Association Between Antipsychotic Use and Risk of Acute Myocardial InfarctionCLINICAL PERSPECTIVE
A Nationwide Case-Crossover Study
Background—Antipsychotic medications have been increasingly and more widely prescribed despite continued uncertainty about their association with the incidence of acute myocardial infarction (AMI).
Methods and Results—We investigated the risk of AMI associated with antipsychotic treatment in 56 910 patients with schizophrenia, mood disorders, or dementia first hospitalized or visiting an emergency room for AMI in 1999 to 2009. A case-crossover design was used to compare the distributions of antipsychotic exposure for the same patient across 1 to 30 and 91 to 120 days just before the AMI event. Adjustments were made for comedications and outpatient visits. The adjusted odds ratio of AMI risk was 2.52 (95% confidence interval, 2.37–2.68) for any antipsychotics, 2.32 (95% confidence interval, 2.17–2.47) for first-generation antipsychotics, and 2.74 (95% confidence interval, 2.49–3.02) for second-generation antipsychotics. The risk significantly increased (P<0.001) with elevations in dosage and in short-term use (≤30 days). Male patients, elderly patients, and patients with dementia were at significantly increased risk (all P<0.001). Physically healthier patients with no preexisting diabetes mellitus, hypertension, or dyslipidemia were at significantly greater risk (P<0.001), largely because they had been exposed to higher doses of antipsychotics (P<0.001). A study of the selected binding of antipsychotics to 14 neurotransmitter receptors revealed only dopamine type 3 receptor antagonism to be significantly associated with AMI risk (adjusted odds ratio, 2.59; 95% confidence interval, 2.43–2.75; P<0.0001).
Conclusions—Antipsychotic use may be associated with a transient increase in risk for AMI, possibly mediated by dopamine type 3 receptor blockades. Further education on drug safety and research into the underlying biological mechanisms are needed.
Recent evidence suggests that psychiatric patients may be at increased risk for cardiovascular events. Large community studies report that people with mental illnesses have a 2- to 5-fold greater risk of coronary heart disease and a 2- to 3-fold greater risk of cardiac mortality than the general population.1–3 This increased risk might be explained by risk factors commonly noted in these patients, including smoking, obesity, and unhealthy lifestyles.4–6
Clinical Perspective on p 243
The association of risk for acute myocardial infarction (AMI) with antipsychotic treatment remains unclear because earlier analyses have methodological issues (eg, residual confounding and limited statistical power and external validity).3,7–11 If the link exists, more attention might be paid to monitoring drug safety, especially because of current increased use of antipsychotics and the widening range of disorders for which they are being indicated.12
This study tapped a large nationwide population-based data set to investigate the association between antipsychotic treatment and risk of AMI in patients with mental disorders. A case-crossover design was used to eliminate known or unknown within-subject time-invariant confounders while examining the acute effects of various antipsychotics on subsequent AMI events.13 The moderating effects of patient characteristics (eg, age, sex, psychiatric diagnoses, and baseline medical conditions) and the relationships of antipsychotic binding to various neurotransmitter receptors to the risk of AMI were also investigated.
Data were obtained from Taiwan’s National Health Insurance Research Database (NHIRD), a large database provided by Taiwan’s single-payer, universal, compulsory healthcare program. This program covered >99% of 23 315 822 residents in Taiwan in 2012. Its database contains inpatient/ambulatory claims, prescription claims, and demographic data for all beneficiaries. The present study used medical claims data and registration data retrieved from the NHIRD for 1997 to 2009. The protocol for this study was approved by the Institutional Review Board of Kaohsiung Medical University Hospital (KMUH-IRB-980174).
