# European Heart Journal

### Editor-In-Chief

Thomas F. Lüscher

### Deputy Editors

Jeroen BaxBernard J. GershGerhard HindricksUlf LandmesserChristian M. MatterFrank RuschitzkaWilliam Wijns

15.064

Published on behalf of

## Risk of arrhythmia induced by psychotropic medications: a proposal for clinical management

Søren Fanoe, Diana Kristensen, Anders Fink-Jensen, Henrik Kjærulf Jensen, Egon Toft, Jimmi Nielsen, Poul Videbech, Steen Pehrson, Henning Bundgaard
1306-1315 First published online: 18 March 2014

## Abstract

Several drugs used in the treatment of mental diseases are associated with an increased risk of sudden cardiac death (SCD). A general cause-relationship between the intake of these drugs and SCD is unattainable, but numerous case reports of drug-induced malignant arrhythmia and epidemiological studies, associating the use of specific drugs with SCD, strongly support the presence of an increased risk. Whereas the absolute risk of drug-induced life-threatening arrhythmia may be relatively low, even small increments in risk of SCD may have a major health impact considering that millions of patients are treated with psychotropics. In subgroups of pre-disposed patients, e.g. patients with cardiac diseases or other co-morbidities, the elderly or patients treated with other negatively interacting drugs, the absolute risk of drug-induced arrhythmia may be considerable. On the other hand, several of the major mental disorders are associated with a large risk of suicide if untreated. The observed risk of malignant arrhythmia associated with treatment with psychotropic drugs calls for clinical guidelines integrating the risk of the individual drug and other potentially interacting risk factors. In this review, data from various authorities on the risk of arrhythmia associated with psychotropic medications were weighted and categorized into three risk categories. Additionally, we suggest a clinically applicable algorithm to reduce the risk of malignant arrhythmia in patients to be treated with psychotropic medications. The algorithm integrates the risk categories of the individual drugs and pre-disposing risk factors and suggests a prudent follow-up for patients with an increased risk. We believe this clinically manageable guideline might improve safety in the many and rapidly increasing number of patients on psychotropic drugs.

• Acquired long QT
• Pro-arrhythmia
• Psychotropics
• Torsade de Pointes ventricular tachycardia
• Drug-induced
• Sudden cardiac death

## Introduction

Several studies have reported an increased incidence of sudden cardiac death (SCD) in patients treated with anti-psychotic or anti-depressant drugs.13 Ray et al.4 demonstrated a two-fold increase in the incidence–rate ratio for SCD in current users of antipsychotics compared with non-users and former users and Weeke et al.5 showed that treatment with certain anti-depressants was associated with up to a 5- to 6-fold increase in the incidence of out-of-hospital cardiac arrest. On the other hand, untreated depression per se more than doubles the risk of SCD,6,7 which adds to the overall risk of SCD in these patients. In the study by Whang et al.1 depression was associated with an increased risk of fatal ischaemic heart disease even when controlling for cardiovascular risk factors, although an unfavourable cardiovascular risk profile is frequent in several subgroups of patients with psychiatric disorders.8,9 Most recently, Khan et al.10 reported a three–four-fold increased mortality risk associated with schizophrenia, major depression, and bipolar mood disorder.

Mental disorders account for about one-third of non-communicable diseases in Europe, and in the EU population the estimated cumulative risk for developing a mental disorder up to the age of 65 years is about 50%11 and others have shown that about one-third receive pharmaco-therapy.12 Thus, even a small risk of SCD associated with psychotropic medications may be a major health issue.

Several drugs associated with SCD have the propensity of prolonging the QT interval,1315 which is considered a substrate for the potentially life-threatening ventricular tachycardia, Torsade de Pointes (TdP). Therefore, drug-induced QT prolongation is generally used as a proxy for an increased risk of TdP, i.e. risk of SCD. Torsade de Pointes may also present as palpitations, dizziness, or syncope. Drug-induced QT-interval prolongation is most often due to a dose-dependent inhibition of the cellular Ikr current through channels coded by the hERG gene.16,17

In 2005, the International Conference on Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) published guidelines to evaluate QT interval prolongation and pro-arrhythmic potential of non-Antiarrhythmic Drugs (ICH-E14).18 With few exceptions, all new drugs have to undergo a ‘thorough QT study’ (TQT) defined by the ICH-E14. However, the majority of the presently used drugs were marketed before TQT studies were mandatory. Additionally, even a well-performed TQT study cannot rule out pro-arrhythmia when the drug is used in large scale in the clinical situation, where patients often receive multiple drugs, have co-morbid substance abuse or even existing heart diseases. Most importantly, several antipsychotics and antidepressants on the market are known to induce QT prolongation. On this basis, psychiatrists and other physicians need to be able to assess and handle a potential risk of drug-induced QT prolongation.

