Amitriptyline, Amoxapine, Clomipramine, Desipramine, Dothiepin, Doxepin, Imipramine, Lofepramine, Nortriptyline, Protriptyline, Trimipramine
TCAs are one of the most common causes of death from poisoning (Crome, 1993). Deaths normally occur outside of hospital (Callaham, 1985). In Australia they are the number one cause of fatality from drug ingestion and 90% of successful TCA suicides do not reach hospital but die at home (Buckley et al, 1995).Tricyclic antidepressants are the second commonest cause of fatal drug poisoning, in developed countries, accounting for 20-25% of fatal drug poisoning in the United Kingdom and United States.
Half the in-hospital fatalities will have trivial toxicity on arrival to hospital but develop major toxicity within 1 hour.
This reflects their rapid absorption and onset of cardiac and central nervous system toxicity. While a number of ECG abnormalities are predictive of serious toxicity a normal ECG does not exclude serious toxicity. Decontamination, management of airway and adjustment of pH to a mild systemic alkalosis with sodium bicarbonate are the mainstays of successful management.
The ingestion of 15-20 mg/kg or more of a TCA is potentially fatal, although there are significant differences in toxicity within the TCA drug class. Tetracyclics and bicyclics appear to be less toxic in overdose than TCAs.
The exact mechanism of action in depression has not been clearly established but involves blocking reuptake of noradrenaline and serotonin and possibly the resetting of serotonin receptors. These effects are probably unimportant in overdose except in combined overdose with selective serotonin reuptake inhibitors (SSRIs).
However, the drugs are pharmacologically “dirty” and bind to many other receptors including histamine (H1 & H2), alpha 1 & 2, GABA-A and muscarinic receptors. Blockade of histamine receptors leads to sedation, alpha receptor blockade leads to vasodilatation, GABA-A blockade may contribute to seizures and anticholinergic effects result from muscarinic receptor blockade.
The drugs block sodium and other membrane ion channels. Thus, they appear to have the antiarrhythmic and proarrhythmic effects of a class Ia antiarrhythmic drug. In overdose, proarrhythmic effects are dominant.
The influx of sodium is the major event responsible for the zero phase of depolarisation in cardiac muscle and Purkinje fibres. This initiates cardiac muscle contraction (systole). The duration of phase 0 in the heart as a whole is measured indirectly as the duration of the QRS complex on the ECG. Thus, blockade of the Na+ channel can be indirectly measured by estimating QRS width. Prolongation of the QRS is a reflection of TCA tissue concentrations and is predictive of both seizures and cardiac arrhythmias.
TCAs block voltage gated Na+ channels in a use dependent manner (i.e. block increases with heart rate). As the degree of Na+ channel block increases with use, the QRS width will increase with increasing heart rates.
However, the Na+ channel blockade also slows the heart rate. The presence of a very wide QRS complex without tachycardia is a sign of severe cardiotoxicity.
Other cardiac channel effects include reversible inhibition of the outward potassium channels responsible for repolarisation giving a mechanism for QT prolongation and arrhythmia generation (Teschemacher et al, 1999) TCAs demonstrate a dose dependent direct depressant effect on myocardial contractility that is independent of impaired conduction (Heard et al, 2001) While the mechanism is not well defined it is known that TCA alter mitochondrial function and uncouple oxidative phosphorylation (Weinbach et al, 1986).
Animal models have shown direct vasoconstrictive effects on pulmonary vasculature with rupture of capillaries and alveolar epithelium causing pulmonary oedema (Liu et al, 2002) These changes are linearly related to dose over the range of 0.01-1.0 mM and have been demonstrated with wide range of cyclic antidepressants (Dahlin et al, 1997).
Tricyclic antidepressants are rapidly absorbed following oral administration and undergo first pass hepatic metabolism. They are highly protein bound (up to 95%) and are highly lipid-soluble, resulting in a very large volume of distribution (10-20 L/kg) and a long elimination half-life (10-81 hours).The elimination half-life may be prolonged following overdose.
Metabolism of TCA drugs involves hepatic microsomal enzymes that cause demethylation of the aliphatic side chain or hydroxylation of the ring nucleus. “Tertiary amines” such as amitriptyline and imipramine undergo hepatic demethylation to form “secondary amines”, which are pharmacologically active metabolites e.g. nortriptyline and desipramine. These are conjugated to form inactive glucuronides, which are excreted by the kidneys. The parent tricyclic antidepressant drug and its metabolites are partially secreted in bile and may be reabsorbed from the bowel before being excreted by the kidneys.
