OUP user menu

Psychotropic medications and the risk of sudden cardiac death during an acute coronary event

Jussi Honkola, Eeva Hookana, Sanna Malinen, Kari S. Kaikkonen, M. Juhani Junttila, Matti Isohanni, Marja-Leena Kortelainen, Heikki V. Huikuri
DOI: http://dx.doi.org/10.1093/eurheartj/ehr368 745-751 First published online: 14 September 2011


Aims Psychotropic medication increases cardiac mortality, but the reasons for this association are not clear. We studied the role of psychotropic drugs as a triggering factor of sudden cardiac death (SCD) during an acute coronary event.

Methods and results The use of medication was compared between victims of SCD and survivors of an acute coronary event in a case–control study including a consecutive series of victims of SCD (n= 1814, mean age 65 ± 11 years) verified to be due to an acute coronary event at medico-legal autopsy and consecutive series of patients surviving an acute myocardial infarction (AMI; n= 1171, mean age 66 ± 12 years). The medication history was obtained from autopsy/hospital records and interviews with relatives of SCD victims and AMI patients. The use of antipsychotics [9.7 vs. 2.4%, odds ratio (OR) 4.4, 95% confidence interval (CI) 2.9–6.6; P< 0.001] and antidepressants (8.6 vs. 5.5%, OR: 1.6, 95% CI: 1.2–2.2; P= 0.003) was more common in the SCD than AMI group, but the use of benzodiazepines did not differ between the groups (11.7 vs. 13.2%; P= 0.270). The use of antipsychotics remained as a significant risk factor for SCD after adjustment for confounding variables (OR: 3.4, 95% CI: 1.8–6.5; P< 0.001). Combined use of phenothiazines and any antidepressant was associated with a very high risk of SCD (OR: 18.3, 95% CI: 2.5–135.3; P< 0.001).

Conclusion The use of psychotropic drugs, especially combined use of antipsychotic and antidepressant drugs, is strongly associated with an increased risk of SCD at the time of an acute coronary event.

  • Myocardial infarction
  • Psychiatric disorder
  • Sudden death

See page 687 for the editorial comment on this article (doi:10.1093/eurheartj/ehr405)


Sudden cardiac death (SCD) accounts for one out of every two deaths from cardiovascular diseases.1 An acute coronary event has been considered to be the most common pathophysiological mechanism leading to SCD.1 Prevention of SCD caused by an acute coronary event has remained a challenge for clinicians, because recognizable risk factors for SCD appear to be partly the same as those predisposing to a non-fatal coronary event. In fact, there is relatively little information available on possible differences in risk profiles between the subjects who survive and those who die suddenly during an acute coronary event.2

Mental disorders have been associated with an increased risk of cardiovascular mortality.37 Recently, depression has been shown to predict specifically the occurrence of SCD.811 In addition, there is growing evidence that also the psychotropic drugs used to treat psychiatric disorders could increase the risk of SCD.1114 Despite the epidemiological evidence of an association between depression, anxiety, and increased risk of cardiovascular mortality, the exact pathways, and pathophysiological mechanisms of these associations are not well established. It is not known, if either the mental disorders or the psychotropic drugs used to treat them predispose to the occurrence of an acute coronary event or if they increase the vulnerability to a fatal outcome during such an event.

A prospective case–control study, the Finnish Genetic Study of Arrhythmic Events (FinGesture)15 was designed to compare genetic and other risk profiles of the victims of SCD with the survivors of an acute coronary event. In this study, we have compared the medications being used by these groups. Our primary focus was to study possible differences in the use of psychotropic drugs between the subjects who died suddenly and those who survived an acute ischaemic event.


Study populations

The FinGesture study was started in 1998 and by the end of 2009 it had gathered data on 2732 consecutive victims of out-of-hospital sudden death from a defined geographical area, i.e. Oulu University Hospital District in northern Finland. All the victims of sudden death were autopsied in the Department of Forensic Medicine, University of Oulu, Finland. Sudden cardiac death caused by an acute coronary event has been previously defined in detail.2,15,16 In each case of sudden death, the mechanism of death was clarified and all patients who were considered to have died due to any cause other than SCD during an acute coronary event were excluded from the study. With respect to out-of-hospital SCD victims, those with (i) a witnessed sudden death within 6 h of the onset of symptoms or within 24 h of the time that the victim was last seen alive in a normal state of health and (ii) evidence of a coronary complication, defined as a fresh intracoronary thrombus, plaque rupture or erosion, intraplaque haemorrhage, or critical coronary stenosis (>75%) in the main coronary artery were included in the SCD group. Cases suggesting severe heart failure, e.g. those with massive myocardial necrosis (>50%) and signs of acute pulmonary oedema, and history of hospitalization due to congestive heart failure were not defined as ischaemic SCDs. Toxicologic investigations were made in cases of any suspicion of intoxication (60%), and cases with suicide were excluded. After exclusions, 2082 of 2732 victims of sudden death were defined as having died due to an acute coronary event.

