to improve the manic and depressive symptoms of the illness, as well as to augment the therapeutic efficacy of other therapies that have failed to achieve symptom remission.^{16,45,53,78,81,108,111,131} Investigations on the use of lithium for compulsive gambling have also demonstrated beneficial results,⁹³ and it is under investigation for use in neurodegenerative disorders such as Alzheimer disease, stroke, amyotrophic lateral sclerosis, Huntington disease, multiple sclerosis, and traumatic brain injury.^{23,47,142,199} In most industrialized nations, approximately one in 1000 persons is prescribed one or more of the various lithium formulations.^84,121,122

PHARMACOLOGY

Lithium, the simplest xenobiotic in the modern pharmacopoeia, has a complex mechanism of action that has not been completely elucidated even after more than 50 years of clinical use. Psychotropics modifying monoamine neurotransmission form the basis of modern psychopharmacology. The current paradigm of neuropsychiatric therapeutics implies that because xenobiotics that affect the neurotransmission of dopamine are most efficacious, this receptor-drug exemplar defines the disease. Multiple lines of investigation have shown that beyond simple monoamine balance, the interaction between multiple signaling cascades, neurotrophic and neuroprotective systems is not only involved in the pathogenesis of mood disorders, but also the cellular and molecular means of the action of lithium.¹⁵⁵ Additionally, the understanding that a particular receptor only undergoes a single agonist-receptor activation has been replaced by the growing comprehension that agonist binding causes activation of multiple pathways of downstream signaling, with a myriad of results.²⁵ Lithium illustrates this new paradigm through its multiple direct and indirect effects.

Clinically, the therapeutic effects of lithium and similar mood-stabilizing pharmaceuticals become evident only after chronic administration, so their mechanism of action is unlikely solely the result of acute biochemical interactions. One of the central processes involved in the therapeutic effects of lithium, and potentially the pathogenesis of mood disorders, is its interaction with the β form of glycogen synthase kinase 3 (GSK-3). GSK-3β is a serine/threonine kinase originally described with the regulation of glycogen synthesis in response to insulin.²⁵ Subsequently it has been shown to influence multiple systems, including gene transcription, neuronal function and neurogenesis, synaptic plasticity, and the circadian cycle, as well as cellular structure, apoptosis, and cell death. In fact, GSK-3β phosphorylates more than 100 substrates, with a suggestion of many more.¹⁰¹ Each of these systems are implicated in the pathophysiology of mood disorders.^25,155

Postmortem tissue from the ventral prefrontal cortex of patients with major depressive disorder showed elevated GSK-3β activity as well as decreased activity in the GSK-3β modifying enzyme, Akt (also known as protein kinase B).^102,103 GSK-3β is inadequately inhibited in association with mood disorders and is inhibited in humans treated with lithium.¹⁰¹ Overactivity of this enzyme is associated with neuronal degeneration and sensitivity to apoptotic stimulation. Dysregulation of GSK-3β is implicated in tumor growth and the neurofibrillary tangles of Alzheimer disease.^56,191 It also is a key regulator of neuronal cell fate, with a proapoptotic effect in many settings.^{5,22,24,26,101,118,131,161} It is involved in regulating the activity of β-catenin, Jun, and cyclic adenosine monophosphate (cAMP) response element-binding protein (CREB), transcription factors important in embryonic patterning, cell proliferation, neuronal modeling and plasticity, neuronal signal transduction, memory consolidation, and cytoskeletal remodeling. Its other targets include transcription factors such as c-Jun, nuclear factor activated T-cells, proteins bound to microtubules (Tau, microtubule-associated protein 1B, kinesin light chain), cell cycle mediators (cyclin D), and metabolic regulators (glycogen synthase, pyruvate dehydrogenase).^153,155 Hypoxia contributes to increased GSK-3β activity, which may be counteracted or inhibited with mood-stabilizing drugs. Vascular depression, or depression after stroke, is an organic model of major depression.^94,133 The finding that this depressive state responds similarly to intervention with mood stabilizers lends further support to the GSK-3β hypothesis.^{70,79,90,150,190}

Many older generation antipsychotics are now known to affect GSK-3β signaling in mice.¹¹⁹ GSK-3β activity is regulated not only by first- and second-generation antipsychotics but also through 5-HT neurotransmission and activation of 5-HT_2A receptors as well as through monoamine-affecting antidepressants which modify these neurotransmitters.^{22, 23, 24, 25, 26, 27, and 28,117} Atypical antipsychotics and antagonists of D₂ dopamine receptors, as well as 5-HT_2A serotonin receptors, may exert some of their utility through inhibition of GSK-3β activity.^22,25