The drugs are highly lipid soluble and therefore rapidly absorbed. Peak drug concentrations occur early and this is the reason why most deaths from poisoning occur outside of hospital.
However, the anticholinergic side effects associated with TCAs may sometimes cause delayed emptying of unabsorbed tablets from the stomach and delayed peak concentrations leading to a deterioration after some hours. Delayed release formulations of amitriptyline can produce peak concentrations and delayed effects up to 42 hours after ingestion (Greene et al., 2002).
The large volume of distribution reflects high concentrations in tissues. To some extent, their effects in overdose may be more pronounced early due to the initially high concentrations that are achieved in plasma and highly perfused organs (heart, brain) before they redistribute extensively into peripheral tissues.
Less than 10 % of TCA circulates as free drug, the rest is bound to circulating proteins (albumin and alpha1 acid glycoprotein) or dissolved in circulating free fatty acids. Alpha1 acid glycoprotein (AAG) has high and low affinity binding sites for TCA. AAG is a more important binding protein than albumin, additional AAG can reverse TCA toxicity whereas albumin has no affect. The “bound” fraction is sensitive to changes in pH with acidosis causing an increase in free fraction.
Alkalinisation causes significant decrease in the percentage of free amitriptyline; with a drop of 20% when pH rises from 7.0-7.4 and 42% over a pH range of 7.4-7.8. (Levitt et al., 1986)
TCAs are metabolised by the hepatic microsomal enzyme system. In overdose, this system is normally overwhelmed and so the half-life of the drug becomes prolonged. Amitriptyline elimination half life in overdose ranges from 25 to 81 hours. Many TCAs have active metabolites. Both the parent drug and the active metabolites may undergo enterohepatic circulation.
Renal excretion is low (3% - 10%).
The pharmacological effects of tricyclic antidepressant drugs at therapeutic doses are complex and include:
Symptoms and signs at presentation depend upon the dose and the time since ingestion. The rapid absorption of TCAs can cause a patient with initially trivial symptoms to deteriorate and develop life threatening toxicity within an hour (Callaham M & Kassell, 1985) Patients who are asymptomatic at three hours post ingestion of normal release medication do not normally develop major toxicity.
In acute TCA overdose there a three major toxic syndromes. These are:
Death in TCA overdose is usually due to CNS and cardiotoxic effects.
Patients at high risk of death are normally identified by a history of high ingested dose, and early onset of deteriorating level of consciousness and the presence of ECG conduction abnormalities.
Tricyclic antidepressants are very toxic in overdose. Ingestion of >15mg/kg can be expected to result in serious and potentially life-threatening toxicity.
The anticholinergic syndrome seen in tricyclic antidepressant, neuroleptic and antihistamine poisonings is often less florid than that seen in the classic anticholinergic syndromes from antimuscarinic plants (Datura, Brugmansia) and anticholinergic drugs such as benztropine. This may be due to the lack of any other sedating drug effects in those poisonings. Those other poisonings often have an agitated delirium with associated management problems.
In TCA poisoning, the pupils may be dilated, but are often found to be mid range. Paralysis of accommodation may lead to some blurring of vision. The pupils react relatively poorly to light. The other anticholinergic effects will lead to a dry mouth and tongue, hot dry skin, and a tachycardia, occasionally urinary retention. Bowel sounds may be absent; this may be associated with an ileus.
Severely poisoned patients (who are generally unconscious on presentation) very commonly develop an anticholinergic delirium late in the course of their illness and this may persist for some days (see also serotonin toxicity). The patient may have visual and auditory hallucinations or patients with relatively mild delirium may just appear to be hypervigilant and suspicious. Thus, it is often useful to ask patients when they regain consciousness whether they're hearing or seeing anything strange and reassure them that this is a drug effect. A severe delirium may make it difficult to communicate with the patient at all and it may interfere with their psychiatric assessment.
Anticholinergic symptoms or signs are a sensitive indicator for ingestion of tricyclic antidepressants but are a poor predictor for life threatening toxicity.
There is a wide spectrum of toxic effects ranging from trivial to life threatening.
Tricyclic antidepressants slow phase 0 cardiac depolarisation by inhibiting sodium channels. The resulting delay in propagation of depolarisation in the atrioventricular (AV) node, His-Purkinje fibres and ventricular myocardium leads to prolongation of the PR and QRS interval. Abnormal atrial and ventricular repolarisation may give rise to ECG changes mimicking myocardial infarction (ST segment elevation and T wave inversion).