We identified all patients ≥18 years of age who were diagnosed with schizophrenic disorders (International Classification of Disease, Ninth Revision, Clinical Modification [ICD-9-CM] code 295), mood disorders (ICD-9-CM code 296), or dementia (ICD-9-CM codes 290 and 331.0) and later found to have an AMI incident (ICD-9-CM code 410). Patients with these disorders were selected because they are commonly prescribed antipsychotic medications.10,14 For patients to be included, they had to have received no myocardial infarction–related diagnoses (ICD-9-CM codes 410, 411.0, and 412) from January 1, 1997, to the date of AMI event leading to hospital admission or emergency department visit in 1999 to 2009. To ensure that new cases of AMI were being identified, patients with concomitant diagnosis of cardiomyopathy (ICD-9-CM code 425), pericarditis (ICD-9-CM code 420), aortic dissection (ICD-9-CM code 441), or coronary aneurysm (ICD-9-CM code 414.11) were excluded. Patients were also excluded if they had been prescribed only prochlorperazine, an agent commonly used as an antiemetic rather than an antipsychotic.
A case-crossover design was used to compare antipsychotic exposure for the same patient within a case period (the 1- to 30-day period before the AMI event) with exposure within the control period (the 91- to 120-day period before the event).11 This self-controlled method reduces the possibility of within-person time-invariant confounding (eg, lifestyles, genetic profiles, and body characteristics affecting the pharmacokinetics of antipsychotics) and separates short-term effects from longer-term effects of an antipsychotic exposure.13 This approach also avoids control selection biases associated with traditional case-control designs.13 Since 1988 when the case-crossover design was originally developed to study immediate triggers for AMI,15 several studies have used it to study the risk of stroke after use of antipsychotics or antidepressants and the risk of AMI after the use of analgesics.16–18 For comparison purposes, this study uses the same definitions of time window (30 days) and time frame as those adopted in previous studies of antipsychotics.11,16
Exposure to Antipsychotics
Antipsychotic use was defined as the patient having at least 1 antipsychotic prescription (Anatomic Therapeutic Chemical [ATC] code N05A) for 1 day during the case or control period.19 Dividing the cumulative prescribed doses into the total number of days supplied during the case or control period allowed the calculation of mean daily dose. Risk of AMI was examined with 3 mean daily dose ranges: <90, 90 to 180, and >180 chlorpromazine-equivalent doses.20
Within-Patient Time-Variant Confounding Factors
For each subject, factors that could potentially change across the case and control periods were identified and analyzed. These included number of outpatient visits (a measure of healthcare use) and the prescriptions of concomitant medications potentially related to AMI, including antihypertensive agents (ATC codes C02 and C08), β-blockers (ATC code C07), angiotensin-converting enzyme inhibitors (ATC code C09AA) or angiotensin receptor blockers (ATC code C09CA), antidiabetic agents (ATC code A10), diuretics (ATC code C03), statins (ATC code C10AA), nonsteroidal anti-inflammatory drugs (ATC code M01A), low-dose aspirin (ATC code B01AC06), nonaspirin antiplatelet agents (ATC code B01AC except B01AC06), and vitamin K antagonists (ATC code B01AA). A detailed list of the comedications is presented in Table I in the online-only Data Supplement.