The link between neuropsychiatric illness and these centrally placed mediators of signaling comes from exploration of dopamine neurotransmission (Fig. 72–1). Scaffolding proteins, the β-arrestins, are traditionally associated with the termination of G protein–coupled receptor (GPCR) signaling and desensitization. After a receptor is stimulated, GPCRs are phosphorylated, leading to the β-arrestin recruitment that uncouples the receptor from the G protein and leads to GPCR internalization. β-Arrestins also act as scaffolds for the formation of a protein complex that allows for GPCRs to signal independently from G proteins.^22,24,26,71 Dopaminergic neurotransmission through GPCRs is mediated through a complex that involves Akt, β-arrestin 2 (βArr2), and protein phosphatase 2A (PP2A). Akt is a serine/threonine kinase that is additionally regulated through phosphatidylinositol-mediated signaling.²⁵ Akt activity results from an equilibrium between phosphorylation (activation) and dephosphorylation (inactivation).²⁷ Regulation of GSK-3β activity is achieved through phosphorylation of its N-terminal serine residue, leading to inhibition.¹¹⁸ The Akt-βArr2-PP2A complex, when activated, dephosphorylates (deactivates) Akt, which leads to the activation of GSK-3β through loss of inhibition (loss of phosphorylation.)^25,27,70

Lithium is a direct inhibitor of GSK-3β that can also inhibit its activity indirectly through a mechanism involving Akt activation.^26,27,48,101 Lithium activates Akt by disrupting assembly of the Akt-βArr2-PP2A complex through displacement of the magnesium cofactor required for assembly.^{25, 26, and 27} Through disruption of this complex, which normally promotes Akt inactivation, lithium promotes Akt activation, leading to increased phosphorylation (inactivation) of GSK-3β. This is demonstrated in β-arrestin knockout mice, which, unable to create this complex, are resistant to the behavioral effects of lithium.^25,27,70,101

Identification of the Akt-βArr2-PP2A signaling complex as a target of lithium lends credence to the use of lithium as an adjunct to enhance the action of atypical antipsychotics and antidepressants in poorly responsive subjects by acting through Akt–GSK-3β signaling.^25,27,101

Inhibition of GSK-3β by lithium is thought to be neuroprotective by modifying the downstream targets and effectors of GSK-3β activity.^{5,23,27,70,91,110,150,162} Lithium is implicated in the neuroprotective modulation of the bcl-2 gene, which is known for its role in preventing apoptosis and in downregulation of the proapoptotic protein p53. Lithium increases bcl-2 concentrations in cultured nervous tissue of both rats and humans.^49,129 Additional support comes from patients undergoing long-term therapy with either lithium or valproic acid (VPA) who show prefrontal cortex volumes significantly greater than in patients not treated with either agent, suggesting a protective effect in humans.^47,162 Further evidence points toward a neuroprotective and neurotrophic effect of lithium with evidence to support benefit in such diverse neurodegenerative conditions as Parkinson disease (in a mice model of Parkinson disease using N-methyl-4-pheynyl-1,2,3,6-tetrahydropyridine {MPTP}),¹⁹⁸ Huntington disease,¹⁶⁸ amyotrophic lateral sclerosis (ALS),⁶⁸ stroke and multiple sclerosis,^23,47 traumatic brain injury,¹⁹⁹ and Alzheimer disease.^{5,19,22,25,42,131,136}

It would be naive to assume that lithium has a single mechanism of action, and it is plausible that other mechanisms, such as the inhibition of inositol monophosphatases, also contribute to its pleiotropic effects on behavior. Compelling evidence has bridged a link between the current molecular targets and downstream regulators of the effects of lithium with the classic proposed mechanism of action, the inositol-depletion hypothesis.

Among the first proposed mechanisms of action of lithium was the inositol-depletion hypothesis.⁹⁰ Inositol is a six-carbon sugar that forms the backbone of a number of cellular signaling mechanisms. Lithium treatment results in a decreased myoinositol (the most biologically active stereoisomer of inositol) concentration in the cerebral cortex.^34,90,131 Abnormalities in regional brain myoinositol concentrations are thought to occur in bipolar patients. This theory is partially supported by experimental magnetic resonance spectroscopy data.^173,197 Myoinositol is phosphorylated to form phosphatidyl inositol (PIP), which is further phosphorylated and combined with diacylglycerol (DAG) to form phosphatidyl 4,5-bisphosphate (PIP₂). Upon stimulation of a cell, G protein–coupled receptors activate phospholipase C (PLC), which hydrolyzes PIP₂ to release the secondary messengers DAG and inositol 1,4,5-trisphosphate (IP₃).^34,92 Each of these secondary messengers in turn initiates a cascade of events, including activation of protein kinase C, which is important for calcium homeostasis and neurotransmitter release,^131,173 as well as independent mobilization and regulation of intracellular calcium.^149,162,173 Many extracellular signals, including some serotonin receptor subtypes, neurotrophin signaling pathways, receptor tyrosine kinase pathways, and G protein–mediated signaling, activate PLC to exert their actions.^80,154,155