Information about the victims' latest medications, including doses of psychotropic medications, was collected from the medico legal autopsy reports and from standardized questionnaires asked from the victims' closest relatives. If the information about medication was not described reliably in the autopsy report or questionnaires, the case was excluded (n= 268). After these exclusions, a total of 1814 SCD cases were included in the study. Information about the prior medical history was obtained from a combination of the available electronic medical records, existing from 1995 in the area of present patient population, and from standardized questionnaires asked from the victims' closest relatives as described earlier.16

The control group consisted of survivors of AMI, collected between 1996 and 2004 (n= 1256) at the Department of Internal Medicine, University Hospital of Oulu.17,18 This sample included all AMI patients from those periods from the same geographical area from which the SCD victims were gathered. Survivors of AMI were recruited to participate during the first 7 days after the diagnosis of AMI, which was confirmed by using the contemporary guidelines.17,18 Information about the patients' medications and prior cardiovascular history was obtained from the medical records during their hospital stay. If no reliable information about medication was available, the patient was excluded from the study (n= 55). Additional 30 patients were excluded, who experienced AMI at the post-operative phase after surgery. The qualifying diagnostic criteria of the patients have been described in detail elsewhere.17,18 Forty-two per cent of the patients had ST-segment elevation AMI and 58% had non-ST elevation AMI. Acute myocardial infarction was anterior in 45% of patients, inferior in 41%, and some other location in 14% of the patients. A total number of 1171 AMI patients were included in the study.

Antipsychotic drugs were grouped as typical, first generation, antipsychotics, including phenothiazines, thioxanthenes, butyrophenones, and benzamides, and atypical, second generation antipsychotics, including aripiprazole, clozapine, risperidone, quetiapine, olanzapine, and drugs similar to these agents. Antidepressants were categorized into three groups; tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs), and other antidepressants, including venlaflaxine, mianserin, mirtazapine, and drugs similar to these agents. The doses of used antipsychotic and antidepressant drugs were categorized as (i) low dose, (ii) normal or intermediate dose, and (iii) high dose according to previously reported recommendations.1921 Circadian rhythm of SCD was also compared between those using and not using psychotropic drugs.

The study conforms to the principles outlined in the Declaration of Helsinki and the study protocol was approved by the Ethics Committee of The Northern Ostrobothnia Hospital District. All living subjects had an opportunity to refuse to participate and gave written informed consent.

Statistical analysis

Differences in continuous clinical variables between study groups were tested with a two-sided t-test. Normal distribution of data was confirmed by using the Kolmogorov–Smirnov test. Differences in categorical characteristics between study groups were tested using a χ2 test. Logistic regression analysis was used to assess the risk of SCD related to different psychotropic medications after adjustment for demographic risk factors, such as male gender, a history of angina pectoris, hypertension, diabetes, a history of prior AMI, hypercholesterolaemia, and the use of cardiovascular medications, i.e. lipid-lowering therapy, statins, nitroglycerin, acetosalicylic acid, β-blockers, Ca-channel blockers, diuretics, ACE-inhibitors or AT2-antagonists and flecainide, amiodarone or sotalol. All analyses were performed with the Statistical Package for Social Studies version 16.0 (SPSS, Inc., Chicago, IL, USA). P< 0.050 was considered significant.


Patient characteristics

Demographic data and the common cardiac risk factors are presented in Table 1. The victims of SCD were more often males and current smokers, while a history of angina pectoris, hypertension, diabetes (type I or II), and hypercholesterolaemia were more common among survivors of AMI. A higher percentage of AMI survivors had a history of a prior myocardial infarction.

View this table:
Table 1

Characteristics of the study populations

CharacteristicSCD (n= 1814)AMI (n= 1171)P-value
Age, years65.2 ± 1165.5 ± 120.393
Sex, males1449/1814 (79.9)832/1171 (71.1)< 0.001
BMI26.8 ± 5.227.6 ± 4.4< 0.001
Smoking status0.001
Never102/292 (34.9)315/945 (33.3)
Ex66/292 (22.6)316/945 (33.4)
Current124/292 (42.5)314/945 (33.2)
Hypercholesterolaemia162/1681 (9.6)721/1120 (64.4)<0.001
Diabetes type I or II333/1720 (19.4)299/1161 (25.8)<0.001
Hypertension634/1665 (38.1)631/1162 (54.3)<0.001
Angina pectoris165/1670 (9.9)590/1129 (52.3)<0.001
Prior AMI184/1695 (10.9)264/1153 (22.9)<0.001
  • BMI, body mass index. Values are expressed as mean (SD) or number of subjects (%). Probability values refer to two-sided t-test, and χ2 test comparing victims of SCD to AMI patients.