To identify which of the receptors known to bind to antipsychotic drugs might be most associated with AMI risk, we calculated and compared the assumed degrees of receptor occupancy of 14 receptors, including the dopamine receptors (D1, D2, D3, and D4), serotonin receptors (5-HT1A, 5-HT2A, 5-HT2C, 5-HT3, 5-HT6, and 5-HT7), and adrenergic (α1 and α2), muscarinic (M1), and histamine (H1) receptors occupied by the antipsychotics prescribed during the study periods, following an equation derived from the operational model in pharmacological receptor theory:21
where [L] is the blood concentration of a prescribed antipsychotic and Kd is the dissociation constant value of that drug. Blood concentration (molar) was estimated by multiplying the mean daily drug concentration (here total human blood volume was assumed to be 5 L)22 of an antipsychotic by its oral bioavailability, defined as the fraction of an orally administered dose of unaltered drug that would reach systemic circulation after absorption, distribution, metabolism, and elimination. We excluded patients who had received thioridazine, trifluoperazine, loxapine, clothiapine, flupentixol, chlorprothixene, clopenthixol, zuclopenthixol, thiothixene, levomepromazine, and pipotiazine, drugs for which their Kd values or bioavailability data are currently unavailable from the National Institute of Mental Health Psychoactive Drug Screening Program23 and Micromedex Healthcare Series Internet Database (Truven Health Analytics).24 Therefore, only the receptor-binding profiles (Table II in the online-only Data Supplement) of the 14 antipsychotics (amisulpride, aripiprazole, chlorpromazine, clozapine, fluphenazine, haloperidol, olanzapine, paliperidone, pimozide, quetiapine, risperidone, sulpiride, zotepine, and ziprasidone) were used to investigate the association between risk of AMI and binding of antipsychotics (high receptor occupancy [>50%] versus low receptor occupancy [≤50%]) to 14 receptors.25
Conditional logistic regression was used to estimate the association between risk of AMI, use of antipsychotics, and binding to receptors. The crude odds ratio (OR) and the adjusted OR (AOR) were calculated, comparing the odds of antipsychotic exposure between 1 to 30 days (case period) and 91 to 120 days (control period) before the AMI event. The adjustments controlled for the effects of outpatient visits and concomitant medications across these 2 periods. Several sensitivity analyses were performed to examine the robustness of the results. First, the present model was compared with other models of different control periods and time windows: time frame 1 (case/control period=1–30/61–90 days), time frame 2 (case/control period=1–30/121–150 days), 14-day window (case/control period=1–14/91–104 days), and 7-day window (case/control period=1–7/91–97 days). Second, because diabetes mellitus, hypertension, or hyperlipidemia status might change within the study time frames, models with and without adjustments for the presence of these diagnoses in the case and control periods were compared. Third, we evaluated the impact of changes in threshold for the definition of AMI on our results using different criteria for subject enrollment. Because AMI and unstable angina belong to the same spectrum of clinical manifestations attributed to obstruction of the coronary arteries, we conducted a separate analysis for patients without a history of unstable angina (ICD-9-CM code 411.1) before the incident AMI. Because an AMI event requires immediate medical attention, we also repeated the models for only patients who received several treatments specifically indicated for AMI within 30 days after an AMI incident. Those treatments included percutaneous coronary intervention, fibrinolysis, bypass surgery, and antiplatelet/anticoagulant therapy. To ensure study quality, we analyzed the risk of 2 negative controls, benzodiazepines (ATC codes N05CD and N05BA) and nonbenzodiazepine hypnotics (ATC code N05CF), both with no known association with AMI. Subgroup analyses were performed by stratifying by various patient characteristics: age group, sex, Charlson Comorbidity Index (a measure of health status),26 psychiatric disorders, number of metabolic disorders, level of insurance premiums (a proxy of socioeconomic status; premiums less than US $700, US $700–US $1000, and more than US $1000 per month reflected low, moderate, and high socioeconomic status, respectively),27 and cumulative days of antipsychotic prescriptions in the year before the AMI. The interactions between these characteristics and antipsychotic use were examined by likelihood ratio tests. To determine the most significant receptor binding associated with an antipsychotic-related AMI, a stepwise regression procedure was used with variable selection criteria set at P<0.05 for both entry into and removal from the regression model. All statistical operations were performed with SPSS statistical software (version 17.0) or SAS statistical software (version 9.13). Statistical significance was defined using the 95% confidence interval (CI) or a P threshold of 0.05.
We identified 59 806 patients with schizophrenia (n=5748), mood disorders (n=21 768), or dementia (n=32 290) who subsequently had a first-ever diagnosis of AMI in 1999 to 2009. Those with comorbid cardiomyopathy (n=514), pericarditis (n=34), aortic dissection (n=317), or coronary aneurysm (n=11) were excluded. Those who had taken only prochlorperazine (n=2020) were further excluded. We were left with 56 910 patients subsequently treated for AMI. Patient demographic and clinical characteristics are summarized in Table 1. Mean age at AMI onset was 71.5 years (SD, 13.2 years); mean cumulative days of antipsychotic prescriptions in the previous year were 102 days (SD, 93.0 days).