In the TQT studies, the drug-induced QT-interval prolongation is used as a surrogate risk marker and the threshold level of regulatory concern is an increase of around 5 ms above baseline, while a QT duration above 500 ms or an increase ≥60 ms above baseline are commonly used as threshold values in clinical practice. The exact criteria for the evaluation in the TQT study depends on the indication for the tested drug.19,20

Several common cardiovascular conditions including ischaemic heart disease, arterial hypertension, heart failure, and brady-arrhythmias, but also a broad range of less common conditions like cardiomyopathies and primary arrhythmia disorders, predispose to the development of drug-induced arrhythmia. Interactions with other concomitantly used drugs, including potassium- and magnesium-wasting diuretics, CYP3A4 inhibitors, and other QT prolonging drugs, e.g. antibiotics and antifungals may also increase the likelihood of serious arrhythmia. In order to improve patient safety, clinical guidelines integrating these many potentially interacting factors are warranted and collaboration between psychiatrists and cardiologists needed. We propose a clinically manageable guideline for reduction of the risk of arrhythmia induced by drugs administered to patients with psychiatric disorders (Figure 1). The guideline also suggests a balanced risk–benefit relation in patients in great need of using these drugs. The proposal was developed in collaboration between the Danish Society of Cardiology and the Danish Psychiatric Society21 (see Appendix).

Figure 1

An algorithm for lowering the risk of cardiac arrhythmia during treatment with psychotropic medications. When a class B or B* drug (Table 2) is chosen, assessment of the cardiac risk profile is recommended. If cardiac risks are identified—the cardiac risk factors should be optimized and/or a drug with a more favourable risk profile should be chosen. Re-evaluation of the ECG and symptoms should take place within 1 to 2 weeks after (≅ 5 half lives) initiation of treatment with class B/B* drugs.

## Methods

In order to assess the risk of arrhythmia associated with psychotropic medications, we classified each drug according to the reported effects at therapeutic levels on the QT interval, induction of arrhythmia, and cardiac conduction disturbances. Data were collected from the following five sources:

1. European Medicines Agency, EU (EMEA),22

2. The US Food and Drug Administration (FDA),23

3. The Maudsley prescribing guideline,24

4. The independent databases Micromedex (Tuven Health Analytics),25 and

5. The Arizona Center for Education and Research on Therapeutics (http://crediblemeds.org) (26 February 2014, date last accessed).

Cardiac data from these five sources were reviewed and additionally, for each drug we reviewed the accessible publications of thorough QT studies. For each drug, the findings reported in each of these sources were carefully reviewed and the combined data weighted; the main—and most important—distinction was made between drugs without an associated risk of arrhythmia (Class A drug) and drugs with such a risk (Class B drug). The latter group was further subdivided into Class B* drugs. These three categories (Table 1) are modified from qtdrugs.org—now http://crediblemeds.org: drugs with neither QT-prolongation nor TdP risk corresponds to Class A. The category ‘possible TdP risk’ corresponds to Class B. The categories ‘drugs to be avoided by congenital long QT’, ‘Drugs with conditional TdP risk’, and ‘drugs with known TdP risk’ are merged into Class B*. In Table 2, psychotropic drugs are classified according to these categories.

View this table:
Table 1

Classification of psychotropic medications according to the risk of QT prolongation and arrhythmia

 Class A A drug considered to be without any risk of QT prolongation or TdP Class B A drug with a propensity of inducing QT prolongation Class B* A drug with pronounced QT prolongation, documented cases of TdP, or other serious arrhythmias
View this table:
Table 2

Categorization of psychotropic medications according to the reported risks for induction of cardiac arrhythmia