Serial dephosphorylation of IP₃ leads to regeneration of myoinositol and recycling of the inositol pool. Two enzymes involved in this pathway are inhibited by lithium. The first enzyme, inositol 1,4-bisphosphate 1-phosphatase (IPPase), dephosphorylates the bisphosphate to inositol monophosphate (IMP). The second enzyme, inositol 1-monophosphatase (IMPase), dephosphorylates IMP to myoinositol⁴ (Fig. 72–2).

The inhibition of IMPase is interesting and important. Lithium inhibits IMPase by “uncompetitively” binding to the enzyme–substrate complex and preventing the release of phosphate. It performs this function by ­displacing a magnesium ion from the active site after hydrolysis. This is the same mechanism by which lithium directly inhibits GSK-3β.⁴⁸ Uncompetitive refers to the inhibitor binding to the enzyme-substrate complex; the higher the concentration of the substrate, the more the enzyme is inhibited.^70,86,127 Uncompetitive inhibitors only bind to the enzyme-substrate complex, inhibiting the reaction at that point. Uncompetitive inhibitor kinetics are related proportionally to the concentration of the enzyme–substrate complex and cannot be overcome by increasing the concentration of the substrate, unlike a competitive inhibitor.¹⁵⁰ This supports a theory about the pathophysiology of bipolar disorder involving an excess of myoinositol and is one reason why the mood-stabilizing effects of lithium are thought to occur only in bipolar patients.⁹⁰ The nature of the action of lithium serves as a regulator to preferentially block pathologic signaling caused by excessive myoinositol while leaving the normal signaling intact. As described, IMPase is an important step in the cellular recycling of the inositol pool and is inhibited by lithium.

Myoinositol is also generated de novo from glucose-6-phosphate by inositol synthase, which forms IMP. The inhibition of IMPase by lithium subsequently leads to myoinositol depletion by preventing the conversion of the newly synthesized IMP to myoinositol. Interestingly, VPA also inhibits inositol synthase, illustrating a potential mechanism for the synergy of these complementary mood stabilizers.^60,115 A third mechanism of intracellular diminution of inositol by lithium (as well as VPA and carbamazepine) is the effect of lithium on reducing activity and transcription of the sodium myoinositol transporter, preventing the uptake of exogenous myoinositol by the cell. This mechanism of inhibition may be overcome by increased extracellular concentration of myoinositol.⁹⁰

The result of these effects is depletion of the inositol pool available to the cell, causing a series of events at different points in the signal transduction cascade that leads to differential gene transcription and expression. This sequence ultimately is responsible for some of the observed clinical effects of lithium on the central nervous system (CNS).¹⁷⁰ Experimental data using dextroamphetamine as a model for clinical mania demonstrate increased regional inositol signaling in the human brain that is attenuated by pretreatment with lithium, lending support to the hypothesis.³¹

The inositol depletion hypothesis, while an original attempt to explain in molecular terms the therapeutic effects of lithium, does not fully elucidate nor replicate the clinical disease or response to therapy. In vivo studies with more drastic inositol depletion than that which occurs from lithium therapy fail to replicate predicted behavioral patterns. Studies in knockout mice lacking various isoforms of inositol monophosphatase fail to replicate the antidepressant or antimanic effects of lithium. The model remains an attractive one, and the flaws may represent species variation with validity more in humans with bipolar disorder than the mouse model illustrates.^23,131,169 However, there may be a synergy with other proposed mechanisms of lithium’s effect. Using a yeast model, it was shown that GSK-3β is required for de novo inositol synthesis, and that loss of GSK-3β activity leads to inositol depletion. This finding links the two targets of not only lithium, but other mood ­stabilizing pharmaceuticals.¹²

In summary, although the precise mechanism of action is unknown, some common features of investigation have emerged. The potential targets, widely found and disparate in function, all seem to be inhibited by lithium in an uncompetitive fashion, most commonly through displacement of a divalent cation, usually Mg²⁺. The systems affected by this inhibition vary widely. Downstream targets seem to modulate secondary cell messengers and intracellular signal transduction, transcription factors and gene expression, and neuronal plasticity and cellular differentiation. Further study is needed to elucidate the complex interaction of these pathways with the action of lithium to form an integrated hypothesis.