The most specific electrocardiographic sign of tricyclic antidepressant toxicity is right axis deviation of the terminal 40-ms vector of the QRS complex in the frontal plane (T 40-ms axis). An R wave in aVR, with an S wave in lead I is an indicator of a rightward T 40-ms axis.
The Brugada wave (downsloping ST segment elevation in leads V1-V3 in association with right bundle branch block) has been reported in tricyclic antidepressant overdose.
ECG changes include:
Minor ECG changes
There may be an increase in the PR interval and dimpling of the T-waves.
Narrow complex tachycardia
Most patients with significant TCA poisonings have a tachycardia, which is due to a varied contribution of anticholinergic effects and peripheral vasodilatation due to alpha blockade. Persistent tachycardia after regaining consciousness is most frequently due to persisting anticholinergic effect or volume depletion. Other possible etiologies that should be considered include anxiety, delirium and drug withdrawal.
Broad complex tachycardia
It may be difficult to distinguish between a supraventricular tachycardia with QRS widening and a ventricular tachycardia.
Both are a poor prognostic sign as the extent of the QRS duration correlates with TCA blood levels Amitai et al, 1993) Patients with acute poisoning and QRS widening are normally unconscious. If at presentation the patient is conscious and has QRS widening consider chronic toxicity or other cardiac disease.
The most common arrhythmia is sinus tachycardia which is due to anticholinergic activity and/or inhibition of norepinephrine uptake by tricyclic antidepressants. In more severe poisoning, tachycardia may be caused by myocardial depression, hypovolaemia or hypotension.
A wide variety of other cardiac arrhythmias may develop: supraventricular tachycardias (AV junctional tachycardia, atrial fibrillation, atrial flutter) with or without aberrant conduction , ventricular arrhythmias (ventricular tachycardia, torsade de pointes or ventricular fibrillation) ,junctional and idioventricular rhythms, second and third degree heart block and asystole may also occur.
Tricyclic antidepressant-induced delayed repolarisation and QT interval prolongation may predispose to torsade de pointes. However this is uncommon as torsade de pointes is a bradycardia or pause dependent arrhythmia and sinus tachycardia is usually the underlying rhythm in tricyclic poisoning.
Major arrhythmias usually occur in patients with marked ECG changes or complications such as hypotension, respiratory depression, coma and seizures but they can occur in asymptomatic patients with minor ECG changes or without prior sinus tachycardia.
A broad complex bradycardia associated with hypotension is a marker of severe toxicity; untreated the likelihood is that the patient will die in the next 10 minutes
The blood pressure may be elevated in the early stages after overdose, presumably due to the inhibition of norepinephrine uptake. Subsequently, the blood pressure is reduced, often to very low levels, because of hypovolaemia, decreased peripheral resistance due to alpha-adrenergic blockade, impaired myocardial contractility and cardiac output and catecholamine depletion.
Hypotension may be due to a number of causes. TCAs themselves can cause direct myocardial depression. However, in practice, the hypotension usually relates to relative volume depletion and alpha-receptor blockade induced vasodilatation. Thus, it usually responds rapidly to intravenous fluids.
The use of inotropes, in particular alpha agonists, should be used with caution. These prolong the effective refractory period (as do TCAs) and thus may be proarrhythmic. Noradrenaline has been shown to predictably induce VT in a rat model of TCA poisoning.
Most patients with significant ingestions of TCAs who are likely to have cardiac complications or seizures have a significantly impaired level of consciousness prior to those complications. Patients will often have a rapid onset of decreasing level of consciousness and coma because of a very rapid absorption of the drug. Patients should be assessed on admission to see if they are hyperreflexic or have myoclonic jerks or any evidence of seizure activity. Some patients who are likely to have seizures may be noted to have relatively brisk reflexes compared to the normal hyporeflexia seen with coma from other causes. This can be a marker of high seizure risk.
Seizures themselves are associated with an increased mortality. Acidosis affects the partitioning of TCAs between the cell membrane and the Na+ channel binding site and increases TCA induced Na+ channel blockade. There may also be a small increase in free drug concentration from pH-mediated changes in protein binding but this is a less important effect.
A number of TCAs (dothiepin, desipramine, and amoxapine) cause seizures more frequently. Thus, they may cause seizure at lower drug ingestions with less ECG abnormalities and occasionally in conscious patients.
Common features include ataxia, nystagmus, divergent squint and drowsiness which may lead to deep coma and respiratory depression. Increased tone and hyperreflexia may be present with extensor plantar reflexes. In deep coma all reflexes (including brain-stem reflexes) may be abolished. Convulsions may lead to haemodynamic compromise.
