Elsevier

Heart, Lung and Circulation

Volume 29, Issue 10, October 2020, Pages 1427-1432
Heart, Lung and Circulation

Review
Sudden Cardiac Death in Schizophrenia: A Review

https://doi.org/10.1016/j.hlc.2020.07.003Get rights and content

In patients with schizophrenia, cardiovascular disease accounts for nearly 50% of deaths and decreased life expectancy, and the incidence of sudden cardiac death is about four times higher than in the background population. While the majority of sudden deaths are due to ischaemic heart disease and its recognised risk factors, about 10% of sudden deaths are unexplained and are thought to be due to cardiac arrhythmias. This review discusses various factors that might contribute to this increased mortality, such as the effect of antipsychotic drugs on potassium and sodium channel function, increased incidence of Brugada pattern in patients with schizophrenia and the role of the autonomic nervous system. It stresses the control of traditional coronary risk factors and discusses various noninvasive tests to identify patients at risk. It also mentions the reported association for nonsynonymous genetic polymorphism rs10503929 within the neuregulin 1 gene (NRG1) and the minor allele C and its role in the risk of sudden cardiac death in schizophrenia.

Introduction

Life expectancy in people with schizophrenia is estimated to be 10 to 25 years less than the general population and the incidence of sudden cardiac death (SCD) in patients with schizophrenia is about four times higher than in the general population. In a large Danish retrospective study of 5,178 deaths over a period of 6 years (2000–2006) in people aged 18–35 years, there were 395 SCD. The incidence of SCD was 14.8 death per 100,000 person-years in psychiatric individuals as opposed to 3.8 per 100,000 person-years in nonpsychiatric subjects. Psychiatric patients had an SCD incidence rate ratio of 4.3 and the incidence rates were highest in schizophrenia spectrum disorders [1]. While the majority of SCD in schizophrenia patients are due to ischaemic or structural heart disease, suicides etc., autopsy studies have reported that a significant proportion of SCD remain unexplained [2,3].

A large Australian study of autopsies over a 10-year period by Sweeting et al. found that 683 of 19,478 (3.5%) post-mortems were performed in subjects with a history of schizophrenia. In these cases, the mean age of death was 51 years (range 18–93 years), with 43% between the ages of 41 to 60 years, and 67% were males. The primary cause of death was cardiovascular in 23%, suicides in 20% and drug toxicity in 17%. In 11% of cases (n=72), no definite cause of death was found, and the death was presumed to be caused by cardiac arrhythmias [2].

Another study by Ifteni et al. used autopsy findings to determine the cause of death in schizophrenia patients admitted between 1989 and 2013. Medical records identified 57 patients who had died suddenly during hospitalisation. Fifty-one (51) out of 57 (89.5%) had autopsies (age range 55.9±9.4 years, males 56.9%). Most deaths were due to cardiovascular disorders (62.8%) of which 52.9% had myocardial infarctions, 11.8% were due to pneumonia, 7.8% due to airway obstruction, 5.9% were due to myocarditis and the remaining were due to miscellaneous conditions. Unexplained sudden deaths accounted for 6 out of 51 or 11.8% although three out of six patients had evidence of coronary atherosclerosis [3].

Sun et al. performed a retrospective analysis of 391 autopsy cases investigated in Maryland, USA. The mean age was 49.5 years with a male:female gender ratio of 1.5:1. Ethnic diversity was seen with 48.8% being White, 47.3% African American and the remainder were Hispanic or Asian. The majority (64.2%) of deaths were caused by natural disease, 12.0% were due to accidents, 11.5% were due to suicides and 9.7% were due to homicides. In 9.7%, the manner of death remained undetermined. Of the 251 natural deaths, 198 cases (78.9%) were due to cardiovascular diseases. Cause of death was attributed to ‘cardiac arrhythmias’ in 11 cases. The diagnosis of cardiac arrhythmia was made by exclusion after complete autopsy and toxicology. This study showed a high fatality rate caused by drug intoxication and cardiovascular diseases [4].

Most patients with schizophrenia are treated with antipsychotic drugs. Typical antipsychotic drugs (thioridazine, haloperidol, chlorpromazine etc.) developed in 1950s have largely been replaced by atypical antipsychotic drugs (clozapine, risperidone, quetiapine, olanzapine, ziprasidone etc.) introduced into practice in the 1990s. While these drugs are effective in the management of patients with schizophrenia and serious mental disorders, the question arises as to whether they contribute to the increased incidence of SCD in these patients.

Ray et al. calculated the incidence of SCD among current users of antipsychotic drugs in a retrospective cohort study of Medicaid enrollees in Tennessee, USA. They compared 44,218 users of single typical and 46,089 users of atypical antipsychotic drugs with 186,660 matched non-users who had a low incidence of schizophrenia but a high incidence of depression. To avoid the confounding factors associated with the indication of schizophrenia, they also included a propensity-matched control group who did not have schizophrenia or related psychosis. This group included 116,069 nonusers and 67,284 users of antipsychotic drugs mainly for depression or bipolar disorder.