Table 2 shows that use of any antipsychotic within the 30-day period before the AMI event was significantly associated with increased risk for AMI after adjustment for outpatient visits and comedications (AOR, 2.52; 95% confidence interval [CI], 2.37–2.68; P<0.001). There was a significant dose-dependent increase in AMI risk (P for trend <0.001; Wald χ2=317.67; df=1). Use of either a first-generation antipsychotic or a second-generation antipsychotic (SGA) significantly increased the risk of AMI, although there was a higher risk associated with SGAs compared with first-generation antipsychotics (AOR, 1.63; 95% CI, 1.29–2.05; P<0.001). Reanalysis of the data using 2 other time frames and time windows, as well as different input variables and subject inclusion criteria, generated similar results (Tables III and IV in the online-only Data Supplement). There was no important interaction of antipsychotic use with comedications on the risk of AMI. The association between AMI risk and use of 2 negative controls (benzodiazepine drugs and nonbenzodiazepine hypnotics) was not significant (AOR, 1.03; 95% CI, 0.92–1.15; and AOR, 1.06; 95% CI, 0.95–1.18; respectively).
As shown in Table 3 and Figure I in the online-only Data Supplement, the highest risk of AMI was found with amisulpride (AOR, 5.65; 95% CI, 2.97–10.76). Table V in the online-only Data Supplement shows the confounding effects of various covariates on the association with AMI risk.
Table 4 shows that the antipsychotic-associated AMI risk was significantly greater among male patients, elderly patients, patients with dementia, patients who had no metabolic disorder, patients who had the lowest Charlson Comorbidity Index scores, and patients who had less prior exposure to antipsychotics (all P<0.001). Risk estimates were uniform among subgroups of different insurance premiums. Baseline comparisons between subgroups revealed that only patients with the lowest Charlson Comorbidity Index scores and those with no metabolic disorder received significantly higher average daily doses of antipsychotics during their case periods (P<0.001) and significantly greater dosage increases across their control and case periods (P<0.001).
Table 5 shows that several receptor-binding characteristics, including binding to D1, D3, 5-HT2A, 5-HT2C, 5-HT6, α2, and H1 receptors, were significantly associated with risk of AMI. After considering the main effects of antipsychotic binding to all 14 receptors and their interaction terms in the conditional logistic regression model and adjusting for outpatient visits and comedications, we performed stepwise regression analysis that showed that only D3 receptor antagonism actually predicted the risk for AMI (AOR, 2.59; 95% CI, 2.43–2.75; P<0.0001).
This study found antipsychotic use in patients with schizophrenia, mood disorders, or dementia to be significantly associated with the risk for AMI. The risk was dose dependent and increased in short-term users, male patients, elderly patients, and patients with dementia. Physically healthier patients with no preexisting diabetes mellitus, hypertension, or dyslipidemia had greater risk, largely because physicians were less likely to restrict common dosages of antipsychotic prescription for these patients. The AMI risk of antipsychotic use appeared to be specifically associated with D3 receptor blockades by antipsychotics.
Our findings of this association between the risk of AMI and the use of antipsychotics are consistent with previous cohort and case-control studies.3,7–9,11 This case-crossover study provides new evidence that antipsychotics may pose significantly higher risk for AMI in short-term users (≤30 days in the year before AMI). This finding may partially explain the reason that 1 previous case-control study,28 which used a long study time window (90 days), did not find an association between antipsychotic use and AMI risk. Such a study design could possibly mix the immediate, severe, rapidly passing effects of antipsychotics with their long-term effects, which would reduce the risk estimates. The relatively lower risk of AMI observed in more long-term antipsychotic users might be related to the effects of tolerance and cross-tolerance to antipsychotic drugs.29 We suggest a prospective cohort study be conducted to further investigate whether such time-dependent risk exists and to determine safer ways of prescribing antipsychotic medications.
