|
A correction has been published: N Engl J Med 2009;361(18):1814.
|
Background Users of typical antipsychotic drugs have an increased risk of serious ventricular arrhythmias and sudden cardiac death. However, less is known regarding the cardiac safety of the atypical antipsychotic drugs, which have largely replaced the older agents in clinical practice.
Methods We calculated the adjusted incidence of sudden cardiac death among current users of antipsychotic drugs in a retrospective cohort study of Medicaid enrollees in Tennessee. The primary analysis included 44,218 and 46,089 baseline users of single typical and atypical drugs, respectively, and 186,600 matched nonusers of antipsychotic drugs. To assess residual confounding related to factors associated with the use of antipsychotic drugs, we performed a secondary analysis of users of antipsychotic drugs who had no baseline diagnosis of schizophrenia or related psychoses and with whom nonusers were matched according to propensity score (i.e., the predicted probability that they would be users of antipsychotic drugs).
Results Current users of typical and of atypical antipsychotic drugs had higher rates of sudden cardiac death than did nonusers of antipsychotic drugs, with adjusted incidence-rate ratios of 1.99 (95% confidence interval [CI], 1.68 to 2.34) and 2.26 (95% CI, 1.88 to 2.72), respectively. The incidence-rate ratio for users of atypical antipsychotic drugs as compared with users of typical antipsychotic drugs was 1.14 (95% CI, 0.93 to 1.39). Former users of antipsychotic drugs had no significantly increased risk (incidence-rate ratio, 1.13; 95% CI, 0.98 to 1.30). For both classes of drugs, the risk for current users increased significantly with an increasing dose. Among users of typical antipsychotic drugs, the incidence-rate ratios increased from 1.31 (95% CI, 0.97 to 1.77) for those taking low doses to 2.42 (95% CI, 1.91 to 3.06) for those taking high doses (P<0.001). Among users of atypical agents, the incidence-rate ratios increased from 1.59 (95% CI, 1 .03 to 2.46) for those taking low doses to 2.86 (95% CI, 2.25 to 3.65) for those taking high doses (P=0.01). The findings were similar in the cohort that was matched for propensity score.
Conclusions Current users of typical and of atypical antipsychotic drugs had a similar, dose-related increased risk of sudden cardiac death.
Less is known about the cardiac safety of the atypical antipsychotic drugs, which have largely replaced the older agents in clinical practice. Several atypical antipsychotic drugs block repolarizing potassium currents2 and prolong ventricular repolarization,1,13 and the electrophysiological effects of some of these drugs are similar to those of the older agents. However, although torsades de pointes has been reported in persons using atypical antipsychotic drugs,14,15,16 whether these drugs increase the risk of sudden cardiac death to the same extent as the older medications is unknown. We therefore conducted a large retrospective cohort study that was designed to compare the risk of sudden cardiac death associated with the use of the two classes of antipsychotic drugs.
Methods
Primary Cohort
We obtained the study data from computerized files of Tennessee Medicaid, which have been used extensively for pharmacoepidemiologic research.17,18 Each person-day of Medicaid enrollment from January 1, 1990, through December 31, 2005 (the study period), was evaluated to determine whether it qualified for inclusion in the analysis. The cohort was restricted to persons 30 to 74 years of age, because among persons younger than 30, sudden cardiac death is very rare and may have a different cause,19 and among persons older than 74, we found death certificates to be less reliable for identifying sudden cardiac deaths. Inclusion in the cohort required the person to have been enrolled in Tennessee Medicaid for at least 730 days (gaps of <7 days were allowed) and to have been eligible for full pharmacy benefits and made regular use of medical care (defined as having had at least one filled prescription and one outpatient visit in each of the 2 preceding years). Patients at high risk for death from noncardiac causes were excluded from the cohort (see Appendix 1 in the Supplementary Appendix, available with the full text of this article at NEJM.org).
The cohort included every eligible Medicaid enrollee with at least 1 qualifying day of use of antipsychotic drugs during the study period. The first day of follow-up was defined as the first qualifying day. The cohort also included two controls for each user of antipsychotic drugs, matched for age, sex, and first day of follow-up, who were randomly selected from qualifying nonusers of antipsychotic drugs on the first day of follow-up.
Follow-up extended from the first qualifying day until the end of the study period, the death of the person, the termination of Medicaid enrollment, or the date on which eligibility criteria for inclusion in the cohort were no longer met. Controls could become users of antipsychotic drugs, and patients who left the cohort could reenter it. The follow-up time did not include time during hospitalization and the 30 days after discharge from the hospital because in-hospital deaths were not considered to be study end points and Medicaid files do not include drugs dispensed in the hospital.
The Vanderbilt Committee for the Protection of Human Subjects approved the study and waived the requirement for informed consent. Tennessee's Bureau of TennCare and Department of Health also approved the study. The study was funded by grants from federal agencies, which had no role in study conduct or reporting.