DrugEMAFDAMicromedexMaudsleyArizonaThorough QT studyThe weighted recommendations
Antipsychotic drugs
AmisulprideYesYesLittleNoB
AripiprazoleNoNoNoNoNoA
ChlorprothixenYesNoNoB
ClozapineNoNoNoLittlePossibleNoB
FlupenthixolYesNoLittleNoB
HaloperidolYesYesYesMuchYesNoB*
LevomepromazineYesNoB
OlanzapineYesNoYesSomeNoA
PaliperidoneYesYesYesNoPossibleYes104B
PerphenazineYesNoLittleNoA
PimozideYesYesYesMuchYesNoB*
QuetiapineYesNoNoSomePossibleYes105B
RisperidoneYesNoYesLittlePossibleNoB
SertindoleYesYesMuchPossibleNoB*
SulpirideYesNoNoB
ZiprasidoneYesYesYesSomePossibleYes105B*
ZuclopenthixoleYesNoNoA
TCA and MAO inhibitors
AmitriptylineBlockYesSomeConditional riskNoB
ClomipramineYesNoNoSomeConditional riskNoB
DoxepinBlockNoNoSomeConditional riskYes106B
ImipramineYesYesSomeConditional riskNoB
IsocarboxazidNoNoNoNoA
MoclobemideYesNoNoB
NortriptylineYesBlockYesSomeConditional riskNoB
Neurotransmitter uptake inhibitors
CitalopramYesYesYesLittleConditional riskYes39B
EscitalopramYesYesYesYes39B
FluoxetineNoNoNoConditional riskNoA
ParoxetineNoNoNoNoConditional riskNoA
SertralineNoNoNoNoConditional riskNoA
DuloxetineNoNoNoYes107A
ReboxetineNoNoNoYes108A
VenlafaxineYesNoYesLittlePossibleNoB
MianserinNoNoNoA
MirtazapineNoNoNoNoNoA
AgomelatineNoYes109A
BupropionNoNoNoNoA
Mood stabilizers
LithiumBlockBlockNoLittlePossibleNoBa
CarbamazepineNoBlockNoNoNoA
LamotrigineNoNoNoNoYes110A
ValproateNoNoNoNoNoA
Anxiolytic drugs
BenzodiazepinesNoNoNoNoA
GabapentinNoNoNoYes111A
PregabalineNoNoNoNoA
Anticholinergic drugs
BiperidenNoNoNoA
Opioid substitutionNo
BuprenorphineNoNoNoNoA
• The cardiac risk reported from the selected sources.

• ‘Yes’ indicates that the source in question has reported a risk of arrhythmia or QT prolongation, ‘No’ indicates that no such risk has been reported. ‘Block’ indicates that the AV block has been reported. ‘Conditional risk’ indicates that these drugs prolong QT and have a risk of inducing TdP under certain conditions. ‘Thorough QT study’ indicates whether a thorough QT study has been published or described in the summery of product characteristics for the particular medicament. The weighted recommendations represent the classification of psychotropic medications according to Table 1.

• aRegarding a QT prolonging effect of lithium, reports are divergent,44,45 but bradycardia, T wave changes and AV block have been described.46

The accessibility of TQT studies was investigated for each drug through searching the following sources: the summery of products characteristics, www.clinicaltrials.gov (26 February 2014, date last accessed), and the US National Library of Medicine National Institutes of Health.

The class of recommendation (Table A1) and class of evidence (Table A2) for the present guideline recommendations are given in Table A3 (see Appendix).

## Psychotropic drugs

### Antipsychotics

An association between treatment with the older, typical antipsychotic drugs and development of QT-interval prolongation, TdP, and SCD was established several years ago.26,27 The newer atypical antipsychotic drugs were considered safer, but in a registry study Ray et al.4 found that users of typical and atypical antipsychotic drugs had similarly increased risks of SCD. For both groups, the risk increased dose dependently with adjusted incidence-rate ratios from 1.31 to 2.42 and from 1.59 to 2.86 in users of typical and atypical antipsychotics, respectively. For individual drugs, the highest incidence-rate ratio was 5.05 (thioridazine). Importantly, former users had no increased risk. The risk was twice as high in men. For persons 70–74 years of age, the risk was 10-fold higher than in those 30–35 years of age. A meta-analysis of randomized trials demonstrated a higher death rate in patients with Alzheimer disease randomized to antipsychotic drugs compared with placebo (OR = 1.54). There were no differences in death rates between treatment groups among patients dropping out of these trials.28 The mortality associated with different antipsychotics has been explored in head to head randomized trials of sertindole vs. risperidone in the SCoP study29 (n = 14 147 patient-years) and of ziprazidone vs. olanzapine in the ZODIAC study30 (n = 18 154 patients followed for 1 year). No difference in all-cause mortality was found in the two studies. However, treatment with sertindole was associated with an increased cardiac mortality compared with risperidone (HR: 2.84; CI: 1.45–5.55).

Haloperidol is widely used and prolongs the QTc interval ∼5 ms, but it is the drug with most documented cases of TdP-VT. This may reflect the extensive use of haloperidol in somatic settings in patients with concurrent somatic conditions, predisposing to the development of arrhythmias.15,31,32

### Antidepressants

#### Tricyclic anti-depressants (TCAs)

Most tricyclic anti-depressants (TCAs) seem to prolong the QT interval, but reports of TdP are few and especially related to treatment with amitriptylin and maprotillin.33 Tricyclic anti-depressants have been shown to delay the AV-node conduction resulting in AV block.34 In a registry study, treatment with TCA was associated with an increased risk of cardiac arrest (OR = 1.69).5 The mean age of this population was 67 years (95% CI 58–75) and several had serious somatic comorbidity.