PHARMACOKINETICS AND TOXICOKINETICS

The volume of distribution of lithium is between 0.6 and 0.9 L/kg. It has no discernible protein binding and distributes freely in total body water, with the exception of the cerebrospinal fluid (CSF), from which it is actively extruded.^62,81,163 The extrusion is believed to occur through an active transport process involving sodium/lithium exchange at the arachnoid processes.⁵⁹ The immediate-release preparations of lithium are rapidly absorbed from the gastrointestinal (GI) tract. Peak serum concentrations are achieved within 1 to 2 hours. Sustained-release products demonstrate variable absorption, with a delay to peak of 4 to 5 hours.^81,182 In overdose, a longer delay to reach peak concentrations or multiple peaks may occur.^58,182 Chronic therapy prolongs the elimination of lithium, as does advancing age.⁸¹ Although lithium is rapidly absorbed, tissue distribution is a complex phenomenon, with a significant delay in reaching a steady state. Lithium exhibits preferential uptake into the kidney, thyroid, bone, and other organs and tissues such as the liver and muscle. Lithium distribution into the brain can take up to 24 hours to reach equilibrium. Lithium is concentrated in red blood cells (RBCs) by both passive diffusion and active transport. The RBC concentration may correlate closely with the brain concentration, although this does not appear to be clinically useful.⁴³ The pharmacokinetic profile of lithium is described as an open, two-compartment model.^62,81

Each 300 mg lithium carbonate tablet contains 8.12 mEq of ­lithium.⁸¹ Ingestion of a single 300 mg tablet is expected to acutely increase the serum lithium concentration by approximately 0.1 to 0.3 mEq/L (assuming a volume of distribution of approximately 0.6–0.9 L/kg and a patient weight of 50–100 kg).

Lithium undergoes no metabolic transformation and is eliminated almost entirely (95%) by the kidneys, with a small amount eliminated in the feces.⁶² Lithium is also found in sweat, saliva, and breast milk.^57,185 In an adult with normal kidney function, lithium clearance varies from 10 to 40 mL/min.^81,181 At steady-state equilibrium, total body clearance equals renal clearance.

Lithium is handled by the kidneys much in the same way as sodium. Lithium is freely filtered, and more than 80% is reabsorbed by the proximal tubule. Evidence also indicates a small amount of reabsorption by the loop of Henle and distal tubule.^{36,69,112,181} Lithium excretion is therefore dependent on factors that affect the glomerular filtration rate (GFR) or decrease sodium concentration. Any condition that makes the kidney sodium avid such as volume depletion or salt restriction increases lithium reabsorption in the proximal tubule.^82,181 Thus, risk factors for development of lithium toxicity include advanced age with its decrease in GFR; use of thiazide diuretics, nonsteroidal antiinflammatory drugs (NSAIDs), or angiotensin-converting enzyme (ACE) inhibitors; decreased sodium intake; and low-output heart failure.^101,113

The therapeutic index of lithium is narrow. The generally accepted steady-state therapeutic range of serum lithium concentrations is 0.6 to 1.0 mmol/L, although much disagreement exists about whether this serum concentration truly reflects therapeutic efficacy.^81,121 Both in therapeutic and overdose situations, clinical signs and symptoms seem to be a more valuable indicator of brain lithium concentrations.^17,140

CLINICAL MANIFESTATIONS

Similar to other xenobiotics having prolonged redistributive phases and tissue burdens, lithium exposure can be divided into three main categories of toxicity: acute, acute-on-chronic, and chronic. In acute lithium toxicity, the patient has no body burden of lithium present at the time of ingestion. The toxicity that develops depends on the rates of absorption, distribution, and elimination. In chronic toxicity, the patient has a stable body burden of lithium as serum concentration is maintained in the therapeutic range, and then some factor disturbs this balance, either by enhancing absorption, or more commonly, decreasing elimination. For chronic users of lithium, small perturbations in the equilibrium between intake and elimination may lead to toxicity. In acute-on-chronic toxicity, the patient ingests an increased amount of lithium (intentionally or unintentionally) in the setting of a stable body burden. With tissue saturation, any additional lithium leads to signs and symptoms of toxicity.

Acute Toxicity

Acute ingestions of lithium-containing preparations produce clinical findings similar to that of ingestions of other metal salts, with predominant early GI symptoms. Nausea, vomiting, and diarrhea are prevalent. Significant volume losses may result from these symptoms. Patients may complain of lightheadedness and dizziness, and they may be orthostatic. Neurologic manifestations are a late finding in acute toxicity as the lithium redistributes slowly into the CNS.