The medications being used by the study groups are presented in Table 2. The use of cardiovascular or antihypertensive medication was more common among AMI survivors. The use of any psychotropic medication, including benzodiazepines, antidepressants, and antipsychotics, was significantly more common in the SCD than in the AMI group. In particular, the use of antipsychotics (P< 0.001) and antidepressants (P= 0.003) was significantly more common in the victims of SCD. In the group of antipsychotics, the use of phenothiazines, thioxanthenes, butyrophenones, and newer atypical antipsychotics was more common in the victims of SCD compared with AMI survivors. In the subgroups of antidepressants, there was no significant difference between the groups in the use of SSRIs or newer types of antidepressants, but the use of TCAs was more common in the SCD group. There was no difference in the use of benzodiazepines between the victims of SCD and the survivors of AMI (P= 0.270). The combined use of antipsychotics and antidepressants was associated with an increased risk of SCD (P< 0.001). A very high risk of SCD was seen in those subjects who had been prescribed both phenothiazine and antidepressant medications (OR: 18.3, 95% CI: 2.47–135.3; P< 0.001) (Figure 1).

View this table:
Table 2

Medications of the study populations

MedicationSCD (n= 1814)AMI (n= 1171)P-value
Cardiovascular medication
 Any cardiovascular or antihypertensive medication1039/1648 (63.0)814/1165 (69.9)<0.001
 Lipid-lowering therapy, other than statins207/1499 (13.8)74/1125 (6.6)<0.001
 Statin189/1483 (12.7)289/1124 (25.7)<0.001
 Nitrate/nitroglycerin355/1519 (23.4)394/1138 (34.6)<0.001
 ASA323/1491 (21.7)460/1144 (40.2)<0.001
 β-Blocker444/1493 (29.7)552/1148 (48.1)<0.001
 Ca-channel blocker150/1490 (10.1)216/1129 (19.1)<0.001
 Diuretic275/1491 (18.4)326/1138 (28.6)<0.001
 ACE inhibitor or AT2-antagonist292/1489 (19.6)330/1132 (29.2)<0.001
 Warfarin123/1504 (8.2)97/1126 (8.6)0.689
 Flecainide/amiodarone/sotalol15/1488 (1.0)24/1120 (2.1)0.018
 Digoxin81/1489 (5.4)78/1129 (6.9)0.119
 Insulin79/1495 (5.3)105/1127 (9.3)<0.001
 Metformin/sulphonylurea140/1492 (9.4)157/1135 (13.8)<0.001
 Asthma or chronic obstructive pulmonary disease medication163/1497 (10.9)148/1129 (13.1)0.081
Psychotropic medication
 Any psychotropic medication322/1506 (21.4)199/1131 (17.6)0.016
 Antipsychotic145/1492 (9.7)27/1120 (2.4)<0.001
  Phenothiazine82/1490 (5.5)8/1118 (0.7)<0.001
  Thioxanthene16/1490 (1.1)–/11170.001
  Butyrophenone26/1490 (1.7)6/1118 (0.5)0.006
  Benzamide1/1490 (0.1)1/1117 (0.1)
  Atypical40/1490 (2.7)15/1117 (1.3)0.018
 Antidepressant128/1492 (8.6)62/1120 (5.5)0.003
  Tricyclic antidepressant43/1490 (2.9)19/1118 (1.7)0.049
  SSRI53/1490 (3.6)26/1117 (2.3)0.070
  Other37/1490 (2.5)19/1116 (1.7)0.174
 Benzodiazepine175/1494 (11.7)148/1125 (13.2)0.267
  • Values are expressed as number of subjects (%). Probability values refer to χ2 test comparing victims of SCD to AMI patients.

Figure 1

Odds ratios of experiencing sudden cardiac death at the time of an acute coronary event in relation to use of psychotropic medications. The 95% confidence intervals are in parentheses and absolute numbers of patients are presented below the medication classes.