#### Selective neurotransmitter re-uptake inhibitors (SSRI and SNRI)

Selective serotonin re-uptake inhibitors (SSRI) and serotonin–norepinephrine reuptake inhibitors (SNRI) are widely used. The single most commonly used antidepressant, the SSRI citalopram reached about 452.6 million defined daily doses in the UK in 2009.35 These drugs are generally regarded as safe even in overdoses.3638 However, as a consequence of recently reported QT studies, both FDA39 and EMA40 have limited the recommended maximum doses of citalopram and escitalopram. For patients older than 60 years of age, the maximum recommended dose is further reduced. A QT-interval prolonging effect of certain antidepressants was most recently confirmed in a large-scale pharmacovigilance study.41 In a Danish nationwide registry study, Weeke et al.5 assessed the treatment with anti-depressants and the risk of out of hospital cardiac arrest. Overall, the treatment with any antidepressant was significantly associated with cardiac arrest (OR = 1.23). Tricyclic anti-depressants and SSRIs significantly increased the risk of cardiac arrest (OR 1.69 and 1.21, respectively), whereas no association was observed for SNRI. Patients treated with SSRI were older than those on TCA (74 vs. 67 years).

#### Mood stabilizers

Carbamazepine, lamotrigene, valproate, and lithium are used as mood stabilizers. The anti-convulsants have generally not been associated with severe arrhythmia. Lithium has been used in the treatment of bipolar disorders for many years and it is well-known that treatment should be monitored.42 Caution should be taken in patients treated concomitantly with anti-arrhythmic drugs.43 Regarding a QT prolonging effect of lithium, reports are divergent,44,45 but bradycardia, T wave changes and AV-block have been described.46

#### Anxiolytic agents

Benzodiazepines and pregabalin, which are bound more selectively to the GABA receptors, are widely used in treatment of anxiety. Benzodiazepines comprise a heterogeneous group of drugs. In vitro studies have shown both inhibition and activation of potassium currents during exposure to benzodiazepines,47,48 but no changes in QT duration have been reported in clinical use.

#### Medications for treatment of opioid addiction

Methadone causes pronounced QT-prolongation4951 and several cases of TdP have been reported.52 The Substance Abuse & Mental Health Services Administration (SAMHSA) has proposed a cardiac risk management plan for methadone maintenance treatment programs.53 SAMHSA recommends that a baseline ECG be recorded if the patient has risk factors and that ECGs for such patients should be recorded annually or when the daily dose exceeds 120 mg. However, both methadone dose and baseline QT length are predictors of QT prolongation.54 Thus, we recommend (Figure 1) baseline and follow-up ECG's, including an additional evaluation if the daily dosage exceeds 100 mg, as recommended in the guideline by Krantz et al.55

Compared with methadone, the alternative buprenorphine causes far less prolongation of the QT interval56 and TdP has not been reported. Unfortunately, buprenorphine is a partial μ-receptor agonist, which makes it less effective compared with methadone in patients requiring high doses.57

## Polypharmacy

In North America and Europe, ∼20% of patients treated for schizophrenia receive more than one antipsychotic drug.58,59 The efficacy and safety of combination therapy regimens have only to a very limited extent been evaluated in randomized trials.60 Concomitant use of different drugs with the propensity of prolonging the QT interval may lead to unpredictable QT prolongation and should, from a drug safety point of view, be avoided if possible.61 Similarly, it is important to recognize that drugs interacting with the metabolism of a QT-prolonging drug can result in higher plasma concentrations and thereby increase the risk of arrhythmia. Thus, it has been shown that concomitant use of erythromycin and other inhibitors of the hepatic CYP3A isoenzyme is associated with a higher risk of SCD.62 Furthermore, a number of other legal or illegal substances, e.g. alcohol and cocaine, may also affect the repolarization.63,64 It is therefore important to know all co-medications, including over-the-counter medications, and it is generally recommended to avoid concomitant use of more than one drug with the propensity of prolonging the QT interval or drugs with pharmacokinetic interactions.61

## Heart diseases pre-disposing to arrhythmia

Ischaemic heart disease is the most prevalent heart disease in developed countries and attributable to ∼15% of all deaths.65,66 Both acute and chronic ischaemic heart diseases, including previous myocardial infarcts, are associated with SCD.67 The substrate for arrhythmia is considered to be re-entry due to changes in conduction velocities and abnormal automaticity.68

The prevalence of heart failure is ∼2% in Western countries69 and these patients are pre-disposed to malignant arrhythmia. A large proportion of the patients with heart failure die suddenly (SCD) due to arrhythmia.70 This is especially the case in patients with heart failure due to ischaemic heart disease.71 Subgroups of patients with inherited cardiomyopathies, e.g. hypertrophic cardiomyopathy, arrhythmogenic right ventricular cardiomyopathy, and dilated cardiomyopathy, are at increased risk of SCD.72,73