We found male patients, older patients, and those with dementia to be at higher risk for AMI when prescribed antipsychotics (Table 4). This elevated risk might be explained by sex- and age-related differences in blood levels of estrogen/cholesterol.30 We also found physically healthier patients with no preexisting diabetes mellitus, hypertension, or dyslipidemia to be at possibly increased risk of AMI when treated with antipsychotics (Table 4), although this finding was likely to be explained by the fact that these patients had been exposed to significantly higher doses of antipsychotics (P<0.001) with greater increases in dosing adjustments (P<0.001) before the AMI events. We observed a dose-dependent effect on AMI risk (Table 2 and Tables III and IV in the online-only Data Supplement), possibly suggesting that the increased amount of drugs that these healthier individuals received could have increased their risk for AMI. Our finding of a lower immediate risk for AMI in patients who received lower doses in the short term but who had a higher baseline metabolic risk indicates that our case-crossover design indeed enabled us to separate the short-term effects of the antipsychotic on AMI risk from its long-term effects mediated by a metabolic disorder.
In addition to AMI, ventricular arrhythmia and sudden cardiac death have been reported to be associated with antipsychotic use.31 Because most sudden cardiac deaths are caused by ventricular fibrillation secondary to AMI,32 we hypothesize that an antipsychotic-related AMI could possibly serve as a sentinel event setting in motion a chain of these serious adverse cardiovascular events. However, further study is needed to investigate the relationship between antipsychotic use and sudden cardiac death, which could also represent an additionally great diversity of cardiac events associated with antipsychotic drugs.
Although it is not known through which mechanism antipsychotic use may influence the risk for AMI, we found D3 receptor antagonism to be the most likely pharmacological property associated with AMI (AOR, 2.59; 95% CI, 2.43–2.75) when studying antipsychotic binding to 14 neurotransmitter receptors. In fact, animal and cell studies have suggested that D3 receptor blockades could potentially contribute to the development of AMI through several pathophysiological processes. For example, in mice, aberrant D3 receptor expression in heart and in peripheral vascular systems may increase intimal permeability, vascular remodeling, atherosclerosis formation, or the possibility of plaque rupture by interacting with renin, angiotensin II, insulin, endothelin, and cytokine vascular permeability factor/vascular endothelial growth factor.33–38 Additionally, there are several blood coagulation pathways involving D3 receptors systems, including the activation of platelet aggregation and the secretion of the procoagulant von Willebrand factor.39,40 Hence, D3 receptors blockade could also predispose the patient to the formation of acute thrombosis in stenotic coronary arteries, contributing to AMI.
Another possible pathogenic pathway in which antipsychotics could increase risk of AMI is through prolonging the QT interval. However, recent evidence from a 20-year prospective cohort study of healthy individuals (n=15 558) does not support the link between QT interval and AMI risk.41 Another possibility is that antipsychotic use could predispose a patient to increased metabolic risk, accelerating the development of AMI, although this seems unlikely in the context of our study. Had this been the case, our results would have shown a progressive increase in AMI risk over the durations of antipsychotic exposure and a specifically greater risk for patients with a metabolic disorder (Table 4). Nevertheless, these hypotheses should be examined through the use of biological models.
This study found that use of amisulpride, which is generally thought to be cardiovascular safe,42 had the highest AMI risk of all the medications studied (Table 3 and Figure I in the online-only Data Supplement). Amisulpride is a unique dopamine receptor antagonist. It is different from other antipsychotics in that it has the highest selective affinity for D3 receptors,43 a property found by this study to be associated with the greatest AMI risk. Although amisulpride has in fact been associated with AMI in 1 postmarketing survey,44 its risk should not be overemphasized because this study included only a small number of patients taking amisulpride. After all, the present study found an association between several antipsychotics and AMI risk.
Possibly as a result of their favorable extrapyramidal side-effect profiles and an increase in official approvals for their application in a wider variety of mental disorders, SGAs are rapidly supplanting first-generation antipsychotics in their use in clinical psychiatry in the United States.12 The present study, however, found SGAs to carry greater risk for AMI than first-generation antipsychotics (AOR, 1.63; 95% CI, 1.29–2.05). That SGAs have a higher affinity for D3 receptors may partially explain this increased risk.45 In addition, because SGAs have been associated with obesity and increased risk of stroke,16 more research may be undertaken to reevaluate the cardiovascular safety of these increasingly popular drugs.