Pulmonary complications include aspiration, cardiac and noncardiac oedema (ARDS). Animal models have shown direct vasoconstrictive effects on pulmonary vasculature with rupture of capillaries and alveolar epithelium causing pulmonary oedema.(Liu et al, 2002) These changes are linearly related to dose over the range of 0.01-1.0 mM.(Dahlin et al, 1997)
Rhabdomyolysis may occur following seizures or pressure necrosis. Hyperthermia occurs rarely the aetiology is multifactorial and includes central temperature dysregulation, seizures, sepsis and reduced heat loss.
The following investigations should be performed:
Electrolytes are normally assessed but are rarely of much assistance with the exception of patients who are on other medications that may affect electrolytes and thus their risk for arrhythmia.
All unconscious patients require arterial blood gases to access adequacy of ventilation and to ensure they are not acidotic. Arterial blood gases assist in monitoring treatment with systemic alkalinisation. In severe poisonings a mixed respiratory and metabolic acidosis is common. Respiratory acidosis is an absolute indication for ventilation. Hypoxia may be due to a number of the pulmonary complications seen in TCA poisoning including aspiration, cardiac and non-cardiac pulmonary oedema.
An ECG should be performed on admission and at 6 hours after the self-poisoning. Minor ECG changes are common and include an increase in the PR interval and dimpling of the T-waves and a narrow complex sinus tachycardia.
The more serious changes reflect altered conduction through Purkinje fibres due predominately to sodium channel blockade. The ECG is the most accurate predictor of toxicity for the majority of tricyclic antidepressant poisonings. Patients with abnormal ECGs require further monitoring.
ECG prediction of complications
Measurements that predict major toxicity include:
The majority of the studies have been too small to allow a confident estimate of sensitivity/specificity and have used arbitrary cut points defined during post hoc analysis. Invariably the prediction has not been as definitive when others have tried to replicate the findings. In addition, the observer variation (inter-rater agreement) has not been defined, particularly as it applies to these arbitrary cut points (Buckley et al, 1996)
The majority of patients at significant risk for developing cardiac or neurological toxicity will have a QRS complex greater than or equal to 100 ms or a rightward shift of the terminal 40 ms of the frontal plane QRS complex vector (Harrigan and Brady, 1999) However a normal ECG does not exclude major complications (Buckley et al ,1994) All patients with abnormal ECGs or altered level of consciousness require ECG monitoring.
Boehnert & Lovejoy (1985) described the association of prolonged QRS >= 100 ms with the development of seizures and arrhythmias. While subsequent experience has confirmed this as a sensitive marker it is also clear that a QRS < 100 ms can not be used to definitely exclude major complications (Foulke and Albertson, 1987).
A right axis deviation and/or RBBB producing a terminal 40 ms frontal plane axis > 120 degrees is often observed in TCA poisoning. The association is not absolute but appears stronger than that observed with QRS width (Foulke and Albertson, 1987).
The height of R wave and R/S ratio in aVR also has predictive value. The sensitivity of an R in aVR of 3 mm or more was 81% and that of an R/S ratio in aVR of 0.7 or more was 75% for predicting seizures or arrhythmias (Liebelt et al, 1995) A group of patients who would satisfy these criteria but have additional ST segment abnormalities have been described as a TCA induced Brugada syndrome. The Brugada syndrome is a genetically determined sodium-channel dysfunction with the ECG characteristically showing a right bundle-branch block and unusual ST-segment elevation in the right precordial leads. It was reported in 12 of 95 patients presenting with overdose of tricyclic antidepressants. The ECG changes resolved when plasma TCA concentrations dropped below 1000 ng/mL. It is not clear whether the presence the Brugada pattern has additional prognostic importance (Goldgran-Toledano et al, 2002).
Reliance on ECG findings on presentation as the sole predictor of subsequent problems cannot be recommended. The majority of complications occur within the first six hours and in patients who are sedated. An alert patient with a normal ECG six hours after overdose who has had gastrointestinal decontamination is extremely unlikely to develop major complications.
These are unhelpful in aiding acute management. Patients with plasma TCA concentrations greater than 450 ng/mL tend to develop cognitive or behavioral toxicity (agitation, disorientation, confusion, memory impairments, fragmented speech, pacing, decreased concentration). Major toxicity and death is associated with concentrations above 1000 ng/mL.
TCAs should be considered along with other drugs with membrane blocking effects in patients with seizures, QRS prolongation and/or ventricular arrhythmias. A presentation with coma in the presence of anticholinergic signs make TCAs the most likely drug ingested.