Chronic Toxicity

Lithium is primarily a neurotoxin. The earliest case reports of lithium toxicity described predominantly neurologic symptoms.^175,187 Most importantly, neurotoxicity does not correlate with serum concentrations. The initial clinical condition of the patient and the duration of exposure to an elevated concentration seem to be more closely predictive of outcome than the initial serum lithium concentration.^3,9,88,167

The tremor of lithium is likely one of the most commonly encountered drug-induced tremors in clinical practice, with approximately 27% of patients developing it.¹⁴¹ The tremor can diminish over time but may increase with toxicity. Other findings of chronic toxicity include fasciculations, hyperreflexia, choreoathetoid movements, clonus, dysarthria, nystagmus, and ataxia.¹⁸² Mental status is often altered and may progress from confusion to stupor, coma, and seizures.^121,134 Electroencephalographic changes are most frequently reported as “slowing.”¹⁷¹ The progression of these symptoms follows no order, and any patient undergoing chronic therapy may have one or any combination of these features.

The syndrome of irreversible lithium-effectuated neurotoxicity (SILENT) is a descriptive syndrome of the irreversible neurologic and neuropsychiatric sequelae of lithium toxicity.^3,97,151 SILENT is defined as neurologic dysfunction that is caused by lithium in the absence of prior neurologic illness and persists for at least 2 months after cessation of the drug. Case reports support these findings and this definition. However, as is true in most case reports, confounders make wide applicability of the findings difficult. Because of the polypharmacy prevalent in psychiatry, long-term neurologic sequelae attributed to lithium are frequently described in patients using lithium in combination with other xenobiotics, such as haloperidol, chlorpromazine, carbamazepine, phenytoin, aspirin, VPA, amitriptyline, β-adrenergic antagonists, calcium channel blockers, ACE inhibitors, diuretics, and NSAIDs.^{3,65,89,128,138} However, there exist reports of patients using lithium without coingestants and no comorbid illness who sustained lasting dysfunction as a result of lithium toxicity.^{9,105,145,167} Cerebellar findings seem to predominate in patients with SILENT.^{1,2, and 3,130,151} One of the predictors of persistent neurologic dysfunction may be the concomitant finding of hyperpyrexia, an ominous finding in patients with lithium toxicity.^83,130 The mechanism of the persistent dysfunction is unclear, but demyelination and cellular loss are proposed.^3,130,165

Acute-on-Chronic Toxicity

Patients undergoing chronic therapy who acutely ingest an additional amount of lithium (either intentionally or unintentionally) are at risk for signs and symptoms of both acute and chronic toxicity. These patients may display prominent GI and neurologic symptoms and may be difficult to diagnose and manage. Serum lithium concentrations in cases of acute or chronic toxicity may be difficult to interpret, and therapy should be guided by the clinical status of the patient.

Other Systemic Manifestations of Chronic Lithium Therapy

The most common adverse effect of chronic lithium therapy is nephrogenic diabetes insipidus which can develop within weeks of the initiation of therapy and affects up to 40% of patients (Chap. 19).^85,144

Lithium inhibits the transport of sodium through the amiloride-sensitive epithelial Na⁺ channel (ENaC), which is also the main route of entry for lithium. Lithium has a twofold higher affinity than sodium for entry through this channel, without a corresponding means of extrusion and thus becomes concentrated intracellularly.^106,181,182

Lithium is believed to inhibit magnesium-dependent G proteins that activate vasopressin-sensitive adenylate cyclase, leading to decreased generation of cAMP in the cell membranes of distal tubular cells.^{35,50,180,186}

Decreased cAMP leads to reduced expression and translocation of the vasopressin-regulated water channel aquaporin-2 (AQP2), making the distal tubules resistant to the action of vasopressin, and thus unable to make a concentrated urine.^6,35,120,188 Additionally, there is potential mechanistic overlap with GSK-3β. GSK-3β exhibits tonic inhibition of cyclooxygenase-2 (COX-2). With GSK-3β inhibition by lithium, this inhibition is removed and increased COX-2 activity leads to increased prostaglandin expression in the renal medulla. Increased prostaglandin expression is important in nephrogenic diabetes insipidus through regulation of glomerular blood flow,^156,157 as well as increasing the degradation of AQP2, further decreasing the ability to form a concentrated urine.^{106,107,156,157}

Chronic lithium therapy is also associated with chronic tubulointerstitial nephropathy, as manifested by the development of renal insufficiency with little or no proteinuria and biopsy findings of tubular cysts. This association was demonstrated in a biopsy-based study of 24 chronically treated patients, although the overall prevalence of this condition is low.⁸⁵