This study has several limitations. One limitation is that the study might still include recurrent cases with an AMI history before 1997. Another is that it did not exclude patients prescribed long-acting antipsychotic drugs (<0.1% of the study population), which have pharmacokinetic properties that show delayed onset and prolonged action. The study might also have included some non-AMI patients as a result of miscoding in the NHIRD, although 1 previous validation study found the positive predictive value and sensitivity of the diagnoses in NHIRD for major cardiovascular events to be high (88.4% and 97.3%, respectively).46 To improve the validity of AMI diagnoses, we identified only patients who had been hospitalized or treated at an emergency department for an incident AMI. We also found that 83% of cases enrolled in our final analysis had been treated with percutaneous coronary intervention, fibrinolysis, antiplatelet/anticoagulant therapy, or bypass surgery within 30 days after the AMI events. Another limitation might be that, because NHIRD reports only prescribed doses, not doses actually taken by the patients, patient nonadherence might have led to underestimates of the actual risk of antipsychotic use. Still another is that, because only adult patients with schizophrenia, mood disorders, or dementia were studied, our results might not be generalized to other populations such as nonpsychiatric patients and children with autism or attention deficit/hyperactivity disorders. In addition, fatality cases resulting from AMI may not be completely included in our analysis because we did not link to the National Mortality File Database in Taiwan. We might not have included all the cases of incident AMI initially presenting with out-of-hospital cardiac arrest, which has been found to constitute 0.3% of all the emergency visits in Taiwan.47 However, because this is a case-only study, missing cases with AMI would simply decrease the representativeness of the study instead of substantially biasing the results. Another limitation might be that, because we used receptor-binding profiles obtained from in vitro studies, our findings on the association between effects on certain receptors with AMI risk might not reflect the actual relationships observed in the human body. Finally, the estimated AMI risk associated with antipsychotic use might be related to the indications (eg, agitation, aggression, and risk-taking behaviors) for which an antipsychotic was prescribed. Hence, our results might be confounded by indications. However, for each psychiatric patient, the past and recent reasons associated with antipsychotic treatment were highly correlated. Confounding related to those reasons, including those unmeasured biological and psychosocial variables, should be able to be eliminated by the self-matching in case-crossover design. Moreover, we used benzodiazepine drugs and nonbenzodiazepine hypnotics, sedatives that are often used to manage agitation and aggression either alone or in combination with antipsychotics, as negative controls to evaluate whether our findings were complicated by confounding by indications. If the association between antipsychotic use and acute risk for AMI found by our study design was spurious, then we should be able to observe that sedative use was also linked to the risk for AMI. In fact, our findings showed that use of benzodiazepine drugs and nonbenzodiazepine hypnotics was not significantly associated with AMI risk (AOR, 1.03; 95% CI, 0.92–1.15; AOR, 1.06; 95% CI, 0.95–1.18, respectively).
This study analyzed data of patients in Taiwan. Asians may be more vulnerable to adverse drug effects because they metabolize drugs more slowly than other ethnic groups,48 so there is a need for caution when our results concerning Asian-specific responses are extrapolated to antipsychotic drugs.
This study found a significant association between antipsychotic treatment and risk of AMI. Given the widespread use of antipsychotics in primary care and a lack of perception of such risk, we suggest that clinicians be educated to start from low dosage of antipsychotics when such treatment is clearly indicated, followed by a close monitoring for the signs/symptoms of AMI, especially within the first 30 days of a treatment period. Special caution should be used when antipsychotics are prescribed to aged individuals with dementia. Because the D3 receptor was found to be the most likely molecular target responsible for AMI associated with antipsychotic use, further research is needed to investigate the possible underlying biological mechanisms of antipsychotic-related AMI.