Cohort Matched for Propensity Score
To assess residual confounding by factors associated with the use of antipsychotic drugs, we performed a secondary analysis, using propensity scores20 (i.e., the predicted probability that a person would be a user of antipsychotic drugs) to identify a nonuser control group with a similar psychiatric-illness profile. This matched cohort excluded users of antipsychotic drugs who had a diagnosis of schizophrenia or related psychosis in the 730 days before the first day of follow-up, since treatment with antipsychotic drugs is the standard of care for these conditions. Thus, this group of qualifying users of antipsychotic drugs primarily had mood disorders (a growing reason for the use of antipsychotic drugs), for which there are alternative medications. For each qualifying user, up to two controls were adaptively matched for propensity score (i.e., controls were selected so as to reduce differences in running mean propensity scores) (Appendix 1 in the Supplementary Appendix), within strata defined by the first day of follow-up, birth year, sex, and a marker of severe psychiatric illness (previous hospitalization for psychiatric cause, diagnosis of bipolar disorder, or lithium therapy).
Exposure to Antipsychotic Drugs and Other Medications
Antipsychotic drugs and other study medications were identified from Medicaid pharmacy files. Data in the files included the date the prescription was dispensed, the drug, the quantity, the dose, and the number of days of supply (edited to resolve infrequent discrepancies with quantity). Computerized pharmacy records are an excellent source of medication data because they are not subject to information bias17 and have high concordance with medication use as reported by patients.21,22,23 The residual misclassification should be limited, and any bias is likely to be in favor of the null hypothesis.17
Each person-day of study follow-up was classified according to the probable use of antipsychotic drugs. Current use included the interval between the time the person filled the prescription and the end of the days of supply (up to a 7-day carryover from previous prescriptions), when the person was most likely to be taking the drug. Indeterminate use included the period up to 90 days after the last current use, and former use included any subsequent person-time that was not classified as current or indeterminate use. Nonuse referred to person-days with no prescribed use of antipsychotic drugs on those days or at any time in the past. Current use was further classified according to the doses of the drugs, which were expressed as approximate equivalents of 100 mg of chlorpromazine (Appendix 1 in the Supplementary Appendix)24,25 and then categorized as low-dose (<100 mg), moderate-dose (100 to 299 mg), or high-dose (
300 mg). The individual drugs that were analyzed were thioridazine and haloperidol, the most frequently prescribed typical agents, as well as atypical antipsychotic drugs with 3000 person-years or more of current use in the primary cohort (for which 5 cases of sudden death were expected under the null hypothesis): clozapine, quetiapine, olanzapine, and risperidone.
Study End Point
The study end point was sudden cardiac death occurring in the community.26,27,28 Sudden cardiac death was defined as a sudden pulseless condition that was fatal, that was consistent with a ventricular tachyarrhythmia, and that occurred in the absence of a known noncardiac condition as the proximate cause of the death.27 The end point excluded deaths of patients who had been admitted to the hospital, deaths that were not sudden, and deaths for which there was evidence of an extrinsic cause (e.g., drug overdose), a noncardiac cause (e.g., pneumonia), or a cardiac cause that was not consistent with a ventricular tachyarrhythmia (e.g., heart failure).
End points were identified from computerized death certificates linked with computerized Medicaid records. The case definition was developed from and validated by our previous study,8,29,30 in which medical records were reviewed for deaths that occurred between 1988 and 1993. Qualifying deaths occurred outside the hospital or other institution and had an underlying cause of death that our previous study had determined to be compatible with sudden cardiac death (Appendix 1 in the Supplementary Appendix). These deaths were further restricted to those for which there was no evidence of care in the emergency department on the day of death that was inconsistent with care for sudden cardiac death. In our previous study,8 we reviewed medical records for 616 of such qualifying deaths that occurred in the present cohort. Of these deaths, 530 (86.0%) were confirmed cases of sudden cardiac death (unpublished data). As long as the accuracy of the definition of sudden cardiac death did not vary according to the use of antipsychotic drugs, the residual misclassification should bias the results toward the null hypothesis.31
Statistical Analysis
The relative risk of sudden cardiac death according to current use, former use, or nonuse of antipsychotic drugs, adjusted for dose (Appendix 1 in the Supplementary Appendix), was estimated with the incidence-rate ratio, as calculated from Poisson regression models. The models (Appendix 1 in the Supplementary Appendix) included demographic characteristics and variables reflecting coexisting conditions at baseline and subsequent changes during follow-up. Baseline coexisting conditions included cardiovascular and other somatic disease as well as psychiatric and neurologic illness.