Current users of typical and atypical antipsychotic drugs had a sudden death adjusted incidence rate ratio of 2.00 and 2.27 (1.89–2.73), respectively. Moreover, the risk increased with increasing dose. For typical antipsychotic drugs the risk was 1.31 for low dose and 2.42 for those taking high doses. Among users of atypical drugs, the ratio increased from 1.59 to 2.86 (p=0.01). There was no significant difference between typical and atypical agents [5].

In an accompanying editorial and analysis of Ray's data, Schneeweiss and Avron estimated the risk of SCD at 2.9 events per 1,000 patient-years and 3.3 events per 1,000 patient years for treatment with high doses. They conclude that this risk level is moderate to low but not rare [6].

One of the main mechanisms by which antipsychotics increase the risk of SCD is by prolongation of the QT interval. The clinical antipsychotic trials of intervention effectiveness study (CATIE) showed that 3% of patients with schizophrenia who were treated with atypical antipsychotics had prolongation of the QT interval [7].

Among the typical antipsychotics, thioridazine causes most marked prolongation of the QT interval and Torsades de Pointes (TdP). QT prolongation can also occur with other typical and atypical antipsychotic drugs. In clinical practice, the QT prolongation is modest (less than 30 msec) [8]. Rasch et al. summarised the public FDA event report system (FAERS) which characterises different antipsychotics by their torsadogenic profile depending on the number of reported cases of TdP, SCD and QT prolongation (Table 1) [9]. Table 2 lists, from Crediblemeds.org, the effects of antipsychotic drugs in patients with a combined list of drugs that prolong QT and/or cause TdP, dividing them into known risk (KR), conditional risk (CR) and possible risk (PR) [10].

Depression is quite common in patients with schizophrenia and selective serotonin reuptake inhibitors (SSRIs) are the most commonly prescribed antidepressant drugs. SSRIs may also reduce potassium current through their effect on the hERG channel and cause QT prolongation [11]. Prolongation of QT interval is reflective of reduced repolarisation reserve and places patients at risk of developing polymorphic ventricular tachycardia in the form of TdP. This can occur in patients with structurally normal hearts as well as in those who have ischaemic heart disease or cardiomyopathy.

Giradin et al. determined the prevalence of drug-induced long QT syndrome in hospitalised psychiatric patients over a 5-year period. They compared 62 patients with drug-induced long QT to 143 control patients with a normal electrocardiograph (ECG). Of the long QT patients, 85.5% had two or more comorbidities such as hypokalaemia, human immunodeficiency virus (HIV) or hepatitis C (HCV) infection, abnormal T wave morphology and treatment with methadone. The prevalence of drug-induced long QT interval (≥500 ms) in schizophrenia patients treated with antipsychotics was reported as between 0.9% to 2.6% [12].

It is important for patients starting on antipsychotic drugs to have an ECG recorded before and after starting these drugs and the dose reduced, or drug regimen altered, if the QT interval prolongs to 500 ms or 60 ms more than the baseline value. Acquired long QT syndrome related arrhythmias are more likely to occur when psychotropic medication is combined with other QT prolonging drugs, in the presence of hypokalaemia or hypomagnesaemia, in patients with cardiac failure, in females, among the elderly and those harbouring potassium channel polymorphism or mutation, and in those on high doses of antipsychotics or with intravenous administration [12,13].

The atypical antipsychotic drug clozapine is used in refractory schizophrenia patients and has the potential for serious side effects such as agranulocytosis, myocarditis, seizures, sedation, obesity, diabetes mellitus and other metabolic abnormalities. The incidence of myocarditis associated with clozapine in an Australian study was 0.7–1.2% [14]. Although a larger study from a Canadian study has reported an incidence of 3.16% [15]. Clozapine has been classed as having significant torsadogenic potential and confers the highest risk of SCD, with an odds ratio of 3.67 compared to conventional agents [9]. Agranulocytosis, pulmonary embolus and pneumonia caused by clozapine can also cause sudden death.

Blom et al. compared ECG findings of 275 patients with schizophrenia with two sets of matched controls. A Brugada pattern was more prevalent in patients with schizophrenia (11.6%) compared to 1.1% in 179 controls from the Netherland study of Depression and anxiety (NESDA). They also used a second control group (HOORN) consisting of subjects who were 20 years older as patients with schizophrenia are biologically older. The incidence of Brugada pattern in this cohort was 2.4%.