This study is based on data from the NHIRD provided by the Bureau of National Health Insurance of the Department of Health, Taiwan, and managed by the National Health Research Institutes, Taiwan. The interpretation and conclusions expressed in this article do not represent those of the Bureau of National Health Insurance, the Department of Health, or the National Health Research Institutes. Drs Lin and P. Yang had full access to all the data in the study. Dr Lin takes responsibility for the integrity of the data and the accuracy of the data analysis.
Continuing medical education (CME) credit is available for this article. Go to http://cme.ahajournals.org to take the quiz.
The online-only Data Supplement is available with this article at http://circ.ahajournals.org/lookup/suppl/doi:10.1161/CIRCULATIONAHA.114.008779/-/DC1.
- Received January 12, 2014.
- Accepted May 8, 2014.
- © 2014 American Heart Association, Inc.
- Pratt LA,
- Ford DE,
- Crum RM,
- Armenian HK,
- Gallo JJ,
- Eaton WW
- Haupt DW,
- Newcomer JW
- Penttinen J,
- Valonen P
- Maclure M
- Mittleman MA,
- Maclure M,
- Tofler GH,
- Sherwood JB,
- Goldberg RJ,
- Muller JE
- Shau WY,
- Chen HC,
- Chen ST,
- Chou HW,
- Chang CH,
- Kuo CW,
- Lai MS
- 19.↵World Health Organization, Collaborating Centre for Drug Statistic Methodology. Guidelines for ATC Classification and DDD Assignment. Oslo, Norway: World Health Organization; 2009.
- Starr C,
- Taggart R,
- Evers C,
- Starr L
- 24.↵Truven Health Analytics, Inc. Micromedex® Healthcare Series [Internet database]. Greenwood Village, CO: Thomson Healthcare. Accessed October 16, 2013.
- Goudie AJ,
- Cole JC
- Jousilahti P,
- Vartiainen E,
- Tuomilehto J,
- Puska P
- Huang H,
- Han Y,
- Wang X,
- Chen C,
- Yu C,
- He D,
- Wang H,
- Zhou L,
- Asico LD,
- Jose PA,
- Zeng C
- Yu C,
- Yang Z,
- Ren H,
- Zhang Y,
- Han Y,
- He D,
- Lu Q,
- Wang X,
- Wang X,
- Yang C,
- Asico LD,
- Hopfer U,
- Eisner GM,
- Jose PA,
- Zeng C
- Kotan Z,
- Ertepe B,
- Akkaya C,
- Sarandol E,
- Ozkaya G,
- Kirli S
- Ravizza L
- Bressan RA,
- Erlandsson K,
- Jones HM,
- Mulligan R,
- Flanagan RJ,
- Ell PJ,
- Pilowsky LS
- 47.↵Taiwan Emergency Management Association. Introduction to Emergency Management Association in Taiwan. January 19, 2011. http://www.taem.org.tw/about.aspx. Accessed March 7, 2014.
Recent evidence suggests that patients with mental illnesses have a 2- to 5-fold greater risk of coronary heart disease and a 2- to 3-fold greater risk of cardiac mortality than the general population. These patients are at risk not only because of their unfavorable health profiles for cardiovascular disease but also possibly because of the drugs that they are taking. To date, the association between the use of antipsychotics and the risk of acute myocardial infarction (AMI) remains unclear. Studying 56 910 patients with schizophrenia, mood disorders, or dementia first hospitalized or visiting an emergency room for AMI, we found a significant association between AMI risk and the antipsychotic drugs used before the AMI event. The risk was dose dependent and increased in short-term users (≤30 days), male patients, elderly patients, and patients with dementia. We also found dopamine type 3 receptor blockading to be the most likely pharmacological property responsible for AMI associated with antipsychotic use. Because clinicians frequently prescribe antipsychotics for various mental health conditions with little perception of potential risk for AMI, we suggest that antipsychotics be initiated starting with low dosages when such treatment is clearly indicated, followed by a close monitoring for patient signs or symptoms of AMI, especially within the first 30 days of a treatment period. Special caution should be used when prescribing antipsychotics to elderly individuals with dementia. Further research is needed to investigate the underlying biological mechanisms of antipsychotic-related AMI.