In the structural congenital heart diseases, the abnormalities as well as chirurgical sequelae may predispose to development of arrhythmia.74 Pre-existing QT-interval prolongation is seen in patients with genetically determined changes in ion-channel function, i.e. inherited long QT syndrome.75 The degree of prolongation of the QT interval is in general considered to reflect the risk of serious cardiac events in patients with inherited long QT syndrome.7577 There is a considerable overlap between the QT of healthy individuals and the QT of carriers of a long QT mutation.78,79 The inherited arrhythmic syndromes, e.g. Brugada syndrome and long QT syndrome, may be triggered by electrolyte disturbances or exposure to specific drugs.80,81 In addition to the arrhythmia associated with several heart diseases per se, the medical treatment of these diseases may add to the risk for development of arrhythmias. This may relate to diuretic-induced hypokalaemia, effects of CYP-system inhibitors, e.g. verapamil, drugs with intrinsic pro-arrhythmic side effects, e.g. flecainide, or drugs with the propensity of prolonging the QT-interval, e.g. sotalol or drugs associated with TdP.

## Cardiac risk factors in the psychiatric patient

Heart diseases are more prevalent in patients with psychiatric diseases.82 This has in part been attributed to life style factors like smoking, poor treatment, and poor treatment compliance,8 but several other factors have been proposed including genes common for both disease groups.83 On the other hand, the prevalence of depression is high among patients with ischaemic heart disease and is associated with a higher mortality.84 Most importantly, a large proportion of the patients treated with psychotropic medications are the elderly, and the risk of ischaemic heart disease increases dramatically with age.69 Elderly people with ischaemic heart disease have the highest rate of SCD—and must be considered to be a high-risk group when exposed to drugs with a pro-arrhythmic potential.

## Management of the individual patient

A rational approach to reduce the risk of serious arrhythmia in a patient considered for treatment with a pro-arrhythmic drug is to make a structured risk assessment before commencement of treatment. The assessment needs to include a medical history of heart disease, present and former cardiac symptoms with focus on chest pain, dyspnoea, palpitations, near-syncopes or syncopes, and family history of SCD; a list of medications to assess for possible drug interactions and other QT-prolonging drugs or potassium-wasting drugs; an ECG for assessment of signs of heart disease, conduction disorders, or prolonged QT interval.

The decision to commence treatment with psychotropic medications includes several steps: evaluation of the severity of the psychiatric condition and the need for pharmaco-therapy; selection of the appropriate psychotropic medication—at the recommended dose; assessment of the cardiac risk associated with the chosen medication; if a class A drug has been chosen (Table 2), the treatment can be commenced without any further cardiac risk assessment. If a class B or B* drug (Table 2) is chosen, assessment of the cardiac risk is needed according to the proposed algorithm (Figure 1); if cardiac risks are identified, the cardiac risk factors should be optimized and/or a drug with a more favourable risk profile should be chosen if possible. In case of structural heart disease, QT prolongation, electrolyte disturbances, or cardiac symptoms referral to a cardiologist should be considered. Follow-up should be organized according to the algorithm (Figure 1). Re-evaluation of the ECG and symptoms should take place within 1 or 2 weeks (i.e. at steady-state >5 drug half lives) after initiation of treatment with class B/B* drugs. Similarly, a significant increase in dose of these drugs necessitates re-evaluating symptoms and a new ECG. A QTc-interval above 500 ms or an increment above 60 ms when compared with baseline are generally considered associated with a definitely increased risk of TdP85 and should in most cases lead to discontinuation of the drug (Figure 1).

If the psychiatric condition is invalidating or life threatening, a higher cardiac risk may be accepted, but necessitates reductions of all reversible risk factors and a close follow-up.

## Discussion

Several drugs used in the treatment of mental diseases are associated with an increased risk of SCD, even though a direct causative link between the drug intake and SCD is difficult to prove. However, the association has repeatedly been reported from various sources and therefore, the risk need to be taken into consideration in daily clinical practice. The incidence of drug-induced long-QT has been estimated to about 11/1 000 000 in France. This is probably an underestimate, since fatal cases were excluded.86

The magnitude of the SCD risk is of major importance, but the true incidence of SCD in the general population is not known. The estimated US annual incidence of SCD varies from 180 000 to >450 00087 out of a total of around 2.5 million deaths per year,88 i.e. 7–18% (or 60–150 per 100 000 person-years) of all deaths across all age-groups are categorized as SCD. In most of these cases, the cause of death is ischaemic heart disease in older patients, where the annual incidence of SCD for 75 years old men reaches 800 per 100 000.73 Additionally, it has been estimated that out-of-hospital cardiac arrest occurs in about 80 per 100 000 person-years in subjects without heart disease.69 In children and in young adults (1–35 years), the SCD rate is generally considered to be much lower, i.e. 2–3 per 100 000 person-years.89 Thus, the SCD rate increases considerable with age and in the study by Ray et al.,4 the unadjusted rate for persons from 30 to 74 years of age (mean 45 years) was 179 per 100 000 person-years, and 10-fold higher in the age-group 70–74 years. The incidence was twice as high for men when compared with women. The rate of 179 SCD per 100 000 person-years represents the average for the group, but in a heterogeneous group of patients some may have had co-morbidities, e.g. ischaemic heart disease, heart failure, pre-existing long QT-interval, or electrolyte disturbances and these patients would have been treated at a much higher risk of SCD. Identification and optimized management of such high-risk patients could probably decrease the SCD rate significantly. Thus, in the otherwise healthy young a doubling of the risk of SCD may represent a very rare event since the absolute risk is low, whereas in older patients and in patients with pre-existing risk factors for SCD the absolute risk associated with treatment with psychotropic medications may reach a dramatic level.