There is some variation in the toxic dose between the drugs in this class. The number of deaths per million defined daily doses (DDDs) prescribed in the UK (a fatal toxicity index) varied from 2 -7 for amoxapine, dothiepin, desipramine and nortriptyline to less than 0.5 for lofepramine and clomipramine (Henry et al 1995). The majority of these deaths occurred outside of hospital.
Large differences in the likelihood of producing major complications have also been noted in clinical studies with amoxapine, dothiepin and desipramine having significantly higher rates of seizures when taken in overdose (Buckley et al, 1994; Wedin et al, 1986).
Dosulepin (dothiepin) and amitriptyline appear to be particularly toxic in overdose. Dosulepin overdose is associated with an increased risk of arrhythmias and seizures.
A worse outcome is associated with any of the following
However the in-hospital mortality is low (<1% in most centres) and therefore even patients from these groups have a reasonable prognosis after reaching hospital.
All patients should have assessment of the adequacy of their airway protection and ventilation. Ventilate any patient who can't protect their airway, or has a respiratory acidosis. Comatose patients require management in ICU and will need to be intubated if they are to have gastrointestinal decontamination safely. Ventilation should not be the primary method of achieving systemic alkalinisation (see sodium bicarbonate); mild hyperventilation can maintain a patient in alkalosis but the CO2 should not go below 25 mmHg.
Patients should have continuous ECG monitoring for at least 6 hours after ingestion.
All patients should have intravenous fluids (normal saline).
After cardiac arrest, prolonged resuscitation may be successful and should be continued for at least 1 hour.
Most patients with massive ingestions will be unconscious or have a deteriorating level of consciousness by 2 hours and should be intubated. If the patient is unconscious and requires intubation to protect the airway insert an orogastric tube, aspirate stomach contents then give activated charcoal.
If patients are alert and co-operative and have ingested > 5 mg/kg, charcoal may be administered orally. In practice, this is only relevant for patients who present within 1 - 2 hours of ingestion.
There are no readily available specific antidotes for the most serious manifestations.
Seizures should, initially, be treated with diazepam 5-20 mg IV followed by phenobarbitone 15-18 mg/kg IV in the event of ongoing seizures and elective intubation and ventilation. If neuromuscular blockade is required for management, EEG monitoring is mandatory.
The major complication of seizures is increased acidosis precipitating the subsequent development of major cardiovascular toxicity. All patients who have seized should be assessed for the presence of indications to receive sodium bicarbonate.
Phenytoin should be avoided because of its sodium-channel blocking properties which may exacerbate tricyclic toxicity.
Mild delirium can often be managed with reassurance plus or minus benzodiazepines. Severe delirium generally requires large doses of benzodiazepines; neuroleptics (most of which have significant anticholinergic activity) should be avoided. Although physostigmine is effective, the short half-life of this drug and its occasional life threatening adverse effects (particularly in TCA overdose) make it relatively contraindicated in patients with evidence of sodium channel blockade.
It is often very difficult to distinguish whether the patient is having a supraventricular arrhythmia with aberrant conduction or primary ventricular tachycardia. Most arrhythmias, especially if they are associated with low output, are treated in a standard cardiac arrest protocol manner. The main difference is the requirement for early and large doses of NaHCO3 (bicarbonate). Hypertonic saline can be considered in refractory cases.
Arrhythmias are best treated by correction of hypoxia and acidosis. This is achieved by administering oxygen and by alkalinisation with bolus injections of intravenous sodium bicarbonate (1 to 2 meq of 8.4%). Further doses of sodium bicarbonate may be given cautiously depending on clinical response to achieve an arterial pH of 7.5-7.55 but care is needed to avoid a sustained alkalosis as pH>7.65 is potentially fatal.
Antiarrhythmic agents should be avoided unless arrhythmias are unresponsive to the aforementioned measures. In these situations, magnesium sulfate or lidocaine may be used to treat refractory ventricular arrhythmias. Class Ia (quinidine,disopyramide,procainamide) and class Ic antiarrhythmic drugs (flecainide,propafenone) are contraindicated as they may worsen sodium channel blockade and exacerbate arrhythmias.
Both sodium loading and alkalinisation have been show to be effective in reversing TCA induced conduction defects and hypotension (McCabe et al, 1998), both are supplied by hypertonic sodium bicarbonate. Sodium bicarbonate is the drug of choice for the treatment of ventricular dysrhythmias and/or hypotension due to TCA poisoning (Albertson et al, 2001).