It should be noted however, that only one subject had spontaneous Type 1 Brugada pattern and 31 had nonspecific type 2 or 3 Brugada pattern. Twenty-three (23) of these 31 had an ajmaline test and 10 had a positive test. Thus 11 patients (4%) had either spontaneous or drug induced Brugada pattern which was not related to the use of sodium channel blocking drugs. Only one patient had a rare variant in the SCN5A gene [16]. The incidence of Brugada pattern in the general population is said to be 0.05%. The incidence of Brugada pattern in schizophrenia is about 4% [17].

The extent to which this increased incidence of Brugada pattern contributes to increased mortality is uncertain, but according to Roden it may play a role [18]. In addition to a Brugada pattern on the ECG, polymorphism or subclinical Na channel mutation may also cause drug induced LQT syndrome and TdP [19,20].

Hattori et al. evaluated parasympathetic activity in 53 Japanese patients with schizophrenia between the ages of 60 and 70 years. All patients had heart rate variability at entry in 2007. During the follow-up period 11 patients died. The parasympathetic activity of non survivors was significantly lower than the survivors as assessed by multiple logistic regression analysis [21]. It has been suggested that this parasympathetic dysfunction leads to sympathetic dominance and that this effect may be dose dependent and may also affect the severity of schizophrenia. Reduced heart rate variability has also been considered a risk factor in causing SCD [22].

Figure 1 illustrates various causes of SCD in schizophrenia. Up to 60% or more of sudden deaths in schizophrenia are from cardiovascular causes. Ischaemic heart disease accounts for the majority of these deaths and is largely related to recognised coronary risk factors like smoking, obesity, diabetes, dyslipidaemia and sedentary lifestyle.

Daumit et al. recently published an 18-month randomised clinical trial incorporating behavioural counselling and care management to reduce cardiovascular risk in adults with serious mental illness including schizophrenia. There were 132 participants and 137 controls with at least one cardiovascular risk factor (hypertension, diabetes, dyslipidaemia, current tobacco smoking and obesity). The mean age of the participants was 48.8±11.9 years and 59.1% had schizophrenia or schizo-affective disorder. A health coach and a nurse provided individually tailored behavioural counselling, collaborating with physicians to implement appropriate risk management. Compared with the control group, the intervention group experienced a 12.7% (p=0.02) relative risk reduction in the Framingham Risk Score which predicts the 10-year probability of a cardiovascular event [23]. Hence, the recognition and correction of coronary risk factors would likely reduce total mortality as well as the incidence of SCD.

Several investigations may also help to identify patients at increased risk of SCD [24].

As mentioned earlier, a 12-lead ECG should be recorded routinely in patients with schizophrenia before starting psychotropic medication. While not very sensitive in detection of subclinical ischaemic heart disease, it may show nonspecific ST-T changes, left ventricular (LV) hypertrophy, Q waves suggesting previous myocardial infarction or bundle branch block. The presence of QRS fragmentation and notching within the QRS morphology may indicate myocardial scar.

An important measurement in the ECG is the corrected QT interval (QTc). Rarely, it detects previously undiagnosed congenital LQTS but its main value is in monitoring the effect of antipsychotic drugs on the QT interval. In general, prolongation of the QT interval to ≥500 ms or an increase in QT interval by ≥60 ms are accepted upper limits to reduce the dose or change the antipsychotic drug regimen.

Another potential marker of risk on the ECG is T peak-end (interval between the peak of the T wave and the end of the T wave) which correlates with transmural dispersion of ventricular repolarisation. Prolongation of this interval suggests an increased risk of TdP [25].

Early repolarisation (ER) is defined as ≥0.1 mV J point elevation in two contiguous inferior and/or lateral ECG leads. ER was previously thought to be benign, until Haissaguerre et al. in 2008 reported an ER pattern in 31% of survivors of idiopathic ventricular fibrillation [26]. ER is quite a common finding in otherwise healthy subjects and is particularly common in young subjects and in athletes. In a large study of 10,000 middle-aged Finnish subjects followed up for 30±11 years, an ER pattern was found in 5.8%. ER patterns in inferior leads, horizontal or descending J point segment and ER of ±2 mV were markers of increased risk (HR 1.43) [27]. Because of its high prevalence in the general population, ER is a poor indicator of risk for SCD.

Many modern ECG machines incorporate the ability to do SAECG which is a diagnostic test that digitises multiple high-resolution QRS complexes (at least 250) over a 10-minute period using three orthogonal leads to detect low amplitude signals called late potentials (LP) in the terminal part of the QRS in sinus rhythm. The presence of LP generally represents delayed ventricular activation reflecting slow conduction into myocardial scar and may indicate the presence of a substrate for re-entrant ventricular tachycardia. SAECG was originally introduced in risk assessment of survivors of myocardial infarction. It had low predictive value and did not gain wide acceptance but was included as a minor criterion for the diagnosis of ARVC [28,29].