Lithium is associated with a number of endocrine disorders. The most prevalent endocrine manifestation of chronic lithium therapy is hypothyroidism.³⁹ The etiology is multifactorial. Lithium is selectively concentrated in the thyroid gland and competes for iodide transport, synthesis of triiodothyronine (T₃), responsiveness of the gland to thyroid-stimulating hormone (TSH), release of T₃ and tetraiodothyronine (T₄), and peripheral conversion of T₄ to T₃.^39,73 Additionally, lithium decreases responsiveness of peripheral tissues to T₃ and increases antithyroglobulin antibodies if already present, it does not induce development de novo (Fig. 56–1).^13,110,143 Although hypothyroidism is most common, hyperthyroidism and frank thyrotoxicosis are also reported.^33,41,110 However, hyperthyroidism, by altering proximal tubule function, leads to decreased lithium excretion.²⁰ Thus, hyperthyroidism may lead to chronic lithium toxicity through impaired elimination, and the elevated lithium concentrations may mask the manifestations of hyperthyroidism.^18,143

The combination of hyperparathyroidism and hypercalcemia is frequently reported with chronic lithium therapy, most commonly in older women. The mechanism is thought to be modification of calcium feedback on parathyroid hormone release through alteration of a calcium-sensing receptor that prevents suppression of parathyroid hormone release in response to elevated calcium,¹⁵⁹ although stimulation of parathyroid hormone release, parathyroid gland hyperplasia, and adenomas are alternative suggested mechanisms.^{11,114,132,178}

Lithium is associated with a number of electrocardiographic (ECG) abnormalities, although the evidence for significant toxicity is ­lacking. Most reports are uncontrolled case reports without corroborating experimental data or biologic plausibility. The most commonly reported manifestation has been T-wave flattening or inversion, primarily in the precordial leads,¹⁴⁷ although prolongation of the QT interval is also noted.^109,184,192 One study associated elevated serum lithium concentrations with QT prolongation above 440 msec, although the number of patients studied was small.⁹⁶ Associations have been made between lithium and sinoatrial dysfunction, with resultant bradycardia.^{10,77,146,172,179} However, many of these cases had either electrolyte disturbances or multiple other cardioactive xenobiotics involved that would be more likely causative.^{44,172,179,189} Numerous case reports have linked lithium with an ECG pattern that is consistent with a myocardial infarction, without evidence of biochemical markers of injury.^104,152 Lithium blocks cardiac sodium channels in a transfected Chinese hamster ovary cell. This is the putative link between lithium therapy and the unmasking of a Brugada pattern as well as cardiomyopathy.^{7,46,55,193,196} For the most part, lithium has few consequential effects on cardiac function, even in overdose, and malignant dysrhythmias or significant dysfunction is very uncommon.

Developmentally, in utero exposure to lithium increases the incidence of congenital heart defects, specifically Ebstein anomaly.^57,150,195 Additionally, many effects similar to those that occur in patients undergoing chronic therapy are found in infants exposed in utero, including thyroid disease and neurotoxicity.⁷⁵

Lithium causes a leukocytosis and an increase in neutrophils. It has been proposed as an adjunct to chemotherapy-induced neutropenia, other marrow suppressive therapies, and acquired immunodeficiency syndrome (AIDS).⁵⁴ Although lithium increases the total neutrophil count, no improved clinical outcomes are documented, and its use has been superseded by recombinant colony-stimulating factors.^66,150,160

DIAGNOSTIC TESTING

Because of the prevalence of lithium use, therapeutic drug monitoring is readily available in most settings, and concentrations should be readily obtainable.¹⁹⁴ A lithium concentration should be requested upon patient presentation and serial measurements requested or considered in most instances, especially after ingestion of sustained-release preparations. Emphasis should be placed on the lithium concentration as a marker of exposure and response to therapy but not necessarily as a determinant of toxicity or treatment. The history, clinical signs, and symptoms rather than the absolute lithium concentration should guide therapy. The sample must be sent in an appropriate lithium-free tube, because use of lithiated-heparin tubes may lead to clinically false-positive results.¹¹³ Serum electrolyte concentrations and kidney function should be monitored because kidney function is important in determining the need for more aggressive therapy, including enhanced elimination techniques such as hemodialysis. If the patient is hypernatremic, nephrogenic diabetes insipidus should be suspected, and determinations of serum and urine osmolarity and electrolytes help confirm the diagnosis (Chap. 19). If thyroid disease is suspected, thyroid function tests should be obtained. If a deliberate ingestion has occurred, a serum acetaminophen concentration should be obtained. An ECG is also indicated. The complete blood count may indicate a leukocytosis, as a stress response or due to the hematologic effects of lithium.