Treatment with plasma alkalinisation to a pH of 7.50 - 7.55 using sodium bicarbonate (to alter both pH and sodium) and/or hyperventilation is effective for all TCA induced arrhythmias. The initial treatment in critically ill patients is often titrated against clinical response with bolus injections of 1-3 mEq of sodium bicarbonate per kg body weight repeated at 3-5 minute intervals. When the clinical situation allows it, arterial blood pH should be checked. The target pH is 7.50 - 7.55, sustained elevations of pH greater than this are associated with impaired oxygen dissociation from hemoglobin. As the patient is usually ventilated the pH can be maintained with mild hyperventilation (pCO2 = 30 mmHg).
Alkalosis causes a decrease in the free drug concentrations by increasing protein binding. In addition, and more importantly, alkalosis affects the partitioning of TCAs between the cell membrane and the Na+ channel binding site and decreases TCA induced Na+ channel blockade. It has also been shown that the administration of hypertonic saline in any form will also improve cardiac conduction. This may be through some indirect effects on angiotensin.
Administration of hypertonic saline had greater efficacy than alkalinisation in improving cardiac conduction and hypotension in a swine model.(McCabe et al, 1998). The dose used in this study was 10 mL/kg of hypertonic saline solution in the form of a 7.5% sodium chloride solution (15 mEq/kg Na Cl). The mean peak serum sodium after treatment was 157 mEq/L. This dose was selected as it had been the maximal dose used previously in the treatment of trauma in humans.
As standard supportive care and sodium bicarbonate have been so effective, the clinical use of hypertonic saline is rarely reported. A 34 year old man’s dothiepin induced ventricular tachycardia and hypotension responded to 170 mmol of hypertonic saline given over 5 minutes, subsequent episodes responded to bolus doses of 100 mmol (Hoegholm & Clementsen, 1991) A 29 year old woman's refractory hypotension and conduction defects responded to 200 mL of 7.5% saline infused over 3 minutes.(McKinney & Rasmussen, 2003) A 6 year old with imipramine had no response to a slow infusion of 60 meq of hypertonic saline (Dolara & Franconi, 1977). Currently its role is undefined but it could be considered in situations of refractory hypotension and arrhythmia. It would appear that it should be administered as a bolus injection rather than infusion. The potential risks of this treatment include fluid and sodium overload and complications of rapidly increasing plasma sodium such as central pontine myelinosis.
Further drug treatment of arrhythmias
All other treatments are of questionable efficacy and safety and therefore controversial.
All class 1a antiarrhythmic drugs are contraindicated and lignocaine and phenytoin (class 1b drugs) while they may be used may still exacerbate Na+ channel blockade and potentially exacerbate arrhythmias (e.g. convert VT into asystole).
Magnesium is normally the drug of choice for treating torsade de pointes and is used for refractory arrhythmias in other settings. However, its calcium channel blocking activity may aggravate the hypotension and heart block that can complicate TCA poisoning. Overdrive pacing should be considered for refractory ventricular tachycardia (Peters et al, 1992).
Second or third degree heart block should be treated with bicarbonate and isoprenaline followed by a pacemaker.
Intralipid has been reported as superior to sodium bicarbonate in a rat model of Clomipramine poisoning (Harvey & Cave 2007), and several case reports advocate its use in humans. At this stage its use should (arguably) by reserved for salvage situations not responding to better evidenced interventions.
This usually responds to volume expansion and pH correction. Hypertonic saline can be considered in refractory cases.
Refractory hypotension may require drugs with alpha agonist properties (e.g. adrenaline & noradrenaline) but these should be used cautiously, if at all, in this setting as they may precipitate ventricular tachycardia.
Hypotension should be treated by administration of intravenous colloids to expand the intravascular volume. Central venous pressure monitoring is advisable to avoid fluid overload.
Alkalinisation with repeated intravenous bicarbonate (1 to 1.5 meq of 8.4%) may correct hypotension. Further doses of sodium bicarbonate may be given cautiously depending on clinical response to achieve an arterial pH of 7.5-7.55 but care is needed as alkalosis to a pH>7.65 is potentially fatal. Administration of hypertonic saline (7.5%) has also been reported to correct hypotension but the evidence base for this treatment is weaker and there is a risk of causing hypernatraemia.
If hypotension is refractory to the above measures and adequate filling pressures, vasopressor support with pure alpha agonist activity such as norepinephrine should be initiated.
If hypotension does not improve, non-catecholaminergic inotropic agents such as glucagon (10 mg iv bolus) and vasopressin (infusion) have anecdotally been successful at improving refractory hypotension.
If available, mechanical support with intra-aortic balloon pump insertion or ECMO(extra-corporeal membrane oxygenation) should be considered.