We calculated a summary cardiovascular risk score from the large number of baseline cardiovascular and somatic variables. The variables included prescribed medications and recorded diagnoses, as well as utilization of medical care and a measure of compliance with drugs (Appendix 1 in the Supplementary Appendix) for long-term use. The summary risk score was defined for the entire cohort as the predicted probability of sudden death, conditional on no exposure to antipsychotic drugs (estimated with Poisson regression analysis among nonusers of antipsychotic drugs), and then expressed as 20 equal parts. This technique permits more parsimonious models when there are numerous covariates and facilitates description of baseline cardiovascular risk.32
We performed several supplementary analyses to test key assumptions. These included an analysis that permitted only one cohort entry per person, as well as analyses in which additional baseline and time-dependent variables were included in the model (Appendix 1 in the Supplementary Appendix). The findings were essentially identical to those reported here.
All analyses were performed with the use of SAS software, version 9.0 (SAS Institute). All reported P values are two-sided.
Results
Characteristics of the Study Cohorts
The primary cohort included 93,300 users of antipsychotic drugs and 186,600 matched controls. There were 44,218 and 46,089 users of single typical and atypical antipsychotic drugs, respectively, at cohort entry. The cohort that was matched for propensity score included 67,824 users of antipsychotic drugs and 116,069 nonusers.
In the primary cohort, users and nonusers of antipsychotic drugs had similar baseline demographic characteristics (Table 1). The mean age was 45.7 years; 65.2% of the cohort members were women, 70.5% were white, and 56.9% were urban residents. Users of antipsychotic drugs were more likely than nonusers to be enrolled in Medicaid because of disability (62.9% vs. 37.4%), but they had a slightly lower mean baseline cardiovascular risk score (9.2 vs. 9.6 on a scale of 0 to 19, with higher scores indicating increased risk). As expected, users of antipsychotic drugs had a higher prevalence of coexisting psychiatric conditions at baseline than did nonusers; however, there was a substantial prevalence of coexisting conditions, particularly affective disorders, among nonusers as well. In the cohort matched for propensity score, users and nonusers of antipsychotic drugs had identical mean propensity scores and similar baseline rates of coexisting psychiatric conditions.
|
Sudden Cardiac Death
During the 1,042,159 person-years of cohort follow-up, there were 1870 sudden cardiac deaths, or 17.9 per 10,000 person-years. The unadjusted rate increased from 4.7 deaths per 10,000 for persons 30 to 34 years of age at baseline to 47.6 per 10,000 for those 70 to 74 years of age and was more than twice as high for men as for women (27.1 vs. 12.9 per 10,000).
Current users of typical antipsychotic drugs had an adjusted rate of sudden cardiac death that was twice that for nonusers (incidence-rate ratio, 1.99; 95% confidence interval [CI], 1.68 to 2.34) (Table 2). A similar increased risk was seen for current users of atypical antipsychotic drugs, who had a rate of sudden cardiac death that was more than twice that for nonusers (incidence-rate ratio, 2.26; 95% CI, 1.88 to 2.72) and that did not differ significantly from the rate for users of the typical agents (incidence-rate ratio for users of atypical as compared with users of typical antipsychotic drugs, 1.14; 95% CI, 0.93 to 1.39). The rates of sudden cardiac death for both current users of typical antipsychotic drugs and current users of atypical drugs were greater than those for former users (P<0.001). Former users did not have a significantly increased risk of sudden cardiac death as compared with nonusers (incidence-rate ratio, 1.13; 95% CI, 0.98 to 1.30). Users of each of the six frequently prescribed antipsychotic drugs had a significantly increased rate of sudden cardiac death (Table 2).
|
300 mg) had the greatest increased risk (incidence-rate ratio, 5.05; 95% CI, 3.09 to 8.27).
|
|
|
Discussion
The frequent occurrence of serious movement disorders in persons taking typical antipsychotic drugs limited the use of these drugs.24 Because atypical antipsychotic drugs are less likely to have this adverse effect, they have been considered a safer treatment alternative34 and have rapidly replaced the older drugs in clinical practice. Overall, the use of antipsychotic drugs has increased, with the number of outpatient visits related to the prescription of an antipsychotic drug nearly doubling between 1998 and 2002.34
Although a link between the use of typical antipsychotic drugs and both torsades de pointes and sudden cardiac death has been established,5,35 this risk was thought to be lower with the use of atypical drugs.36 However, the limited data available on the surrogate markers for torsades de pointes — inhibition of the potassium current IKr and prolongation of the QT interval — suggest that commonly used atypical drugs have electrophysiological effects that are similar to those of the typical antipsychotic drugs.1,13 There are now case reports that document the occurrence of torsades de pointes among users of several atypical antipsychotic drugs.14,15,16 Our data show that in a large retrospective cohort of adults, current users of the atypical antipsychotic drugs had a dose-dependent increase in the risk of sudden cardiac death that was essentially identical to that among users of the typical agents.
