Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

Antipsychotics and sudden cardiac death


Sudden cardiac death has been reported with use of antipsychotics since the 1960s, though there have been few studies, and those we have were small and took little account of potential confounding factors. The evidence was therefore weak. A large case-control study from a high-quality database with full patient information [1] provides more information.


A large Dutch electronic database of about half a million patients of 150 general practitioners provided the information. The records contain details on demographics, symptoms, diagnoses, and outpatient and hospital records, including tests, and drug prescriptions. All patients aged 18 years or older, except those with cancer, or those with death by suicide, formed the source population.

The study sought deaths occurring over six years (1995-2001), and experts blinded to patient exposure to antipsychotic drugs reviewed those that could possibly be regarded as sudden cardiac death. Sudden cardiac death was defined as death occurring within one hour after onset of acute symptoms, and if the death was recorded as sudden or acute cardiac death, or similar. If death was not witnessed, unexpected death of anyone seen in a stable medical condition less than 24 hours previously and with no evidence of a non-cardiac cause was used. For each case of sudden death up to 10 controls matched for age, sex, and practice were chosen.

Exposure was defined as current use, past use (longer than 30 days since end of last prescription), or non-use of antipsychotics. Type of antipsychotic, dose (DDD defined as recommended daily dose for an adult for schizophrenia), and duration of use were also noted. Known risk factors for sudden cardiac death were also collected from databases, and used in adjusting results.


There were 582 cases of sudden cardiac death in a population of 250,000 adults over the period, an incidence of 1 per 1,000 per year. Controls (4,463) were available for 554 cases, of which 334 deaths were witnessed. The mean age was about 72 years and 60% were men. Known conditions and behaviours (heart failure, diabetes, smoking), and cardiovascular drugs were all associated with increased sudden cardiac death.

Most of the cases and controls did not use antipsychotics, or had done so in the past (Table 1). Only 19 (3.4%) cases and 34 (0.8%) controls were current users of antipsychotics, and in these the adjusted odds ratio for increased sudden cardiac death was 3.3 (95% CI 1.8 to 6.2; Figure 1). There was a somewhat higher association between use of antipsychotics and sudden cardiac death in witnessed deaths than unwitnessed deaths.

Table 1: Numbers of cases of sudden cardiac death and controls, and use of antipsychotic medicines

Use of antipsychotic medicine
Total patients
Past use
Current use
Of which
Others (atypical)
At doses of
≤0.5 DDD
>0.5 DDD

The numbers of users of different types of antipsychotics was small for each group (Table 1), but for users of older antipsychotics (haloperidol, for instance) the odds ratio was 7.3 (2.8 to 19) based on 12 cases and 13 controls using this group. No other group reached statistical significance, but based on small numbers of cases and controls using them. Higher doses were also associated with more frequent cardiac death (Figure 1), but only five cases and four controls were using doses above 0.5 DDD (Table 1). There was no difference between longer or shorter periods of use.

Figure 1: 95% confidence interval of the odds ratio associating antipsychotic use or dose with sudden cardiac death


This is a detailed study, looking at a large population with excellent recording of patient details, and with a large number of sudden cardiac death events. Only 19 of these events occurred in people using antipsychotics, on which the whole thesis rests. For any further analysis, by type of drug, by dose of drug, or by duration of use of drug, the numbers of cases mainly fall to single figures. Most of these analyses contain so few cases or controls that there must be a risk that they will be wrong just by the random play of chance.

So the best evidence we have is still limited, despite the quality and validity, and size of the population. We can be reasonably sure than antipsychotics are associated in something like a threefold increase in sudden cardiac death, and perhaps that older antipsychotics may be worse. We cannot be sure that newer antipsychotics, or lithium, or phenothiazines, are without effect.


  1. SM Straus et al. Antipsychotics and the risk of sudden cardiac death. Archives of Internal Medicine 2004 164: 1293-1297.

previous or next story