Repeated doses of activated charcoal may increase the clearance of the drug, but there is no evidence of a change in clinical outcome and it should not be routinely used. It may be considered for modified-release preparations.
Patients are medically fit for discharge if they have no symptoms or signs of toxicity (including no anticholinergic features such as tachycardia) and a normal ECG six hours following the overdose (especially if they have passed a charcoal stool).
Patients who still have an isolated tachycardia generally should be kept in hospital and observed. As the usually cause is volume depletion, IV fluid to ensure adequate volume replacement should be given.
Patients with a QRS complex equal to or greater than 100 milliseconds should be monitored until this has returned to normal.
Reports of occasional late cardiac arrests have occurred in patients with persistently abnormal ECGs. Continued drug absorption or exposure may be the cause for late deterioration in this group and markedly delayed toxicity (> 24 hours) has only been reported in patients who did not receive gastrointestinal decontamination or, more recently, in modified release amitriptyline overdose.
Antiholinergic delirium is relatively common post resolution of acute toxicity, and serotonin syndrome may be seen where a serotonergic drug has been co-inigested.
Neurological sequelae and other end organ injury may occur as a consequence of hypotension and cellular hypoxia in the setting of cardiac arrest or profound toxicity. They are not expected as a consequence of elevated drug levels per se, and are not seen post resolution of lesser toxicity.
Hoffman JR, Votey SR, Bayer M et al. Effect of hypertonic sodium bicarbonate in the treatment of moderate-to-severe cyclic antidepressant overdose. Am J Emerg Med 1993; 11: 336-341.
Tran TP, Panacek EA, Rhee KJ et al. Response to dopamine vs norepinephrine in tricyclic antidepressant-induced hypotension. Acad Emerg Med 1997; 4: 864-868.
Thanacoody HKR, Thomas SHL. Tricyclic Antidepressant poisoning: Cardiovascular Toxicity. Toxicological Reviews 2005; 24(3):205-14.
Bradberry SM, Thanacoody HKR, Watt BE, Thomas SHL, Vale JA. Management of the Cardiovascular Complications of Tricylic Antidepressant Poisoning. Role of Sodium Bicarbonate. Toxicological Reviews 2005; 24(3):195-204.
Albertson TE, Dawson A, de Latorre F, Hoffman RS, Hollander JE, Jaeger A et al. TOX-ACLS: toxicologic-oriented advanced cardiac life support. Ann Emerg Med 2001; 37(4 Suppl):S78-S90.
Amitai Y, Erickson T, Kennedy EJ, Leikin JB, Hryhorczuk DO, Noble J et al. Tricyclic antidepressants in red cells and plasma: correlation with impaired intraventricular conduction in acute overdose. Clin Pharmacol Ther 1993; 54(2):219-227.
Ansel GM, Coyne K, Arnold S, Nelson SD. Mechanism of ventricular arrhythmia during amitriptyline toxicity. Journal of Cardiovascular Pharmacology 1993;22:798-803.
Boehnert MT, Lovejoy FH. Value of the QRS duration versus the serum drug concentration in predicting seizures and ventricular arrhythmias after an acute overdose of tricyclic antidepressants. N Engl J Med 1985;313:474-79
Buckley NA, Dawson AH, Whyte IM, Henry DA. Greater toxicity of dothiepin in overdose than of other tricyclic antidepressants. Lancet 1994;343:159-162.
Buckley NA, Dawson AH, Whyte IM, McManus P & Ferguson N. Self-poisoning in Newcastle, 1987-1992. Med J Aust 1995;162:190-193.
Buckley NA, O'Connell DL, Whyte IM, Dawson AH. Interrater agreement in the measurement of QRS interval in tricyclic antidepressant overdose: Implications for monitoring and research. Ann Emerg Med 1996; 28(5):515-519.
Buckley NA, McManus PR. Can the fatal toxicity of antidepressant drugs be predicted with pharmacological and toxicological data? Drug Saf. 1998 May;18(5):369-81.
Callaham M & Kassell D. Epidemiology of fatal tricyclic antidepressant ingestion: Implications for management. Ann Emerg Med 1985;14:29-37.
Caravati EM, Bossart PJ. Demographic and electrocardiographic factors associated with severe tricyclic antidepressant toxicity. J Toxicol Clin Toxicol 1991;29(1):31-43.
Chevalier S, Buckley NA, OConnell D, Leitch JW. ECG predictors of arrhythmias after tricyclic and thioridazine overdose. Presented at the Cardiac Society of Australia and New Zealand Annual conference, Canberra, August 1995.