MANAGEMENT

The initial management and stabilization should begin with assessment and, if necessary, support of airway, breathing, and circulation. Lithium rarely, if ever, affects the patient’s airway or breathing, although coingestants may. Emesis, which occurs with significant frequency after acute exposure, may lead to aspiration and respiratory compromise. After the patient is stable, the characteristics of the exposure should be determined while the physical examination and laboratory assessment commence. The formulation and nature of the product should be ascertained and most importantly, identified as immediate-release or sustained-release. Also, whether or not lithium is part of the patient’s medication regimen may help determine whether the ingestion is acute, acute-on-chronic, or chronic.

Gastrointestinal Decontamination

For patients who present after an acute overdose or an acute-on-chronic overdose, a risk benefit analysis of GI decontamination must be undertaken. Two factors should be considered. With an acute overdose and predominance of early GI symptoms, including emesis, self-decontamination may already have started. Second, immediate-release preparations are often rapidly absorbed and may not lend themselves to GI evacuation.

Few GI decontamination options are available. Lithium is a monovalent cation that does not bind readily to activated charcoal.¹²³ Because of the danger of a depressed level of consciousness, potential loss of protective airway reflexes, and generally prominent emesis, no beneficial effect from activated charcoal is expected unless indicated for treatment of a potential coingestant. Induced emesis is no longer recommended. Orogastric lavage has essentially no role in the acute management of a patient with a lithium overdose unless indicated for a coingestant. Whereas immediate-release preparations of lithium are rapidly absorbed and typically produce emesis, sustained-release formulations of lithium (ie, controlled-release tablets) compounded in a slowly dissolving film-coated formulation often make the tablet too large to fit through even the largest lavage tube.

Sodium polystyrene sulfonate (SPS) is a cationic exchange resin often used for the treatment of severe hyperkalemia. It binds potassium in exchange for sodium, allowing elimination of excess potassium in the feces. Because of the similarity between potassium and lithium, use of SPS has been proposed for decontamination of patients being treated for lithium toxicity. A number of models have examined the effectiveness of this technique.^{123, 124, and 125} Use of SPS has many theoretical benefits, including demonstrated effectiveness of lithium binding compared with activated charcoal and the ability of orally administered SPS to reduce serum concentrations of intravenously administered lithium in mice.¹²³ Unfortunately, the finding that doses used to increase lithium elimination also lead to significant hypokalemia in human subjects limits the application of this technique.¹⁶⁴ In a murine model, potassium supplementation with SPS was found to mitigate this process but only at the expense of elevating lithium concentrations.¹²⁵ Two reports in the literature demonstrate increased ­lithium elimination with SPS, one in a healthy volunteer and another in a patient with an acute overdose. However, the serum potassium concentration was not reported in either case.⁷⁴ In a retrospective cohort review of 12 chronically poisoned patients, it was suggested that SPS reduced the elimination half-life of lithium by 50%. However, many of the patients had altered kidney function that was corrected at the same time as the intervention (which was not standardized), and no clinical outcomes were reported.⁷⁶ At present, use of SPS in the management of the lithium-poisoned patient cannot be routinely recommended, until further clinical trials can be performed.

Whole-bowel irrigation (WBI) is the only GI decontamination that has some demonstrated efficacy in eliminating lithium from volunteer human subjects. In one of the few clinical trials of WBI, the serum lithium concentrations of 10 normal volunteers who had ingested sustained-release lithium carbonate were plotted against time over a 72-hour period. In the second phase of the trial, when the volunteers received 2 L/h of polyethylene glycol solution one hour after the ingestion, there was a significant reduction (67%) in the serum concentration, even as early as one hour after the therapeutic intervention.¹⁷⁴ WBI is recommended for patients manifesting significant toxicity (ie, neurologic dysfunction) and who have ingested sustained-release preparations and have no contraindications (eg, protected airway, no obstruction or ileus) (Antidotes in Depth: A2).

Fluid and Electrolytes

The critical initial management of the lithium-poisoned patient should focus on restoration of intravascular volume, both in acute poisonings with GI losses and in chronic poisonings with toxic effects that are often the result of disturbances of kidney function and lithium elimination. Many patients with lithium toxicity have volume-responsive decreases in kidney function,¹⁵⁸ which can be managed by infusion of 0.9% sodium chloride solution at 1.5 to 2 times the maintenance rate. This therapy increases renal perfusion, increases the GFR, and lithium elimination. Urine output must be closely monitored and any electrolyte abnormalities corrected. Caution must be used in patients with prerenal acute kidney injury (AKI), chronic kidney disease (CKD), and congestive heart failure. Monitoring for the development of hypernatremia in patients suspected of having nephrogenic diabetes insipidus is critical.^120,181,182