An increased incidence of LP on SAECG has been reported in patients with serious mental disorders including schizophrenia. Antoniou et al. studied 52 patients and compared them with 30 healthy subjects and reported that late potential prevalence was higher than in controls (31% vs 7%, p=0.012) [30].

Fitzgerald et al. recently reported a cross sectional study on the point prevalence of ECG and SAECG abnormalities in 80 patients with serious mental illness. Half (49%) of these patients were male with an age range of 39±17 years. Including all who showed one, two or three LP criteria, they found SAECG abnormality in 34%, while 19% showed an early repolarisation pattern. The authors acknowledge that they did not have a control group and did not have echocardiograms to exclude structural cardiac disease. There was no clinical follow-up in either of these studies and hence the significance of these findings remains speculative [30,31].

There has been only one publication reporting genetic variants in schizophrenia and epilepsy linked to SCD, by Huertas-Vazquez et al. This paper was a part of the ongoing Oregon sudden unexplained death population study. A total of 682 SCD subjects were included (340 cases and 342 controls from The Harvard Cohort SCD study). The age range was 63.4±12.5 years. There were only three subjects with schizophrenia and seven with epilepsy amongst 340 SCD subjects in the Oregon group and two subjects with schizophrenia and 11 with epilepsy in the Harvard group. The authors found strong evidence for the association of the nonsynonymous single nucleotide polymorphism rs 10503929 within the neuregulin 1 gene (NRG1) and the minor allele C was associated with an increased risk of SCD (OR 1.94, 95% CI 1.5–2.5) [32].

NRG1 is a signalling protein that mediates cell-to-cell interactions and is involved in important biological process of schizophrenia, epilepsy, heart development and function. It was previously identified as a candidate gene for schizophrenia in a linkage analysis of Icelandic families and is considered one of the leading candidate genes in schizophrenia.

The logic behind attributing NRG1 as a cause for SCD is that if schizophrenia increases the risk of SCD then NRG1 is a risk factor for SCD. If it is assumed that rs10503929 has an effect on NRG1 at the cardiac level, it is possible that carriers of the rs10503929 minor allele have relative parasympathetic insufficiency and therefore lower protection from catecholamine-dependent arrhythmogenesis. Napolitano recommends confirmation of these findings by independent studies, careful phenotype analysis of schizophrenia SCD subjects and further confirmation of the role of the NRG1 gene role by functional study [33].

Section snippets

Conclusion

Patients with schizophrenia have a reduced life expectancy and have a higher incidence of SCD compared to the general population. The high incidence of SCD is mainly due to ischaemic heart disease. A small percentage, about 8 to 10%, are unexplained and are presumed to be due to cardiac arrhythmias.

Various other factors may contribute to the increased incidence of SCD. Antipsychotic drugs have effects on potassium and sodium ion channels and may cause TdP by QT prolongation, also drug toxicity,

Declarations of Interest

None.

Funding Sources

I have not received any grant support for this work.

Competing Interest Statement

There is no competitive interest.

Conflicts of Interest

There is no conflict of Interest.

References (33)

  • W.A. Ray et al.

    Atypical antipsychotic drugs and the risk of sudden cardiac death

    N Engl J Med

    (2009)
  • S. Schneeweiss et al.

    Antipsychotic agents and sudden cardiac death-How should we manage the risk?

    N Engl J Med

    (2009)
  • J.A. Lieberman et al.

    Effectiveness of antipsychotic drugs in patients with chronic schizophrenia

    N Engl J Med

    (2005)
  • J.C.R. Diaz et al.

    The relationship between antipsychotic drugs, QT prolongation and torsades de pointes: implications for clinical use

    Expert Opin Drug Saf

    (2020)
  • E. Raschi et al.

    Torsadogenic risk of antipsychotics: combining adverse event reports with drug utilization data across Europe

    PLoS One

    (2013)
  • R.L. Woosley et al.
  • Cited by (21)

    • Use of potassium ion channel and spliceosome proteins as diagnostic biomarkers for sudden unexplained death in schizophrenia

      2022, Forensic Science International
      Citation Excerpt :

      Generally, schizophrenia patients have a life expectancy of approximately 25 years shorter than the normal lifespan [12]. Risk factors for SUD in this cohort are more complex compared with the general population, since antipsychotic medications, unhealthy lifestyles, central nervous system pathology can independently contribute to cardiovascular or respiratory diseases [13]. Although most of sudden deaths in schizophrenia are due to cardiovascular causes, such as ischemic heart disease, myocarditis and cardiomyopathy, about 11% of sudden deaths are still unexplained after comprehensive postmortem examination and are thought to be due to cardiac arrhythmia [14–16].

    • Asphyctic Deaths

      2022, Encyclopedia of Forensic Sciences: Volume 1-4, Third Edition
    View all citing articles on Scopus
    View full text