The primary limitation of our study is the potential for confounding by factors associated with the use of antipsychotic drugs. For persons with serious mental illness, these factors include cardiovascular and other somatic disease; concurrent use of other proarrhythmic medications; mood disorders; behavioral risk factors, including substance abuse, poor self-care, and smoking; and other effects of mental illness.12 However, both the study design and analysis included several provisions to manage confounding.
We controlled for an extensive set of cardiovascular disease variables. In the Medicaid population studied, users of antipsychotic drugs had a slightly lower baseline prevalence of diagnosed cardiovascular disease than did comparable nonusers, reflecting the fact that many nonusers qualified for Medicaid because of somatic illness. The requirement that cohort members have regular use of medical care, defined by at least one outpatient visit in each of the 2 years before baseline, should reduce the bias from underdiagnosis of cardiovascular disease in patients with mental illness. The analysis also controlled for concurrent use of other proarrhythmic medications, as well as for diagnosed or treated mood disorders.
With regard to behavioral risk factors, the cohort excluded persons with recorded diagnoses of substance abuse and those who did not have regular medical care. Although study data on smoking were limited, the analysis controlled for cardiovascular diseases caused by smoking,37 diseases that mediate much of the increased risk of sudden death. Furthermore, a sensitivity analysis (Appendix 1 in the Supplementary Appendix) suggested that residual confounding by smoking had at most a minor effect on estimates of relative risk. Although unmeasured behavioral factors may influence the study findings, the absence of a significantly increased risk of sudden death among former users of antipsychotic drugs and the marked dose–response relationship are evidence of a drug effect per se.
An analysis of the cohort that was matched for propensity score provided an additional check as to whether the study findings were due to confounding by factors associated with the use of antipsychotic drugs. This cohort excluded persons with a baseline diagnosis of schizophrenia or related psychoses, for whom such confounding is of greatest concern, and had a similar distribution of coexisting psychiatric conditions at baseline among users and nonusers of antipsychotic drugs. Findings were very similar to those for the primary cohort. However, some point estimates of relative risk were slightly lower; these differences, although not significant, underscore the fact that in this observational study residual confounding cannot be entirely ruled out.
Our study did not assess the mechanisms by which either class of antipsychotic drugs increased the risk of sudden cardiac death. Although antipsychotic drugs have long-term adverse cardiovascular effects,12 the risk of sudden death was elevated in an analysis that excluded long-term users, which suggests that acute drug effects are involved. We believe that the most plausible explanation is that antipsychotic drugs increase the risk of serious ventricular arrhythmias, probably through blockade of potassium channels and prolongation of cardiac repolarization. However, other mechanisms may be involved, including autonomic effects, inhibition of other ion channels, and other acute cardiotoxic effects, such as the myocarditis associated with the use of clozapine.38
In conclusion, current users of typical antipsychotic drugs and of atypical antipsychotic drugs in the study cohort had a similar dose-related increased risk of sudden cardiac death. This finding suggests that with regard to this adverse effect, the atypical antipsychotic drugs are no safer than the older drugs.
Supported by grants from the National Heart, Lung, and Blood Institute (HL081707) and the Agency for Healthcare Quality and Research Centers for Education and Research on Therapeutics (cooperative agreement, HS1-0384).
Dr. Ray reports receiving research support from Pfizer and consulting fees from plaintiffs' attorneys in litigation involving rofecoxib and zoledronic acid and from insurance companies regarding rofecoxib and hormone-replacement therapy; and Dr. Stein, consulting fees from insurance companies regarding antipsychotic drugs. No other potential conflict of interest relevant to this article was reported.
We thank the staff of the Tennessee Bureau of TennCare and the Department of Health, who gave us access to the study data.
Source Information
From the Division of Pharmacoepidemiology, Department of Preventive Medicine (W.A.R., K.H.), the Divisions of Rheumatology (C.P.C., C.M.S.), Cardiology (K.T.M.), and Clinical Pharmacology (K.T.M., C.M.S.), Departments of Medicine and Pharmacology, Vanderbilt University School of Medicine; and the Geriatric Research, Education, and Clinical Center, Nashville Veterans Affairs Medical Center (W.A.R.) — both in Nashville.
Address reprint requests to Dr. Ray at the Department of Preventive Medicine, Village at Vanderbilt, Suite 2600, 1501 21st Ave. South, Nashville, TN 37212, or at cindy.naron{at}vanderbilt.edu.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Related Letters:
Atypical Antipsychotic Drugs and the Risk of Sudden Cardiac Death
Baldessarini R. J., Price L. H., Ray W. A., Murray K. T., Stein C. M.
Extract |
Full Text |
PDF
N Engl J Med 2009;
360:2136-2138, May 14, 2009.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |
Previous