Lithium-induced nephrogenic diabetes insipidus may be reversed by discontinuation of the drug and repletion of electrolytes and water. However, nonreversible effects are reported.^106,181 Clinical application of amiloride to mitigate lithium-induced polyuria is described, although the potential for volume contraction and stimulation of lithium reabsorption limits recommendation of this drug as a routine adjunct to acute care.^{30,64,106,158}

Any attempt to enhance elimination of lithium by forced diuresis using loop diuretics (furosemide), osmotic agents (mannitol), carbonic anhydrase inhibitors (acetazolamide), or phosphodiesterase inhibitors (aminophylline) should be avoided. An initial small increase in elimination may be achieved, but typically salt and water depletion subsequently develop followed by increased lithium retention. Sodium bicarbonate for urinary alkalinization should also be avoided because it does not significantly increase elimination over volume expansion with sodium chloride and may lead to hypokalemia, alkalemia, and fluid overload.

Extracorporeal Drug Removal

Debate surrounds the efficacy and practicality of using enhanced elimination techniques in cases of lithium poisoning. Lithium does have pharmacokinetic properties that make it amenable to extracorporeal removal,^{21,63,95,100,112} and with these characteristics, it would seem to be an ideal candidate for hemodialysis. In fact, hemodialysis is often recommended for treatment of acute, acute-on-chronic, and chronic lithium toxicity.^{14,63,98,148,182,199} However, some characteristics of lithium make extracorporeal elimination difficult. Lithium is predominantly localized intracellularly and diffuses slowly across cell membranes.⁶³ When traditional intermittent hemodialysis is used for chronic exposures, clearance of the blood compartment is often followed by a rebound phenomenon of redistribution from tissue stores, leading to recurrent increased serum concentrations. The clinical significance of this rebound is uncertain, but without additional elimination this can redistribute back into target tissues.³⁸ There have only been three cases reported of clinical deterioration after rebound from an extracorporeal elimination method, likely as a result of failed decontamination in an extended release ingestion.^37,40,71 An additional complicating factor is that the brain, the “target organ” of toxicity, is not amenable to a rapid artificial elimination process. Attempts have been made to correlate the serum concentration with the lithium concentration in the CSF and brain. But in the few studies where CSF concentrations were obtained, although serum and CSF lithium concentrations seemed to correlate, brain concentrations and toxicity did not.^92,99 Magnetic resonance spectroscopy studies of bipolar patients with steady-state lithium concentrations demonstrated a significant variability between brain and serum concentrations, especially within the therapeutic range.⁶⁷ In addition, because of the toxicokinetic profile of lithium, serum concentrations do not correlate well with toxicity.^81,99,182

Some consensus is emerging regarding when to perform an extracorporeal elimination technique (extracorporeal treatment {ECTR}). The Extracorporeal Treatments in Poisoning Workgroup (ExTRIP) has performed an extensive analysis of the data available regarding lithium. Their most recent consensus provides the following analysis: “ECTR should be performed in severe Li poisoning,” graded (1, D). The “1” indicates that more than 85% of their members strongly recommend, and the D that there is a very low level of evidence (no observational studies or randomized controlled trials).⁶¹ Hemodialysis or an alternative extracorporeal technique is clearly indicated for three groups of patients. The first group consists of patients who are manifesting severe signs and symptoms of neurotoxicity, such as alterations in mental status or seizures. The second group consists of patients who have AKI or CKD and show signs or symptoms of lithium toxicity. Such patients are unable to eliminate their lithium burden on their own and should undergo hemodialysis. The third group consists of patients who show little or no sign of toxicity but who cannot tolerate sodium repletion therapy; these patients who should be considered for early hemodialysis include those with congestive heart failure or such “redistributive diseases” as liver failure, pancreatitis, or sepsis.

For a patient who belongs to this last group, the next step is to determine the probability that the patient will develop toxicity if elimination is not enhanced. The criteria for hemodialysis originated from a case series of 23 patients and a review of 100 other patients published in 1978, which was before the introduction of sustained-release products.⁸⁸ These recommendations have never been prospectively evaluated (or subsequently reevaluated).^15,63,72 The ExTRIP group proposes the consensus recommendations based on the classification of the overdose and categories based on serum concentrations⁶¹ (Table 72–1).

Interestingly, the above serum concentrations and recommendations are irrespective of symptomatology. Additional consensus recommends ECTR for acute neurologic dysfunction such as confusion, decreased level of consciousness, and seizures. It is also suggested to use ECTR for hyperthermia. Whether hemodialysis diminishes or enhances the risk of permanent neurologic sequelae is a subject of debate.^176,177 Although no controlled studies have analyzed this important management question, the preponderance of evidence suggests a reduced risk.^{3,9,15,63,95,99,137,148,151,166,182}