In severely ill patients suffering from psychiatric diseases such as schizophrenia, depression, or bipolar disorders, a treatment-associated risk of SCD may be considered to be unavoidable.90 Furthermore, some of these conditions may, if left untreated themselves, increase the risk for heart diseases or death, but all possible precautions to reduce the cardiac risk should be taken, including administering the lowest effective dose, optimization of the treatment of co-morbidities, if possible discontinuation or reduction of dosages of negatively interacting drugs, reduction of the administration of un-documented combination regimens to a minimum—and establishing a prudent follow-up for detection of side effects in time.

The recent study by Khan et al.10 of adult patients (mean age of 37–45 years) with psychiatric diseases included in clinical placebo-controlled trials by pharmaceutical companies for US FDA showed that 3–4 months exposure to modern psychotropic agents did not worsen the mortality risk. The exposure period was 23.711 and 2.183 years for psychotropic agents and placebo, respectively. A total of 265 patients died, 109 (41.1%) from suicide. Suicide rates were generally lower in patients on active treatment—with patients on heterocyclic antidepressants as the only exception. However, it should be noted that in the actively treated groups, 44 patients (18%) died due to cardiovascular causes and additional 20 (8%) died suddenly. Comparable figures for the placebo groups were not given. Thus, the authors concluded that psychotropic agents did not increase all-cause mortality, but no conclusions were made regarding the risk of drug-induced SCD. Additionally, old patients, patients with cardiac co-morbidities, patients treated simultaneously with more than one psychotropic agent etc. are generally less likely to be included in such trials. This reduces the generalizability of these trials.

Generally, we are as physicians obliged to inform the patient about known side effects including a risk of SCD to the extent the patient is capable of understanding the message. If the patient is not capable of understanding the risk and especially in the few compulsory treated patients the necessary acute therapy should be prioritized and the cardiac assessment and follow-up postponed until the patient is able to cooperate and capable of understanding the information.

Lethal arrhythmia does not leave traces and cannot be diagnosed post mortem91 and often autopsy is not even performed to rule out more common causes of death such as myocardial infarction. As a result, the exact incidence of death caused by arrhythmia will be unattainable. The QT/QTc is currently generally accepted as a risk marker for arrhythmia. However, the risk related to QT prolongation is difficult to predict since a certain prolongation induced by one drug may not be associated with the same risk as when the QT prolongation was induced by another drug. For example amiodarone is known to prolong the QT interval, but amiodarone-induced TdP is relatively rare.92 This demonstrates that the QT interval is only an imperfect surrogate for measuring the risk of arrhythmia and clearly shows the limitations of an even well-conducted TQT study.

At present, several drugs on the market are capable of prolonging the QT interval far beyond the limits formulated by the ICH-E14. Some of these drugs might not have been approved and marketed today. The risks associated with these drugs should ideally be evaluated in phase IV studies. The responsibility to conduct such studies is difficult to place since patents may have expired, and heavy financial burdens on producers of older drugs might close the production, and reduce the spectrum of available pharmacological treatments. However, some post-marketing trials have been carried out. Thus, in 1998, sertindole was temporarily suspended from the European marked due to suspected treatment-associated unexplained deaths and as an authority demand for returning to the market in 2001, a randomized trail comparing sertindole and risperidone was performed and published in 2010.29 A total of 9858 patients were randomized and after 14 147 person-years no effect on all-cause mortality was identified between the two groups, but an independent committee blinded to treatment groups identified higher incidence of SCD in the sertindole group (HR:2.84 CI: 1.45–5.55), irrespective only sertindole-treated patients were needed to be ECG monitored and withdrawn if QTc exceeded 500 ms. Therefore, safe treatment with sertindole demands similar ECG monitoring as in the SCoP study.

Whereas the association between psychotropic medications and risk of malignant arrhythmia has been known for many years, the knowledge about the extent and prevention of such pro-arrhythmia seem to be limited. At present, only guidelines concerning safe use of anti-depressants and drugs for heroin addiction have been proposed.53,93 Finding the optimal psychotropic drug for the patient should always include evaluation of the perceived effectiveness and side-effects. Risk of suicide and psychotropic-induced life shortening conditions, such as obesity and diabetes, should be included in the evaluation.94

In the absence of precise knowledge about the individual drugs propensity of inducing arrhythmia and precise risk markers, the classification of psychotropic medications in this review must be considered more as a consensus than based on high-class evidence.