Crome P. The toxicity of drugs used for suicide. Acta Psychiatrica Scandinavica, Supplementum 1993; 371:33-37.
Dahlin KL, Lastbom L, Blomgren B, Ryrfeldt A. Acute lung failure induced by tricyclic antidepressants. Toxicol Appl Pharmacol 1997; 146(2):309-316.
Dolara P, Franconi F. Hypertonic sodium chloride and lidocaine in a case of imipramine intoxication. Clin Toxicol 1977; 10(4):395-398.
Faber TS, Zehender M, Just H. Drug-induced torsade de pointes. Incidence, management and prevention. Drug Safety 1994;11(6):463-76.
Foulke GE, Albertson TE, Walby WF. Tricyclic antidepressant overdose: Emergency department findings as predictors of clinical course. Am J Emerg Med 1986;4(6):496-500.
Foulke GE, Albertson TE. QRS interval in Tricyclic antidepressant overdosage: Inaccuracy as a Toxicity indicator in Emergency Settings. Ann Emerg Med 1987;16(2)-:39-42.
Goldgran-Toledano D, Sideris G, Kevorkian JP. Overdose of cyclic antidepressants and the Brugada syndrome. N Engl J Med 2002; 346(20):1591-1592
Greene SL, Dargan PI, Antoniades H, Wyncoll D, Jones AL. Delayed clinical effects in modified-release amitriptyline poisoning. A case report with toxicokinetic data. J Toxicol Clin Toxicol 2002; 40(3):EAPCCT abstract no 138.
Harrigan RA, Brady WJ. ECG abnormalities in tricyclic antidepressant ingestion. Am J Emerg Med 1999; 17(4):387-393
Heard K, Cain BS, Dart RC, Cairns CB. Tricyclic Antidepressants Directly Depress Human Myocardial Mechanical Function Independent of Effects on the Conduction System. Acad Emerg Med 2001; 8(12):1122
Henry JA, Alexander CA, Sener EK. Relative mortality from overdose of antidepressants [published erratum appears in BMJ 1995 Apr 8;310(6984):911]. BMJ 1995; 310(6974):221-224.
Hoegholm A, Clementsen P. Hypertonic sodium chloride in severe antidepressant overdosage. Journal of Toxicology - Clinical Toxicology 1991; 29(2):297-298.
Knudsen K, Abrahamsson J. Effects of magnesium sulfate and lidocaine in the treatment of ventricular arrhythmias in experimental amitryptiline poisoning in the rat. Critical Care Medicine 1994;22(3):494-498.
Liebelt EL, Francis PD, Woolf AD. ECG lead aVR versus QRS interval in predicting seizures and arrhythmias in acute tricyclic antidepressant toxicity. Ann Emerg Med 1995; 26(2):195-201.
Liu X, Emery CJ, Laude E, Herget J, Gill G, Cope G, Barer GR. Adverse pulmonary vascular effects of high dose tricyclic antidepressants: acute and chronic animal studies. Eur Respir J 2002; 20(2):344-352.
McCabe JL, Cobaugh DJ, Menegazzi JJ, Fata J. Experimental tricyclic antidepressant toxicity: a randomized, controlled comparison of hypertonic saline solution, sodium bicarbonate, and hyperventilation. Ann Emerg Med 1998; 32(3 Pt 1):329-333.
McKinney PE, Rasmussen R. Reversal of severe tricyclic antidepressant-induced cardiotoxicity with intravenous hypertonic saline solution. Ann Emerg Med 2003; 42(1):20-24.
Niemann JT, Bessen HA, Rothstein RJ, Laks MM. Electrocardiographic criteria for tricyclic antidepressant cardiotoxicity. Am J Cardiol 1986;57: 1154-1159.
Teschemacher AG, Seward EP, Hancox JC, Witchel HJ. Inhibition of the current of heterologously expressed HERG potassium channels by imipramine and amitriptyline. Br J Pharmacol 1999; 128(2):479-485.
Wolfe TR, Caravati EM, Rollins DE. Terminal 40-ms frontal plane QRS axis as a marker for tricyclic antidepressant overdose. Ann Emerg Med 1989;18:348-
Weinbach EC, Costa JL, Nelson BD, Claggett CE, Hundal T, Bradley D et al. Effects of tricyclic antidepressant drugs on energy-linked reactions in mitochondria. Biochemical Pharmacology 1986; 35(9):1445-1451
Author NAB 21/12/99 Reviewed IMW 24/5/02, AHD 3/8/2003, AHD 18/10/03