Risk analysis

The results of both univariate and multivariate analysis for potential SCD risk factors are summarized in Table 3. In the multivariate model, the use of antipsychotics remained as a significant risk factor for SCD after adjustment for demographic parameters (P< 0.001). In addition, the use of TCAs (P= 0.011), phenothiazines (P< 0.001), and butyrophenones (P= 0.037) remained as significant predictors of SCD. The combined use of phenothiazines and antidepressants (adjusted OR: 9.62, 95% CI: 1.15–80.8; P= 0.037) was strongly associated with an increased risk of SCD in the multivariate model.

View this table:
Table 3

Univariate and multivariate analysis of the risk of sudden cardiac death

OR95% CIP-valueAdjusted OR95% CIP-value
Atypical antipsychotics2.031.113.690.0181.300.543.110.555
Tricyclic antidepressants1.721.02.970.0493.411.338.770.011
Newer antidepressants1.470.842.570.1741.110.492.520.795
  • In multivariate modelling, medications are adjusted for male gender, a history of angina pectoris, hypertension, diabetes (type I or II), a prior AMI, hypercholesterolaemia and lipid-lowering therapy, statins, nitroglycerin, acetosalicylic acid (ASA), β-blockers, Ca-channel blockers, diuretics, ACE-inhibitors or AT2-antagonists, and flecainide, amiodarone or sotalol.

Doses of psychotropic drugs, concomitant use of alcohol, and circadian rhythms

Among all users of psychotropic drugs, high dose of antipsychotic drugs had been used more commonly among the victims of SCD than the AMI survivors (20 vs. 0%, P= 0.03). The doses of antidepressant drugs did not differ between the groups. Blood alcohol concentration was measured in cases with a suspicion of alcohol use in SCD victims (42%). Among the victims of SCD, alcohol was observed in blood (>0.1‰) more commonly among those using psychotropic drugs (69%) compared with those not using these drugs (57%) (P= 0.002). The exact time of SCD could be defined in 996 of cases. In these cases, the circadian rhythm of SCD was different among those using psychotropic drugs with a higher incidence during the night-time among the drug users (P= 0.040) (Figure 2).

Figure 2

Circadian distribution of sudden cardiac deaths among the users and non-users of psychotropic drugs. Significantly higher incidence of sudden cardiac death was observed during the night hours among the users of psychotropic drugs (P= 0.040).


The use of psychotropic drugs was more common among the victims of SCD compared with survivors of AMI. Victims of SCD were more commonly users of both antipsychotics and antidepressants. The difference was most prominent in the use of antipsychotics, especially the phenothiazines. In the subgroups of antidepressants, the use of TCAs was more common in the SCD group, whereas there were only borderline differences in the use of SSRIs or newer types of antidepressants. The combined use of both antipsychotic and antidepressant drugs was associated with a very high risk of SCD during a coronary event. The use of cardiovascular drugs, such as aspirin, beta-blocking medication, and angiotensin-converting enzyme inhibitors, was less common among the victims of SCD. However, the difference in the use of psychotropic medication, which was the main focus of the present study, remained significant after adjusting for the use of other medications and other confounding variables.

Previous studies have shown that mental disorders increase the risk of cardiovascular mortality,37 but it has not been clearly established if psychiatric disorders, such as depression or schizophrenia, predispose to the occurrence of cardiovascular events or whether they modify the outcome during such events. The present results support the latter proposal, since it was found that the subjects using psychotropic drugs had a higher risk of dying during an acute coronary event, with fatal arrhythmia being the most probable cause of SCD during the ischaemic event. This concept is also supported by studies showing that depression increases specifically the occurrence of SCD but not non-fatal cardiac events.9,11

Previously, two large cohort studies have documented an increased risk of SCD among users of both typical and atypical antipsychotic drugs.13,14 In addition, a dose-dependent risk of SCD has been earlier found for the TCAs.22 In one study,11 the risk of SCD was most prominent in women who used antidepressants, whereas the rate of depression did not confer an elevated risk in the fully adjusted models. In the present study, the victims of SCD used more commonly both TCAs and antipsychotics but the excess use of SSRIs and newer antidepressants was not significant, pointing to a direct drug influence underpinning the mechanism of SCD. The risk of SCD was most strongly associated with the use of phenothiazines. This group includes drugs such as thioridazine, which is the most frequently reported cardiotoxic antipsychotic drug.12 The risk of SCD tended also to be higher in those who used newer types of antipsychotic drugs, although the difference did not remain significant after adjustments for confounding variables, partly supporting the previous data showing that the newer atypical antipsychotic drugs also possess a proarrhythmic potential.14

The causal relationship between the use of psychotropic drugs and SCD during an ischaemic event was further supported by an observation of higher doses of antipsychotic drugs among the victims of SCD as well as a special circadian rhythm of SCD among those using psychotropic drugs with a higher incidence during the night hours. Higher evening doses of psychotropic drugs are commonly used because of their sedative effects. Many patients with psychiatric problems also experience disturbances in their sleep-awake cycle and insomnia, which are reasons to administrate higher doses of psychotropic drugs in the evening. Concomitant alcohol use with psychotropic drugs was also relatively common among the victims of SCD.