This review provides the first overview including a coordinated risk classification of all psychotropic drugs. Additionally, we provide the first clinically manageable guideline to increase the safety in the administration of these drugs.

Conflict of interest: J.N. has received research support from Pfizer and Hemocue and received honoraria from Astra-Zeneca, BMS, Hemocue and Lundbeck. P.V. reports no biomedical financial interests or potential conflicts of interest in relation to the topic of this article. S.P. received honorario from Lundbeck, Eli Lilly, Servier, and Janssen Pharmaceuticals. Other authors have no conflicts of interest to report. S.F., D.K., A.F., H.J., E.T. and H.B. have no conflicts of interest.

## Appendix A

### The development of the guideline

The spark initiating this work was experiences of serious cardiac events in patients treated with psychotropics—and the realization of a striking lack of clear clinical recommendations within the field.

View this table:
Table A1

Class of recommendation112

Classes of recommendationsDefinitionSuggested wording to use
Class IEvidence and/or general agreement that a given treatment or procedure is beneficial, useful, effectiveIs recommended/is indicated
Class IIConfliction evidence and/or a divergence of opinion about the usefulness/efficacy of the given treatment or procedureIs recommended/is indicated
Class IIaWeight of evidence/opinion is in favour of usefulness/efficacyShould be considered
Class IIbUsefulness/efficacy is less well established by evidence/opinionMay be considered
Class IIIEvidence or general agreement that the given treatment or procedure is not useful/effective, and in some cases may be harmfulIs not recommended
View this table:
Table A2

Level of evidence112

 Level of evidence A Data derived from multiple randomized clinical trials or meta-analyses Level of evidence B Data derived from a single randomized clinical trial or large non-randomized studies Level of evidence C Consensus of opinion of the experts and/or small studies, retrospective studies, registries
View this table:
Table A3

Recommendations for treatment with drugs with the propensity of prolonging the QT interval

 Class I If the QT-interval or QTc reaches a length >500 ms or increases by >60 ms compared with baseline, treatment with the particular drug should be ceased or dose reduced (Level of Evidence: C) Hypokalaemia should be avoided during treatment with drugs capable of prolonging the QT interval (level of evidence: C) Concomitant treatment with more than one drug with the propensity of prolonging the QT interval should be avoided if possible (level of evidence: C) Class IIa: Before initiation of treatment, assessment of cardiac risk is needed (level of evidence: C) The QT interval should be evaluated before initiation of treatment and during titration of dose (level of evidence: C) In elderly, treatment should be done with caution (level of evidence: C) If cardiac risks are identified, the cardiac risk factors should be optimized and/or a drug with a more favourable risk profile should be preferred if possible in the clinical situation (level of evidence: C) In case of structural heart disease, QT prolongation or cardiac symptoms referral to a cardiologist should be considered (level of evidence: C)

In order to establish the most useful recommendations, we decided at an early stage to establish a formal working group between the two national specialties—cardiology (Danish Cardiac Society, DCS) and psychiatry (Danish Society of Psychiatry, DPS). A short description of the purpose was agreed on formally by the boards of the two specialties and members from both specialties were elected to the working group. The working group had a total of four meetings. Before each meeting, each member was given a topic to describe in writing and all the proposals were emailed to the other members of the working group before the next meeting. At the meeting, the proposals were presented, discussed, and consensus reached. The specific topic was accordingly rephrased by the member covering this special topic—and then presented again on the next meeting. This process was continued until final consensus was reached. During the process, the individual proposals were included in the report document. Following the final acceptance of the report by the working group, the report was formally in hearing among all members of the two specialties (DCS – 1300, DPS 900 members) before the final and formal approval by the boards of the two specialties. Thus, the process for developing the recommendations was a combination of a modified Delphi process and consensus.

### Evaluation of the QT interval

The QT interval is defined as the time from the start of depolarization (the beginning of QRS) to the end of repolarization of the ventricles (the end of the T-wave). The end of the T-wave can be difficult to determine, not least when there is a partial superimposition of the T- and U-waves.95 Physiologically, the duration of the QT interval mainly depends on the concerted action of sodium (Na+), potassium (K+), and calcium (Ca2+) ion channels. Certain drugs may affect the QT-interval length and T-wave morphology (Figure A1). The end of the T-wave can be difficult to determine, not least when there is a partial superimposition of the T- and U-waves95 (Figure A2).