A very high risk of SCD was observed in those individuals who were using both antipsychotic and antidepressant medication. The risk was particularly high with phenothiazines combined with antidepressant medication, suggesting that the combined use of these drugs may potentiate their proarrhythmic effects at the time of an acute ischaemic event. The use of benzodiazepines, a class of drugs which do not have any recognized proarrhythmic potential,12 was similar between the groups, suggesting that psychiatric disorders, at least anxiety, does not increase the vulnerability to fatal events, whereas the drugs used for mental disorders have a more marked effect.

The risk of SCD caused by antiarrhythmic drugs has usually been linked to their proarrhythmic effects. These effects have been especially linked to TCAs and antipsychotics,12 but the cascade leading to the fatal arrhythmia induced by these drugs is still not totally clear. Some antipsychotic drugs as well as TCAs have been shown to cause prolongation of the QT-interval in the electrocardiogram, which can lead to malignant polymorphic ventricular arrhythmia, torsades de pointes, and ultimately to sudden death. At the cellular level TCAs, such as imipramine and amitrioptyline, and antipsychotics, such as thioridazine, are associated with inhibition of potassium channels encoded by HERG, which correlates with prolonged QT intervals and a risk of torsades de pointes.23 Differences in risk between classes of psychotropic drugs may result from additional cellular effects of particular drugs, such as from effects on sodium and calcium channels especially at the time of acute myocardial ischaemia and concomitant sympathetic activation. Currently, there is not much data on the electrophysiological effects of psychotropic drugs from ischaemic cellular models, which should obviously be a focus of future research.

Limitation of the study is that we did not have specific diagnoses of mental disorders, and therefore it is possible that depression or schizophrenia may well be important denominators and pathway variables of the observed association between psychotropic drugs and SCD. The case–control design of the study is another potential limitation, because the baseline variables differed significantly between the groups. Furthermore, due to the study design, causality cannot be directly attributed to psychotropic drugs. The time period of collecting the study populations differed slightly, since the timeframe for the control group ended ∼5 years prior to the collection of cases. Nonetheless, the sample size consisted of a consecutive series of AMI survivors from two studies and consecutive victims of SCD from the same geographical area as the AMI survivors, providing information on the characteristics of those subjects dying and those surviving an acute coronary event. The strength of the study is the careful diagnosis of the mode of SCD. The definition of SCD without the autopsy data in previous epidemiological studies may mean that several mechanisms have been involved in the SCD. For example, suicide is a common reason of death in psychotic and depressive patients,24 and it may be difficult to differentiate suicide from SCD in epidemiological studies that do not have access to autopsy evidence with toxicological examinations, which are very comprehensive and reliable in Finland.25

Psychoses and other severe mental disorders are common, and their public health and human importance is enormous. The results of this study provide some insight into the extent and potential mechanisms behind the association between cardiac mortality and various psychiatric disorders and have also some practical implications. Almost 1 out of 10 victims of SCD (9.7%) was using antipsychotic drugs in the present population, whereas the prevalence was only 2.4% in survivors of AMI. The life-time prevalence of psychoses has been reported to be ∼3% in Finland.26 High prevalence of the use of antipsychotics in this study supports the view that the target population for using antipsychotics is wider than just psychoses, perhaps behavioural symptoms, anxiety, depression, etc. Therefore, liberal off-label use of psychotropic drugs, also TCAs, e.g. in the treatment of pain and sleep disorders, should be restricted whenever possible. Patients using both antipsychotic and antidepressants seem to be at a very high risk of dying suddenly during a coronary event, suggesting that if possible this combination should be avoided, especially in those patients who carry other cardiovascular risk factors. In addition, attempts should be focused on preventing coronary events in the subjects suffering from psychiatric disorders and who require treatment with antidepressant and/or antipsychotic medications.


This study was supported by Sigrid Juselius Foundation (E.H. and H.V.H.), Helsinki, Finland, and the Foundation Leducq (M.J.J.), Paris, France.

Conflict of interest: none declared.


All authors have contributed to the design of the study, interpretation of results, revising the manuscript, and approve the final version of the manuscript.


View Abstract