Figure A1

Two cases of drug-induced QT prolongation. Patient A: A 24-year-old woman with prolongation of the QT-interval (QT = 405 ms, QTcB = 510 and QTcF = 489) (correction according to Bazett's (B) or Fridericia's (F) formula) during treatment with sertindole at a dose of 16 mg per day. After sertindole treatment was ceased, the QT interval normalized (QT = 370 ms, QTcB = 455 and QTcF = 425). Patient B: A 53-year-old man with prolongation of the QT interval (QT = 500 ms, QTcB = 495 and QTcF = 497) during the treatment with methadone at a dose of 180 mg per day. The QT interval normalized after treatment with methadone was ceased (QT = 400 ms, QTcB = 396 and QTcF = 397). Standard limb lead II (left) and pre-cordial lead V2 (right) are shown.

Figure A2

Measurements of the QT-interval, examples of T-wave morphology, and Torsade de Pointes. (A) Measurements of the QT interval: the beginning of the QT-interval is usually well defined as the first deflection from the isoelectric line after the P-wave. The end of the T-wave is more difficult to define. One method is to identify the intercept between the steepest tangent at the descending part of the T-wave and the isoelectric line. The morphology of the T-wave can be entirely different even though regarded in the same lead. The three examples represent (B) the well-defined T-wave with normal QT interval. (C) The unambiguous prolonged QT interval. (D) A flattened and prolonged the T-wave which makes estimation of the QT interval very difficult. (E) Typical Torsade de Pointes proceeded by ventricular ectopies.

The QT interval prolongs with decreasing heart rate and shortens at higher heart rates and therefore measurements of the QT-interval length are often normalized or ‘corrected’ to a heart rate of 60 bpm denoted as the corrected QT or QTc. Several formulas for correction of QT intervals measured at higher or lower heart rates have been developed. Usually, the QT is corrected using either Bazett's formula96 or Fridericia's formula:97$$mathtex$$\eqalign{ & \hbox{Bazett:}\quad \hbox{QTcB} = \sqrt {\displaystyle{{\hbox{QT}} \over {\hbox{RR}}}} \cr & \hbox{Fridericia:}\quad \hbox{QTcF} = \root 3 \of {\displaystyle{{\hbox{QT}} \over {\hbox{RR}}}} }$$mathtex$$

Bazett's formula is simplest to use, but overcorrects at higher heart rates (>∼80 bpm) and undercorrects at lower heart rate.91 Especially at high heart rates, the use of Fridericia's formula is recommended.98 The normal upper QTc value in men is 450 ms, in women 460 ms,99,100 and in children <440 ms.101 However, it is of major importance to note that these limits represent the 95th percentile values.100 QTc values from 440 to 470 ms are considered ‘grey zone’ due to a considerable overlap between affected and controls in this range.100

A validation study comparing automated QT-interval measurements with approved measurements from previous QT-studies suggests that automated methods are feasible.102 In principle, the American Food and Drug Administration (FDA) accepts the use of fully automatic methods to measure the QT interval in TQT studies as long as the sensitivity is confirmed by including a positive control.103 It is essential to emphasize that this is a special situation with ECGs recorded in healthy subjects. In the clinical routine, it is only recommended to rely on automated QT interval measuring if the ECG is otherwise normal. Therefore, all physicians need to be able to measure the QT interval manually (Figure A2).

1. 1.
2. 2.
3. 3.
4. 4.
5. 5.
6. 6.
7. 7.
8. 8.
9. 9.
10. 10.
11. 11.
12. 12.
13. 13.
14. 14.
15. 15.
16. 16.
17. 17.
18. 18.
19. 19.
20. 20.
21. 21.
22. 22.
23. 23.
24. 24.
25. 25.
26. 26.
27. 27.
28. 28.
29. 29.
30. 30.
31. 31.
32. 32.
33. 33.
34. 34.
35. 35.
36. 36.
37. 37.
38. 38.
39. 39.
40. 40.
41. 41.
42. 42.
43. 43.
44. 44.
45. 45.
46. 46.
47. 47.
48. 48.
49. 49.
50. 50.
51. 51.
52. 52.
53. 53.
54. 54.
55. 55.
56. 56.
57. 57.
58. 58.
59. 59.
60. 60.
61. 61.
62. 62.
63. 63.
64. 64.
65. 65.
66. 66.
67. 67.
68. 68.
69. 69.
70. 70.
71. 71.
72. 72.
73. 73.
74. 74.
75. 75.
76. 76.
77. 77.
78. 78.
79. 79.
80. 80.
81. 81.
82. 82.
83. 83.
84. 84.
85. 85.
86. 86.
87. 87.
88. 88.
89. 89.
90. 90.
91. 91.
92. 92.
93. 93.
94. 94.
95. 95.
96. 96.
97. 97.
98. 98.
99. 99.
100. 100.
101. 101.
102. 102.
103. 103.
104. 104.
105. 105.
106. 106.
107. 107.
108. 108.
109. 109.
110. 110.
111. 111.
112. 112.
View Abstract